Professional Documents
Culture Documents
Barriers - Hedrich Model
Barriers - Hedrich Model
At Massey University
Albany
New Zealand
Elizabeth MacColl
2014
Abstract
Background
Falls are an unacceptable cost to the patient and their family/whānau and to the health care
environment. A risk assessment tool to identify which hospital in-patients have a high risk of
a fall enables staff to implement targeted fall prevention strategies. The tool should have
good specificity and sensitivity, be clear and quick to complete and be acceptable to the staff
members who use it. This study aimed to identify barriers for nurses and midwives using the
Hendrich II Fall Risk Model.
Method
A non-experimental descriptive survey design was selected to explore the research questions.
The validated questionnaire, the Barriers and Facilitators Assessment Instrument, was used to
identify the barriers for nurses and midwives (n = 404) from medical, surgical and obstetric
settings, in five hospitals and two continuing care facilities in one New Zealand District
Health Board.
Results
An overall response rate of 31% was achieved. The barriers found were insufficient
supportive staff, a lack of equipment, poor design of space, the specificity and flexibility of
the Hendrich II Fall Risk Model, lack of care provider knowledge and motivation and that
patients do not cooperate with their falls prevention plan. Results demonstrated that
respondents work according to procedures, are able to adapt their practice to incorporate new
routines and use the Hendrich II Fall Risk Model as a beginning point for falls prevention
planning. Analysis of the responses of nurses in medical and surgical areas was different
from responses from hospital staff in outpatient, paediatric, obstetric and emergency
department areas.
Conclusion
The study identified a number of barriers to the use of the Hendrich II Fall Risk Model. The
recommendations to the District Health Board included professional development for nurses
about how to incorporate clinical judgment as part of falls risk assessment, to improve patient
education regarding falls prevention, to review the Upright training and use other methods of
assessing falls risk in specific areas of practice. Further research into acceptable tools to
assess risk is required in short stay, outpatient clinics, paediatric and obstetric areas.
ii
Acknowledgements
A very special thanks to all the nurses and midwives who participated in this study. I am very
grateful that they gave their time to respond to this survey.
I wish to thank my friends and colleagues for their encouragement and interest, especially
Cynthia Ronaldson.
I would also like to thank Susan McHugh, Julie Daltry and Veronique Gibbons for their
support to help me get started and their ongoing interest.
Last and certainly not least, thanks to my family for their patience and understanding.
iii
Contents
Abstract ............................................................................................................................... ii
iv
Chapter Four – Results..................................................................................................................49
References .....................................................................................................................................83
Appendices ...................................................................................................................................99
Appendix 4 Permission to use Barriers and Facilitators Assessment Instrument ............................. 103
v
Appendix 8 DHB authorisation ........................................................................................................ 108
List of Tables
List of Figures
vi
Chapter One – Introduction
1.1 Introduction
The nature of a patient fall is complex and consequently research into falls is challenging. A
fall is influenced by patient-specific factors such as diagnosis and acuity of condition, by
aspects of the environment and staffing, and by organisational resources and procedures.
Falls are the most common serious adverse event reported in New Zealand hospitals (Health,
Quality and Safety Commission, 2012). Falls may result in extended treatment or serious
psychological or physical morbidity, or death (Joint Commission Resources, 2005). Research
into falls in New Zealand is needed as much as it is in other parts of the world.
Falls in hospital and when a serious injury is sustained are predominately in the population
who are over 60 years of age (Hitcho et al., 2004). As this is the same population that is
predicted to rise in coming years (World Health Organization (WHO), 2007), there is likely
to be an increase in incidence and prevalence of falls. Without effective measures to prevent
falls, there will be increasing costs to the district health boards, to patients and to their
families. Studies in New Zealand have looked at in-patient falls but none have identified the
barriers to the use of a risk assessment tool to plan preventive care for falls for nurses and
midwives.
Most literature pertaining to falls prevention in a hospital setting supports the use of a
validated tool to assess risk, staff knowledge regarding falls prevention, provision of a well-
resourced and safe environment, leadership support, facilitation of a multidisciplinary
approach and a robust method of reporting and learning from fall events (Miake-Lye,
Hempel, Ganz, & Shekelle, 2013). Despite the wealth of literature, studies to prevent falls are
largely inconclusive about which single or multi-factorial strategies to implement (Cameron
et al., 2012; Hempel et al., 2013; Oliver, Daly, Martin, & McMurdo, 2004). It is therefore
difficult to apply an effective hospital-wide programme based on best-evidence literature.
Many interventions to improve the rate of falls begin with introduction of a new falls risk
assessment (Hempel et al., 2013), although others have advocated that they are not essential
to a prevention strategy (Oliver, 2008). This study investigated the barriers to the validated
Hendrich II Fall Risk Model (H2FRM), a replacement in 2010 to the locally developed falls
risk assessment.
In this chapter, the background to this project and an outline of the purpose of this study is
provided. A description of the locality of the study is followed by definitions of the research
participants and of a fall. Issues related to research into falls prevention are identified and the
1
motivators for these studies are highlighted. In order to focus on barriers to a falls risk
assessment tool, a model is presented to conceptualise and provide boundaries to the large
number of studies in this area. The research question and aims are then stated. The five
chapters of the study are briefly overviewed before the conclusion to the chapter.
1.2 Background
The large District Health Board (DHB) where this study was carried out has five hospitals,
seven birthing units, a mental health facility and 20 community bases accounting for a range
of primary, secondary and tertiary health services (Waikato District Health Board, 2013a).
There are over 370,000 people in the DHB area, but tertiary services are provided for a larger
population of 846,600 people (Waikato District Health Board, 2013a). According to the New
Zealand Ministry of Health Manatū Hauora (2013a), when comparisons are made between
the DHB population and the national average, the age structure is comparable, there is a
slightly higher proportion of Māori, a smaller proportion of Pacific peoples and a slightly
higher number of people who are classified as deprived.
In New Zealand, the DHB is one of the twenty DHBs funded by the National Health Board,
which in turn, fund hospitals and services in their board area. The National Health Board sets
health targets and priorities for the DHBs to implement, monitors performance and
undertakes funding accountability checks. Each DHB sets priorities based on those from the
National Health Board (Ministry of Health Manatū Hauora, 2013a).
The twenty DHBs also comply with the Health and Disability Services (Safety) Act (2001).
One the guiding principles of this Act is to provide safe services and to continuously improve
such services (Health & Disability Services, 2001). The Health, Quality and Safety
Commission (HQSC) is charged with ensuring that the Act is implemented. Since 2007,
DHBs have submitted the number of in-patients sustaining a fall as part of the reporting of
serious and sentinel events (SSE) (Health Quality and Safety Commission, 2012). Five
percent of admissions are associated with an adverse event in hospital settings, falls
comprising the largest event type (Davis, Lay-Yee, Briant, & Scott, 2003). The number of
SSE falls increased from 2007/8 until 2011/12, when there was a slight decrease, although
still comprising 47% of the reported serious and sentinel events. This decrease was despite all
of the DHBs being asked to report falls more consistently. Previous to 2011/12, some boards
reported on preventable falls only, that is, where a reason for the fall was known (Health
Quality and Safety Commission, 2012). In that year all boards were asked to include both
preventable and non-preventable falls.
2
In 2009, the DHB has instigated a patient safety initiative that prioritised reducing the
number of hospital falls by ten per cent (Waikato District Health Board, 2012b) and
introduced a wide-ranging package to meet this goal. This included an updated definition of a
fall; leadership such as the appointment of a Lead Facilitator for falls prevention and
promotion of falls champions in wards; purchase of a validated tool to assess risk and
supporting media; resourcing of equipment such as ultra low beds and non-slip socks; and
continued the use of incident forms to monitor the rate of falls. The DHB also has hospital
policy and procedure documents such as the Falls Minimisation and Management policy on
the intranet which includes actions to take following the fall of a patient and an ordering
system for additional patient equipment. However, the statistics for the falls risk assessment,
minimisation and management (inpatient) procedure showed that the ten per cent decrease in
falls was difficult to achieve.
One of the measures introduced into the DHB to improve provision of care and reduce
adverse events including falls, is the Productive Ward quality improvement programme
which was developed in England by the National Health Service Institute for Innovation and
Improvement (NHS Institute). It is an initiative which the Ministry of Health has provided
support to the DHBs to implement (Moore & Blick, 2013). The Productive Ward House of
Modules include foundation modules such as the Well Organised Ward and process modules
such as Shift Handover (NHS Institute, 2012). Each module has a structured approach for
nursing staff to follow in order to examine practices and the ward or unit environment,
identify the gaps and implement improvements. Falls prevention can be the focus of the Well
Organised Ward module. It was reported that the modules completed in wards in the DHB
have resulted in a gain of 19% direct care time (Moore & Blick, 2013).
A specific intervention purchased to achieve the hospital falls reduction goal was a validated
falls assessment tool, the H2FRM (Appendix 1). The H2FRM document has the patient risk
scoring system on one side and Risk Factor Operational Definitions on the back. The
H2FRM is completed on patient admission, whenever there has been a change in condition
(including after a fall), and at least every week. A score of five or more indicates that a
patient is likely to have a fall. A high score equates to a higher risk as the assumption is that
when more risk factors are present, the patient is more likely to fall (Hendrich, Bender, &
Nyhuis, 2003). When the assessment of risk is made, an individualised plan is written by the
patient’s nurse or midwife on the Patient Care Plan (PCP) (Waikato District Health Board,
2012a). This informs the extra interventions to minimise a fall and according to the DHB
policy document, includes liaison with the multidisciplinary team (MDT). Usual cares may
3
be on the PCP and these include factors such as a de-cluttered environment, call bell within
reach and good lighting to toilets.
The aim of the online training package associated with the H2FRM is to “establish a pro-
active approach to avoid adverse events” (Hendrich, n.d., para. 1). The H2FRM package,
Upright, includes monitoring the completion of the online training. The statistics indicated
that the completion rate was very good (Waikato District Health Board, 2013b). The training
has to be completed at least once and takes approximately two hours.
In addition to the document and the online training, the H2FRM is operationalised in the
DHB hospitals as a Stop Falls poster for the bed area, a flip chart with information and
suggested interventions and a dedicated webpage on the DHB intranet, all using the Upright
logo. In 2012, the H2FRM was the only fall risk screen used across the DHB hospitals,
convalescent care facilities and birthing units. While the literature states that the H2FRM was
developed in “an acute care tertiary facility” (Hendrich et al., 2003, p.9), so might be
expected to only be useful for acutely ill patients, rehabilitation populations were included
and obstetrics and paediatrics excluded in the developmental study (A. Hendrich, personal
communication, 7 December 2012, Appendix 1). This seems to justify the use of the H2FRM
throughout all the areas in the DHB hospitals with the possible exception of obstetrics and
paediatrics.
The staff who routinely complete the one page H2FRM scoring document, have been defined
for the purpose of this study as clinical nurses (both registered and enrolled) and clinical
(registered) midwives. Clinical nurses and midwives work across many settings and have
different duties but all are governed by the Health Practitioners Competence Assurance
(HPCA) Act (2003). The Act states that the scope of practice for the nurses is regulated by
the Nursing Council of New Zealand and midwives by the Midwifery Council of New
Zealand. The term Registered Nurse (RN) includes different qualifications, usually achieved
with three years of education, so the scope defines the individual’s nursing practice (Ministry
of Health Manatū Hauora, 2013c). The term RN is equivalent to baccalaureate or
professional nurses in the literature. Enrolled Nurses (ENs) have different qualifications and
at present, the qualification is gained after 18 months of study (Ministry of Health Manatū
Hauora, 2013b). The EN title is the equivalent to associate nurse in the literature. Both RN
and EN scopes of practice include levels of expertise. Similarly, Registered Midwives (RMs)
hold different qualifications but are registered by the Midwifery Council of New Zealand to
comply with the principle to protect the health and safety of women during childbirth.
4
In the DHB hospitals, falls are notified by the completion of incident forms by RNs, RMs or
ENs. The statistical analysis of these is reported regularly to management and board
members. In 2012, the reporting of falls had increased and the patient injury rate from falls
was down (Waikato District Health Board, 2013b). An increase in the rate of an event has
been attributed to raising awareness leading to increased reporting (Mardon, Khanna, Sorra,
Dyer, & Famolaro, 2010; Smith, 2005).
When an innovation such as a risk assessment model is introduced, the fall rate is usually
used to evaluate effectiveness so defining what is and is not a fall is required. Many
definitions of a fall are based on that of the World Health Organisation, “inadvertently
coming to rest on the ground, floor other lower level, excluding intentional change in
position to rest in furniture, wall or other objects” (World Health Organization (WHO), 2007,
p.1). Internationally, the definition of a hospital fall is not standardised. Cummings et al.
(2008) commented that unpreventable falls, such as an unanticipated physiological fall,
should be included in the data. Others state that assisted lowering to the floor needs to be
included, while others exclude certain patient groupings such as an unconscious patient or
toddler in their definition (Morse, 2009). In the DHB hospitals where this study was
undertaken, a fall is defined as “a sudden unintentional or unanticipated change in body
position with the individual landing at a lower level” (Waikato District Health Board, 2012,
para. 1). A systematic review by the Cochrane Collaboration concluded that a common
definition would aid understanding of the issues by researchers, would help consistency
across studies and comparisons between studies which aim to identify what interventions
make a difference (Cameron et al., 2012). Even so, studies on this topic often record falls as
injurious or non-injurious.
Further issues using the rate of falls to measure effectiveness of an intervention include the
low rate of falls and associated ethical issues. The rate of falls was reported as 2.7 per 1000
bed days in a large Swiss study (approximately 236,000 hospital days), but varied from 0%
5
(obstetrics, gynaecology, paediatric intensive care) to 12% (Halfona et al., 2001). A higher
rate of 20 falls per 1000 bed days was reported in sub-acute settings (Haines & Waldron,
2011). The falls per 1000 bed days are not recorded by HQSC partly because of the
inconsistencies in measurement across the DHBs.
The falls rate is dependent on the patient case-mix and is high for those areas who have
patients with cognitive impairment (Haines & Waldron, 2011). A study to measure the actual
rate of fall events found relative increases of 13% to 125% when compared to using incident
reports alone, with significant differences depending on speciality (Shorr et al., 2008). Morse
(2006) suggested that the fall injury rate is a more valid statistic for judging the efficacy of a
new intervention protocol as incident reports are almost always filed when there is a fall with
injury. However, Shorr et al. found there were 28% more falls, although this was not found
to be statistically significant, when using a fall researcher to record all fall-with-injury events.
This variability, the few number of falls per 1000 bed days and differences in event
recording, mean that consistent reporting and many patients over a lengthy time, are required
for an innovation to have sufficient power to be statistically proven as effective (DiBardino,
Cohen, & Didwania, 2012).
Another issue with research into falls prevention is that it is unethical to ask staff to cease
providing the usual falls prevention cares when a study is being undertaken (Myers &
Nikoletti, 2003). The authors of this literature critique concluded that where staff become
aware of the study and increase the effectiveness of their usual falls prevention care, the
result from the intervention on the falls rate is more difficult to interpret. These issues need to
be discussed when the research is published.
In New Zealand a survey of safety processes and measures was completed in 18 DHBs
(Raymont et al., 2012). It was found that most of the DHBs had a falls risk assessment and
protocol and use of both were mandatory, however, compliance was not assessed in many of
them. All DHBs had a system for reporting adverse events but many of the Quality and Risk
personnel surveyed did not rate the willingness of staff to engage in open disclosure as high.
Raymont et al. stated that some DHBs monitored nursing staff ratios and had systems for
maintaining staffing but few adjusted these for the patient case-mix and even fewer for the
nurse skill level. It seems unlikely that New Zealand DHBs can have safe systems for
preventing falls even when there is a high level of commitment and teamwork, if the nursing
staff numbers and level of expertise are not suited to the needs of the patients.
Patient safety involves interconnecting levels within an organisation. The DHB management
are accountable, with the board, for leadership, resourcing and ensuring that a safe
6
environment is provided to optimise the health of patients. Management write and review
policy documents which define and describe the roles and responsibilities of different
positions. The Nursing Council of New Zealand Nurses’ Code of Practice, Principle 4, states
that nurses uphold professional standards regarding safe health care (Nursing Council of New
Zealand, 2012). The Midwifery Code of Practice states that midwives will adhere to
legislation and work in partnership with women (Midwifery Council of New Zealand, 2010).
As any new protocols developed to improve the rate of patient falls require organisational
support to effect change (Fixsen, Scott, Blase, Naoom, & Wagar, 2011), both codes of
practice need to be taken into account when an innovation is proposed that affect nursing and
midwifery practice. Since RNs, ENs and RMs are obliged to comply with the relevant code,
it is incumbent on the organisation to facilitate this by direction from the strategic level,
management commitment, regular education and support, monitoring and resourcing.
The financial costs of a fall are often stated as one of the reasons an organisation seeks to
reduce falls. The no-fault Accident Compensation Corporation (ACC) in New Zealand meets
financial costs related to a fall when certain criteria are met (Accident Compensation
Corporation, 2012). In the U.S.A., two of the major health insurers have stopped funding
hospitals for never events, such as in-patient falls (Shever, Titler, Mackin, & Kueny, 2011).
Consequently there are calls for more economic evaluation of efficacious interventions
(Miake-Lye et al., 2013). While there is a very different system of health care in New
Zealand, if the same expectation that falls are preventable and therefore the costs of them
should not fall on ACC, the insurer, the decrease in funding would impact on hospital
budgets.
Further costs of a fall are to the patient and often to family members. Psychological effects
for the patient have been noted as fear, decreased self-efficacy (Fortinsky et al., 2004), loss of
confidence (Hendrich et al., 2003) and avoidance of activity (Zecevic, Salmoni, Speechley, &
Vandervoort, 2006). The fear of a fall may be more disabling than the fall itself as the
decreased activity increases health issues and affects the person’s quality of life. Increased
hospital stays and increased discharge to long-term facilities are costs for the patient, family
and hospital (Oliver et al., 2004). Effective falls reduction measures are important for many
reasons.
When using the number of falls as a measure of the efficacy of a fall risk tool, it is not only
the tool that is measured, but also the fall prevention procedure. A decrease in the fall rate
may be due to the accurate identification of at-risk patients and consequent implementation
of effective strategies, but could also be a result of factors which could not be controlled. In
2005 a staff survey of Singaporean nurses identified barriers to the implementation of a fall
7
prevention strategy (Koh, Manias, Hutchinson, Donath, & Johnston, 2008). The goal of that
study was find ways to better target implementation of a Clinical Practice Guideline (CPG) to
reduce the number of falls. This was not dissimilar to the issue in the DHB in that the number
of falls had not shown the expected decrease since the introduction of the H2FRM. The focus
of both studies was on the identification of barriers rather than effectiveness of the prevention
procedure as measured by the rate of falls. The result from a similar study could help the
DHB management decide through an increased understanding of the barriers and facilitators
for the clinical staff using the tool, whether to retain and budget for the H2FRM package as
part of the falls prevention strategy.
The barriers found in the study by Koh at al. (2008) could be organised using the Safety
Platform model developed by Dykes et al. (2011, p.440). This model has the patient at the
centre and surrounded and kept on the platform by the care providers, the protective
environment and the accurate fall risk assessment with a current and individualised care plan
(Figure 1). The care providers are registered nurses (RNs) and certified nursing assistants
(CNA). They are described by Dykes et al. as having competence (skills and knowledge),
self-efficacy (self-belief in ability to achieve goals), capability (affected by priorities and
workload) and commitment (motivation to do task). Lack of any of these elements could be
understood as personal barriers affecting falls prevention by clinical staff. The protective
environment includes assistive devices in good working order, safe facilities and
organisational support, policies and procedures. The care plan, based on correctly identified
falls risk factors, includes explicit interventions and evaluation of the effectiveness of these.
The four concepts of patient, risk assessment, care provider, and context have formed the
basis of the research questions.
Falls
Accurate fall risk assessment
&
Current & individualized care plan
injuries
Patient
Safety Platform
8
1.4 The aim and research questions of the study
The aim of this study was to identify barriers for clinical nurses and clinical midwives when
using the Hendrich II Fall Risk Model to plan preventive care for falls in the hospitals,
continuing care facilities and birthing units of a DHB.
x What are the Hendrich II Fall Risk Model barriers for clinical nurses to plan falls
prevention care?
x What are the Hendrich II Fall Risk Model barriers for clinical midwives to plan falls
prevention care?
x What personal barriers for falls prevention care do clinical nurses and clinical
midwives hold?
x From the clinical nurse or clinical midwife perspective, what patient characteristics
are barriers to falls prevention care?
x What are context barriers for planning falls prevention care?
9
Chapter Four: Results
This chapter provides the analysis of the data. This analysis includes the barrier calculation
by the authors’ of the survey instrument, frequencies and measures of central tendency, and
comparisons within and between different subgroups. Further analysis using correlation
coefficients is stated. The information is presented using text, tables and figures.
1.6 Conclusion
Research into the barriers for nurses and midwives when using a tool to plan preventive care
for falls is a worthwhile area of study and has not been addressed in the New Zealand
context. This chapter has provided the background and rationale for the study in the DHB.
Falls prevention has been described as affecting the interconnecting systems at the board,
management, care provider and patient levels. The mandate to implement strategies to reduce
falls has been stated as arising from legislature and national directives, from within the DHB
and from the patient and their family. The difficulty of monitoring the effectiveness of
interventions has been highlighted by the definition of a fall and the scarcity of fall events. A
model was introduced in order to focus this study on the risk assessment, the care provider,
the patient and the context. In this chapter, the aim and research questions have been
articulated, and an overview of the chapters to follow summarised.
10
Chapter Two – Literature Review
2.1 Introduction
The previous chapter presented the context and background to the study and provided an
overview of the issues related to effective fall prevention strategies. This chapter examines
the literature related to barriers for nurses and midwives using a falls assessment tool to plan
for preventive care. A critical literature review is important to establish the topic within the
body of material on the subject and to help define the research questions (Coughlan, Cronin,
& Ryan, 2007).
This chapter begins by setting out the strategy used to access the literature included in this
review. The review is discussed under the four areas of the safety platform model presented
on page 8. Firstly four tools and nurse clinical judgement to identify the risk of a fall are
discussed, then, the other parts of the safety platform model. They are the characteristics of
the care provider (nursing and midwifery staff), patient risk factors and the protective
environment or context elements. The elements discussed are safety culture, teamwork and
organisational development. Finally, research on the adoption of an innovation, such as a
new falls risk tool, is reviewed in order to place studies on the implementation of the H2FRM
into context.
The rationale for this research was that leaders in the DHB wanted to identify the barriers and
facilitators for clinical staff when using the H2FRM. As the risk model was introduced
throughout all of the hospitals in the DHB, a search was made into articles and information
on acute, short-stay, convalescent and long-term hospital care across all specialities, such as
medical, surgical, oncology, obstetrics and paediatrics.
The literature search was carried out using the terms hospital fall risk, fall risk assessment,
Hendrich II Fall Risk Model; organisation safety and safety culture were added once the
research questions had been identified. Searches were made of the Cumulative Index for
Nursing and Allied Health Plus, Web of Science, Scopus and Cochrane Library and relevant
policy publications. The search commenced with date restrictions between 2005 and 2013
then broadened to include seminal articles on falls prevention, and was limited to the English
language. Reference lists were also reviewed for cited papers. Articles included primary
research studies, systematic and clinical reviews. Unpublished material from the DHB,
reports, guidelines, and personal communication from experts were also considered. The
search was carried out in 2012 and again in the first six months of 2013.
11
2.3 Prevention planning
The knowledge of where and when a patient is likely to fall aids prevention planning. Most
in-patient falls happen at the bedside either transferring from bed to chair, or while
undertaking activities of daily living (Ang, Mordiffi, & Wong, 2011), before the weekend
and after eating (especially lunch) (Grenier-Sennelier, Lombart, Jeny-Loeper, Maillet-
Gouret, & Minvielle, 2002). A higher risk of falling is associated with unfamiliar
surroundings (Hsiao, 2012) so the longer the patient is in hospital, the lower the risk of a fall
(Haines et al., 2011; Schwendimann, Bühler, De Geest, & Milisen, 2006). The median
length of stay before a fall is three days, with oncology and surgical wards having the highest
median stay before the first fall (7 and 8 days) (Hitcho et al., 2004). A fall where an injury is
sustained is even more important to prevent. According to Hitcho et al. (2004), a fall
occurring when attempting to meet elimination needs increased the risk of an injury,
especially if the patient was older than 65 years. An increased risk of a fall with injury was
found to occur during evenings and overnight, and moving without assistance (Hendrich,
n.d.). While studies can identify increased risk at specific times and places, when deciding
how to use these findings in an area of practice, consideration needs to be given to
differences between the study and the local context regarding the culture, environment,
patients and staffing.
In fall prevention studies, criticisms have been made of the lack of description of the
intervention and comparison groups and the way that innovations are introduced and
sustained (Haines & Waldron, 2011; Hempel et al., 2013) as without such information the
usefulness of results are limited. The difficulty of translating evidence into clinical practice is
demonstrated by the following examples. The suggested plan for high-risk patients usually
includes placing the patient close to the nurses’ station to enable increased monitoring
(Spoelstra, Given, & Given, 2012). While that can improve vigilance, Tzeng et al. (2011)
found that the single variable, quietness, as assessed by the patient, accounted for 16% of the
variability in the fall rate. Increased noise increased the number of falls but quietness is
threatened in bed-spaces near to the nurses’ office where vigilance is enhanced. Bed alarms,
fall-risk alert signage, ultra low beds and non-slip footwear, are commonly used
interventions yet studies do not indicate that they significantly decrease the rate of falls
(Cameron et al., 2012). Another common strategy is regular toileting and timed observations
(Hempel et al., 2013), but staff reported difficulty in adhering to such falls prevention
interventions (Tucker, Bieber, Attlesey-Pries, Olson, & Dierkhising, 2012; Tzeng, 2011).
Many studies use multi-factorial interventions so it is not known whether it is the
combination or a single intervention that has an effect on the rate of falls (Cameron et al.,
12
2012). It is a challenge for management and clinical staff to institute evidence-based
interventions that are appropriate and effective.
Discussions of risk assessment tools centre on the development, risk factors and the
usefulness of the tool in clinical settings. A review of the tools used for assessing falls risk
found that most were developed using literature review and clinical information, and few
included case control (Haines, Hill, Walsh, & Osborne, 2007). Those tools which used face
and content validity to select factors of importance were criticised by Morse (2006) as being
imprecise. According to Myers and Nikoletti (2003), confounding variables such as lack of
blinding and treatment paradox meant that no tool could be recommended. It was stated that
the tools were lacking in statistical and methodological rigour, such as the process to identify
risk factors, the controlling of variables, and that there were few tests of validity on the tools
once they were developed.
Two statistics are commonly used to gauge the effectiveness of a tool. Sensitivity is a
measure of how well a risk model will predict a fall for those identified as high risk, and
specificity of how often a person with low risk did not fall (Oliver, Papaioannou, et al.,
2008). A sensitivity of over 80% and specificity over 75% have been stated as acceptable
statistics (Perell et al., 2001). The Youden Index has been developed to combine the two
statistics, specificity and sensitivity, so that one figure can be used to compare different tools
(Haines et al., 2007). A zero Youden index indicates a relationship through chance, and a one
indicates perfect accuracy.
Tools to assess the risk of falling have been developed so that nurses or midwives correctly
identify a patient to receive targeted interventions in order for that patient to not sustain a fall.
Development of the screening or assessment tools have identified risk factors, either through
prospective or retrospective analysis of fall and non-fall data, or literature review, and they
are then summed with the assumption that this will more accurately predict risk (Haines et
al., 2007). According to Oliver (2008), such assessments, where fall risk is reduced to a
single figure, prompt interventions based on the figure rather than the needs of the patient
13
and that this was the underlying reason that the number of patient falls had not decreased.
Similarly, Shever et al. (2011) suggested that interventions based on identified risk factors
would be more appropriate than a single score and generalised strategies. High risk scores are
not enough to protect a patient from a fall but whether this is because the tools are inaccurate,
the nursing and the involvement of the multi-disciplinary team is inadequate, the organisation
has an ineffective strategy or as a result of patient non-compliance, are issues analysed in this
review.
There are so many tools to assess falls risk that Perrell et al. (2001) asserted that there is no
need for organisations to develop their own. However, it is difficult for organisations
intending to use a validated tool to choose the best when there are differences between the
psychometrics of the developmental population and those of later studies using the same tool.
Three of the oldest and widely cited tools are the Morse Fall Scale (MFS), the St Thomas
Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY) and the Downton Index.
The next section discusses these, the H2FRM and nurse clinical judgment, just five of the
many measures to ascertain the risk of falling.
The psychometric studies of 36 risk screening tools were examined using the Youden Index.
The two tools that had the most studies able to be calculated were STRATIFY with 11
studies, and MFS, four studies (Haines et al., 2007). These tools had a similar Youden Index
14
to each other and to that of the pooled studies (mean = 0.2, 95% confidence interval 0.1 -
0.25). The two prospective studies on the Downton tool had the highest predictive accuracy
in the study. The authors noted that even those tools where the Youden Index indicated
greater accuracy in the prediction of falls, other factors such as the completion rate of the tool
in clinical practice and the costs associated with the tool needed to be taken into account
when making comparisons.
The Youden index for the H2FRM, just one study included for the calculation, was good
(mean = 0.5, 95% confidence interval 0.4 - 0.6). However the authors found that tools that
were developed on retrospective data over-predicted their accuracy and the most accurate
tools had few studies associated with them. These concerns need to be considered when
identifying the H2FRM as a good predictor of falls.
Studies have endorsed the use of the H2FRM. Two studies found the sensitivity to be 70%,
and 86% and specificity 61.5% and 43% respectively (Ang, Mordiffi, Wong, Devi, & Evans,
2007; Ivziku, Matarese, & Pedone, 2011). Both studies recommended the use of the H2FRM
as the statistics were similar enough to those of the developmental population. The
populations studied were 80 adult patients who had fallen in Singaporean acute medical and
surgical units (Ang et al., 2007) and patients older than 65 years in an Italian acute care
geriatric setting who had fallen over in an eight month period (Ivziku et al., 2011). A
randomised control study in an acute care hospital which used the H2FRM for identifying
patients to receive additional targeted interventions, showed a significant reduction in the
number of falls (Ang, Mordiffi, & Wong, 2011).
A different study did not recommend the H2FRM. Using the earlier version, the Hendrich
Fall Risk Model and 560 patients recruited from a German geriatric hospital, the authors
considered the sensitivity and specificity values to be good (Heinze, Dassen, Halfens, &
15
Lohemann, 2008). However, the recommendation was to use a general care plan, as the fall
prediction rate was similar and the workload less, when completing only one form.
Morse (2006) differentiated between tools used for risk and assessment. Risk scores are
quick and provide little information for further care planning. An assessment takes longer, is
thorough, involves the multi-disciplinary team (MDT) and provides a plan for the fall
prevention and protection measures that need to be taken based on physiological,
psychological and socio-cultural attributes (Morse, 2006). On this basis, the H2FRM is not
an assessment as it is rapid and the sum of the risk factors do not provide the rationale as to
why a patient is at risk of having a fall (Hendrich, 2007). However, some tools incorporate
both a risk screen and risk assessment. One such tool escalates interventions from two simple
Yes/No questions through to education and referral as the risk of a fall becomes more likely
(Äberg, Lundin-Olsson, & Rosendahl, 2009).
According to Perrell et al. (2001) a good risk tool should be developed on a similar
population to that where it is implemented, have an explicit written outline on its use, have
good inter-rater reliability, can be completed in a reasonable timeframe and the score for high
risk is clear. A study found that the H2FRM is easy to use, quick and clearly identifies
patients who have a high risk of falling (Chapman, Bachand, &Hyrkas, 2011). However,
other authors state that a good risk tool will identify modifiable risk factors (Haines et al.,
2007), but this is not the aim of the H2FRM (Hendrich et al., 2003).
Specific criticisms have been made of the implementation of the H2FRM. A retrospective
study in U.S.A. used the H2FRM on 73 matched faller and non-faller patients who had a
diagnosis of diabetes mellitus, acute heart failure or a stroke (Swartzell et al., 2013). For the
53 patients who fell and had diabetes mellitus, the H2FRM score was related positively to
high risk at predicting a fall. However, a similar relationship was not seen for the smaller
number (10) of patients with heart failure. As 44% of patients who fell did not have a
H2FRM high falls risk, the authors considered this error in risk score too high. Missing
information in the data meant that 41% of charts had to be removed from the analysis. This
non-compliance and that patients fell when the tool did not identify a risk, suggested to the
authors that the H2FRM may not be valued by staff. Other criticisms included the apparently
objective ‘up-and-go test’ which has no published reliability and validity data, inter-rater
reliability which appeared to be threatened by clinical judgements and the lack of
standardised measures for confusion and depression (Swartzell et al., 2013). Criticisms have
also been raised concerning the lack of clarity about why certain factors are part of the
H2FRM and how the weighting for the factors was calculated (Morse, 2006).
16
The implementation of the H2FRM was assessed in a pre- and post-audit study in a New
Zealand hospital by examining the completion of documentation. After a focused education
package, the post-audit showed an increase in accuracy and frequency of completing the risk
score and 80% documentation of targeted intervention plans, but no improvement in the
evaluation of the effectiveness of the care plan nor of the falls education to patient and family
(Gibbons et al., 2013). It was noted that this did not mean that such interventions were not
occurring. Another H2FRM implementation study suggested specific education in the scoring
of the H2FRM scale, as it was found that 2.3% of H2FRM scores had errors in assessment or
addition (Chapman et al., 2011).
The H2FRM has many factors in common with other falls risk assessments. Mental state, gait
or mobility, the falls history, and medications were in most of the 32 fall risk assessments
examined by Myers and Nikoletti (2003). The H2FRM includes altered cognition and
requires a sit-to-stand test, but does not include a patient’s falls history and lists specific
medications. A study on medications found that contrary to earlier findings, it was not the
number of drugs a patient was on, it was the type of medications, in particular those that
affected the central nervous system, that were correlated with a fall (Lamis, Kramer, Hale,
Zackula, & Berg, 2012). The medications listed in the H2FRM as risk factors are anti-
epileptics (anticonvulsants) and benzodiazepines (or benzodiazepine like drugs) which is
consistent with this finding.
According to Myers and Nikoletti (2003), the next most commonly shared factors are
changes in elimination, poor vision, diagnosis, older age, poor hearing and mood. The
regression analysis for the H2FRM proved a strong relationship with altered elimination and
weighted this accordingly. Vision and hearing loss, age and diagnosis are not factors included
in the H2FRM. Dizziness, vertigo and male gender are categories in the H2FRM but were not
listed in the most common categories (Myers & Nikoletti, 2003). As varied methodologies
were used to develop the 32 risk assessments, the most common factors are likely to have
face validity as they have been derived from literature or analysis of falls data. The H2FRM
includes lesser known factors and does not include some common factors, so it may lack this
type of validity.
17
who fell from the 226 patients in the sample (Myers & Nikoletti, 2003). Nurse clinical
judgement (NCJ) was required in a qualitative study involving 23 nurses as the MFS was
believed “to be incomplete” (Dykes, Benoit, Carroll, Middleton, & 2009, p. 6). This finding
suggests a complementary relationship between a tool and NCJ.
Six studies were used to compare the accuracy of falls prediction using the Youden Index
calculation with the result that NCJ was similar to the average pooled Youden Index across
all of the studies, and the results for the STRATIFY and MFS tools (mean = 0.2, 95%
confidence interval 0.1 - 0.3) (Haines et al., 2007). The authors stated that NCJ had moderate
levels of predictive accuracy across many different settings, an issue for all tools, and met
many other criteria for a tool, such as time taken to identify high risk. They suggested that
new tools could use NCJ as the standard for comparison (Haines et al., 2007). The
accessibility of NCJ would be of assistance both when developing a falls risk tool and
whenever a hospital team wish to validate a tool for their specific patient population.
A synergistic relationship could exist between the use of a tool and NCJ as neither has
perfect sensitivity and specificity. For many nurses, their clinical judgement and usual care
may be as effective as an intervention, but a tool promotes consistency of risk score between
staff, allows for reassessment of the same patient using the same parameters and facilitates
coordinated care (Joint Commission Resources, 2005). The interaction between using a tool
and NCJ is apparent when a protocol states that the falls risk is reassessed when there is a
change of the patient’s condition, and that decision is made through clinical judgement.
18
Falls risk assessment tools have been reviewed. The development and factors included in
different tools have been compared and contrasted, and nurse clinical judgement has been
compared to the use of tools with the conclusion that both are required. The aim of the risk
assessment is to implement extra cares to prevent a patient from falling. Intrinsic patient
factors, including the issue of effective patient education, socio-cultural and extrinsic risk
factors are examined in the next section.
The major intrinsic, physiological, patient risk factors are cognitive impairment, especially
delirium, dementia and depression; altered elimination and/or excretion; musculoskeletal
deficits affecting balance and strength; sensory deficits and recent injury or infection
(Hendrich et al., 2003; Hsaio, 2012; Oliver et al., 2004). Contributing and adding to these
risk factors are medications that affect the function of the neurological, digestive and urinary
systems causing fluid or electrolyte imbalances and deficits (Shever et al., 2011). Age,
gender, most diagnoses and comorbidities, and severity of illness were not found to be
significantly related to falling (Titler, Shever, Kanak, Picone, & Qin, 2011), similar to
findings by Hendrich (2003) but different to the Myers (2003) summary of domains found in
risk assessment tools. While research into ways to improve physical strength and balance
have been shown to reduce the likelihood of a fall by a patient (Cameron et al., 2012),
interventions may not be successful as, for many patients, the intrinsic factors are hard to
modify (Cumming et al., 2008).
Falls for obstetric patients are primarily physiological. Predisposing factors include weakness
and fainting, and most occur when patients are ambulating unassisted to the toilet (Chen,
Chen, & Su, 2010). A search for fall prevention tools for obstetric patients indicated that
there has been little development in this area, despite the U.S.A. National Safety Goals
including obstetric patient falls (Frank, Lane, & Hokanson, 2009). An assessment for a
postpartum mother who has had an epidural delivery has been developed by Frank et al.
(2009) and a general assessment by Heafner (2013), and both acknowledge that it is difficult
as falls are few and that, generally, the women are healthy. However, another study identified
that all women in labour or delivery are at a high risk for a fall (Johnson et al., 2011). The
aim of an obstetric tool is similar to other falls risk assessments – to identify risk factors and
plan care to reduce risk, such as supporting blood pressure, treating infection or providing
education, in order to prevent a fall from occurring. In a hospital, prevention for the newborn
infant from sustaining a fall is in procedures outlining where an infant can be fed, changed
and bathed (Waikato District Health Board, 2012a).
19
As well as physiological, intrinsic risk can be categorised as psychological and socio-
cultural. The psychological risk, fear of falling, is evidenced by behaviour of being unwilling
and slow to mobilise after a previous fall (Jellesmark, Herling, Egerod, & Beyer, 2012). A
study claimed that health professionals seek information of factors before and after the fall,
while the patient’s main concern was the emotional distress associated with a fall, and what it
means for their health (Zecevic et al., 2006). A near-miss will be considered of lesser
importance and may not be reported by staff, yet to the patient any loss of balance may
damage their self-confidence and self-efficacy leading to avoidance of mobilisation. This
study concluded that patients felt stigmatised when labelled a ‘faller’ with its connotations of
fragility and aging. A qualitative synthesis of 11 studies stated that many older people did not
want to accept the stigma of being a ‘faller’ and may rationalise it as a ‘small trip’ rather than
a non-injurious fall (McInnes, Seers, & Tutton, 2011). A fall was seen as threatening to the
person’s competence, independence and sense of control. It was suggested that in order to
motivate patients for whom this holds true, health professionals need to listen carefully to the
patient. This synthesis concluded that an injurious fall is a motivator to increase knowledge
and self-management such as allowing time to mobilise and taking care to avoid hazards.
Adult teaching strategies on falls prevention also involve motivators. Strategies acknowledge
the patient’s prior knowledge and the involvement of the patient in developing their care plan
and the setting of goals (Grol & Grimshaw, 2003). A literature review found a dissonance
between the patient’s and an institution’s primary concerns regarding quality of care. The
patient’s concerns were the care provider’s competence, knowledge and ability, but quality of
care studies concentrated on the staffing, skill mix and nurse autonomy (Currie, Harvey,
West, McKenna, & Keeney, 2005). The latter studies may not elucidate the need for care
providers to involve patients to participate in and be responsible for their falls prevention. A
qualitative interview-based study of nine patient participants found that the in-patients had
emotional obstacles around asking for help and stated that they needed to be part of the
individualised plan of care (Carroll, 2010). The participants in this study said that they were
not aware of their risk of falls and received inconsistent messages regarding how to reduce
their risk, which affected their sense of safety and autonomy. Care provider knowledge about
falls prevention needs to be consistently conveyed to patients so that they can participate in
their preventive care as patients themselves are an important part of falls prevention
management.
Education is one way that a person can gain control and confidence. A study used a script
developed by staff to provide education to reduce the risk of a fall and asked the patient to
repeat the message back after the education (Quigley et al., 2009). This teach back method
was rated as very successful by the nursing staff. It was part of the multifactorial intervention
20
which reduced the number of falls in the study when compared to the baseline data. A study
which followed 102 patients up after discharge to measure the effectiveness of the education
and pamphlets regarding prevention of falls, found that the intervention did not affect the fall
rate, nor fear-of-falling, after hospitalisation (Rucker et al., 2006). Education by health
professionals on falls prevention has to be sensitive to the patient’s psychological and
physiological blocks, as well as their knowledge and understanding.
Patient education can be effective to reduce the fall rate. A randomised control trial was
undertaken in Singapore using the H2FRM to identify patients who were at a high risk of
falling. Education was provided to these patients (and their family where appropriate) for 30
minutes on additional targeted fall prevention activities (Ang et al., 2011). There was a
significant reduction in the number of falls for the intervention group. An Australian trial
with 1206 patients used multimedia education with individual follow-up by a trained health
professional for one group (complete package), just the multimedia for the second and usual
cares for the control group (Haines et al., 2011). Overall there was no significant reduction in
the rate of falls for the three groups, but the authors found that falls were least in those
cognitively-intact patients who had the complete package.
A patient may have a high falls risk solely based on their socio-cultural status so that staff
members cannot communicate easily with them. It was found that education on falls
prevention was more difficult when there is a nurse-patient cultural barrier (Koh et al., 2008).
Patient ethnicity (classified as black or non-black) did not have an effect on falls in a survey
of nurses in the U.S.A. (Lucero et al., 2010). Male gender was noted by Hendrich et al.
(2003) to be related to impulsiveness and increased risk of falling. New Zealand is a multi-
21
cultural society with a bi-cultural heritage. All nurses are encouraged to comply with
Principle 2 of the Nurses Code of Conduct which states that the cultural needs and values of
health consumers are to be respected (Nursing Council of New Zealand, 2012). This includes
knowledge of tikanga Māori and incorporating Māori models of health into preventive care
for falls, such as involving family/whānau when planning care, and respecting the dignity and
privacy of the patient. A patient’s socio-cultural status is likely to affect their understanding
of hospital policies and practices, including those related to falls prevention.
Extrinsic patient risk factors, those factors that are not inherent in the individual, include poor
building design, being alone, inadequate staffing and environmental risks such as clutter,
poor lighting and trip hazards. Modifying the environment has been shown to be effective
(Cameron et al., 2012) and some falls risk checklists include external risk factors such as safe
footwear and call bell at hand (Schwendimann et al., 2006) which could be considered
standard nursing practice or usual cares. Extrinsic factors such as building layout, floor
surfaces and illumination may not be able to be easily ameliorated, whereas a patient’s
improper use or poorly maintained mobilisation device may be easily improved.
This section has reviewed patient intrinsic and extrinsic risk factors for falls. It has been
reported that safety involves minimising physiological risk factors, patient education to
reduce risk of falling and increase self-confidence, and reducing trip hazards in the
environment. Nurses, the health professionals who spend most time with a patient, hold a key
position to ensure a patient’s safety. The next section discusses competence and self-efficacy
from the Safety platform model by Dykes et al. (2011) (see Figure 1, page 8), focussing on
the knowledge, attitudes and motivation of the care provider.
Nurses and midwives are described by the title care provider. The care provider may be the
person who feels most responsible for a patient who has a fall. The aim of this thesis is to
identify the barriers for clinical care providers when using the H2FRM to plan preventive
care for falls. As a comprehensive discussion on attitudes, beliefs, values, readiness to work
with others, and use of evidence in practice, are outside of what is able to be discussed for
this study, only research which is primarily about nurses’ knowledge of, and attitude to, falls
prevention and implementation of strategies is examined. This has been summarised as care
provider competence, capability and commitment (Äberg et al., 2009).
2.6.1 Knowledge
There are few studies that specifically assess fall prevention knowledge. One pre and posttest
study involved 40 nurses (El Enein, El Ghany, & Zaghloul, 2012). Knowledge of
22
environmental factors affecting falls did not significantly improve following the education.
However, after training, all nurses, except those who held a professional qualification,
improved their knowledge on health education and the impact on fall prevention. The authors
suggested that the nursing school curricula should be updated to reflect current knowledge. A
survey of 1830 nurses to identify barriers to implementing fall prevention Clinical Practice
Guidelines found that the highest barrier (83%) for care providers was knowledge and
motivation (Koh et al., 2008). The follow-up study after strategies were instituted to reduce
this barrier, concluded that the nurses’ completion of the Clinical Practice Guideline form
was a good marker of success of the new protocol, even when there had been no decrease in
the fall rate (Koh et al., 2009). This study reported that 99% of the nurses completed the
form compared to 50% before education.
The five root causes of falls, as identified by the Joint Commission, were used to analyse
interviews in a study which concluded that nurses lacked knowledge or overestimated
knowledge of evidence-based specific fall prevention measures (Tzeng & Yin, 2008). By far
the most frequent barrier that staff identified was the unsafe environment. The authors stated
that this was expected as people blame external events rather than their own capability - the
focus of the four other root causes. As none of the nine staff nurses interviewed in the study
mentioned the root cause staff training and orientation, it indicated to the authors that there
was lack of fall prevention knowledge.
Other studies have highlighted the importance of knowledge. The recommendation from a
small qualitative study was that site and patient specific knowledge are needed for nurses in
regard to falls prevention and use of restraint (Williams et al., 2011). A study in the United
States over three years demonstrated a decrease in the total number of falls (Johnson et al.,
2011). In this study, the Helping Hands programme, falls champions were utilised to educate
and engage nurses to undertake best practice for falls prevention. The programme began with
staff education and over the three years introduced six more falls prevention measures, such
as shift-change handovers and post-fall huddles to discuss patient safety and ensure
continuity of care. These studies showed that context-specific fall prevention knowledge is an
important precursor to good fall prevention practice.
An integrative review identified that nurses use colleagues for much of their tacit knowledge
for assessment and clinical care (Spenceley, O'Leary, Chizawsky, Ross, & Estabrooks,
2008). The lack of a knowledge culture, of which tacit knowledge is a part, is one of the
reasons that some authors give to explain the lack of effectiveness of a falls prevention
strategy (Dykes et al., 2010; Schwendimann et al., 2006; Semin-Goosens, van der Helm, &
Bossuyt, 2003; van Harten-Krouwel, Schuurmans, Emmelot-Vonk, & Pel-Littel, 2011). On
the other hand, studies which examine nurses’ judgement of risk suggest that knowledge of
23
falls prevention is good as there is little difference between scores of high risk by a validated
model or by clinical judgement (Cina-Tschumi, Schubert, Kressig, De Geest, &
Schwendimann, 2009; Myers & Nikoletti, 2003; Vassallo et al., 2004). The dynamics of the
nursing team and use of tacit knowledge may be as important as the attributes of the
individual care provider.
The knowledge culture was explored in a Dutch study. The five-month study which surveyed
nurses and made 82 observations on nurse interventions for 15 post-fall patients, found that
nurses survey responses indicated knowledge as a minor barrier (8%) to the application of a
prevention protocol (van Harten-Krouwel et al., 2011). Consequently, it seemed inconsistent
that the prescribed protocol for the delirium/confusion intervention was seldom observed to
be used (24% of the time). This “forgetfulness” was attributed to the ward customs. The
authors concluded that nurses needed to be educated about prevention of falls, especially
those that require more time and knowledge such as using a protocol or contacting the
family.
While an integrative review recommended staff training in falls prevention (Spoelstra et al.,
2012), the Cochrane Database of Systematic Reviews did not find that studies involving staff
training decreased the number of falls (Cameron et al., 2012). Improved knowledge of
aspects of fall prevention, particularly health promotion and safety culture relevant to the
speciality, rather than of environmental hazards may be required. This knowledge could be
gained through orientation or on-going site-specific training and methods other than the rate
of falls may be more sensitive to the effectiveness of improved knowledge.
24
statistics (O'Connell & Myers, 2001). Both of these studies detailed the design and
methodology enabling later researchers to learn from their methods such as leadership and
the preparedness for staff to make changes.
A study undertaken in Australia to improve nurses’ attitudes towards falls prevention through
professional development on decision-making, found no change in the psychological and
social factors that were measured (Dempsey, 2009). There was however, an enhanced use of
fall prevention activities although the reduction in falls was not significant. This mixed
methods study identified a number of confounding factors which impacted on the attitudes of
the nurses. An attitude is an individual trait and while the goal is for nurses to individually
have an attitude towards safety including falls prevention (Joint Commission Resources,
2005), in the Dempsey (2009) study groups of nurses were worked with. They found that
team relationships affect individual nurse attitudes. A recommendation from the study was
that decision-making on falls prevention improvements should be made within teams and at
the local level.
Avoidance of stress can be a motivator to prevent further falls. A qualitative study with 15
participants on acute care nurses’ experiences of a fall, found that even when it was an
omission or action of the patient that contributed to the fall, the aftermath was stressful and
time consuming for all parties involved (Rush et al., 2008). In a qualitative study of 23 health
care professionals, fear of legal action seemed to result in compliance to safety guidelines
(Stenberg & Wann-Hanson, 2011). Extra fall prevention interventions for all patients were
implemented, including those who were identified as low-risk. Such actions resulted in
nurses feeling confident that patients under their care were safe, while at the same time being
criticised for being wasteful of resources. ACC, the New Zealand insurer for accidents, has a
no-fault mandate. In New Zealand approximately 5% of hospital admissions are associated
with a preventable adverse event which is similar to data from U.S.A. and Australia (Davis et
al., 2003). Avoidance of the potential stress arising from a malpractice suit does not seem to
have an effect on the adverse event data. It appears that it is not a simple relationship
between stress avoidance and improved compliance. Nevertheless, Haines and Waldron
(2011) in a review of literature on prevention of patient falls, summarised that overall, as
different areas have different rates of falls, nurses will have different motivations and stress
avoidance behaviours which necessitates a local approach to preventive practices.
Attitudes and motivation were found to be very high barriers for staff to implement a fall
prevention guideline (Koh et al., 2008). While a patient assessment for each shift and taking
of blood glucose and were missed the least (17% and 26% respectively), other cares that
pertain to risk of falling were found to be omitted regularly in a study in three Michigan
25
hospitals (Kalisch et al., 2009). Ambulation, toileting within five minutes and response to call
lights were missed 80%, 70% and 66% of the time respectively. There was similarity
between the events in the three hospitals and reasons given for the omission of nursing care.
They included lack of time, inadequate staffing, and that associate-level trained nurses
reported more missed care than those who held diploma or degree certifications. The authors
suggested the latter finding may be a result of associate nurses providing nursing care as a
series of tasks, whereas falls prevention involves the integration of regular assessment,
environmental safety and mitigation of patient risk factors in a timely way rather than by
scheduled task.
In summary, nurses may have an attitude that many falls cannot be prevented and this may be
a reflection of the culture of the workplace. There is evidence that behaviours that could help
prevent falls are being omitted. The motivation of a care provider to prevent a patient from
falling results in the completion of prevention tasks and is affected by the nurse’s self-
efficacy, need to keep the patient safe and teamwork. The protective environment is
discussed next.
2.7 Context
26
provider and organisation. These elements are part of the Safety platform model (see Figure
1, page 8), as the safe patient, care provider competence and self-efficacy and the protective
physical environment (Dykes et al., 2011).
Äberg et al. (2009) identified the three dimensions of a safety culture as organisation, staff
and leadership. Organisation components include staff who have safety as a priority and who
automatically apply safety-promoting behaviours, staff and leadership who learn from events,
updated policies and guidelines and lastly, a well-resourced environment. Staff and
leadership have five components which are commitment, competence, capability, support and
control. The authors identify organisational learning as the use of evidence-based practices,
but where local knowledge reframes these to aid adoption and facilitate an innovation into
becoming routine practice.
Magnet® hospitals affirm the local knowledge of nurses. Magnet hospitals were a concept
developed in the U.S.A. in the 1980s which has subsequently been adopted by hospitals
across the U.S.A. and in other countries. The aim of the first Magnet association was to fully
utilise the expertise and knowledge of nurses. The result is that high quality care is provided,
that there are better outcomes for all staff, not just nursing staff, and that there is stability of
staffing (Parsons & Cornett, 2011). Magnet status is achieved, in part, by employing nurses
with degree or baccalaureate-level and specialist qualifications and having safety patient-
focused initiatives for events such as falls, hand-washing, pressure ulcers, restraint and
infections (Swanson & Tidwell, 2011). A comparison of fall rate and staffing in Magnet
versus non-Magnet hospitals found that there were fewer falls across all specialities in the
Magnet hospitals (Lake, Shang, Klaus, & Dunton, 2010). Analysis of the difference found
that it was not the higher registered nurse staffing ratio that accounted for the 5% lower fall
rate, but that it was an organisational effect. This finding suggests that management have to
recognise and value nurses and nursing, have quality the goal throughout the organisation
and improve the safety culture to reduce the fall rate.
The relationship between patient safety and adverse events, including patient falls, was the
focus of an exploratory study in 179 hospitals. The hospitals in the study were a sample of
those registered with the American Hospital Association, and were not named as holding
Magnet status. The patient safety data was correlated with the perceptions of nurses and
allied health professional staff to the safety culture in the hospital that respondents worked in
(Mardon et al., 2010). The statistics from the Hospital Survey of Patient Safety found a
significant inverse relationship. The largest negative regression coefficients were for
handoffs and transitions, teamwork across units and frequency of events reported, that is,
better handovers of care were correlated with lower numbers of falls. A high reporting of
adverse events was seen as an indicator of a safety culture. Staff members were able to report
27
and the organisation and care providers learn from the errors. This improved the coping
resilience of individuals and of the group. Mardon et al. concluded that after controlling for
size, ownership and teaching status, hospitals with fewer patient adverse events had a higher
safety culture.
Patient handovers aim to share information about the patients and the ward environment and
improve the continuity of care for patients. Ways to improve information between team
members about a patient’s falls risk include the individualised falls care plan discussed at
shift handover (Mardon et al., 2010; Spoelstra et al., 2012), signage at the patient’s bedside
so that all stakeholders are aware of the general care the patient requires (Johnson et al.,
2011) and clear delegation to members of the team (Dykes et al, 2009). Wilson (2013) found
that improved safety organising behaviours were associated with nurses’ holding a positive
attitude to members of their team and to actions promoting safety made by the manager. A
computer-generated care plan for identified risk factors significantly reduced the rate of falls
partly because of the improved communication from one caregiver to the next and throughout
the interdisciplinary communication network (Browne et al., 2004).
The results from an Australian study were that even the most experienced individual nurse
did not have the capability or capacity to enact cultural change and at the ward and hospital
level many factors could influence context (Schultz & Kitson, 2010). This finding is in
accordance with that of Wilson (2013) who found that safety organising behaviours were best
interpreted at the local level. The Schultz and Kitson (2010) study utilised the Alberta
Context Tool (ACT), and the context element of the Promoting Action on Research
Implementation in Health Services (PARIHS) framework in an acute care hospital.
Intervention wards undertook professional development on improving the care of older
persons, including the topic of patient falls. Responses to the ACT questionnaires were used
to identify wards and areas where the context was amenable for development and change.
28
The study concluded that intervention should be targeted at ward level as there is significant
variation in context factors between wards. While that study focused on unit level leadership,
another which interviewed 15 Chief Nursing Officers of Magnet hospitals identified that it
was at executive level where quality improvement changes were sustained (Parsons &
Cornett, 2011). This finding may reflect the participants positional power, and not preclude
that ward level managers also have a role in introducing and maintaining new strategies.
Teams have been described as the natural organisation of health care providers. The
sustainability of a teamwork project was followed up one-year after the completion of a
quality improvement study. The Collaborative Breakthrough Series on the prevention of falls,
was conducted to determine the progress of the 37 teams who had been involved, and to
ascertain if the programme had been maintained after the researchers left (Neily, Howard,
Quigley, & Mills, 2005). Over 90% of teams said that they had continued to collect data and
maintain the gains made after the initial year-long quality improvement cycle. Four of the
sites that were used for falls data collection had a non-significant increase in fall rate and all
had a zero major injury rate. Leadership support, as part of the supportive staffing initiative,
was cited as the success factor behind the continued implementation of the project. Leaders
of successful teams nurture relationships, support creativity and accomplishments while
controlling the group tasks and goals (Holleman, Poot, Mintjes-de Groot, & van Achterberg,
2009), particularly when innovations are introduced.
29
were developed and used found that they were best used as an interactive tool that established
minimum standards and where strategies for staff to be able to receive feedback were
embedded in the guideline (Bahtsevani, Willman, Stoltz, & Östman, 2010). The authors
found that participant narratives included the PARIHS context element. Leadership and a
reiterative process to review and refine the CPG were used until the practice was routine and
the CPG not used. The participants stated that too many CPGs decreased critical thinking and
lead to stagnation in professional development. This study highlighted that leaders need to
allow time from the introduction of an innovation for it to become routine practice, for new
staff to be well-orientated to the culture and practices of the area and for challenges to
initiatives to be expected.
The PARIHS framework has been used to map the interconnecting systems at the policy,
ward and individual level, and the feedback systems. Such mapping can highlight gaps in
processes when undertaking organisational development. An example is the inclusion of the
interdisciplinary team to ensure that physiological, psychological and socio-cultural risk
factors are identified (Joint Commission Resources, 2005). Some fall risk studies state how
the medical team, physiotherapists and occupational therapists will be involved
(Schwendimann et al., 2006; van Harten-Krouwel et al., 2011), and a nurse referral to other
health professionals is seen as a necessary action to prevent falls (Johnson et al., 2011), but
many do not state how different disciplines are included (Pentland et al., 2011).
Organisational development involves people at all levels and in all positions.
A quality initiative, the Productive Ward quality improvement programme, which is based on
an organisational change model, has been evaluated. The programme engages ward staff to
30
make local changes in order to increase the time they have for direct care, increase patient
satisfaction and improve efficiency. A mixed methods study of its uptake in the U.K.
indicated that it has a high positive impact on teamwork (86%) and staff experience (82%)
(Robert, Morrow, Maben, Griffiths, & Callard, 2011). This study attributes the positive
adoption of the programme to locally adapted modules, resourcing, formal vertical
dissemination channels and that it is owned and controlled largely by ward staff. A report on
the Well Organised Ward, a module which can have a focus on reducing patient falls, states
that mortality from falls decreased after completion of the programme (Dean, 2010).
Structural empowerment is the removal of dependence and powerlessness and has similarities
to the Productive Ward model. A study to investigate this and its relationship to patient and
nurse satisfaction in Canada surveyed 679 medical and surgical nurses (Purdy, Laschinger,
Kerr, & Olivera, 2010). The study established that group processes are mediated by team-
level structural empowerment and that this empowerment significantly reduces the number of
patient falls. The rate of patient falls was inversely proportional to the percentage of
registered nurses in the staffing roster. There was also a non-significant but clear positive
relationship between structural empowerment and nurse attributes of psychological
empowerment, quality of care and job satisfaction. Manager led structural empowerment
included access to information for care, time, equipment, support, guidance and opportunities
for professional growth for the nurses. The individual care provider and team characteristics
are closely linked in improvement initiatives.
Time for the adopter to interact with, learn about and become competent in its use, is
essential for an innovation to become effective. A lengthy time will allow for embedding of
practices after the initial launch and promotion period of a new preventive regime (Haines et
al., 2011). An organisation can help with the development of competency through good
resourcing for training, coaching and performance assessment (Fixsen et al., 2011). During
31
this time indicators are required to monitor progress towards the goals and provide feedback
(Grol & Grimshaw, 2003). The measurement of progress for a falls risk tool can take the
form of identifying barriers and facilitators to the uptake of the innovation or measuring the
effectiveness of the implementation through improved preventive care.
Economic evaluations are difficult for the reasons given earlier such as deciding what a fall is
and the accuracy of the monitoring process. An analysis of data from 9,000 patients found
that when there was a lower than median level of RN staffing, there were increased costs to
the hospital due to the increase in the number of falls (Titler et al., 2005). The analysis of the
costs of implementing an effective falls reduction multi-media programme in cognitively
intact patients, found that there had to be more than 4% of such patients having a fall to pay
for the additional cost of the programme (Haines et al., 2013). A study did not recommend
the use of the Downton Index risk assessment tool after an economic analysis (Meyer et al.,
2009). The Cochrane Collaboration of Systematic Reviews stated that no economic-based
conclusions could be drawn from the trials reviewed (Cameron et al., 2012). While economic
studies are important for resourcing decisions, an economic analysis which includes the non-
financial costs of the decreased quality of life of the patient, their families, and the costs
related to the effects that a fall has on health care staff, is required for the best analysis.
Interaction between the adopter (clinician) and innovation affect the sustainability of change.
Adopter traits of motivation, intelligence, tolerance of ambiguity and compatibility of values
have been found to be associated with the successful adoption of innovations (Wadensten &
Carlsson, 2007). Greenhalgh, Robert, MacFarlane, Bate and Kyriakkidou (2004) described
features of an innovation which help make change possible. The features are that the
innovation will have a relative advantage over the previous model, will be compatible with
the adopters’ and organisations values, there will be organisational systems to implement the
innovation which have simple ‘hard core’ and fuzzy periphery processes, will be feasible and
easy-to-use, will have support through training or a helpdesk, the benefits can be seen, and it
will be able to be adapted by the adopters. From the earlier discussion, adoption of the
32
H2FRM should be acceptable as most organisations have a falls risk assessment as part of the
prevention of falls procedure, it is simple, easy-to-use and supported by the Upright training.
However, as it is a validated and trademarked model, it is not able to be altered or reframed
without further revalidation and the permission of the author.
As well as the debate on the appropriate sensitivity and specificity of risk assessment tools
and whether a single figure indicating risk of falling gives sufficient information, there are
three areas relating to the adoption of a tool in the context of a health organisation which are
included in this review. They are the goal for falls reduction, specifically to zero falls, and
two examples of operationalization of the tool, the “hard core, irreducible elements of the
innovation” (Greenhalgh et al., 2004, p. 597). The examples are whether any tool is accurate
enough to use across all specialities and whether it is accurate over the length of a patient’s
admission.
A goal of a particular falls rate, taking into account the patient mix and resources available
may be better to use as a measure of effectiveness upon the introduction of a new falls
assessment tool (Äberg et al., 2009; Haines & Waldron, 2011). The Joint Commission (2005)
33
state that design of safe systems is a priority rather than error-free individual performance.
Therefore, indicators for a falls prevention programmes could be safe systems rather than a
zero fall rate as demonstration of such systems will account for variations in nurse workload,
a patient’s risk of falling and the context of the provision of care. One example of a safe
system is evidence that the risk of falling has been assessed. Such documented completion
has been used in some studies as a measure of the effectiveness of an education intervention
aimed at nurses (Gibbons, Esselink, & McHugh, 2013; Koh, Hafizah, Lee, & Muthu, 2009).
Haines and Waldron (2011) assert that it is useful and advisable to use a falls risk tool that
has been previously validated as effective especially if there is not the capability or capacity
in the organisation to validate a locally designed one. While many hospitals use a validated
tool, in one survey 15% of the hospitals were using a locally developed risk assessment
(Shever et al., 2011) and many research articles detail the development of their local risk
assessment (Äberg et al., 2009; Browne, Covington, & Davila, 2004; Giles et al., 2006;
Johnson, Veneziano, Green, Howarth, & Malast, 2011; Lohse et al., 2012). A model
validated for the specific patient population is the preferred tool but it seems that a tool that is
34
credible to clinical staff is important when an innovation is introduced into any area,
especially if education is not readily available.
The timeliness of the reassessment protocol was challenged in a study to understand the
efficacy of pre-operative patient education. The control group (n = 172) had seven fallers,
only two of which were classified as high risk by the H2FRM (Clarke, Timm, Goldburg, &
Hattrup, 2012). The ward policy was that 12 hourly assessments of risk were made (H2FRM
document states that reassessments occur at a change of patient’s condition or location, after
fall, or every 7 days) but the study did not report on adherence to the protocol. It was
reported by the researchers that A. Hendrich stated the poor ability of the tool to identify a
patient who subsequently fell was probably due to the lengthy time between assessments. A
study in rehabilitation areas found that the risk of falling can change more rapidly than a risk
tool can be completed (Äberg et al., 2009). Even when a tool can be completed in a
reasonable timeframe, in clinical practice it is one of many tasks that need to be prioritised
for the quality provision of care.
2.9 Conclusion
In this chapter the H2FRM and screening tools were discussed. Research shows that the
H2FRM is clear, simple and quick to complete, but there are issues with its use in
populations that are not medical and surgical, its face validity and the degree to which it adds
value to professional clinical judgement. It was suggested that a falls risk assessment tool is
best used in conjunction with clinical judgement. Education on scoring and on the factors
using local scenarios could improve the inter-rater reliability. During the time innovations
take to become routine, it is expected that there will be reframing to meet the site-specific
35
needs but as the H2FRM is a trademarked and validated model, changes cannot be made to
the H2FRM without permission and revalidation.
Many intrinsic patient attributes that raise falls risk are difficult to ameliorate. Patient
education was identified as an issue and one that seems to feature in studies of missed care.
Increased knowledge and giving consistent messages were discussed as challenging, and a
patient who has altered cognition has particular difficulties with understanding fall
prevention strategies. Identification and treatment of the causes of patient risk factors
requires involvement of the multidisciplinary team.
Usual cares need to include general risk reduction measures and clear well-communicated
individualised prevention strategies. The involvement of the patient, family and multi-
disciplinary team in such strategies was recommended. Context-specific falls prevention
knowledge is seen as a precursor to good falls prevention practices. This may change the
attitudes of nurses and empower them towards better fall prevention although the dynamics
of the team may be more important than changing the individual nurse attitude. Leadership is
vital to facilitate change but transformational leadership is not easy to apply and sustain in
complex hospital settings.
The aim for the safety culture is to have safe systems to reduce the fall rate rather than the
goal of zero patient falls. It was stated that zero falls was not possible over the longer term
but may be possible over shorter terms because of the variability in the fall rate. Falls with
injuries may be a better indicator than falls without injury, but other methods of judging the
effect of a new falls risk tool are required such as documentation of assessment, planning and
evaluation of care. Care provider competence, capability and commitment affect the falls rate
but are difficult to quantify compared to using the fall rate as a measure of the success of an
intervention.
This review highlighted the need for further research to be conducted to examine the barriers
to the implementation of a falls risk tool. A study could find out whether the barriers are
related to the tool itself, personal attitudes of the nurses and midwives towards falls
prevention, or attributes of the patients, or whether the barriers arise from factors in the
environment or organisational practices. The next chapter will present an overview of the
non-experimental descriptive study design which was used for this study.
36
Chapter Three – Research Design
3.1 Introduction
The previous chapter reviewed the literature related to the prevention of in-patient falls under
the headings of prevention planning, risk assessment tools and clinical judgement of falls
risk, care provider characteristics, patient risk factors and context. The review highlighted
that there is a need for further research to identify barriers for nurses and midwives when
using the H2FRM to plan preventive care for falls. In particular, the research questions were:
x What are the Hendrich II Fall Risk Model barriers for clinical nurses to plan falls
prevention care?
x What are the Hendrich II Fall Risk Model barriers for clinical midwives to plan falls
prevention care?
x What personal barriers for falls prevention care do clinical nurses and clinical
midwives hold?
x From the clinical nurse or clinical midwife perspective, what patient characteristics
are barriers to falls prevention care?
x What are context barriers for planning falls prevention care?
This chapter explains the research design and methodology of a survey to research the
questions. In this chapter the measures, participants and procedures and the ethical
considerations are presented.
Research is the basis for evidence-based practice. Good research design will facilitate
worthwhile results to be translated into clinical care. The approach and philosophy that is
used to research a question directs the sampling techniques and how data is collected,
analysed and presented (Creswell, 2014). The research question for both qualitative and
quantitative designs can begin with “What are …?” The researcher critically analyses the
strengths and weaknesses of possible methodologies and their underpinning assumptions to
decide on the best approach to answer the research question (Creswell, 2014). A quantitative
design was chosen for this research as a numeric description of the attitudes and opinions of
the staff members who complete the H2FRM was sought.
The assumptions for a quantitative design are that there is a systematic empirical observation
of a phenomena and that the results are statistically analysed to prove or disprove a
hypothesis, and support or refute a theory (LoBiondo-Wood & Haber, 2006). Empirical
observation states that when observing behaviour the researcher will not influence what is
37
being recorded (Creswell, 2014). The statistical tests which are used for quantitative designs
have assumptions which determine which test can be used (Armitage, Berry, & Matthews,
2002).
The quantitative non-experimental descriptive survey research design can be used to gather
information about a population to allow comparisons of a variable in groups of people or
when little is known in order to guide more research (Kermode & Roberts, 2006). The design
of this survey is non-experimental as there is no group who are receiving an intervention
(Kermode & Roberts, 2006). A survey is descriptive as it seeks to describe characteristics
such as attitude, belief, behaviour or knowledge (Minichiello, Sullivan, Greenwood, &
Axford, 2004). The above characteristics of a survey are compatible with the research focus
to describe the attitudes and knowledge barriers that care providers hold towards the H2FRM
when planning falls prevention care.
A high response rate is desirable in order to reduce sampling error and bias, and increase the
representativeness and therefore generalisability of results to the population from which the
sample is drawn (Browne, 2005). Browne stated that voluntary participation can be a source
of bias as the reason for respondents purposely completing the questionnaire is not known. A
sample, even if large, drawn in a way so that many participants could not access the survey
may mean that the results are not representative (Armitage et al., 2002).
Non-sampling error is a function of the survey instrument not measuring the attitude, belief,
behaviour or knowledge it is designed for and is more difficult to detect than sampling error
(Browne, 2005). One type of non-sampling error is when respondents give answers that they
think are acceptable to the researcher. This may be minimised by assurances of
confidentiality, so that respondents answer truthfully rather than giving socially desirable
answers. Poor validity and reliability of the survey instrument are also described as non-
sampling errors.
This research used the same survey instrument that was used in a study in Singapore
hospitals on falls prevention (Koh et al., 2008). The aim of that study was to identify nurses’
perceived barriers to the implementation of Fall Prevention Clinical Practice Guidelines. A
similar design was used as the conclusions that were reached in that study were useful to the
management of the Singapore hospitals. Despite cultural differences, the predisposing factors
to patients having a fall are similar, so basing a study on another with a similar aim, can
establish or challenge the findings of that study and is less expensive as the design and
methodology have already been determined (Moule & Goodman, 2009). The differences
between the Koh et al. study and this one were the questionnaire, staff members working in
38
acute medical, surgical and geriatric tertiary hospitals were surveyed in Singapore but most
staff in the DHB were included in this survey, and the method of delivering and returning
questionnaires was different.
3.3 Measures
The study questions direct the research design. The questionnaire consisted of the Barriers
and Facilitators Assessment Instrument and demographic questions. In this section the
instrument that was used and the psychometric tests on it are discussed, then the demographic
questions are considered.
Studies by Wensing and Grol (as cited in Harmsen et al., 2005) have used the instrument to
identify barriers to the adoption of changes to clinical practice such as implementation of a
new cerebrovascular accident prevention protocol. The instrument is said to be useful as it
has been drawn from a range of different theoretical perspectives – organisational theory,
39
marketing, social learning, adult learning, behavioural and cognitive theories or models
(Harmsen et al., 2005). The authors gave permission for its use and requested feedback on
the results (Appendix 4).
The Barriers and Facilitators Assessment Instrument (Peters et al., 2002) begins with
questions that are easier to answer; has questions that are clearly stated and unambiguous;
has questions that do not use leading or emotive words; does not have questions that lead
respondents to a particular answer, groups questions logically so that they flow well, and is in
an easy-to-read format. According to Minichiello et al. (2004), these are tenets of good
questionnaire design. The questions on the instrument were answered on a 5-point Likert
scale. This scale enabled respondents who held a neutral position to indicate the mid-point,
while indicating direction and intensity of the respondent’s choice on either side of it.
As per the authors’ recommendations, changes were made to the Peters et al. (2002)
instrument before it was distributed. These changes were to replace “innovation” with the
“Hendrich II Fall Risk Model” and questions that were not relevant were removed such as
those on opening hours and reimbursement. The questionnaire was discussed with three
management staff of the DHB with the result that changes were made. Firstly, two open, or
free form, questions were added. The aim of this was to allow respondents to elaborate, if
they wished to, on the questions “I think parts of the H2FRM are incorrect” (question 8) and
“the H2FRM does not fit into my way of preventing patient falls” (question 16). Secondly,
the question that other health professionals “do not cooperate to apply the H2FRM” (question
10) was divided into three questions: working with doctors, with physiotherapists and with
occupational therapists. Thirdly, minor wording changes were made to questions to improve
interpretation such as, “weigh the wishes of the patient” was changed to “consider the wishes
of the patient”. The authors (Peters et al., 2002) suggested that questions could be added to
elicit more information so two questions were inserted based on the literature. One addition
related to the perception respondents have of the hospital safety culture and the other to the
efficacy of delegation. These changes meant that the original 27 item instrument was
modified into 23 items with one question in three parts and three free form questions.
The questions in the Barriers and Facilitators Assessment Instrument were divided into four
categories and ordered so that all of the questions for one category were not asked in
sequence. The categories were: Innovation (H2FRM), Care provider, Patient and Context.
The characteristics in the H2FRM category concerned the usefulness of the model – time
taken, flexibility in its use, whether the training was adequate. The questions for the Care
provider characteristics were in relation to the knowledge, skills and attitudes towards falls
prevention and the H2FRM. The questions for the three Patient characteristics were
40
pertaining to intrinsic factors. The characteristics of the Context category were measured by
questions about the extrinsic factors in the workplace and attitudes of the respondent to
working with members of the multi-disciplinary team. The survey with the modified
instrument is attached as Appendix 3.
The authors of the instrument placed the group norms, socialisation characteristic in the Care
provider category. The survey questionnaire, when used by Koh et al. (2008) to investigate
the perceived barriers to clinical practice guidelines, placed this characteristic under Context.
As the definition of context includes social characteristics (Peters et al., 2002), the decision
was made to put this characteristic in the same category as the Koh et al. (2008) study in
order to assist with comparison of the results between the two studies.
Psychometric and validation testing for the instrument were carried out in the Netherlands
and the authors stated that it was useful in twelve studies. It has a Cronbach alpha coefficient
between 0.63, care provider characteristics, to 0.68 for patient characteristics. This
coefficient is a measure of internal consistency and for new scales, a coefficient of 0.70 is
acceptable (DeVellis, 2003). However the statistic is less sensitive when there are less than
ten items (Pallant, 2007), which is the case for the four subscales of the instrument.
41
3.3.3 Demographic questions
Demographic questions need to be based on literature and avoid asking for extraneous
information (Browne, 2005). The rationale for the three demographic question are as follows.
A question on the practice setting was included as a risk screening tool has been found to be
of most use when the population it is used on is similar to that of the developmental study
population (Oliver et al., 2004; Spoelstra et al., 2012). The qualification of the respondent
was asked as grey data from the DHB indicated that midwives were not undertaking the
online education on the H2FRM at the same rate as nurses. The third demographic question
asked about the years of experience as new graduate nurses were less able to assess risk
without a tool than other nurses (Myers & Nikoletti, 2003) and experienced nurses are more
likely to have used a variety of frameworks to score fall risk and have a comparison. The
diploma or degree status was not asked, nor was gender or age, as there was no literature to
suggest completing the falls risk assessment related to those variables. The Koh et al. (2008)
study using the same instrument reported no relationships with the qualification, gender and
age demographic variables.
3.3.4 Piloting
As there is no contact when using a questionnaire survey between the researcher and the
participants, misunderstandings regarding questions may arise. Pilot studies provide a check
that the meaning of each question is unambiguous and that the time taken is appropriate (De
Vaus, 2004). De Vaus also stated that piloting helps to ensure that the layout of the
questionnaire is clear and inviting, that respondents are able to make accurate responses and
that support, if required, is available. Piloting gives an opportunity to check systems of
distribution, collection, entering, coding and cleaning data (Boynton, 2004). Lastly, it is a
check that the survey has a clear focus and purpose which relate to the research questions and
that the respondent is engaged and motivated to complete it (Rattray & Jones, 2007). The
results of psychometric tests on a pilot study show that questions which have a high return
rate demonstrate that respondents understand the question, but where less than 20% have
endorsed a response then that the item may be unreliable (Rattray & Jones, 2007).
The nurses who were invited to pilot the questionnaire included one new graduate registered
nurse, one senior registered nurse, one registered nurse who identified as Māori, two
registered nurses who spoke English as their second language, two registered midwives, one
enrolled nurse and the Falls Committee members present at the August 2012 meeting.
Despite the invitation to midwives to pilot the study, no responses were received. The
feedback from this consultation resulted in minor wording changes so that the questions were
easier to understand in the New Zealand context while retaining the meaning.
42
3.4 Participants
The project was carried out in the hospital facilities of a DHB. These included a large
tertiary teaching hospital, four rural hospitals, two rural continuing care facilities and seven
birthing units – a total of 770 beds. The specialities, wards, units and areas, included acute
medical and surgical wards, emergency departments, oncology, paediatrics, rehabilitation and
convalescent care wards, obstetrics units and gynaecology wards, a psychiatric unit, day-stay
medical and surgical areas and outpatient clinics.
Efforts were made before the survey was distributed to identify the number of clinical staff in
the population of nurses and midwives who are expected to routinely complete the H2FRM.
Some areas in the DHB had an exemption from completing the assessment, and others said
that they were intending to apply for an exemption because the manager did not think that the
H2FRM was an appropriate falls assessment tool. Exemptions to the routine completion of
the H2FRM were granted by the DHB Director of Nursing. In these areas, the Manager of the
area used a standardised DHB framework to address why using the H2FRM was not
appropriate e.g. patients do not move, have a nurse at the bedside and have constant
monitoring. The sampling frame excluded those wards exempted but included all the areas
which indicated they were intending to apply for an exemption, except the psychiatric
services where staff requested the researcher to not distribute the survey despite DHB
approval to do so. This area was removed from the sampling frame. Similarly, surveys did
not get distributed to all of the outpatients’ clinics as the Manager stated that this area would
be applying, and subsequently did apply, for an exemption. Lead Maternity Carers (LMC),
midwives, who were not employed by the DHB but only work there when following their
client, are not expected to complete DHB admission documentation, so were not included in
the sampling frame. The participants were clinical nurses and clinical midwives from medical
and surgical wards, day and long stay wards/units, obstetrics and paediatric areas, who
routinely completed the H2FRM on admission, whenever there has been a change of
condition, and at least every week.
43
The pool of potential respondents was 1328 and included 1156 clinical registered nurses, 50
clinical enrolled nurses and 122 clinical midwives employed by the DHB who worked in
wards, clinics and units where they were expected, as part of their patient assessment, to
complete the H2FRM.
3.4.2 Recruitment
The survey was distributed to wards, facilities, clinics and units throughout the DHB
hospitals. The rationale for using the survey procedure is that pen and paper surveys are often
used in the DHB to gather information from nurses. It was considered that it would be
beneficial to use this familiar approach rather than using an online survey design. Each
survey pack included an Information Sheet (Appendix 2), the modified Barriers and
Facilitators Assessment Instrument questionnaire and demographic data questions (Appendix
3) and a small chocolate bar as a token of reciprocity.
3.5 Procedures
The H2FRM is used in all areas of the DHB hospitals and continuing care units. The surveys
were hand-delivered and the number left for distribution was determined by either the
manager or clinical educator of the ward or area, who told the researcher how many clinical
staff, registered and enrolled nurses and registered midwives, were on the roster and the best
way to distribute the surveys. This was by the manager or educator of the ward hand
delivering them, by being left in the ward staffroom for staff to independently access, or by
being named and placed in the staff member’s personal mailbox. Most surveys were
distributed over October to mid-November 2012 but by request of the Clinical Nurse
Managers, the questionnaires for one ward and one unit 1 were distributed in January -
February 2013. The return address was on the back of the last sheet of paper with fold lines
so that respondents could fold and staple the form and put it in the internal mail to the
researcher. A poster, indicating a return date of one month after distribution of the survey,
was displayed in the most appropriate place for the ward or area.
1
Psychiatric unit. Subsequently it was decided to not distribute the surveys.
44
Over the time that the survey was conducted, there were activities around the tertiary hospital
that could have impacted on the return rate (Haines & Waldron, 2011). Several wards were
preparing to shift to new premises and some had shifted within the two months previous to
the survey. An audit and a comprehensive education package on the H2FRM was undertaken
in another area and on-going audits were mentioned by several ward leaders. One manager
asked for the survey distribution to be delayed because of audits.
According to De Vaus (2004) reminders increase participation rates, so weekly emails were
sent to the nominated contact person in the ward or area for three weeks following
distribution of the questionnaire requesting staff to respond. Notes were also written in the
ward Communication book, for staff to be notified directly. Questionnaires were on the ward
or area until the closing date. After that date, a notice thanking participants was put up and
the in the ward, clinic or area visited to retrieve any remaining surveys.
Mail surveys are commonly reported to have a response rate of 25 - 30% (Gills & Jackson,
2002). It was therefore expected that there would be between 330 and 400 responses as 1328
surveys were distributed. Generally a higher response rate means that there is more certainty
that the analysis will be able to be generalised to the population (Neuman, as cited in
Minichiello et al., 2004).
The process of cleansing the data is to detect errors that will affect the analysis of the data
and then correct or delete these (Armitage et al., 2002). Reviewing data may also indicate
why a response is incomplete and provide information on the population who did not respond
to the survey. Data were visually screened before analysis for missing values, coding errors,
responses from participants who were not in the sampling frame and other inaccuracies.
Responses that had at least one answer in the Likert-scale section were retained.
Data were entered into an Microsoft Excel spreadsheet and transferred into SPSS version
18.0 (PASW Statistics Student Version 18.0, 2009) for analysis. Controllers were placed on
what could be entered into the Excel database in order to aid correct entry. The recoding that
was required by the authors of the Barriers and Facilitators Assessment Instrument was
carried out on nominated questions. Some respondents had more than one qualification. The
most relevant or highest qualification was entered such as RM, if the respondent was working
in an obstetrics area, or RN as the higher qualification.
The responses to the questionnaire were on a 5-point Likert scale. Responses have been
collapsed for analyses within questions or groups of questions. For instance, this was
45
necessary for the barrier calculation which was “the percentage who agree and fully agree”
(Harmsen et al., 2005, p.5); certain characteristics were obtained by the average of “the
percentage who agree and fully agree” responses for two or more questions; and ‘day stay’
included medical and surgical day stay. As there were few responses from the post-
anaesthetic care unit, these responses were in day stay as there were similarities in the time a
patient spent in the area.
Analysis used non-parametric tests as they do not have to meet assumptions on the normality
of the distribution or that the scale is ordinal. Gamma, a non-parametric coefficient, can be
used to investigate correlations or relationships on a bivariate tables where there is ordinal
data, such as the Likert scale, and is less sensitive to cells which have a frequency less than
five and rare combinations of responses compared to other correlation tests (Denscombe,
2010). The limitation is that the coefficient does not describe a relationship where there is
one-way dependence and that values of -1 to +1 will not always indicate an association
(Argyrous, 2000). Therefore Gamma is not useful for directionality, cause and effect or an
important causal relationship, but tests the null hypothesis that there is no independent
relationship between two categories. A Gamma of -1 indicates perfect negative association,
of +1 when there are only concordant pairs and 0 when there is as many concordant as there
are discordant pairs (Argyrous, 2000). A - 0.30 Gamma correlation can be considered weak
negative association and + 0.70, a strong positive association (Denscombe, 2010).
Descriptive data analyses include measures of central tendency and variance analysis. Even
though the Likert scale is ordinal, the assumption has been made that the distances between
responses are nearly equal, indicating that the descriptive statistics, mean, median, mode and
standard deviation, can be used (Plitcha & Garzon, 2009).
46
collection, data analysis, writing up, publication and dissemination to potential readers of the
report (Creswell, 2014).
The ethical approval was gained from the Massey University Human Ethics Committee
(Appendix 6). The application included support from Kaumatua Kaunihera of the DHB to
conduct the research (Appendix 7) and DHB authorisation for the research to proceed
(Appendix 8). The study appendices also include the Information Sheet (Appendix 2) and a
letter of approval for the use of the Barriers and Facilitators Assessment Instrument
(Appendix 4).
The researcher tried to minimise impact on the wards by checking initially with the Clinical
Nurse Managers that a survey could be distributed and the best method and timing to do that
within the ward culture. These recommendations were subsequently adhered to.
A pilot of the survey on a similar but different group from the sample can ascertain if there
are issues such as literacy or inappropriate terminology and how to mitigate these while
ensuring the questionnaire remains valid (Minichiello et al., 2004). The responses to the pilot
study showed that the length of the questionnaire was appropriate to the time that staff had to
complete it in their working day. As the participants were required to have completed a
qualification and be currently employed by the DHB in order to be recruited, vulnerabilities
due to age, literacy level and physical condition were not manifested. Neither the authors of
the survey nor Koh et al. (2008) indicated that the survey had questions that were intrusive or
caused offense. The nurses who were involved in the pilot survey were asked specifically if
there were any questions that were difficult to answer for reasons such as these and all
replied that there were no issues. The pilot included both a similar group and potential
respondents.
Consultation about the research started early with the DHB Māori consultation committee,
Kaumatua Kaunihera, to ensure issues of cultural safety could be addressed. No participant
used any of the methods – telephone, direct contact or email, to contact the researcher or the
Supervisor to express concerns about the survey.
Informed consent was presumed after the respondent had read about the survey on the
Information Sheet and decided to complete and post the questionnaire. Confidentiality and
anonymity were maintained as nothing on the questionnaire could be used to identify the
respondent. A few named responses were returned and immediately de-identified. Some
respondents chose to write comments and these were used sparingly and only to demonstrate
a result from the data analysis so that confidentiality was maintained.
47
Responses are kept at the time of writing in a locked cupboard and a password protected
computer and the data will be destroyed at the end of the study by Massey University (in
about 5 years). The Supervisor and the researcher were the only people to see the responses.
While a small incentive (a chocolate bar) to complete the questionnaire was offered, these
relied on an honesty system to collect and no person could feel pressured by the researcher
into completing the questionnaire. No conflict of interest was identified and there were no
financial incentives for the participants or the researcher.
The responses identify position and speciality. The authors of a multi-national study stated
that different countries have different ethical priorities (Dovey et al., 2011). They stated that
as New Zealand has a small population, maintaining and protecting privacy is a greater issue
than in other more populous countries. While there was a low response rate for some
specialities, it was not low enough to be able to identify a particular clinical nurse or clinical
midwife.
3.8 Conclusion
The Research Design chapter has discussed and given the rationale behind the use of the
descriptive non-experimental survey. The rationale for the use of a survey was that it is an
appropriate design to identify attitudes of nurses, specifically the barriers for clinical nurses
and midwives when using the Hendrich II Fall Risk Model to plan preventive care for falls;
that the survey would aid understanding about an issue that had arisen in the DHB; that
survey methodology is familiar to the nurses employed by the DHB and could be completed
during the time they were at work; and that it was relatively inexpensive.
The participants are a sample of clinical nurses and clinical midwives working for a New
Zealand DHB, selected on the basis of willingness to participate in the survey. The validity of
the survey tool, developed for identifying barriers for health care staff using a new preventive
strategy, has been discussed. Some of the issues in the distribution process have been
explained including the reason why not all of the nurses and midwives employed by the DHB
were included in the survey. Data analysis procedures were presented. The ethical
considerations for the design and method, and the informed consent process, have been
discussed and relevant permissions have been referred to. In the next chapter, the results of
the survey are given.
48
Chapter Four – Results
4.1 Introduction
In previous chapters the issues of patient falls were introduced and the reasons for this study
were presented. The study aimed to identify the barriers that clinical nurses and clinical
midwives have when using the H2FRM to plan preventive care for falls. The methodology
chapter explained the descriptive non-experimental survey approach in order to obtain data to
answer the research questions. The survey used the validated Barriers and Facilitators
Assessment Instrument (Peters et al., 2002). The research questions were:
x What are the Hendrich II Fall Risk Model barriers for clinical nurses to plan falls
prevention care?
x What are the Hendrich II Fall Risk Model barriers for clinical midwives to plan falls
prevention care?
x What personal barriers for falls prevention care do clinical nurses and clinical
midwives hold?
x From the clinical nurse or clinical midwife perspective, what patient characteristics
are barriers to falls prevention care?
x What are context barriers for planning falls prevention care?
This chapter presents the results. Firstly, the demographics of the sample are presented,
followed by the barriers identified by calculations for the sample and for subgroups in each
research question. Lastly, the correlation statistics are presented.
The purpose of this section is to understand the data in order to compare it with known
statistics of the population and infer information about the sample. The data includes the
demographic variables relating to qualification, length of time nursing and area of work. This
data comes from the nurse/midwife responses.
49
frequencies of responses per question for the sample were all above 94%. Nineteen of the
twenty three questions were answered by over 97% of respondents.
Table 1
Sample Demographics by Qualification
Number of Percentage
Qualification Population
respondents (%)
Enrolled nurse (EN) 50 21 42
Registered nurse (RN) 1156 366 32
Registered midwife (RM) 122 9 7
Unspecified 8
TOTAL 1328 404 31
Respondents were asked how many years they had been nursing (Table 2). The mean was 13
years 8 months with a standard deviation of 11 years (SD = 11.0). The mean for the
registered nurses years of experience was 12 years (SD = 10.6) which was less than that of
the midwives in the sample, at 18 years (SD = 9.9). The mean for enrolled nurses was much
longer at 27 years (SD = 12.6). The training for enrolled nurses stopped for ten years and
only resumed again in 2003 (Meek, 2009) and in the local training establishment in 2010, so
the experience level for enrolled nurses reflects the length of time since they gained their
qualification. The Nursing Council of New Zealand state that 75% of enrolled nurses are
older than 50 years of age (Nursing Council of New Zealand, 2011) which is in accordance
with the length of service of the EN sample.
Table 2
Sample Demographics by Experience
Survey Response Cumulative
Years of experience Sample Rate Percentage*
(n) (%) (%)
Up to 1 year 24 6 6
2 to 5 years 84 21 27
6 to 10 years 91 23 49
11 to 15 years 45 11 60
16 to 20 years 47 12 72
More than 20 years 82 20 92
Unspecified 31 8 100
TOTAL 404 100 100
50
Half of the respondents have been working 10 years or less, and a fifth have been working
longer than 20 years (Table 2). The Nursing Council report stated that surgical nursing has
the lowest number of older nurses (Nursing Council of New Zealand, 2011) and as surgical
nurses comprise about one third of the respondents, the sample may be from a younger age
cohort of nurses.
Table 3
Sample Demographics by Area of Practice
Population Number of Percentage
respondents of
respondents
Ward or area
(N) (n) (%)
Surgical wards had the highest return rate and the lowest return rate was from the emergency
departments and obstetric areas (including delivery and pre-and postnatal) (Table 3). When
the response rate is above 30% there is more certainty that the analysis will be representative
(Neuman, as cited in Minichiello et al., 2004). Using this rule, the results from surgical and
medical areas can be generalised, the response rate of the longer-stay wards (26%) may be
representative, but the response rate from day-stay areas, outpatient clinics, paediatric wards
and emergency departments was not likely to be representative. As most midwives work in
the pre- and post-natal and the delivery suite, the low response from this area (7%) is
associated with the low response from midwives (6%). Work across wards refers to those
nurses who work in different areas depending on staffing needs. Responses were received
from all of the wards, clinics and areas into which the surveys were distributed as shown by
Table 3.
51
The Barriers and Facilitators Assessment Instrument is used “to measure barriers to and
facilitators for improvement of patient care, with a special focus on preventive care”
(Harmsen et al., 2005, p.2). The barriers will now be discussed firstly by frequency
distributions, then calculations of the percentage and mean barriers.
The protocol for the use of the Barriers and Facilitators Assessment Instrument is to identify
the barriers for the 18 characteristics (Harmsen et al., 2005). A barrier is defined by the
authors of the instrument as “the percentage of those who ‘disagree’ and ‘fully disagree’ of
positive questions 1-3 and 14; the percentage of those who ‘agree’ and ‘fully agree’ of
negative questions 4-13 and 16-22; the average of questions for the specificity, flexibility and
group norms, socialisation characteristics” (Harmsen et al., 2005, p.9). Firstly, the
distribution of the sample is presented. This is followed by a summary of barriers as
calculated by the percentages as above and by descriptive statistics, in particular, the mean.
The purpose of descriptive statistics is to describe and summarise the numerical data in order
to draw conclusions from the analysis. Descriptive statistics include measures of central
tendency and of variability (Plitcha & Garzon, 2009). Measures of central tendency include
the mean, median and mode. A discussion of variability indicates how the data is distributed
and includes the shape of the distribution, outliers, standard deviation, range and interquartile
range.
The rating used for the Peters et al. (2002) questionnaire is a 5-point Likert scale. This scale
is subjective and ordinal but can be used to measure the attitudes in a quantifiable way with
comparisons of rank order (Denscombe, 2010). The coding of responses from 1 to 5
corresponded to strongly disagree through to strongly agree, so a mean response can be
calculated. The 5-point Likert scale has been collapsed rather than calculating the median and
is presented as the percentage barrier. Percentage barriers may be the more appropriate
statistic to use to explain the data as it is recommended by the authors of the instrument and it
does not assume ratio data. As there are many modes, this measure of central tendency has
not been included.
A barrier was indicated when the percentage calculation was 30% or more and/or the mean
score was 3.0 or greater. A higher mean score was a stronger perceived barrier. Low scores
indicated a facilitator to the use of the H2FRM to plan preventive care for falls. Questions 15
and 23, the additional questions, are discussed in section 4.6. In the following tables, barriers
are highlighted in bold.
52
4.3.1 Distribution of sample
The shape of a distribution will indicate information about how the question was answered
(Plitcha & Garzon, 2009). Typically the shapes of frequency distributions are described as
symmetrical normal bell-curve, U-shaped, bi-modal or asymmetrical by a skew to the right or
left (Plitcha & Garzon, 2009). Outliers are values that are much larger or smaller so are
distant from other values.
The frequencies for responses on the Likert scale for each question have been graphed. From
the shapes of these graphs, respondents did not mark many responses as strongly agree or
strongly disagree and tended to use the middle categories of agree, do not agree or disagree
and disagree. It was found that approximately a third of the participant questionnaires had no
response on the Likert scale marked in either the strongly agree or strongly disagree choices,
a third had marked either the strongly agree/disagree at least once, and the final third had
marked both the strongly agree and strongly disagree choices at least once. This indicates
that the questions did not generate an intense reaction in the respondents.
Only one question had responses that had a symmetrical bell-shaped frequency distribution
graph. This was the question about whether nurses had doubts about the H2FRM. As the
most common response was non-committal, agree and disagree frequencies similar, and few
responses at the extremes, there appears to be ambivalence about this characteristic.
Five questions had a bimodal frequency distribution graph. The questions were on the
flexibility, specificity of the H2FRM, involvement, knowledge, motivation, and the health
status. The sample was evenly divided between those who agree and those who disagree as
illustrated in Figure 2.
250
responses
ofresponses
200
150
Numberof
Number
100
50
0
Fully 2 3 4 Fully
disagree agree
A graph with a negative skew frequency distribution for the Likert scale, where 1 is fully
disagree and 5, fully agree, will have the lower tail longer than the upper tail (Armitage et al.,
53
2002) and most responses will be towards the fully agree, that is, a strong barrier. The three
questions that had a negative skew were all in the Context category. They were
nurse/midwife attitudes to the supportive staff they work with and having the facilities and
the building conducive to preventing falls. The negative skew is illustrated in Figure 3.
250
Number of responses
200
150
100
50
0
Fully 2 3 4 Fully
disagree agree
The rest of the questions have a positive skew frequency distribution indicating that most
responses are disagree or fully disagree, and that the characteristic is not a strong barrier and
may be a facilitator. Two questions had distributions that indicated higher numbers of fully
disagree than were seen in other questions. One of these was the characteristic attitude, role
perception (Figure 4) and the other was time investment. The nurses who completed these
questions seemed to completely refute the questions, implying that respondents had no
personal problems with accepting new procedures, and that time to complete the H2FRM
was reasonable.
250
Number of responses
200
150
100
50
0
Fully 2 3 4 Fully
disagree agree
54
There was a weak positive skew to the frequency distributions of the three characteristics,
group norms, socialisation; ethnicity and compatibility. The distribution indicated weaker
barriers to these questions when compared to questions with a stronger positive skew
distribution. A rectangular distribution was noted for the question pertaining to how nurses
rated patient’s motivation to change.
The strongest barriers demonstrated by both the percentage and mean data are the Context
characteristics, lack of physical space, equipment and supportive staff. Despite the strength of
these barriers, there were no free form comments to support these results.
The specificity and flexibility of the instrument was a strong barrier and many free form
comments such as “the model needs to be updated to allow nursing discretion to come into
the equation”, “the form is not specific to post-anaesthetic patients” and “need to have
medium risk”, supported this.
Lesser barriers were doubts about the correctness of the instrument, knowledge of how to
provide falls prevention and involvement of the H2FRM in the way staff prevent falls. The
latter two barriers were not supported by the calculation of the means. Free form comments
supported the barrier that parts of the H2FRM were not correct. They included that the tool
did not include a previous history of falls, age of the patient, vision and other disabilities.
Respondents commented that they wanted more medications such as opioids on the list, the
“foreign environment” acknowledged and disagreed that male was a risk factor. Also “I have
never heard of anyone referring to it once it has been filled in or taken it into consideration in
care planning” and “I don’t find it pertinent to patients real risk of falls”. More specific
comments were that the H2FRM did not provide for patients with a nerve block or in pain,
and also noted that the up-and-go test was limited as it did not account for mobility aids or
unsteadiness after starting to mobilise.
A patient’s motivation to engage in their fall prevention plan was also demonstrated as a
barrier and free form comments such as “taught about call bell but patient still don’t use it,
patients take short cuts” and “can’t prevent the confused from falling”, reflect this issue. The
health status of the patient was demonstrated as a lesser barrier to effective falls prevention.
Of note were the very low barriers, perhaps facilitators. These were that changing routines
and working to procedures were not barriers to the adoption of the H2FRM. Free form
comments on this area stated that respondents were willing to comply with procedures,
55
Table 4
Summary of Barriers (n=404)
Percentage Barrier Mean
Barrier (%) Barrier
H2FRM Characteristics
Compatibility 23 2.2
Attractiveness 19 2.6
Involvement 33 2.7
Education 15 2.4
Patient Characteristics
Ethnicity 26 2.8
Context Characteristics
Facilities 87 4.1
Building 87 4.0
56
especially when they improved patient safety. The results also demonstrated that the H2FRM
is a good starting point to find out more about falls prevention.
The results for the subgroups for each research questions are analysed by the percentage and
mean calculations and are presented in this section.
4.4.1 What are the H2FRM barriers for clinical nurses to plan falls prevention care?
The strongest barrier demonstrated by the data in Table 5 was the specificity, flexibility of the
H2FRM for RNs and ENs. This characteristic is discussed further in 4.4.1.1.
The time to complete the H2FRM and compatibility of the Upright training are barriers for
ENs. The free form comments, “it takes time away from being able to prevent falls”, is
consistent with the EN time investment barrier. Other comments, “the online training is too
long and most people I know that have done the training just skip over as there is just too
much too read”, “the online certification … if you don’t click the ‘expected’ answer its
wrong, waste of money! Waste of time!” and “terrible voiceover” appear to demonstrate a
compatibility barrier. The usefulness, didactic benefit, of the H2FRM to the planning of falls
prevention care interventions is not a barrier.
Table 5
The H2FRM characteristics are analysed further by the area of practice of the RNs and ENs
(Table 6). This breakdown showed few barriers except for the sample of nurses from the
paediatric wards for whom all of these characteristics were barriers, the didactic benefit and
57
time investment for staff working in say stay areas, and the H2FRM attractiveness for nurses
in ED.
Table 6
Subgroup n
Surgical 140 24 2.7 23 2.6 13 2.5 10 2.6
Respondents practising in surgical, medical and rehabilitation settings had a stronger barrier
towards the ability to include the patient’s perception into the care plan than they did to add
their own clinical judgement. This finding was clearer for RNs than for ENs. RMs did not
demonstrate a barrier for this characteristic. The results for the acute surgical and medical
wards, areas similar to the developmental population of the H2FRM, consistently
demonstrated stronger barriers than most of the other areas of practice. Strong barriers were
found for nurses working in rehabilitation and paediatrics, the small number of responses
affected the results from nurses in day stay.
58
Table 7
Subgroup n
RN 366 42 3.0 50 3.3
EN 21 24 2.8 24 3.1
RM 9 0 2.3 0 2.1
ED 9 11 2.6 22 2.7
4.4.2 What are the barriers for clinical midwives using the H2FRM to plan falls
prevention?
A difference from the data for nurses was that RMs responses did not indicate any barriers to
the H2FRM (Table 8). However the free form comments were strong. They included that
there was “very little need to implement this in postnatal area”, “if faint there is usually little
warning”, and that the H2FRM “does not apply to the birthing unit”.
59
Table 8
Compatibility 14 2.9
Attractiveness 14 2.7
4.4.3 What personal barriers for falls prevention care do clinical nurses and
midwives hold?
The data on Table 4 and the breakdown by qualification and years of experience did not
demonstrate barriers for the Care provider characteristics of lifestyle, working style and
attitude, role perception. The subgroup results for the other four characteristics in this
category are presented in Table 9. The data is grouped by the qualification of the respondent
and by area of practice. When the percentage barriers are over 45%, there are also barriers by
use of means, independent of sample number. However, there is poor consistency in the
results between the two methods of calculating the barriers compared to barriers to those of
the subgroups in the H2FRM category. When the numbers in the sample are low the
inconsistency to the identification of barriers is probably a result of the variability in
responses.
The data demonstrated a strong barrier for RNs, ENs and RMs to provide falls preventive
care due to a lack of training and this was especially high for nurses practising in the
paediatric and ED areas.
Nurses working in surgical, medical and paediatric wards had doubts that the H2FRM is
correct, and the data shows that this may be a barrier for nurses in day stay and rehabilitation
areas. The response that RM respondents had no doubts about the veracity of the H2FRM did
not seem to be in accordance with making an informed choice as a barrier for the group was
knowledge, motivation and involvement in the H2FRM and RM responses demonstrated a
lack of knowledge of the H2FRM.
RNs, ENs, RMs and nurses in surgical and rehabilitation areas have strong barriers indicated
by the percentage barrier calculation but not the mean calculation for the fit between the way
60
that falls prevention is provided and the H2FRM. There is a strong barrier for nurses working
in paediatrics for the usefulness of the H2FRM in the way that they prevent falls for children.
There is no strong barrier related to the knowledge of the instrument except that there may be
barriers for RMs and nurses working in paediatric wards.
Table 9
Subgroup n
RN 366 34 2.9 30 2.9 31 2.7 16 2.4
4.4.5 From the clinical nurse or clinical midwife perspective, what patient
characteristics are barriers to falls prevention care?
The previous table (Table 4) demonstrated that ethnicity was not a barrier, however, when
the results are organised by qualifications and areas of practice, strong barriers for this
characteristic and the other characteristics were found in the Patient category and this data is
presented in Table 10.
61
Most nurses and midwives working throughout many specialities believe that the patient’s
ability to change their behaviour in accordance with their fall prevention care is a strong
barrier. The only group where this was not demonstrated was by the staff members working
in day stay.
Table 10
Subgroup n
RN 366 37 3.1 31 2.8 26 2.7
The small number of respondents in the RM, Rehab and ED subgroups demonstrated strong
barriers to providing falls preventive care to patients who seem healthy. Percentage barriers
for this characteristic were found for RNs and nurses in surgical and medical settings, but
were not supported by similar mean barriers calculations.
While the patient’s cultural background was not a barrier for most groups, those nurses in
rehabilitation/convalescent care, have demonstrated that this is a strong barrier. Some results
indicate that nurses in medical wards, RMs and obstetric areas also found ethnicity to be a
barrier.
62
4.4.5 What are the context barriers for planning falls prevention care?
The Context characteristics regarding supportive staff, facilities and building were all
demonstrated as strong barriers without exception (Table 11). These questions were aimed at
finding what factors respondents generally considered important for preventive care rather
than asking if these were important barriers in the area of practice of the respondent.
The context questions relating to group norms, socialisation were not strong barriers except
for nurses and midwives practising in paediatric and obstetric areas. There may be barriers
cooperating with members of the team for nurses working in surgical, day stay, and ED but
the barriers were not demonstrated by both barrier calculations.
Table 11
Context Barriers
Subgroup n
RN 366 89 4.1 87 4.3 83 3.9 15 3.0
The three original survey questions on teamwork were expanded to five questions to find
nurses' perceptions of how specific members of the multidisciplinary team cooperate with the
63
nurse on the application of the H2FRM. The questions on team support, although
consecutive, did not show a pattern of response. ENs rated all five questions contributing to
this characteristic as higher barriers when compared to those of the RNs (Table 12).
Table 12
RN EN RM
Study Response Response Response
(%) (%) (%) (%)
Question n 404 366 21 9
… nurses cooperation 11 11 19 14
… doctors cooperation 33 34 37 17
… physiotherapists cooperation 11 10 26 20
… managers cooperation 10 9 20 0
The highest barrier within this characteristic (with one exception) was on the teamwork with
doctors. Many free form comments were made about lack of doctors’ knowledge of the
nursing fall prevention procedures. Additional comments were that physiotherapists do not
always cooperate. Managers, on the other hand, seem to be considered as supportive and the
comments included that “management are committed”.
RNs of all experience levels had a barrier that patients were not motivated to comply with
their falls prevention plan (mean = 3.1), and RNs of 2 to 5 years of experience indicated that
it is difficult to provide falls prevention care to patients who are healthy (n = 61, percentage =
51%, mean = 3.0). All respondents at all levels of experience demonstrated that lack of space
in the building was a barrier (means from 3.9 to 4.4).
64
4.5 Gamma correlation analysis
As noted earlier, the gamma coefficient does not indicate directionality, cause and effect and
may not even point to an important relationship. The following correlations do not pertain
directly to the research questions as they cross categories of the questionnaire. While there
are many small correlations that seem important, only those with medium to strong
associations are discussed but all of those greater than 0.30 are listed in Appendix 9. The
correlations have been computed on the total sample after being reverse coded.
4.5.1 Education
Four of the strongest relationships were about education. The correlations between education,
the knowledge of the H2FRM and doubts about the H2FRM was 0.98, and between
education and the patient’s motivation to change was 0.89. The correlation between the
H2FRM in the falls prevention process with patient health status was 0.63.
The questions numbered 13 and 24 were not part of the validated survey tool but were added
as they seemed pertinent according to the literature. The return rate for the two questions was
98% and 95% respectively, so the questions show clarity, but less than 20% have endorsed
some responses so the items may be unreliable. The results are presented in Table 13.
65
Table 13
Surgical 11 12
Medical 17 16
Rehabilitation/Convalescent care 25 25
Day stay 3 6
Obstetrics 13 0
Paediatrics 57 14
Using the formula for the validated survey to calculate barriers, it can be seen that both
questions were responded to in the low percentages (except for paediatric nurses and safety
culture). The results show that the H2FRM does engender a safety culture and delegation to
staff about risk of falling was implemented effectively.
4.7 Conclusion
The 31% response rate can be considered a representative sample but there was wide
variation of the response rate from the subgroups. The response rate from surgical, medical
and rehabilitation/convalescent care specialities appears to be representative, but from other
specialities, the response rate and the number of surveys returned were too low to generalise
the results. The barriers were identified by using shapes of the frequency distributions;
percentage barriers and means for the sample and subgroups of qualifications, specialities
and experience levels; and gamma correlation coefficients. There is general accordance
between the different measures used to identify the main barriers in relation to the five
research questions.
The barriers identified by both the percentage barriers and means in this sample of clinical
staff when using the H2FRM to plan preventive care for falls are the barriers for supportive
staff, facilities and building, the specificity, flexibility of the H2FRM, the patient’s motivation
to change and issues with how doctors cooperate in applying the H2FRM. The results for
percentage barrier calculation included the characteristics: knowledge, motivation,
66
involvement, doubts about the H2FRM and the patient characteristic, health status. The EN
results for the means included compatibility and time investment characteristics as barriers.
Analysis of subgroup data found a barrier to nearly all of the other characteristics of the
survey although the numbers in many of the subgroups were too low to make generalisations.
Respondents working in surgical and medical areas, a similar population to that on which the
H2FRM was developed, often had different barriers to respondents from other specialities.
Respondents with six to ten years of experience reported fewer barriers that other experience
level groups. Nurse respondents working in the paediatric wards and emergency departments
and registered midwife respondents had more barriers than other subgroups.
Weak barriers implied that some characteristics were facilitators to the use of the H2FRM
rather than barriers. Managers were perceived as generally showing leadership and co-
operated with providing falls prevention care. The H2FRM is a benefit as demonstrated by
the results to questions on attitude, role perception and didactic benefit. In the next chapter
the results of this study are discussed under the research questions with reference to the
literature.
67
Chapter Five – Discussion
5.1 Introduction
The aim of this study was to identify the barriers for clinical nurses and clinical midwives
when using the H2FRM to plan preventive care for falls. As the identification of barriers is
the basis of planning for change, it was thought that information from the study would be of
use to DHB personnel to assist with their planning to reduce the number of patient falls. Falls
prevention is a widely researched topic as the cost of an inpatient fall is unacceptably high.
However, sustained reduction in the fall rate has not been easy to achieve in any hospital. A
model of falls risk by Dykes et al. (2011) presented falls prevention as the safety platform of
the patient; a current and individualised care plan based on an accurate falls risk assessment;
care provider competence, self-efficacy, commitment and capability; and context of care
including the protective physical environment (see Figure 1, page 8).
The H2FRM for predicting patient falls was validated in acute medical and surgical areas and
further studies have recommended its use in similar areas. While it is expected that no tool
for assessing risk can be 100% accurate, the sensitivity and specificity results for the H2FRM
are similar to that of other commonly used falls risk assessments. In the DHB, where the
present research took place, the H2FRM document for the patient’s notes has supporting
material including an online training module, signage for high risk above the patient’s bed, a
flip chart to assist with planning preventive care, monitoring of fall events and policy
documents. The DHB also has resourced a leader to coordinate the falls prevention
programme and equipment to help prevent falls such as bed alarms. The falls prevention
programme is linked to the DHB strategic plan. The H2FRM is used throughout the hospitals
of the DHB and the few exceptions to this have met stringent falls prevention criteria. The
individualised falls prevention care plan is written for patients who have a high falls risk
identified through the completion of the H2FRM. This does not mean that patient’s with a
low risk do not need re-assessment.
Five research questions were written and a quantitative non-experimental descriptive survey
design was used to address the aim of the study. The nurses and midwives who were
expected to complete the H2FRM, as per the DHB Falls risk assessment, minimisation and
management policy, comprised the sampling frame. The validated Barriers and Facilitators
Assessment Instrument was used to identify attitudes to the H2FRM and to falls prevention.
There were 404 responses used for analysis (366 from registered nurses, 21 from enrolled
nurses, 9 from registered midwives and 8 with no qualification) from the 1300 surveys that
were distributed. The response rate means that findings can be generalised from the sample
68
and RN, EN, surgical, medical and long-stay subgroups, but is less likely to be representative
of other subgroups. Percentage barrier and descriptive statistics calculations were used to
analyse the data. The difficulty of providing falls preventive care when there was insufficient
supportive staff, a lack of equipment and poor design of space, were identified as barriers for
the sample and all subgroups. The results clearly indicated that there were issues related to
the training in and utilisation of the H2FRM and that patients did not cooperate with their
care prevention plan. Respondents indicated that they work according to procedures, are able
to adapt their practice to incorporate new routines and cooperate with members of their team.
This section, 5.2 to 5.6, discusses the implications of the results to each of the five research
questions. The basis for discussing a barrier is a percentage calculation of 30% or more and
supported by the mean of 3.0 or greater.
5.2 Research question: What are the Hendrich II Fall Risk Model barriers for
clinical nurses to plan falls prevention care?
This research question examined barriers related to the H2FRM, namely its specificity and
flexibility, didactic benefit, compatibility and attractiveness. The term ‘nurse’ refers to both
registered and enrolled nurses, staff who have different responsibilities and scopes of
practice. The RN and EN groups responded differently to the survey questions so RN and EN
findings are discussed separately. The gamma correlations are discussed in 5.2.1 as RNs
composed 91% of the sample so had the largest effect on the calculations.
5.2.1 What are the H2FRM barriers for clinical RNs to plan falls prevention care?
The highest characteristic barrier was specificity, flexibility. This was not the highest
characteristic barrier in a similar study in Singapore hospitals and the authors did not indicate
that they had split the characteristic into the two questions (Koh et al., 2008). With few
exceptions, the highest barrier in this study was for the second question in this characteristic
that asked whether the H2FRM leaves enough room to consider the wishes of the patient.
The H2FRM is a validated and trademarked tool so it is not possible to add more risk factors.
However, incorporating the wishes of the patient into the falls risk assessment is supported
by studies that indicated the importance of the patient’s voice as part of planning care
(Currie et al., 2005), that patients requested to be part of the team who make decisions to
prevent them from falling (Carroll, 2010), and that the patient’s perspective is valid even
when it is different from that of staff (Rush et al., 2008). The patient’s attributes such as their
culture, cannot be entered into the falls risk, but the two questions were highly correlated.
Difficulties in communication and understanding of the care plan can arise from cultural
69
differences (Koh et al., 2008), consequently basing a plan only on the H2FRM factors would
not manage this risk. This characteristic is a barrier to effective planning.
The second question for the specificity, flexibility characteristic was whether the H2FRM
leaves enough room for the nurse to make her/his own conclusions about falls prevention.
The nurse can make a judgement call for some factors on the H2FRM such as, confusion/
disorientation/ impulsivity and altered elimination and guidelines are provided in the Upright
training on how to score these factors. However, the application of the nurse’s assessment is
limited. This barrier may be explained by referring to the literature regarding the accuracy of
clinical judgment to assess fall risk.
The use of NCJ to assess falls risk is generally supported by the literature. The results of
studies conclude that NCJ is not good enough to use in acute hospital settings (Webster et al.,
2010); that the specificity of NCJ is good enough for surgical and general medicine patients
who were less than 75 years old (Milisen et al., 2012); that NCJ of experienced nurses is
accurate enough to judge the falls risk (Myers & Nikoletti, 2003); and partly because of its
flexibility, NCJ is good enough to use as a standard for judging the effectiveness of new tools
(Haines et al., 2007). The reason for this barrier may be because there is no factor in the
validated H2FRM which supports the use of NCJ. The H2FRM was designed to be objective
with high inter-rater reliability (Hendrich et al., 2003). However, an individualised care plan
could be based on both the H2FRM and NCJ. Risk factors identified on the H2FRM, NCJ
and the needs of the patient could be given equal weighting when planning care.
The high correlations between this specificity, flexibility question and the questions on the
patient’s ethnicity, the health status of the patient, and involvement are in accordance with the
conclusion to affirm NCJ when planning care. A measure of risk of falling made by using
only the H2FRM is restrictive and clinical judgement can improve the accuracy of the risk
assessment and individualised care plan based on this.
5.2.2 What are the H2FRM barriers for clinical ENs to plan falls prevention care?
Enrolled nurses, or their equivalent in other countries, are seldom differentiated in studies of
falls prevention. In the study there were few enrolled nurse respondents (n = 21) but a good
response rate. The barriers identified were specificity, flexibility, time investment and
compatibility. The specificity, flexibility, barrier has been discussed for RNs and the literature
probably applies similarly to EN respondents.
The time taken to complete the H2FRM has been recorded as one minute (Ivziku et al.,
2011), so it seems that the EN time investment barrier is inconsistent with the literature as this
is a reasonable time to complete the risk score. A similar study, in which 27% of the
70
demographics were enrolled nurses, did not have time investment as a barrier (Koh et al.,
2008). However, there is support in the literature for the time investment related to the
frequency that the H2FRM would have to be completed in acute settings. Correspondence
from the H2FRM author regarding the number of incorrectly H2FRM identified patients who
fell, stated that the risk scores needed to be completed more frequently (Clarke et al., 2012).
It has been suggested that associate nurses are more task orientated than baccalaureate nurses
(Kalisch et al., 2009). If the ENs were considering the time taken to frequently update the
H2FRM document in the patient’s notes, the result would be in accordance with the
literature. Also, if respondents considered the time taken to provide fall prevention care
when giving their response, several studies endorse the findings that there is insufficient time
for care (Lucero et al., 2010; O'Connell & Myers, 2001; Robert et al., 2011).
The compatibility characteristic may be understood as a barrier against the use of a screening
tool. While some authors state that a risk screen or assessment is the beginning of the process
(Dykes et al., 2009), other authors do not propose this as mandatory (Cameron et al., 2012;
Oliver, 2008). This means that whether this is seen as a barrier or not, is in accordance with
the literature. The Koh et al. (2008) study found this characteristic to be a facilitator.
5.3 Research question: What are the Hendrich II Fall Risk Model barriers for
clinical midwives to plan falls prevention care?
The generalisation of the results for midwives is not possible as just 7% of the RM
population responded, nine respondents in total. The low response rate may indicate
opposition to the survey, which is consistent with cognitive stacking where tasks designated
as low priority are not completed (Sitterding, Broome, Everett, & Ebright, 2012). The low
response rate may also be due to the lack of use of the H2FRM in the obstetric areas as
indicated by the DHB grey literature which showed a lack of completion of the Upright
training by RMs.
The highest barriers for the respondents were the personal barriers knowledge, motivation and
involvement (indicated by the percentage barriers) and compatibility with the H2FRM
(indicated by the mean score). These barriers reveal a lack of familiarity with the H2FRM by
RMs which is in agreement with the free form responses that the H2FRM was not suitable
for RM use. The barriers and comments are supported by the literature such as that a specific
fall risk assessment is needed for the obstetrics area of practice (Heafner, 2013). Also the
obstetric area was not part of the population which Hendrich used to develop the model and
studies state that a tool should only be used in populations similar to where they were
developed (Perrell et al., 2001). The barriers need to be acknowledged as there is a gap in the
way falls assessments for obstetric clients are made.
71
5.4 Research question: What personal barriers for falls prevention care do clinical
nurses hold?
This section examined the respondents’ knowledge of the H2FRM, knowledge, motivation
and involvement characteristics; whether they considered the H2FRM to be accurate, doubts
about H2FRM; education and understanding of falls preventive care; and attitudes, attitude,
role perception and lifestyle, working style characteristics.
The knowledge, motivation characteristic was a barrier for respondents particularly for nurses
in the emergency units and paediatric wards when compared to respondents from acute
medical and surgical areas, and ENs. It has been found that this characteristic is typically a
barrier when introducing a new innovation (Grol & Grimshaw, 2003; Johnson et al., 2011;
Koh et al., 2008). However, in the DHB the H2FRM was launched two years before the
survey with an associated training package so it was expected that the H2FRM would be
familiar to all clinical staff.
Knowledge, motivation has been found to be barrier in falls prevention studies and was the
highest barriers in a study using the same survey instrument and with a similar aim. The
barrier was attributed to the difficulty of knowing and applying all of the published best
practice guidelines (Koh et al., 2008) but in this study the barrier is specific to the lack of
knowledge of the H2FRM. Despite the H2FRM supporting documentation, this barrier points
to lack of effective messages. Systematic reviews recommend that staff have a good
knowledge of falls prevention (Hempel et al., 2013; Spoelstra et al., 2012) and staff
education in prevention of falls was seen as an essential pre-requisite to a multi-factorial
intervention study (van Harten-Krouwel et al., 2011). In contrast, the Cochrane Bone and
Joint Collaboration concluded that improving staff knowledge of falls prevention made no
difference to the falls rate (Cameron et al., 2012). Correlations were found between this
characteristic and the context characteristics of supportive staff, good facilities and a safe
building which seems to indicate that the elements of the safety platform model are well
understood. This question is similar to the ‘involvement’ question and both had similar
percentage barriers.
The involvement characteristic was a barrier, It was higher for nurses working in medical and
rehabilitation/convalescent care wards and ENs. This characteristic was a similar percentage
barrier in a study using the same survey instrument, but in that study it was of lesser
importance compared with the other identified barriers (Koh et al., 2008), so not discussed. It
is known that nurses gain much of their knowledge from colleagues (Spenceley et al., 2008),
that professional development is associated with increased self-efficacy (Dempsey, 2009;
72
McKinley et al., 2007) and work satisfaction (Purdy et al., 2010). If knowledge of the
H2FRM is predominantly gained from colleagues, any misunderstandings about the Upright
training may persist with the result that the barriers to knowledge, motivation and
involvement questions would be similar, as was found. The moderate correlation found
between the involvement barrier and the incorporation of the nurse’s judgement into the falls
assessment is in accordance with that conclusion. In addition, the free form comments
pointed to issues with the Upright training. The method of delivery or lack of local content in
the training (Johnson et al., 2011) may have made this characteristic a barrier rather than
there being a resistance to professional development.
Doubts about the H2FRM was the next highest percentage barrier identified by respondents.
Doubts about a part of a risk assessment instrument are not reported in the falls literature.
However, a study using the same survey tool reported this characteristic as an 80% barrier to
the use of a new protocol for a sample of doctors (Wensing & Grol, as cited in Harmsen,
2005). If the barrier signposted that factors are missing from the H2FRM, such as a history of
falls and older age, it is in accordance with literature on falls screening or assessment tools
as many include these factors (Myers, 2003; Spoelstra et al., 2012). Free form comments also
stated that the H2FRM should include such factors. The lack of falls history, older age visual
impairment and diagnosis may give rise to doubts about the H2FRM even though statistically
the extra factors were not important enough to include in the H2FRM when it was developed.
There was a strong correlation between the doubts about the H2FRM characteristic and the
education characteristic. As the provision of safe care is a principle of the Code of Rights
(Health & Disability Commissioner, 2009) it is part of nurse education and training. The
correlation may indicate that nurse education on falls prevention does not include the use of a
falls risk tool to identify risk. The barrier, the literature, the comments and the correlation
suggest that there is a lack of understanding about how the H2FRM was developed and its
recommended use in falls prevention care. The education characteristic was a percentage
barrier for respondents from the paediatric area and mean calculation barrier for ED
respondents. The specific characteristics of these specialities may be the reason respondents
identified education as a barrier rather than a lack of training in providing preventive care.
5.5 Research question: From the clinical nurse or clinical midwife perspective, what
patient characteristics are barriers to falls prevention care?
All three questions regarding the patient were barriers. The means for the motivation to
change characteristic indicated that this was an issue for nurses and midwives in most areas.
Health status was a barrier to RNs working in areas similar to where the model was
developed, and to RMs. The third question asked about the ethnicity barrier. This was a
73
barrier for the study by Koh et al. (2008), but only a barrier in this study for RNs working in
rehabilitation and medical wards, and RMs.
The barrier identified by nurses that patients lack motivation to apply falls risk interventions,
is supported by the literature and by the high correlation with the education barrier. It is in
accordance with a study that found nurses are likely to relate the reasons for a fall to the
patient or environment rather than to themselves (Tzeng & Yin, 2008). While a successful
fall reduction intervention was patient education, targeted at the patient’s individual fall risks
and undertaken by the researchers (Ang et al., 2011), such interventions are not always
successful in reducing the fall rate (Cumming et al., 2008). Effective teaching requires time
to check that a patient understands their care plan. Nurses find that time to provide education
is difficult to fit into their workload (Tucker et al., 2012). Further issues related to this barrier
are that patients do not always follow instructions or ask for assistance (Carroll, 2010;
Fortinsky et al., 2004; Kalisch, Tschannen, & Lee, 2012) and that falls are sometimes beyond
the nurse’s control (Rush et al., 2008). This question was not asked in the study by Koh et al.
(2008). The barrier and literature suggest a gap in nurse education about how to motivate
patients to be responsible for their falls prevention. The very high correlation with the
motivation to change barrier and … I am not trained in giving preventive care supports this
conclusion.
The second barrier, health status, was not specifically asking about the H2FRM. This barrier
is in accordance with the Koh et al. (2008) study where the authors asserted that nurses were
unable to place the healthy patient in the clinical guideline with the result that falls
prevention care to patients who seem healthy was a barrier. A moderate correlation between
health status and the involvement characteristic suggests that the part of the Upright training
where the importance of usual cares is discussed, for instance, maintaining a safe
environment, call bell and walking aids within reach, bed in lowest position, may not be
remembered.
The barrier may also be discussed in relation to patient attributes. It is expected that younger
patients have less risk factors (less cognitive impairment, unaltered elimination and/or
excretion, few musculoskeletal and sensory deficits) and so have a lower falls risk. Yet
statistics from the HQSC state that increasing numbers of patients between 40 and 45 years
old sustain falls (Health Quality and Safety Commission, 2012), and a randomised control
trial found no change to the fall rate for patients who were less than 65 years old, but a
significant decrease for those over 65 years post-intervention (Dykes et al., 2010). The
younger patient, who may appear to be healthier, may be more difficult to provide preventive
care to.
74
Non-preventable falls are, by definition, the unexpected events and risk assessments are less
sensitive to the precipitating factors, so the barrier may reflect this type of unexpected fall
and inability to control the factors associated with it. In such cases where the healthy patient
falls, an in-depth post-fall debrief is suggested in order to prevent recurrence and to aid staff
learning (Gray-Miceli, 2007; Grenier-Sennelier et al., 2002; Quigley et al., 2009). In
accordance with Koh et al. (2008), the patient who seems healthy is difficult to give
preventive care to because of their attributes and because the H2FRM does not identify such
patients as high risk so they do not receive targeted interventions.
According to Koh et al. (2008) ethnicity was also a barrier to effective falls prevention as it
was in this study for some subgroups of RNs. The barrier was discussed by Koh et al. (2008)
in relation to the multi-racial context of Singapore hospitals and lack of readily available
interpreters to communicate with patients about falls education. Listening to the patient,
essential to communication, is the beginning of understanding falls risk behaviour on which
to base individualised education in falls prevention (McInnes et al., 2011). Ethnicity, if it
affects communication, will be a barrier to falls prevention. Education of the nurse will help
find ways around this barrier.
5.6 Research question: What are context barriers for planning falls prevention care?
The exception to the low group norms, socialisation barrier was the cooperation of doctors.
This was a barrier for most groups of respondents. Referrals to physiotherapists and
occupational therapists are routine in many specialities and may use the H2FRM as
assessment information. In contrast, communication with doctors may be supported by
literature that states that the post-fall focus is on the physiological effects from a fall (Zecevic
et al., 2006). Two studies to improve falls prevention targeted inter-professional
communication, including with doctors (Andreoli et al., 2010; Browne et al., 2004). It was
found that a tool improved communication in a wide range of situations, teamwork and
within interdisciplinary units. Communication using the H2FRM as part of providing a
patient assessment may be a gap.
75
The supportive staff and facilities characteristics were also barriers in a similar study in
Singapore hospitals (Koh et al., 2008). The meaning that respondents applied to supportive
staff is unclear as other questions about how health professionals cooperate to apply the
H2FRM were low barriers in this study. In the Koh et al. (2008) study, supportive staff were
explained as “a fall nurse specialist or a change champion” (p. 6) and stated that the lack of
people in these roles was a barrier to effective falls prevention. Similarly this could be an
explanation for this barrier as these are not formal positions in the DHB. There were many
moderate or strong gamma correlations with the three characteristics, supportive staff,
building, facilities, which indicated how important they are to attitudes and ways of
preventing falls. These barriers do not mean that they are in the areas of practice of the
respondents’ but are an acknowledgement that they are issues for effective falls prevention.
As most of the multifactorial studies on fall prevention identify context barriers, the findings
of this study are supported in the literature. Minimising consequences of less than ideal
facilities and building layout (Grenier-Sennelier et al., 2002; Robert et al., 2011), utilising
empowerment and leadership support (Johnson et al., 2011; Quigley et al., 2009) and
organisational goal-setting (Grol & Grimshaw, 2003) are all important in falls prevention.
The questions asked in this survey did not comprehensively include the range of context
factors, for instance there were no specific questions on organisational policies, processes,
resourcing, professional development or specific equipment.
There were methodological limitations of this study similar to all survey designs. A low
return rate for a self-completed postal questionnaire may mean that the sample is not
representative of the population and the findings not generalisable (Browne, 2005). Mail
surveys are commonly reported to have a response rate of 25-30% (Gills & Jackson, 2002).
The return rate for this survey was 31%, so generalisable using that criteria, but it varied
between 6% and 60% for sub-groups of respondents. As the population was defined as the
clinical nurses and midwives from one DHB, the results cannot be generalised to a different
DHB without further understanding of the populations of other DHBs. A further limitation is
that results are threatened as respondents may have provided answers that did not reflect their
thinking and only those who have a favourable attitude to surveys and the H2FRM may have
responded. Finally, the survey was taken at one point of time so if repeated, the findings may
be different.
Questions were altered to use words that are more commonly used in the DHB such as, weigh
the wishes of the patient became to consider the patient’s wishes. This may have altered the
validity of the instrument. According to the literature, the Cronbach alpha of the categories
76
within the Barriers and Facilitators Instrument was not high (DeVellis, 2003) and the
additional questions were not psychometrically tested, so limiting the validity.
Survey data does not allow for a deeper understanding of the reasons behind a respondent’s
answer and topics are limited to the questions in the survey. While there were changes to
wording to make the meaning more clear, it is not known how respondents interpreted words
such as healthy and supportive staff. Communication of falls risk is a common issue found
through studies but this was either not a barrier in the DHB or the survey questions did not
uncover this issue.
It is known that confounding variables affect outcomes (Scott et al., 2003). Educational
programmes and audits were concurrently happening while the survey was distributed. While
during the month that the survey was available, no ward shifted sites as part of the major
building programme, it was not possible to avoid the preparations before a shift or the
reorganisation afterwards. It is likely these ward changes have affected the findings and
response rate.
The statistics used to analyse results were those suggested by the authors of the instrument
and measures of dispersion and central tendency. The mean statistics may not be appropriate
as the 5-point Likert scale does not provide ratio numbers.
The strengths of the survey are that the questions were from a range of disciplines. The good
response rate for both the survey and for individual questions, means that there can be
confidence in the results. The low response rate from some areas can also be considered a
strength as it is in accordance with the resistance that was noted in the DHB grey data and the
free form comments on the surveys. As the H2FRM is particularly relevant to the acute
medical-surgical area, the high response rate from respondents strengthens the
generalisability of the findings for these areas.
5.8 Recommendations
This study set out to investigate the barriers to the use of the H2FRM. Falls are a cost to
patients, families and the health system so reducing the rate is very worthwhile. While the
focus of this study has been on the falls assessment tool, the recommendations are not limited
to the H2FRM. Underpinning the recommendations is that change needs to be owned at the
ward level (Dempsey, 2009; Robert et al., 2011; van Harten-Krouwel et al., 2011) that
leadership is required (Holleman et al., 2009; Parsons & Cornett, 2011; Purdy et al., 2010;
Schultz & Kitson, 2010) and that any intervention has to be mindful of other initiatives and
audits that are occurring at the same time (Scott et al., 2003). The recommendations arising
from the present study are incorporation of nurse clinical judgment, review of the Upright
77
training, increase knowledge of falls prevention, improve patient education in falls
prevention planning, trial specific falls risk screen or tool for the obstetrics and paediatrics
areas and investigate orientation of doctors.
Flexibility could be added through focussed nursing interventions which are specific to the
patient diagnosis and ward culture based on NCJ, not necessarily just on H2FRM. The
current hospital procedures of the Upright training, completion of H2FRM, application of
risk minimisation usual cares and specific cares, documentation and evaluation of
interventions would be reinforced. Interventions for ward/unit/facility issues may be
evidenced as a flowchart for different common conditions, a regular discussion point at ward
meetings, handovers, in-service discussions, or through the communication book. It would fit
with the ward culture of communication and learning. The specific interventions may be
shared with the appropriate Co-ordinator in the DHB for further discussion. If the ward
nurses agree, the module in the Productive Wards, the Well Organised Ward, could be used
for falls prevention (Robert et al., 2011), to utilise the plan, do, study, act cycle with a focus
on increasing flexibility.
The outcome would be that targeted strategies are written in the Fall Minimisation section of
the Patient Care Plan. This care plan has been in place unchanged for several years and the
recommendation is that in the next reprint this section needs to be larger or the form
reformatted to be in accordance with the established flip chart. Furthermore, the DHB policy
would also need to be revisited and updated.
Included in this recommendation is that the following areas use NCJ as the fall risk
assessment until a new risk score has been established. The outpatient clinics are considering
options but clinical judgement alone may be sensitive enough as the falls risk screen in this
area. Day stay medical and surgical units and post-anaesthetic units require additional study
in order to understand how H2FRM fits their patients. Fall risk tools for the emergency
department (ED) have been developed (Terrell et al., 2009) and some psychometric testing
78
has been completed on them. Even though H2FRM used ED data in the developmental
population, the low return rate in this survey seems to reflect a resistance to the use of the
H2FRM, hence a specific tool is needed in this area. The standardised forms that are
currently used could be updated to facilitate this. According to Äberg et al. (2009) two
questions are sufficient for an initial screen of risk and these could be on the form. High risk
would need to be recorded as would documentation and evaluation of the preventive actions.
79
experiences and scenarios at in-service meetings would be timetabled as part of the ward
preventive strategy. Topics to be covered could be how to include relatives, an issue
according to literature and falls education to “patients who seem healthy”, a barrier identified
by the survey.
5.8.5 Trial specific fall risk screen for obstetrics and paediatrics
Both of these specialities indicated lack of fit of the H2FRM to their areas, and were
supported in this by the literature (Heafner, 2013; Razmus et al., 2006), so investigations to
find better instruments for these areas needs to be carried out.
This study has highlighted areas where more research could be conducted.
5.9.2 Efficacy of nurse clinical judgement for short stay areas in the hospital
The clinical judgement of a nurse can be made quickly and so be responsive to the changing
condition of a patient (Haines et al., 2007). In this study, barriers to the use of the H2FRM
were identified in outpatient clinics and the transit lounge. These areas have a high
throughput of patients, so a judgement of falls risk seems to be congruent with NCJ. Future
studies could identify whether clinical judgement is efficacious.
80
Further studies on how to improve this aspect of fall prevention in the New Zealand context
would be worthwhile.
This study highlighted that there are issues with the flexibility of the validated H2FRM
scoring document to include all risks as judged by nurses, the training in the model and the
involvement of patients in their fall prevention plan. Although there were similarities in the
identification of the barriers for registered and enrolled nurses and registered midwives, the
small number of responses from the latter meant that the results need to be interpreted with
caution. A model was used to depict the inter-relationships between the patient, accurate falls
risk assessment with its associated current and individualised care plan, the care provider and
safe environment (Dykes et al., 2011). Two issues with this model were that the accurate
falls risk did not necessarily reconcile well with clinical judgement, especially when NCJ
was not part of the falls risk assessment. Secondly, although the patient is in the centre, the
active involvement of the patient in their fall prevention care was not a clear outcome of this
placement.
It is suggested that the barriers about the flexibility of the model to incorporate site-specific
falls risk issues, can be minimised through the sharing of effective interventions and
evaluations with the team. These need to be documented on the care plan and evaluated in the
notes. Active dissemination using timetabled in-service or use of the nurse-led Productive
Ward module is recommended. The online training should be kept but modified to be more
user-friendly so important messages are retained, such as about the validity of the model.
81
The most consistent and strongest barriers identified were the lack of supportive staff to
implement preventive care, facilities and equipment that reduce ability plan for safety and
building layouts which do not reduce extrinsic risk factors. As these results demonstrated
general issues rather than highlighting specific ward or unit barriers, it is incumbent on
management to continue to reduce these barriers through leadership, resourcing and strategic
direction.
This study found that staff members are willing to accept new procedures, able to adapt their
routines and that they can cooperate with many members of the multi-disciplinary team.
Many other studies have found that team communication is a barrier, but although not
specifically asked about this issue, the nurses and midwives at the DHB have not identified
this as a barrier. Team dynamics are important for effective falls prevention care. Working
with colleagues, managers and the multidisciplinary team was not found to be a barrier, nor
delegation of the risk of falling. Finally, it is the competence of the team and nurse/midwife,
the commitment to reduce risk behaviour in patients or elements of the environment, and the
organisation providing the capacity for these to occur that will reduce the number of patients
sustaining a fall.
82
References
Äberg, A. C., Lundin-Olsson, & Rosendahl, E. (2009). Implementation of evidence-based
prevention of falls in rehabilitation units: A staff's interactive approach. Journal of
Rehabilitation Medicine, 41, 1034-1040. doi: 10.2340/16501977-0452
Accident Corporation Corporation (ACC). (2012). Introduction to ACC. Retrieved from
http://www.acc.co.nz/about-acc/overview-of-acc/introduction-to-
acc/index.htm#P16_1588
Aiken, L. H., Sloane, D. M., Bruyneel, L., Koen, & Sermeus, W. (2013). Nurses' reports of
working conditions and hospital quality of care in 12 countries in Europe.
International journal of nursing studies, 50(2), 143-153. doi:
10.1016/j.ijnurstu.2012.11.009
Andreoli, A., Fancott, C., Velji, K., Baker, R., Solway, S., Aimone, E., & Tardif, G. (2010).
Using SBAR to communicate falls risk and management in inter-professional
rehabilitation teams. Healthcare Quarterly, 13(2010), 94-101.
Ang, E., Mordiffi, S. Z., & Wong, H. B. (2011). Evaluating the use of a targeted multiple
intervention strategy in reducing patient falls in an acute care hospital: A randomized
controlled trial. Journal of Advanced Nursing, 67(9), 1984-1992. doi:
10.1111/j.1365-2648.2011.05646.x
Ang, N. K. E., Mordiffi, S. Z., Wong, H. B., Devi, K., & Evans, D. (2007). Evaluation of
three fall-risk assessment tools in an acute care setting. Journal of Advanced
Nursing, 60(4), 427-435. doi: 10.1111/j.1365-2648.2007.04419.x
Argyrous, G. (2000). Statistics for social and health research with a guide to SPSS. London:
SAGE
Armitage, P., Berry, G., & Matthews, J. N. S. (2002). Statistical Methods in Medical
Research (4th ed.). Massachusetts: Blackwell.
Bahtsevani, C., Willman, A., Stoltz, P., & Östman, M. (2010). Experiences of the
implementation of clinical practice guidelines - interviews with nurse managers and
nurses in hospital care. Scandinavian Journal of Caring Sciences, 24, 514-522. doi:
10.1111/j.1471-6712.2009.00743.x
Boynton, P. M. (2004). Hands-on guide to questionnaire research: Administering, analysing,
and reporting your questionnaire. BMJ: British Medical Journal, 328(7452), 1372.
Browne, J. (2005). Survey design. In J. Green & J. Browne (Eds.), Principles of social
research (pp. 116-127). Maidenhead, Berkshire: Open University Press/McGraw-
Hill.
83
Browne, J., Covington, B. G., & Davila, Y. (2004). Using information technology to assist in
redesign of a fall prevention program. Journal of Nursing Care Quality, 19(3), 218-
225.
Cameron, I. D., Gillespie, L. D., Robertson, M. C., Murray, G. R., Hill, K. D., Cumming, R.
G., & Kerse, N. (2012). Interventions for preventing falls in older people in care
facilities and hospitals. Cochrane Bone, Joint and Muscle Trauma Group. doi:
10.1002/14651858.CD005465.pub3
Carroll, D. l. (2010). Patients' perspectives of falling while in an acute care hospital and
suggestions for prevention. Applied nursing research : ANR, 23(4), 238-241.
Chapman, J., Bachand, D., & Hyrkas, K. (2011). Testing the sensitivity, specificity and
feasibility of four falls risk assessment tools in a clinical setting. Journal of Nursing
Management, 19(1), 133-142. doi: 10.1111/j.1365-2834.2010.01218.x
Chen, K., Chen, L.-R., & Su, S. (2010). Applying root cause analysis to improve patient
safety: Decreasing falls in postpartum women. Quality & Safety In Health Care,
19(2), 138-143.
Cina-Tschumi, B., Schubert, M., Kressig, R., De Geest, S., & Schwendimann, R. (2009).
Frequencies of fall in Swiss hospitals: Concordance between nurses' estimates and
fall incident reports. International Journal of Nursing Studies, 46, 164-171. doi:
10.1016/j.ijnurstu.2008.09.008
Clarke, H. D., Timm, V. L., Goldburg, B. R., & Hattrup, S. J. (2012). Preoperative patient
education reduces in-hospital falls after total knee arthroplasty. Clinical Orthopedic
Related Research, 470(244-249). doi: 10.1007/s11999-011-1951-6
Coughlan, M., Cronin, P., & Ryan, F. (2007). Step-by-step guide to critiquing research. Part
1: Quantitative research. British Journal of Nursing, 16(11), 658-663.
Creswell, J. (2014). Research design: Qualitative, quantitative and mixed methods
approaches (4th ed.). Thousand Oaks: California: SAGE Publications.
Cumming, R. G., Sherrington, C., Simpson, J. M., Vogler, C., Cameron, D., & Naganathan,
V. (2008). Cluster randomised trial of a targeted multifactorial intervention to
prevent falls among older people in hospital. BMJ, 336. doi:
10.1136/bmj.39499.546030.BE
Currie, V., Harvey, G., West, E., McKenna, H., & Keeney, S. (2005). Relationship between
quality of care, staffing levels, skill mix and nurse autonomy: Literature review.
Journal of advanced nursing, 51(1), 73-82. doi: 10.1111/j.1365-2648.2005.03462.x
Davis, P., Lay-Yee, R., Briant, R., & Scott, A. (2003). Preventable in-hospital medical injury
under the “no fault” system in New Zealand. Quality and Safety in Health Care,
12(4), 251-256. doi: 10.1136/qhc.12.4.251
84
De Vaus, D. (2004). Structured questionnaires and interviews. In V. Minichiello, G. Sullivan,
K. Greenwood & R. Axford (Eds.), Research Methods for Nursing and Health
Science (2nd ed., pp. 347-392). French Forest, NSW: Pearson/Prentice Hall.
Dean, E. (2010). Productive ward measures see hospital's mortality rates plummet. Nursing
Standard, 24(8), 15-17.
Dempsey, J. (2009). Nurses values, attitudes and behaviour related to falls prevention.
Journal of Clinical Nursing, 18, 838-848. doi: 10.1111/j.1365-2702.2008.02687.x
Denscombe, M. (2010). The good research guide for small-scale social research projects
(4th ed.). Maidenhead, England: McGraw-Hill/Open University Press.
DeVellis, R. F. (2003). Scale development: Theory and application (2nd ed.). Thousand
Oaks, CA: SAGE.
DiBardino, D., Cohen, E. R., & Didwania, A. (2012). Meta-analysis: Multidisciplinary fall
prevention strategies in the acute care inpatient population. Journal of Hospital
Medicine, 7(6), 497-503. doi: 10.1002/jhm.1917
Dovey, S., Hall, K., & Phillips, R. (2011). Seeking ethical approval for an international study
in primary care patient safety. British Journal of General Practice, 201, 192-204.
Dykes, Carroll, D. L., Hurley, A., Lipsitz, S., Benoit, A., Chang, F., . . . Middleton, B.
(2010). Fall prevention in acute care hospitals: A randomized trial. JAMA, 304(17).
doi: 10.1001/jama.2010.1567
Dykes, P., Carroll, D., McColgan, K., Hurley, A., Lipsitz, S., Colombo, L., . . . Middleton, B.
(2011). Scales for assessing self-efficacy of nurses and assistants for preventing falls.
Journal of Advanced Nursing, 67(2), 438-449. doi: 10.1111/j.1365-
2648.2010.05501.x
Dykes, P. C., Benoit, A., Carroll, D. L., Middleton, B., & Hurley, A. C. (2009). Why do
patients in acute care hospitals fall? Can falls be prevented? Journal of Nursing
Administration, 39(6), 299-304.
El Enein, N. Y. A., El Ghany, A. S. A., & Zaghloul, A. A. (2012). Knowledge and
performance among nurses before and after a training programme on patient falls.
Open Journal of Nursing, 2, 358-364. doi: 10.4236/ojn.2012.24053
Fixsen, D., Scott, V., Blase, K., Naoom, S., & Wagar, L. (2011). When evidence is not
enough: The challenge of implementing fall prevention strategies. Journal of Safety
Research, 42, 419-422. doi: 10.1016/j.jsr.2011.10.002
Fortinsky, R. H., Iannuzzi-Sucich, M., Baker, D. I., Gottschalk, M., King, M. B., Brown, C.
J., & Tinetti, M. E. (2004). Fall-risk assessment and management in clinical practice:
Views from healthcare providers. Journal of the American Geriatrics Society, 52(9),
1522-1526. doi: 10.1111/j.1532-5415.2004.52416.x
85
Frank, B. J., Lane, C., & Hokanson, H. (2009). Designing a postepidural fall risk assessment
score for the obstetric patient. Journal Of Nursing Care Quality, 24(1), 50-54.
Gibbons, V., Esselink, T., & McHugh, S. (2013). Assessing practice relating to fall risk
management among nurses in an acute ward setting: A best practice implementation
report. JBI Database of Systematic Reviews & Implementation Reports, 11(6), 275 -
285.
Giles, L. C., Whitehead, C. H., Jeffers, L., McErlean, B., Thompson, D., & Crotty, M.
(2006). Falls in hospitalized patients - can nursing information systems data predict
falls? Computers, Informatics, Nursing, 24(3), 167-172.
Gills, A., & Jackson, W. (2002). Research for nurses: Methods and interpretation.
Philadelphia, PA: F.A. Davis Company.
Gray-Miceli, D. (2007). Fall risk assessment for older adults: The Hendrich II Fall Risk
Model. Try This: Best Practices in Nursing Care to Older Adults(8), 2.
Greenhalgh, T., Robert, G., MacFarlane, F., Bate, P., & Kyriakkidou, O. (2004). Diffusion of
innovations in service organizations: Systematic review and recommendations. The
Milbank Quarterly, 82(4), 581-629.
Grenier-Sennelier, C., Lombart, I., Jeny-Loeper, C., Maillet-Gouret, M.-C., & Minvielle, E.
(2002). Designing adverse events prevention programs using quality management
methods: The case of falls in hospital. International Journal for Quality in Health
Care, 14(5), 419-426.
Grol, R., & Grimshaw, J. (2003). From best evidence to best practice: Effective
implementation of change in patients' care. The Lancet, 362(9391), 1225-1230. doi:
10.1016/s0140-6736(03)14546-1
Haines, T. P., Hill, A.-M., Hill, K. D., McPhail, S., Oliver, D., Brauer, S., . . . Beer, C.
(2011). Patient education to prevent falls among older hospital inpatients: A
randomized controlled trial Archives of Internal Medicine, 171(6), 516-524
Haines, T. P., Hill, A. M., Hill, K. D., Brauer, S., Hoffman, T., Etherton-Beer, C., &
McPhail, S. (2013). Cost effectiveness of patient education for the prevention of falls
in hospital: Economic evaluation from a randomized controlled trial. BMC Medicine,
11(135). doi: 10.1186/1741-7015-11-135
Haines, T. P., Hill, K., Walsh, W., & Osborne, R. (2007). Design-related bias in hospital fall
risk screening tool predictive accuracy evaluations: Systematic review and meta-
analysis. Journal of Gerontology: Medical Sciences, 62A(6), 664-672.
Haines, T. P., & Waldron, N. G. (2011). Translation of falls prevention knowledge into
action in hospitals: What should be translated and how should it be done? Journal of
Safety Research, 42, 431-442. doi: 10.1016/j.jsr.2011.10.003
86
Halfona, P., Egglib, Y., Van Mellec, G., & Vagnair, A. (2001). Risk of falls for hospitalized
patients: A predictive model based on routinely available data. Journal of Clinical
Epidemiology, 54(12), 1258-1266. doi: http://dx.doi.org/10.1016/S0895-
4356(01)00406-1
Harmsen, M., Peters, M. A. J., & Wensing, M. (2005). Barriers and faciltitators asessment
instrument: Introduction , instructions and instrument Nijmegan, The Netherlands:
Radboud University Nijmegan Medical Centre.
Heafner, L. (2013). Development of a tool to assess risk for falls in women in hospital
obstetric units. Nursing For Women's Health, 17(2), 98-107. doi: 10.1111/1751-
486X.12018
Health & Disability Commissioner. (2009). The Code of Rights. Retrieved from
http://www.hdc.org.nz/the-act--code/the-code-of-rights/the-code
Health & Disability Services. (2001). Health and Disability Services Act. Retrieved from
www//hqsc.govt.nz/assets/Falls
Health Quality and Safety Commission. (2012). Serious and sentinel events. Retrieved from
http://www.hqsc.govt.nz/assets/Reportable-Events/Publications/Serious-Sentinel-
Events-Report-20102011.pdf
Health Quality and Safety Commission. (2013). Serious adverse event reports. Retrieved
from http://www.hqsc.govt.nz/our-programmes/reportable-events/serious-adverse-
events-reports/
Heinze, C., Dassen, T., Halfens, R., & Lohemann, C. (2008). Screening the risk of falls: a
general or a specific instrument? Journal of Clinical Nursing, 18, 350-356. doi:
10.1111/j.1365-2702.2008.02453.x
Hempel, S., Newberry, S., Wang, Z., Booth, M., Shanman, R., Johnsen, B., . . . Ganz, D. A.
(2013). Hospital fall prevention: A systematic review of implementation,
components, adherence, and effectiveness. Journal of the American Geriatrics
Society, 61(4), 483-494. doi: 10.1111/jgs.12169
Hendrich, A. L. (2007). Predicting patient falls. American Journal of Nursing, 107(11), 50-
58.
Hendrich, A. L. (n.d.). Upright. Retrieved from http://www.waikatodhb.health.nz/for-
health-professionals/falls-minimisation/
Hendrich, A. L., Bender, P. S., & Nyhuis, A. (2003). Validation of the Hendrich II Fall Risk
Model: A large concurrent case/control study of hospitalized patients. Applied
Nursing Research, 16(1), 9-21. doi: 10.1053/apnr.2003.YAPNR2
Hitcho, E. B. M. J., Krauss, M. A., Birge, S., Dunagan, W. C., Fischer, I., Johnson, S., . . .
Fraser, V. J. (2004). Characteristics of falls in a hospital setting: A prospective
analysis. Journal of General International Medicine 19, 732-739.
87
Holleman, G., Poot, E., Mintjes-de Groot, J., & van Achterberg, T. (2009). The relevance of
team characteristics and team directed strategies in the implementation of nursing
innovations: A literature review. International Journal of Nursing Studies, 46, 1256-
1264. doi: 10.1016/j.ijnurstu.2009.01.005
Hsaio, Y.-L. (2012). Profile of hospitalized elderly patients treated for falling. International
Journal of Gerontology, 6(1). doi: 10.1016/j.ijge.2012.01.019
Hsiao, Y.-L. (2012). Profile of hospitalized elderly patients treated for falling. International
Journal Of Gerontology, 6(1). doi: 10.1016/j.ijge.2012.01.019
Ivziku, D., Matarese, M., & Pedone, C. (2011). Predictive validity of the Hendrich fall risk
model II in an acute geriatric unit. International journal of nursing studies, 48(4),
468-474. doi: 10.1016/j.ijnurstu.2010.09.002
Jellesmark, A., Herling, S. F., Egerod, I., & Beyer, N. (2012). Fear of falling and changed
functional ability following hip fracture among community-dwelling elderly people:
an explanatory sequential mixed method study. Disability and Rehabilitation, 34(25),
2124-2131. doi: 10.3109/09638288.2012.673685
Johnson, J. E., Veneziano, T., Green, J., Howarth, E., & Malast, T. (2011). Breaking the fall.
Journal of Nursing Administration, 41(12), 538-545.
Joint Commission Resources. (2005). Reducing the risk of falls in your healthcare
organization. Illinois: Joint Commission on Accreditation of Healthcare
Organizations.
Kalisch, B. J., Landstrom, G., & Williams, R. A. (2009). Missed nursing care: Errors of
omission. Nursing Outlook, 57(1), 3-9. doi:
http://dx.doi.org/10.1016/j.outlook.2008.05.007
Kalisch, B. J., Tschannen, D., & Lee, K. H. (2012). Missed nursing care, staffing, and patient
falls. Journal of Nursing Care Quality, 27(1), 6-12. doi:
10.1097/NCQ.0b013e318225aa23
Kermode, S., & Roberts, K. (2006). Quantitative research designs. In B. Taylor, S. Kermode
& K. Roberts (Eds.), Research in nursing and health care: Evidence for practice (3rd
ed., pp. 172-198). Melbourne, Aus: Centage Learning.
Koh, S., Hafizah, N., Lee, J. Y., & Muthu, R. (2009). Impact of a fall prevention programme
in acute hospital settings in Singapore. Singapore Medical Journal, 50(4), 425-432.
Koh, S., Manias, E., Hutchinson, A. M., Donath, S., & Johnston, L. (2008). Nurses' perceived
barriers to the implementation of a Fall Prevention Clinical Practice Guideline in
Singapore hospitals. BMC Health Services Research, 8, 105. doi: 10.1186/1472-
6963-8-105
88
Krauss, M. J., Evanoff, B., Hitcho, E., Ngugi, K. E., Dunagan, W. C., Fischer, I., . . . Fraser,
V. J. (2005). A case-control study of patient, medication, and care-related risk factors
for inpatient falls. Journal of General Internal Medicine, 20(2), 116-122. doi:
10.1111/j.1525-1497.2005.40171.x
Lake, E. T., Shang, J., Klaus, S., & Dunton, N. E. (2010). Patient falls: Association with
hospital Magnet status and nursing unit staffing. Research in Nursing & Health,
33(5), 413-425. doi: 10.1002/nur.20399
Lamis, R. L., Kramer, J. S., Hale, L. S., Zackula, R. E., & Berg, G. M. (2012). Fall risk
associated with inpatient medications. American Journal of Health-System
Pharmacy, 69(21), 1888-1894. doi: 10.2146/ajhp110723
LoBiondo-Wood, G., & Haber, J. (2006). Nursing research: Methods and critical appraisal
for evidence-based practice (6th ed.). St Louis, Missouri: Mosby Elsevier.
Lohse, G. R., Leopold, S. S., Theiler, S., Sayre, C., Cizik, A., & Lee, M. J. (2012). Systems-
based safety intervention: Reducing falls with injury and total falls on an orthopaedic
ward. Journal of Bone & Joint Surgery, 94(13), 1217-1222.
Lovallo, C., Rolandi, S., Rossetti, A. M., & Lusignani, M. (2010). Accidental falls in hospital
inpatients: Evaluation of sensitivity and specificity of two risk assessment tools.
Journal of Advanced Nursing, 66(3), 690-696. doi: 10.1111/j.1365-
2648.2009.05231.x
Lucero, R. J., Lake, E. T., & Aiken, L. H. (2010). Nursing care quality and adverse events in
US hospitals. Journal of Clinical Nursing, 19(15/16), 2185-2195. doi:
10.1111/j.1365-2702.2010.03250.x
Magnussen, L. (2008). Applying the Principles of Significant Learning in the e-Learning
Environment. Journal of Nursing Education, 47(2), 82-86. doi: 10.3928/01484834-
20080201-03
Mardon, R. E., Khanna, K., Sorra, J., Dyer, N., & Famolaro, T. (2010). Exploring
relationships between hospital patient safety culture and adverse events. Journal of
patient safety, 6(4), 226-232. doi: 10.1097/PTS.0b013e3181fd1a00
Matthew-Maich, N., Ploeg, J., Jack, S., & Dobbins, M. (2010). Transformative learning and
research utilization in nursing practice: A missing link? Worldviews on Evidence-
Based Nursing, 7(1), 25-35.
McInnes, E., Seers, K., & Tutton, L. (2011). Older people's views in relation to risk of falling
and need for intervention: a meta-ethnography. Journal of advanced nursing, 67(12),
2525-2536. doi: 10.1111/j.1365-2648.2011.05707.x
McKinley, C., Fletcher, A., Biggins, A., McMurray, A., Birtwhistle, S., Gardiner, L., . . .
Lockhart, J. (2007). Evidence-based Management Practice: Reducing Falls in
Hospital. Collegian, 14(2), 20-25. doi: org/10.1016/S1322-7696(08)60551-X
89
Meek, G. (2009). Second level nurses: a critical examination of their evolving role in New
Zealand Healthcare. (Master's dissertation), Waikato Institute of Technology,
Hamilton, New Zealand. Retrieved from
researcharchive.wintec.ac.nz/962/1/Repository_entry_HLNU902.pdf (HLNU902)
Meyer, G., Köpke, S., Haastert, B., & Mühlhauser, I. (2009). Comparison of a fall risk
assessment tool with nurses' judgement alone: A cluster-randomised controlled trial.
Age & Ageing, 38(4), 417-423. doi: 10.1093/ageing/afp049
Miake-Lye, I. M., Hempel, S., Ganz, D. A., & Shekelle, P. G. (2013). Inpatient fall
prevention programs as a patient safety strategy: A systematic review. Annals of
Internal Medicine(158), 390-396.
Midwifery Council of New Zealand. (2010). Code of conduct. Retrieved
http://www.midwiferycouncil.health.nz/conduct/
Milisen, K., Coussement, J., Flamaing, J., Vlaeyen, E., Schwendimann, R., Dejaeger, E., . . .
Boonen, S. (2012). Fall prediction according to nurses' clinical judgment:
Differences between medical, surgical, and geriatric wards. Journal of the American
Geriatrics Society, 60(6), 1115-1121. doi: 10.1111/j.1532-5415.2012.03957.x
Minichiello, V., Sullivan, G., Greenwood, K., & Axford, R. (2004). Sampling in quantitative
research. In Research methods for nursing and health science (2nd ed., pp. 176-209).
French Forest NSW: Pearson/Prentice Hall.
Ministry of Health Manatū Hauora. (2013a). District Health Boards. Retrieved from
http://www.health.govt.nz/new-zealand-health-system/key-health-sector-
organisations-and-people/district-health-boards
Ministry of Health Manatū Hauora. (2013b). Enrolled nurses. Retrieved from
http://www.health.govt.nz/our-work/nursing/nurses-new-zealand/enrolled-nurses
Ministry of Health Manatū Hauora. (2013c). Responsible authorities under the Act.
Retrieved from http://www.health.govt.nz/our-work/regulation-health-and-disability-
system/health-practitioners-competence-assurance-act/responsible-authorities-under-
act
Moore, D., & Blick, G. (2013). Assessment of the implementation of the Productive Ward
and Productive Operating Theatre programmes in New Zealand. Wellington: Sapere
research group.
Morse, J. M. (2006). The Safety of Safety Research: The Case of Patient Fall Research
Canadian Journal of Nursing Research, 38(2), 74-88.
Morse, J. M. (2009). Preventing patient falls: Establishing a fall intervention program (2nd
ed.). New York: Springer Publishing Ltd.
Moule, P., & Goodman, M. (2009). Development of nursing research. In Introduction to
Nursing Research (pp. 4-25). London: SAGE Publications Ltd.
90
Myers, H. (2003). Hospital fall risk assessment tools: A critique of the literature.
International Journal of Nursing Practice, 9(4), 223-235. doi: 10.1046/j.1440-
172X.2003.00430.x
Myers, H., & Nikoletti, S. (2003). Fall risk assessment: A prospective investigation of nurses'
clinical judgement and risk assessment tools in predicting patient falls. International
Journal of Nursing Practice, 9, 158-165.
Neily, J., Howard, K., Quigley, P., & Mills, P. (2005). One-year follow-up after a
collaborative breakthrough series on reducing falls and fall-related injuries. Joint
Commission Journal on Quality & Patient Safety, 31(5), 275-285.
NHS Institute. (2012). Releasing time for care: the Productive Ward - the story so far ...
Retrieved from http://institute.nhs.uk.productiveseries
Nursing Council of New Zealand. (2011). Workforce statistics Nursing Retrieved from
http://nursingcouncil.org.nz/Publications/Reports
Nursing Council of New Zealand. (2012). Code of Conduct. Retrieved from
http://nursingcouncil.org.nz/Nurses/Code-of-Conduct
O'Connell, B., & Myers, H. (2001). A failed fall prevention study in an acute are setting:
Lessons from the swamp. International Journal of Nursing Practice, 7, 126-130.
Oliver, D. (2007). Preventing falls and fall injuries in hospital: A major risk management
challenge. Clinical Risk, 13, 173-178.
Oliver, D. (2008). Falls risk-prediction tools for hospital inpatients: Time to put them to bed?
Age and Ageing, 37, 248-250. doi: 10.1093/ageing/afn088
Oliver, D., Daly, F., Martin, F. C., & McMurdo, M. E. (2004). Risk factors and risk
assessment tools for falls in hospital in-patients: A systematic review. Age and
Ageing, 33(2), 122-130. doi: 10.1093/ageing/afh017
Oliver, D., Killick, S., Even, T., & Wilmott, M. (2008). Do falls and falls-injuries in hospital
indicate negligent care - and how big is the risk? . Quality Safe Health Care, 17, 431-
436. doi: 10.1136/qshc.2007.024703
Oliver, D., Papaioannou, A., Giangregorio, L., Thabane, L., Reizgys, K., & Foster, G. (2008).
A systematic review and meta-analysis of studies using the STRATIFY tool for
prediction of falls in hospital patients: How well does it work? Age and Ageing, 37,
621-627. doi: 10.1093/ageing/afn203
Page, A. (Ed.). (2004). Keeping patients safe : Transforming the work environment of nurses.
Washington, DC: National Academies Press.
Pallant, J. (2007). Choosing the right statistics. SPSS Survival Manual (3rd ed., pp. 100-125).
Crows Nest, NSW: Allen & Unwin.
91
Parsons, M. L., & Cornett, P. A. (2011). Sustaining the pivotal organizational outcome:
Magnet recognition. Journal of Nursing Management, 19, 277-286. doi:
10.1111/j.1365-2834.2011.01224.x
PASW Statistics Student Version 18.0. (2009). Chicago, USA.
Pentland, D., Forsyth, K., Maciver, D., Walsh, M., Murray, R., Irvine, L., & Sikora, S.
(2011). Key characteristics of knowledge transfer and exchange in healthcare:
Integrative literature review. Journal of Advanced Nursing, 67(7), 1408-1423. doi:
10.1111/j.1365-2648.2011.05631.x
Perell, K. L., Nelson, A., Goldman, R. L., Luther, S. L., Prieto-Lewis, N., & Rubenstein, L.
Z. (2001). Fall Risk Assessment Measures: An Analytic Review. The Journals of
Gerontology Series A: Biological Sciences and Medical Sciences, 56(12), M761-
M766. doi: 10.1093/gerona/56.12.M761
Peters, M. A. J., Harmsen, M., Laurant, M. G. H., & Wensing, M. (2002). Barriers to and
facilitators for improvement of patient care Retrieved from www.iqhealthcare.nl
Plitcha, S. B., & Garzon, L. S. (2009). Descriptive statistics: Getting to know the data.
Philadelphia: Wolters Kluwer/Lippincott, Williams & Wilkins.
Purdy, N., Laschinger, H. K. S., Kerr, M., & Olivera, F. (2010). Effects of work
environments on nurse and patient outcomes. Journal of Nursing Management, 18,
901-913. doi: 10.1111/j.1365-2834.2010.01172.x
Quigley, P. A., Hahm, B., Gibson, W., Powell-Cope, G., Sarduy, I., Tyndall, K., & White, S.
V. (2009). Reducing serious injury from falls in two veterans' hospital medical-
surgical units. J Nurs Care Qual, 24(1), 33-41.
Rattray, J., & Jones, M. C. (2007). Essential elements of questionnaire design and
development. Journal of Clinical Nursing(16), 234–243. doi: 10.1111/j.1365-
2702.2006.01573.x
Raymont, A., Graham, P., Hider, P. N., Finlayson, M. P., Fraser, J., & Cumming, J. M.
(2012). Variation in the adoption of patient safety practices among New Zealand
district health boards. Australian Health Review, 36(2), 163-168. doi:
http://dx.doi.org/10.1071/AH10972
Razmus, I., Wilson, D., Smith, R., & Newman, E. (2006). Falls in hospitalized children.
Pediatric Nursing, 32(6).
Robert, G., Morrow, E., Maben, J., Griffiths, P., & Callard, L. (2011). The adoption, local
implementation and assimilation into routine nursing practice of a national quality
improvement programme: The Productive Ward in England. Journal of Clinical
Nursing, 20(7/8), 1196-1207. doi: 10.1111/j.1365-2702.2010.03480.x
Rogers, J., & Niven, E. (2003). Ethics: A guide for New Zealand nurses (2nd ed.). Auckland:
Pearson Education New Zealand.
92
Rucker, D., Rowe, B. H., Johnson, J. A., Steiner, I. P., Russell, A. S., Hanley, D. A., . . .
Majumdar, S. R. (2006). Educational intervention to reduce falls and fear of falling
in patients after fragility fracture: Results of a controlled pilot study. Preventive
Medicine, 42(4), 316-319. doi: http://dx.doi.org/10.1016/j.ypmed.2006.01.008
Rush, K. L., Robey-Williams, C., Patton, L. M., Chamberlain, D., Bendyk, H., & Sparks, T.
(2008). Patient falls: acute care nurses' experiences. Journal of Clinical Nursing, 18,
357-365. doi: 10.1111/j.1365-2702.2007.02260.x
Schultz, T. J., & Kitson, A. L. (2010). Measuring the context of care in an Australian acute
care hospital: A nurse survey. Implementation Science, 5(60), 1 - 11. doi:
10.1186/1748-5908-5-60
Schwendimann, R., Bühler, H., De Geest, S., & Milisen, K. (2006). Falls and consequent
injuries in hospitalized patients: Effects of an interdisciplinary falls prevention
program. BMC Health Services Research, 6(69), 1-7. doi: 10.1186/1472-6963-6-69
Scott, Mannion, R., Davies, H., & Marshall, M. (2003). Implementing culture change in
health care. International Journal of Quality in Health Care, 15(2), 111-118. doi:
10.1093/intqhc/mzg021
Semin-Goosens, A., van der Helm, J. M. J., & Bossuyt, P. M. M. (2003). A failed model-
based attempt to implement an evidence-based nursing guideline for fall prevention.
Journal of Nursing Care Quality, 18(3), 217-225.
Shever, L. L., Titler, M. G., Mackin, M. L., & Kueny, A. (2011). Fall prevention practices in
adult medical-surgical nursing units described by nurse managers. Western Journal
of Nursing Research, 33(3), 385-397. doi: 10.1177/0193945910379217
Shorr, R. I., Mion, L. C., Chandler, A. M., Rosenblatt, L. C., Lynch, D., & Kessler, L. A.
(2008). Improving the capture of fall events in hospitals: Combining a service for
evaluating inpatient falls with an incident report system. Journal of American
Geriatrics Society . 56, 701-704.
Sitterding, M. C., Broome, M. E., Everett, L. Q., & Ebright, P. (2012). Understanding
situation awareness in nursing work: A hybrid concept analysis. Advances in Nursing
Science, 35(1), 77-92 10.1097/ANS.1090b1013e3182450158.
Smith, I. J. (Ed.). (2005). Reducing the risk of falls in your health organization. Illinois: Joint
Commission on Accreditation of Healthcare Organizations.
Spenceley, S. M., O'Leary, K. A., Chizawsky, L. L. K., Ross, A. J., & Estabrooks, C. A.
(2008). Sources of information used by nurses to inform practice: An integrative
review. International Journal of Nursing Studies, 45, 945-970. doi:
10.1016/j.ijnurstu.2007.06.003
93
Spoelstra, B. L., Given, B. A., & Given, C. W. (2012). Falls prevention in hospitals: An
integrative review. Clinical Nursing Research, 21(1), 92-112. doi:
10.1177/1054773811418106SAGE
Stenberg, M., & Wann-Hanson, C. (2011). Health care professionals' attitudes and
compliance to clinical practice guidelines to prevent falls and fall injuries.
Worldviews on Evidence-Based Nursing, 8(2), 87-95.
Sushil, S., & Verma, N. (2010). Questionnaire Validation Made Easy. European Journal of
Scientific Research, 46(2), 172-178. http://www.eurojournals.com/ejsr.htm
Swanson, J. W., & Tidwell, C. A. (2011). Improving the culture of patient safety through the
Magnet® journey. Online Journal of Issues in Nursing, 16(3), 1-1. doi:
10.3912/OJIN.Vol16No03Man01
Swartzell, K. L., Fulton, J. S., & Friesth, B. M. (2013). Relationship between occurrence of
falls and fall-risk scores in an acute care setting using the Hendrich II Fall Risk
Model. Medsurg Nursing, 22(3), 180-187.
Terrell, K. M., Weaver, C. S., Giles, B. K., & Ross, M. J. (2009). ED patient falls and
resulting injuries. Journal of Emergency Nursing 35(2), 89-92. doi:
10.1016/j.jen.2008.01.004
Titler, M. G., Dochterman, J., Picone, D. M., Everett, L., Xie, X.-J., Kanak, M., & Fei, Q.
(2005). Cost of hospital care for elderly at risk of falling. Nursing Economics, 23( 6).
Titler, M. G., Shever, L. L., Kanak, M. F., Picone, D. M., & Qin, R. (2011). Factors
associated with falls during hospitalization in an older adult population. Research
and Theory for Nursing Practice: An International Journal 25(2), 127-152. doi:
10.1891/1541-6577.25.2.127
Tucker, S. J., Bieber, P. L., Attlesey-Pries, J. M., Olson, M. E., & Dierkhising, R. A. (2012).
Outcomes and challenges in implementing hourly rounds to reduce falls in
orthopedic units. Worldviews on Evidence-Based Nursing, 9(1), 18-29. doi:
10.1111/j.1741-6787.2011.00227.x
Tzeng, H.-M. (2011). Link between patients' perceptions of their acute care hospital
experience and institutions' injurious fall rates. Journal Of Nursing Care Quality,
26(2), 151-160.
Tzeng, H.-M., & Yin, C.-Y. (2008). Nurses' solutions to prevent inpatient falls in hospital
patient rooms Nursing Economics, 26(3), 179-187.
United States Patent and Trademark Organisation, (2014). Mission. Retrieved from uspto.gov
van Harten-Krouwel, D., Schuurmans, M., Emmelot-Vonk, M., & Pel-Littel, R. (2011).
Development and feasibility of falls prevention advice. Journal of Clinical Nursing,
20, 2761-2776. doi: 10.1111/j.1365-2702.2011.03801.x
94
Vassallo, M., Raj, V., C., J., Sharma, J. C., Hallam, H., Binns, K., . . . Allen, S. (2004). The
effect of changing practice on fall prevention in a rehabilitative hospital: The hospital
injury prevention study. Journal of the Amercian Geriatrics Society, 52(3), 335-339.
Vassallo, M., Stockdale, R., Jagdish, C. S., Briggs, R., & Allen, S. (2005). A comparative
study of the use of four fall risk assessment tools in on acute medical wards. Journal
of American Geriatrics Society . 53(6), 1034-1038. doi: 10.1111/j.1532-
5415.2005.53316.x
Wadensten, B., & Carlsson, M. (2007). Adoption of an innovation based on the theory of
gerotranscendence by staff in a nursing home - part III. International Journal of
Older People Nursing, 2(4), 302-314.
Waikato District Health Board. (2012a). Fall risk assessment, minimisation and management
policy.
Waikato District Health Board. (2012b). Priorities for 2011/12. Retrieved from
http:www.waikatodhb.govt.nz/file/filied/42814
Waikato District Health Board. (2013a). The provider arm - Health Waikato. Retrieved
from
http://www.waikatodhb.govt.nz/page/pageid/2145849111/Health_Waikato#sthash.27
Hzdx8u.dpuf
Waikato District Health Board. (2013b). Quality improvement. Retrieved from
http://www.waikatodhb.govt.nz/file/fileid/45992.pdf
Webster, J., Courtney, M., Marsh, N., Gale, C., Abbott, B., Mackenzie-Ross, A., & McRae,
P. (2010). The STRATIFY tool and clinical judgment were poor predictors of falling
in an acute hospital setting. Journal of Clinical Epidemiology, 63(1), 109-113. doi:
http://dx.doi.org/10.1016/j.jclinepi.2009.02.003
Williams, J., Kaasalainen, S., Hadjistavropoulos, T., Scudds, R., Thorpe, L., Neville, S., . . .
Andersen, D. (2011). A qualitative investigation of injurious falls in long-term care:
Perspectives of staff members. Disability and Rehabilitation, 33(5), 423-232. doi:
10.3109/09638288.2010.498555
Wilson, D. S. (2013). Registered nurses' collective safety organising behaviours: The
association with perceptions of patient safety culture. Journal of Research in
Nursing, 18( 4), 320-333. doi: 10.1177/1744987112461781
World Health Organization (WHO). (2007). WHO global report on falls prevention in older
age. Retrieved from http://www.who.int/falls/publications
Wright, K. M. (2006). Falling head over heels: Reducing falls in high risk neurosurgical
inpatients with the implementation of a 'high risk falls room'. Australasian Journal of
Neuroscience, 18(1), 3-7.
95
Zecevic, A. A., Salmoni, A. W., Speechley, M., & Vandervoort, A. A. (2006). Defining a fall
and reasons for falling: Comparisons among the views of seniors, health care
providers, and the research literature. The Gerontologist, 46(3), 367-376. doi:
10.1093/geront/46.3.367
96
Appendix 1 Hendrich II Fall Risk Model
and A. Hendrich Personal communication
Risk
points
Confusion
Disorientation 4
Impulsivity
Symptomatic 2
Depression
Altered 1
Elimination
Dizziness 1
Vertigo
Male 1
Gender
Any
Administered 2
Antiepilectics
Any
Administered 1
Benzodiazepines
©AHI of Indiana Inc. All Rights Reserved. US Patent (US20050182305) has been allowed.
Reproduction and use prohibited except by written permission from AHI of Indiana Inc.
Hendrich, A. L. (2007). Predicting patient falls. American Journal of Nursing, 107(11), 50-58.
97
Personal Communication from A. Hendrich
To:
Elizabeth MacColl
Hello Elizabeth
The numerous studies were done initially in a large, tertiary, medical center in a randomized,
diverse, population. The validity and reliability has been validated in numerous replication
studies in the US and multiple countries. Rehabilitation, ED, behavioral health, ICU, med/surg,
etc. were all randomized in the case-control methodology. The study does not apply to OB and
pediatrics.
Regards,
Ann Hendrich PhD, RN, FAAN
SVP, Quality and Safety
Executive Director, PSO
314-733-8187
Executive Assistant, Gabriele Nash
314-733-8187
Ascension Health
Sent from my iPad
98
Appendix 2 Information Sheet
A Study on the Barriers to the Implementation of the Hendrich II Fall Risk Model
Kia ora,
I am a Masters student at Massey University and a nurse working in orthopaedics at Waikato Hospital. I
would like to invite you to participate in a survey into the use of Hendrich II Fall Risk Model. This research
project is part of my Masters programme of study.
I am distributing the questionnaires for you to complete and return anonymously. Completion and return
of the questionnaire implies that you consent for me to use your responses for my research.
How you complete the survey, or whether you choose not to, will have no impact on your employment
with Waikato DHB. You do not have to answer any or all of the questions.
Patient falls are a concern to all of us. Every RN, RM and EN who is expected to routinely complete the
Hendrich II Fall Risk Model will be receiving the questionnaire. The information gathered will help
identify factors that are not working well, or which are working well, of the model. While it is not
compulsory to complete the questionnaire, the more everybody contributes, the better the data and the
more meaningful the conclusions.
It will take about 15 minutes to complete the questionnaire and I hope you will be able to find that time
within the usual duties that you will be doing today.
The survey is based on a validated questionnaire to identify barriers and enablers for a new assessment
in a health care setting1. If you want to talk to me about any issues regarding this survey, I can be
contacted through Ward 6, or 07 8716716, or at [email protected].
My Supervisor and I will be the only ones who will see your returned questionnaires and I will be the one
analysing them. I will ensure that the paper questionnaires stay locked away until they are destroyed,
and that digitally stored data will be password protected until it too is destroyed.
I hope next year that you will read about the results. I intend to have them on the intranet, for them to
be available in your ward or area, to talk about them at the Nursing Round Table, and eventually, for the
thesis to be published and be available in the library. A report will be provided to the WDHB Board.
Participant’s Rights
You are under no obligation to accept this invitation. If you decide to participate, you have the right to:
x decline to answer any particular question;
x withdraw from the study (before data analysis begins in November 2012);
x ask any questions about the study at any time during participation;
x provide information on the understanding that your name will not be used unless you give
permission to the researcher;
x be given access to a summary of the project findings when it is concluded.
Ka kite anō
This project has been reviewed and approved by the Massey University Human Ethics Committee: Northern, Application MUHECN
12/059. If you have any concerns about the conduct of this research, please contact Dr Ralph Bathurst, Chair, Massey University
99
Human Ethics Committee: Northern, telephone 09 414 0800 x 9570, email [email protected]. 99
Appendix 3 Survey and demographic questions
Do not
Fully agree or
disagree Disagree disagree Agree
The Hendrich II Fall Risk Model ...
100
Do not
Fully agree or
disagree Disagree disagree Agree
Doctor/s
Physiotherapist/s
Occupational Therapist/s
Practice Setting
Acute Medical Acute Surgical (please state)
Position
Registered Nurse Registered Midwife
Enrolled Nurse
Any further comment about how you use the Hendrich II Fall Risk Model …
Thanks
ˈPeters, M. A. J., Harmsen, M., Laurant, M. G. H., & Wensing, M. (2002). Barriers to and facilitators for improvement of patient care
Retrieved from www.iqhealthcare.nl
Appendix 4 Permission to use Barriers and Facilitators Assessment Instrument
4/17/12
[email protected]
to M.Harmsen, me
Dear Liz,
Please find attached an English version of the barriers and facilitators assessment instrument. It
is no problem that you use the instrument, as long as you refer to it properly.
Best wishes,
Mirrian Hilbink
6500 HB NIJMEGEN
Tel: 024-361 96 41
Fax: 024-354 01 66
E-mail: [email protected]
Internet: www.iqhealthcare.nl
103
Appendix 5 Category, characteristic and survey questions
Question
Category Characteristic Number Question in survey
Involvement 16 The Hendrich II Fall Risk Model does not fit into
my way of preventing patient falls.
104
Question
Category Characteristic Number Question in survey
Group norms, socialisation/ Nurses in my team do not cooperate in applying the
9
nurse Hendrich II Fall Risk Model.
Group norms, socialisation/ Doctors do not cooperate in applying the Hendrich
doctor 10 II Fall Risk Model.
Group norms, socialisation/ Physiotherapists do not cooperate in applying the
physio 10 Hendrich II Fall Risk Model.
Group norms, socialisation/ Occupational therapists do not cooperate in applying
OT 10 the Hendrich II Fall Risk Model.
Context Group norms, socialisation/ Managers do not cooperate in applying the Hendrich
manager 11 II Fall Risk Model.
105
Appendix 6 Ethical approval
106
Appendix 7 Māori consultation
1 August 2012
Liz MacColl
Registered Nurse
Waikato District Health Board
Thank you for submitting your proposal on the 10 July 2012 Kaumatua Kaunihera
Ethics Research Committee meeting. As this was a different process, the Kaumatua
Kaunihera Ethics Research committee and accepted your proposal and we are very
pleased to provide you support to your research with an invitation to:
We wish you well with the study and should you have any further queries please contact
me.
Noho ora mai ra
Rose Smith
Kaumātua Kaunihera Sub-Committee representative
Manaaki Nepia
Service Development Manager
Te Puna Oranga-Maori Health Service
[email protected]
07 834 3628
021 415 062
107
Appendix 8 DHB authorisation
108
109
110
Appendix 9 Gamma correlations
The following bullet point are a list the correlations that have been calculated as greater than 0.3.
The characteristics and the questions related to them are in Appendix 5.
* Questions were reverse coded before the correlation calculation was made.
112