RESEARCH REPORT

Innovation in Data-Driven
Health Care

Sponsored by
SPONSOR PERSPECTIVE

With today’s introduction of advanced digital technologies, there is a

significant opportunity to transform health care. The vision is that by
connecting diverse data points, clinical insights will be generated that help
improve patient care, population health, and the ability to identify disease at its
earliest stages. For health care settings, connected operational data can also
generate valuable insights to realize time and cost efficiencies, improving the
overall clinician and patient experience in other important ways.

At the center of the opportunity to transform the diagnosis, treatment, and

prevention of diseases stands the discipline of laboratory medicine, which Corinne Dive-Reclus
analyzes the test results of an individual’s clinical samples to determine health
Global Head of Lab Insights
status. Over the past 125 years, Roche has grown into one of the world’s largest
Roche Information Solutions
biotech companies as well as a leading provider of in vitro diagnostics and
transformative innovative solutions across many disease areas. Roche Diagnostics

Every day, we work with leaders in laboratories and hospitals worldwide to

ensure tests that inform decisions about an individual’s health status are
scientifically accurate and adhere to rigorous standards of quality control.
In 2022, 29 billion tests worldwide were performed with Roche Diagnostic
products, delivering crucial information on disease prevention, diagnosis,
and treatment.1

At Roche, we are also introducing new digital technologies at every step of the
patient care continuum—prevention, diagnosis, treatment, and monitoring of
health and disease. With high worldwide incidence of diseases such as cancer
and heart disease, early detection and treatment can become crucial factors
for improving patient outcomes and survival rates as well as reducing costs
and relieving human suffering. Because health care is complex and local to Matthew Manley
every country, we also believe it takes everyone working together to solve these Vice President of Digital
challenges. The recent pandemic highlighted the importance of collaboration Healthcare Solutions
to connect data for prevention and treatment efforts. It also revealed the Roche Diagnostics Corporation/
current gaps in health care—data and system interoperability challenges—a USA
remaining pain point also highlighted by this report. Roche Diagnostics

The creation of open digital ecosystems is one approach being used by

Roche to address interoperability and to ease data access using the latest
innovations. As strategic partners, we are bringing tech companies and health
care leaders together to address the unique needs and priorities of health
care while creating solutions based on flexible, scalable ideas and principles.
With our multidisciplinary teams from health care and technology, we aim to
bring worldwide best practices and knowledge to local health systems and to a
wider population. To ensure that digital innovations and the data they generate
are used accurately, responsibly, and ethically in health care, we are taking a
holistic and science/health-based approach at Roche, aligning carefully with
data privacy, government, and other regulations, that puts patient needs at the
center of care.

Finally, the introduction of digital health solutions represents a significant step

toward early detection and prevention of disease in an operationally efficient
way. With their accuracy, accessibility, and affordability, we firmly believe that
digital technologies have the potential to transform health care by using data
in the service of better patient care.

1 Roche Annual Report 2022, page 17 and page 35, https://www.roche.com/investors/annualreport22/#c56c7547-02d2-4bfd-8ba7-f1b4f17fd2eb

Innovation in Data-Driven
Health Care

Innovative uses of data in health care are helping

solve the most challenging problems in patient HIGHLIGHTS

health and operational efficiency. Today, many health

care organizations understand that a data-driven 94% of respondents agree that
approach can improve patient health outcomes, enable data-driven health care creates
new opportunities for patients
faster clinical decisions, and improve treatment and and doctors to benefit from
more personalized health care
hospital workflows. approaches.

These benefits of data-driven health care are widely understood and prioritized
today. In January 2023, Harvard Business Review Analytic Services conducted 64% of respondents agree
a global survey of 757 members of the Harvard Business Review audience who that digital technologies for
work in the health care industry or in a health care–related industry and are health care have made data
familiar with their organization’s use of health data and digital technologies integration easier.
for research, diagnosis, and/or treatment decisions. The survey finds that 94%
of respondents agree that data-driven health care creates new opportunities
for patients and doctors to benefit from more personalized health care 43% of respondents say one
of the greatest inhibitors of
approaches. Seventy-four percent of respondents say their organization has
becoming more data driven is
spent more effort on its approach to data integration/interoperability in the disconnected or incompatible
past three years, and 72% say their organization has spent more effort on its systems/data.
approach to data strategy.
Despite this stronger focus on data, most organizations continue to fall short
of data-based decision-making maturity: just 16% describe their organization
as mature in this area. Mature organizations, which are defined in this report
as leaders, are able to access, integrate, and analyze data from diverse sources
and make decisions quickly. Maturing organizations, defined as followers, have
mostly digital data—the electronic presentation of information in a format
Due to rounding, some figures in this report may not
or language that machines can read or understand—but there are still gaps
add up to 100%.
in the workflow that slow down diagnosis and decision making. Followers
comprise 56% of respondents. Immature organizations, defined in this report
as laggards, use some electronic systems or technology to collect and manage

Research Report | Innovation in Data-Driven Health Care

1
 Harvard Business Review Analytic Services

data, but there’s no easy way to identify and pull together

relevant data quickly. Twenty-eight percent of respondents
are categorized as laggards. FIGURE 1
The number of mature organizations has changed only
nominally over the past four years: in a Harvard Business
“You can’t run a profitable health care
Review Analytic Services survey of 742 health care leaders
conducted in 2019, 15% described their organization as mature organization without leveraging data
in this capacity. in novel ways. You can’t see patients
Certain challenges are standing in the way of health care you need to see without leveraging the
organizations becoming more data driven. Disconnected
or incompatible systems and data are making integration
data of the past to find care journeys
and interoperability difficult or impossible. Organizational that are optimal for patients in the
silos are inhibiting collaboration, and funding challenges are future. And you can’t reduce burnout
obstructing innovation and critical infrastructure updates
that support emerging data capabilities. A lack of analytic
in your clinicians unless you are
skills among frontline professionals and concerns about data looking at the data to determine what
security are also factors stalling progress. clinician needs to see what patients,”
As health care organizations continue to accelerate their
use of digital health tools, there’s an urgency to prioritize
says John Halamka, MD, president of
improved data capabilities. This report explores the data Mayo Clinic Platform.
and technology impacts of the past four years on health care
organizations, how they are managing clinical and operational

data across settings today, new and persistent challenges that

organizations are experiencing, today’s enduring privacy and
security concerns, and how health care organizations are
activating the insights derived from data to meet operational
and clinical challenges.
“You can’t run a profitable health care organization without
leveraging data in novel ways. You can’t see patients you
need to see without leveraging the data of the past to find
care journeys that are optimal for patients in the future.
And you can’t reduce burnout in your clinicians unless you
are looking at the data to determine what clinician needs
to see what patients,” says John Halamka, MD, president of
Rochester, Minn.-based Mayo Clinic Platform, a health care
and technology innovation think tank. “This is why data is
so critical.”

Evolving Data-Driven Health Care

Health care organizations have withstood tremendous
challenges over the past three years, emerging from the Covid-
19 pandemic with improved data management strategies
and clearer visions for boosting operational efficiency and
productivity, enabling faster decision making, and enhancing
customer satisfaction.
As all organizations strive to become more data driven,
however, certain obstacles are inhibiting progress. Forty-
three percent of respondents rank disconnected or
incompatible systems/data as a top challenge organizations

Research Report | Innovation in Data-Driven Health Care

2
 Harvard Business Review Analytic Services

face. Organizations are also challenged by a range of other

factors, including insufficient funding (42%), organizational
silos/lack of collaboration across the business (37%), and a
lack of analytic skills among frontline professionals (33%).
The top barrier cited by 40% of leaders and 44% of followers
is disconnected or incompatible systems/data, yet the top
barrier cited by 42% of laggards is insufficient funding. Each
of these, excluding organizational silos/lack of collaboration
across the business, has shown a slight increase since this same
question was asked in an April 2019 survey of the Harvard
Business Review audience. FIGURE 2
Digital technologies and partnerships are key to solving
these challenges. While organizations are finding electronic
medical records (EMRs) across institutions (49%), EMRs
within their institution (46%), and in-hospital digital patient
monitoring (34%) “extremely useful” among the data domains
and tools they routinely use, they’re also partnering with a
number of external entities for capabilities like data analytics,
artificial intelligence (AI), and machine learning (ML). For
some, digital health care startups are filling a niche and
supplying some of the most cutting-edge data capabilities.
As organizations embrace new partnerships and work to
solve these challenges in pursuit of better patient health
and operational efficiency, many are keenly aware of the
data privacy and security risks that data integration and
management can pose. Striking a balance between risk and
reward while exploring new secure and innovative ways to
access and share data will be key.

Digital Transformation Accelerates

through the Pandemic
Health care organizations had been digitally transforming
long before the start of the Covid-19 pandemic. As it swept
across the globe in March 2020, organizations scrambled to
put in place critical infrastructure, technology, and processes.
“Covid didn’t just accelerate digital transformation—
it pushed it off a cliff,” says Philip Bradley, digital health
strategist, validation and analytics services, at the Healthcare
Information and Management Systems Society (HIMSS),
a Chicago-based not-for-profit research and consulting
company. “Health care organizations had been easing toward
all these new technologies and models, and all of a sudden
the land was gone beneath them,” he says.
The sudden pivot from in-person care to telehealth drove
change. Sixty-six percent of respondents agree that the
pandemic allowed their organization to incorporate digital
technologies that enabled better patient care from anywhere.
Incorporating these digital technologies is a hallmark of
leaders; 42% strongly agree that these technologies helped
enable better care, compared with 24% of followers and 13%
of laggards.

Research Report | Innovation in Data-Driven Health Care

3
 Harvard Business Review Analytic Services

“Covid brought us together. It gave us this notion that boy—unless we

work together on problems, we’re not going to make any progress. All
that competitive angst across organizations melted away,” says Mayo
Clinic Platform’s Halamka.

“We’ve had to figure out new ways to deliver care to our Cesar Nomura, director of diagnostics at the hospital, says
patients using digital means. We’ve had to figure out ways that this app was vital to managing the influx of patients
to accelerate our science and discovery programs using during Covid and managing physician availability. “In Brazil,
digital data. And we’ve had to deploy these programs in a physicians don’t just work at one hospital like many do in
way that is respectful to our patients, ethical, and focused the U.S. Here, we have physicians that work in two, three,
on concepts like health equity,” says Jeff Ferranti, MD, chief or four hospitals,” he says. “So if we’re able to separate the
digital officer and senior vice president at Durham, N.C.- patients who truly need to see a doctor from the patients who
based Duke University Health System. can be seen via telehealth, that helps us to better manage
Ferranti says that before the Covid-19 pandemic, Duke things on the physician side.”
University Health System provided care for an average of five The pandemic expedited the adoption of new digital
to 10 patients a week via telehealth. Today, the organization technologies and enabled better patient care from anywhere.
sees approximately 5,000 patients weekly in this fashion. Fifty-six percent of respondents indicate that the pandemic
Evolving telehealth from what he calls a “project on the side” accelerated data integration for their organizations, and 43%
into what now accounts for more than 10% of its clinical agree that because of the pandemic their organization was
volume wasn’t necessarily simple. able to break down data silos. Leaders (52%), in particular,
“It’s a completely different way to deliver care,” he says. agree that their organization was able to break down silos
“We had to think about the data we needed to take care of because of the pandemic, more than followers (46%) and
those patients, how to investigate the efficacy of taking laggards (31%).
care of those patients, and whether telehealth is just as Mayo Clinic Platform’s Halamka says these changes were
good as in-person visits—all these questions and uses of born in part of necessity. “Before Covid, organizations
data came up.” didn’t want to work with one another out of fear of loss of
These considerations prompted many organizations to intellectual property or fear of data leakage,” he says. “Covid
implement new digital technologies, the survey finds. Fifty- brought us together. It gave us this notion that boy—unless
seven percent of respondents say their organization adopted we work together on problems, we’re not going to make any
new digital tools during the pandemic for managing data, progress. All that competitive angst across organizations
of which 31% say their organization significantly improved melted away.”
its operational insights. Among the 64% of leaders that The pandemic also served as an impetus for health care
adopted new tools during the pandemic, 49% say their organizations to examine the role of data and who has access
organization significantly improved its operational insights. to it, says David Rankin, director of clinical governance and
Comparatively, among the 59% of followers that adopted new informatics at Malvern, Australia-based Cabrini Health, a
tools during that time, 31% say their organization significantly not-for-profit private hospital. Before Covid, data specialists
improved its operational insights. Of the 50% of laggards that at Cabrini Health provided management with the data
adopted new tools, 19% say their organization significantly that management thought was important, he says, which
improved its operational insights. predominantly reflected its needs. This data included metrics
At Hospital Sírio-Libanês in São Paulo, Brazil, a hospital like nursing hours per occupied bed, the overall complication
with 10,000 employees and 520 patient beds, clinicians rate, length of stay, and deaths.
needed a new way to triage patients seeking emergency care “Management doesn’t have the ability to change patient
during the height of Covid. They developed and launched an care, though. We’ve matured a lot now, and we’re starting
app that prompts patients with a series of questions to gauge to give data back to frontline staff—the doctors, nurses, and
symptoms and pain level, for example, and then, depending allied health staff—who actually care for the patients and
on the responses, routes the patient to a telehealth visit, to a empower them to understand what can be changed,” Rankin
nurse for additional questioning, to a doctor, or ultimately says. Putting this data into their hands has highlighted areas
to the emergency department. of improvement for clinicians.

Research Report | Innovation in Data-Driven Health Care

4
 Harvard Business Review Analytic Services

Expanding access to data and digital technologies is

important to health care organizations. Ninety-three percent
of respondents agree that researchers, hospital/lab leaders,
and staff want easy, integrated access to operational data,
while 89% agree that providers increasingly want access to
digital tools to help provide care to patients. The challenge
is in integrating and managing these technologies and data.

Data Integration Challenges

and Technology’s Role
One lesson learned during the Covid-19 pandemic is that
speed is of the essence, says Duke University Health System’s
Ferranti. “We don’t have 17 years to come up with a vaccine
or viable treatments. We’ve got to do things much faster and
much more efficiently,” he says. “That requires using our data
in ways that we haven’t traditionally used it before. We have
an opportunity now to use multimodal data from a variety of
sources—not just notes and data points from the [electronic
health record], but also imaging, digital pathology, genomics,
and social determinants of health data.”
Using multimodal data—or data that spans different
types and contexts—has been a primary focus for health useful application and try to extract the data from the back
care organizations since the start of the Covid-19 pandemic. end into our data warehouse so we can collate it up, it’s a
Seventy-four percent of respondents say their organization nightmare and an enormous amount of work. The lack of
has spent more effort on its approach to data integration and data standards around applications is almost crippling our
interoperability in the past three years. The top two reasons ability to intelligently collect data.”
why: to improve operational efficiency and productivity (89%) Health care organizations in the U.S. are facing similar
and to enable faster decision making (67%). challenges, Ferranti says. While there have been efforts to
“The question is, how do you take your institutional data build standards that define data in a certain way, including
assets that are sitting in warehouses around your organization FHIR—Fast Healthcare Interoperability Resources—it hasn’t
and de-identify them and make them available so we can do been enough. FHIR defines how health care information can
research at scale and investigate problems, so instead of taking be exchanged between different computer systems regardless
17 years to develop a solution, it takes 17 days?” Ferranti says. of how it is stored in those systems, allowing health care
That’s a scenario that few health care organizations have information, including clinical and administrative data, to
mastered. While 96% of respondents agree that managing be available securely to those who have a need to access it.
health/clinical data across care settings—including labs, FHIR was established by Health Level Seven International,
hospitals, doctors’ offices, clinics, and patients’ homes—is a not-for-profit organization accredited by the American
very important to the future of health care, only 25% indicate National Standards Institute.
that their organization is very successful at it. Similarly, 89% “There’s a disconnect between the informaticians who
of respondents agree that managing operational data across create standards like that, the day-to-day technical teams,
care settings is very important, while just 22% indicate that and users of that data,” Ferranti says. “There’s a steep learning
their organization is very successful at it. FIGURE 3 curve that as an industry, we just haven’t gotten over yet.”
Health care organizations are encountering a range of That’s not the case everywhere, however. The National
factors that make data integration and management difficult, Health Authority in India, for example, announced the launch
from varying data standards across systems to usability of the Unified Health Interface, a network of open protocols
challenges. Australia-based Cabrini Health, for example, designed to enable discoverability and interoperable digital
is contending with a lack of data standardization, which health service delivery, says Bhagwati Prasad, chief operating
makes integrating data from various systems painstakingly officer at the Koita Centre for Digital Health at the Indian
complicated. Institute of Technology–Bombay in Mumbai. This network
“Each developer uses their own data tables, data standards, will enable patients and health service providers on different
and data processes,” Rankin says. “When we purchase a very systems to seamlessly interact.

Research Report | Innovation in Data-Driven Health Care

5
 Harvard Business Review Analytic Services

“Today, if a doctor has a patient with data at five different records within the institution (64%), in-hospital digital patient
hospitals, it’s not possible to access all that data at once monitoring (48%), and digital tools for clinical decision
because no one wants to share it and there are no protocols support (46%) “extremely useful.” In contrast, followers and
to share it,” Prasad says. laggards that use the same data and tools are less likely to say
Integration and interoperability challenges arise from they are extremely useful—electronic medical records across
other factors, too. At Penda Health, a primary and urgent institutions (49% for followers, 35% for laggards); electronic
care center with 380 employees across 19 branches based in medical records within the institution (47%, 33%); in-hospital
Nairobi, Kenya, data integration difficulties stem from data digital patient monitoring (35%, 22%); and digital tools for
quality problems and laborious—and sometimes pricey— clinical decision support (24%, 26%).
data processes. At Duke University Health System, vital sign data, lab
“Data quality problems happen when you have systems value trends, and blood test data are used in conjunction with
where users don’t know how to enter data, the data isn’t ML capabilities to create Sepsis Watch, a tool that identifies
validated, or people are frustrated using it,” says Rob Korom, patients who are at risk of developing sepsis about 72 hours
Penda Health’s chief medical officer. “Our current vendor before the signs and symptoms of it emerge.
is very willing to make changes, but it’s expensive. All that “When [the system] sees constellations of signals coming
data then has to go through an ETL [extract, transform, load] together—maybe your heart rate is slightly elevated, your
process which, because we’re not the database developer lactic acid is up, and your white blood cell count is higher than
for our [EMR], isn’t a process we control.” This creates usual—the program looks at the risk profile, and if it triggers
limitations on their dashboard and on business intelligence a certain level, it alerts the provider that a patient might be
capabilities, he adds. developing sepsis,” Ferranti says. This feature enables the care
Another challenge organizations face in data management team to administer antibiotics earlier, draw blood cultures
and data interoperability is disorganized data, says Bradley of sooner, and more closely monitor a patient to see if further
HIMSS. Physicians have more data now than ever before but care is needed.
aren’t always able to make it actionable. “There are so many These applications and tools that make data integration
instances where they have all the data in the world but they easier are also improving operational efficiency. At Kenya-
have no idea what they’re supposed to do for the patient,” based Penda Health, integrated data is used to manage patient
Bradley says. “Organizations have to take the time not only wait times, Korom says. When a patient arrives at the clinic,
to integrate the data but do so in a way that’s logical to the the patient management system records the length of time
care provider team.” the patient waits with the receptionist, at the nursing triage
While data integration and managing data across settings station, and for the doctor. This data provides a comprehensive
has been a challenge, 64% of respondents agree that digital view of the length of time patients are waiting versus being
technologies for health care have made data integration easier, attended to and where those bottlenecks are occurring so
and 68% use software applications or tools to pull together management can put in place measures to reduce them.
data from diverse sources. This is especially true for leaders, as The clinic also relates that data to patient satisfaction scores,
81% say they use software applications or tools to pull together retention rates, and churn rates, he says.
data from diverse sources, compared to 73% of followers and Penda Health is also using integrated data to improve
52% of laggards. clinical quality, Korom says. “While the vast majority of our
More than half of respondents say their organization patients report clinical improvement of their urgent care
prefers that its digital solutions be evidence backed (67%) issues, we noticed that within a few days there were some
and medically certified (55%). Respondents say that their outliers that have reshaped our practice,” he says.
organizations prefer evidence-backed digital solutions for By sorting outcomes data by diagnosis, Penda Health
their evidence of value (73%), followed by their evidence found that lower back pain was a common diagnosis with
of safety (66%). When it comes to medically certified tools, one of the lowest improvement rates. The clinical quality
respondents say that their organizations prefer them primarily team discovered that while the diagnosis and treatment were
because of compliance with regulatory bodies (81%). Leaders generally appropriate, there was a gap in how team members
(49%) are more likely than followers (31%) or laggards (35%) were explaining what patients should expect for timelines of
to say that it is an organizational requirement for digital improvement. This discovery prompted new patient education
solutions to be medically certified. materials aimed at this diagnosis, Korom says.
These software applications and tools that enable integration More-advanced health care organizations are including
and interoperability are having direct impacts on clinical care integrated data in their clinical decision support rules, Bradley
and operations. Leaders that use these tools find electronic says. If a patient is diagnosed with an allergy by a clinic down
medical records across institutions (67%), electronic medical the street, for example, doctors at another location will see

Research Report | Innovation in Data-Driven Health Care

6
64%
of respondents
agree that digital
technologies
for health care
have made data
integration easier.
 Harvard Business Review Analytic Services

this allergy flagged if they try to order a medication that could

cause a reaction.
“We’re doing pretty well with connecting the wires—and
we can always do better—but we’re really inching toward that
next step,” he says. “How do we organize this wealth of data “You can have the greatest data
so doctors and nurses can use it better?”
system out there and you can have
great reporting, but if the people
Nurturing a Data-Driven Culture who are running units, running
While data integration, interoperability, and data management
have posed challenges for health care organizations, culture
departments, and running clinical
and leadership have also been factors inhibiting data-driven areas aren’t using them, what use
initiatives. Organizations are contending with insufficient are they?” says Jeff Ferranti, MD,
funding for their data projects (42%), organizational silos/a
lack of collaboration across the business (37%), and a lack of
chief digital officer and senior
analytic skills among frontline professionals (33%). vice president at Duke University
A common misstep fueled by silos and a lack of collaboration Health System.
is adopting a “build it and they will come” mentality for new
data efforts. In these scenarios, IT either builds data capabilities
or purchases a solution without input from its end users.
When systems are built without input, they simply won’t
get used, Ferranti says. “You can have the greatest data system (66% vs. 71% in 2019); process/redesign changes (66% vs. 70%);
out there and you can have great reporting, but if the people formal training to develop and practice new ways of working
who are running units, running departments, and running (56% vs. 54%); investing in new collaboration tools (53% vs.
clinical areas aren’t using them, what use are they? That’s not 51%); and formal change management programs (39% vs. 41%).
a technical problem; that’s a culture problem,” he says. “The Driving a data culture requires collaboration and trust,
projects I’ve seen go off the rails almost always do so because but equally important is having skilled frontline workers
you have IT people doing things separate from clinical and who are comfortable using data to drive decision making.
operational people. You’ve got to have cross-disciplinary teams As organizations look to the future, it’s critical that health
because it helps with adoption and thinking through things care teams prioritize the necessary training and upskilling to
you might not otherwise consider.” support new data advancements, Rankin says.
Collaboration is also key in building trust around the data, “There’s an assumption that all physicians understand
Korom says. “We see challenges with people who say they don’t data and graphs because they’re intelligent and curious, but
really trust a KPI for one reason or another, and sometimes we not all do,” he says. “Data scientists and informaticians just
have to address that to gain their trust,” he says. pump out this data and give it to physicians and think, ‘Well,
One KPI Penda Health tracks, for example, is the turnaround we’ve done our job.’ But if they’re not using the data because
time for lab results. A timer begins when a lab technician they don’t understand it, what’s the point?”
opens a patient’s chart to view the labs the patient needs and Building a strong data culture starts at the top with
ends after the labs have been completed and the technician executives, boards, and management all buying into the value
has entered the results in the system. of data, supporting these projects, and understanding the
“Our lab techs quickly learned that they could game that infrastructure needs. For some organizations, however, the
KPI by asking the patient what labs they needed and opening top is where many budget discussions end.
their chart after they started running the lab, which made the Conversations with administration around funding data
turnaround time look even faster than what was feasible,” he initiatives have been one of the biggest challenges at Cabrini
says. “As much as I’d love to use that as a core KPI in the lab, we Health. “We bend over backwards to produce meaningful
had to revisit it to make sure that people were using the timing reports for management, but to develop a business case for
functionality in the way it was intended for that number to be why we need to redevelop our IT infrastructure, particularly
trusted and actually have meaning.” data warehouses, is extraordinarily difficult at the moment,”
As organizations aim to improve collaboration efforts, the Rankin says. “Administration loves the reports, but they don’t
hierarchy of priorities is similar to the one found when this understand the infrastructure that’s required underneath it.”
same question was asked in a 2019 survey: communication Navigating this disconnect at the top is a challenge for many
from top management about the importance of collaboration leaders, says Peter Daneyrd, MD, CEO of health care consulting

Research Report | Innovation in Data-Driven Health Care

8
 Harvard Business Review Analytic Services

company Daneyrd Consulting, based in Stockholm, Sweden. organizations continue to seek solutions for data integration
“These investments are seen as a cost at the top, especially and interoperability challenges, and as they progress toward
because it’s not immediately apparent to them to what extent becoming more data driven, partnerships with external
they will contribute to making health care better,” he says. organizations grow more important in other ways.
Funding issues are also fueled by negative margins in “Hospitals might have great clinical talent and a lot of
hospitals, Halamka adds. It’s rare that CEOs approach IT wonderful data available in their systems, but they might not
asking for algorithms or analytics dashboards; instead, they necessarily have the same understanding of all the available
ask about new revenue streams, care models, and solutions for tools that technology partners might have,” Ferranti says. “You
staff recruitment, retention, and burnout problems, he says. want to bring those things together and work collaboratively
All that said, funding issues, Halamka believes, are bound on projects where you’re taking the best skills from both sides
to reverse. “Now, post-Covid, the notion of using machine and creating something that neither one of you can create
learning and workflows becomes an imperative because of the alone. That’s the real potential of partnerships.”
economic headwinds that these organizations are facing,” he Health care organizations are partnering with a number
says. “We’re starting to see more cooperation because there’s of external entities in similar capacities. The top entities
more demand for these solutions than ever before.” cited by respondents include regulatory bodies (55%), health
IT companies (51%), health care associations/professional
bodies (50%), consulting companies (48%), and health data
The Role of Partnerships providers (48%). Leaders are more likely to collaborate with
Ferranti says outside expertise may be helpful in dealing with external entities overall and particularly with regulatory
funding challenges that persist. “That’s where I think there’s bodies (63%), health data providers (63%), and other health
increasing opportunity to work with industry partners who care organizations (58%), compared with followers (56%, 52%,
are really expert in some of the new technologies to make and 46%, respectively) and laggards (50%, 34%, and 38%,
biomedical information more computable.” As health care respectively). FIGURE 4

Research Report | Innovation in Data-Driven Health Care

9
 Harvard Business Review Analytic Services

The capabilities that organizations are seeking to gain by

working with external entities reflect current and cutting-
edge data priorities. These top capabilities include data
analytics and other digital tools (65%), health care, clinical
experiences (56%), artificial intelligence, machine learning
(46%), cybersecurity (39%), and data sets for applying machine
learning (34%). Leaders and followers, however, are more
likely than laggards to be seeking to gain capabilities in two
areas: artificial intelligence and machine learning (both 49%
vs. 37%) and data sets for applying machine learning (41% and
36% vs. 26%). FIGURE 5
While partnerships with external entities can be valuable,
they can also introduce challenges. With digital health care
startups in particular, more than half (55%) of respondents
agree that the growing number of these organizations has
made data integration more challenging.
“One of the challenges in creating novel data sources is
that it creates more silos,” Halamka says. “You have to figure
out where you are going to put this data that comes from the
startup if it is a data type not typically recorded in an electronic
health record or other generally operational, administrative,
or clinical system.”
Another challenge organizations face with digital
health care startups is a lack of security maturity. “A lot
of startup organizations have high energy and agility, but
don’t necessarily have competency in regulatory and sub-
regulatory guidance or compliance,” he says. “Do they really
understand what [the Health Insurance Portability and
Accountability Act (HIPAA)] or the general data protection
rules in Europe require?”
While they do present challenges, digital health care
startups are bringing innovative solutions to health care,
74% of respondents agree. Hospital Sírio-Libanês in São Paulo,
Brazil, for example, frequently holds meetings with startups
to vet new tools and capabilities, with varying outcomes.
One of those capabilities was an AI solution designed to
identify pneumonia and pleural effusion—the buildup of
fluid between the layers of tissue that line the lungs and chest
cavity—on X-rays. Offered as a single solution, however, it
was more expensive and didn’t suit the hospital’s budget
requirements or needs. “In the future, I believe some of these
small, very good ideas will move to big companies, where
they’ll be put together and offered as a digital solution for the
entire journey, not just a part of the journey,” Nomura says.
Bradley says that these cutting-edge single solutions are
a sweet spot for digital health care startups today and can
provide differentiating value to organizations looking to strike
on something new.
“If digital health startups are focusing on some of the newest
areas that health care is venturing into, like acute care settings
at home or patient-facing technologies and services, I think
they’re filling the niche of what the community is expecting,”

Research Report | Innovation in Data-Driven Health Care

10
 Harvard Business Review Analytic Services

he says. “Some organizations have a harder time reacting to In the U.S. and Europe, health care organizations are
quick needs like that. So as most niche vendors have over beholden to HIPAA and the General Data Protection
time, they’ll fill in the gaps from the time the need shows up Regulation, respectively—two regulations that focus on
until the big guys can actually start addressing it.” protecting the privacy and security of personal health
information in health care organizations.
Health care organizations expect their technology partners
Enduring Data Privacy and to address these concerns. Among respondents who use cloud-
Security Considerations based services to manage clinical applications and health
While partnerships with external entities can help solve data for individuals and/or population groups, data security
some of today’s most pressing patient health and operational and privacy (54%) is the characteristic they consider most
efficiency challenges, organizations also recognize that these important when choosing cloud-based services, followed
data and technology solutions can bring privacy and security by ease of integration with existing systems (40%), service
risks to health care. Seventy percent of respondents agree reliability (36%), and cost of service (35%). FIGURE 7
that data-driven health care creates new risks for patient More organizations appear to be finding the security
data privacy and security, while 66% agree that integrating affirmation they seek, as cloud-based health care data
hospital/lab operations creates new risks for the privacy and management services have become much more widely
security of institutional data. FIGURE 6 used over the past four years. Forty-one percent say their
The challenge for health care organizations is ensuring organization is using cloud-based services to manage clinical
the data is used ethically, with the most respect for patient applications for individual and population health in 2023,
privacy, Ferranti says. “At academic medical centers in compared with 27% in 2019.
particular, we are working to enable academic freedom and While data privacy and security remain a top consideration
inspire new discoveries while simultaneously protecting for many health care organizations, that’s not the case for all. In
patient privacy and operating with the highest ethics,” he India, for example, there is no regulation for health care data,
says. “That’s what we’re all trying to solve for.” says the Indian Institute of Technology’s Prasad. “A typical

Research Report | Innovation in Data-Driven Health Care

11
 Harvard Business Review Analytic Services

health care organization in India will invest in priorities that

will increase their revenue, and cybersecurity doesn’t do that,”
Prasad says. “It is viewed as an expense.”
Data privacy and security are similarly nebulous at Penda
Health in Kenya. While the organization follows the Data
Privacy and Protection Act, which was assented to by former
Kenyan president Uhuru Kenyatta in 2019, Korom says the
oversight of data security still isn’t very strong. “While we
are guided by the Data Privacy and Protection Act, the area is
evolving so rapidly that we also need to use our judgment on
how best to protect patient data and privacy,” he says.

Conclusion
Health care organizations are turning to data to solve some of
the most pressing challenges in patient health and operational
efficiency. And while many have made strides in becoming
more data driven over the past four years, fewer organizations
have reached maturity in this area.
To achieve maturity, health care organizations must work
to improve data integration and interoperability. Data tools
and partnerships will be key to connecting incompatible and
disconnected systems and introducing new data sources, thereby
enabling richer analytical insights and unlocking innovative
solutions to improve patient health and operational efficiency.
Organizations must also address cultural challenges by
increasing collaborative efforts, upskilling data users, and
continuing to lobby for the necessary funding to support
integral and cutting-edge data efforts. At the heart of these
efforts is a strong chief medical officer, Bradley says. “They’re
the ones championing the value and importance of data.
They’re the ones having these conversations, and they’re the
ones driving these efforts forward,” he says.
As health care organizations progress toward becoming
more data driven, partnerships with external organizations,
particularly for data analytics and other digital technologies,
will become more critical. While these partnerships present
opportunities for collaboration, they may also introduce data
integration challenges, which will need to be managed. Lastly,
as digital transformations progress, privacy and security must
remain top of mind, particularly as more organizations adopt
cloud-based data solutions.
The future of health care is data driven nonetheless, Halamka
says, and organizations that harness these new capabilities and
address these challenges have promising futures. “By 2030,
health care organizations will be data businesses. Their jobs
will be to collect and interpret data from all kinds of locations,
from the patient’s home to the devices they wear, and they’ll be
integrating and interpreting all of the data around a patient’s
lifetime journey,” he says. “But you can’t do this alone. You
need an urgency to change, a guiding coalition, and you need
a senior vision. That’s what will take us into the future.”

Research Report | Innovation in Data-Driven Health Care

12
M E T H O D O LO GY A N D PA R T I C I PA N T P R O F I L E

Harvard Business Review Analytic Services surveyed 757 members of the Harvard Business Review
audience via an online survey fielded in January 2023. Respondents qualified to complete the survey
if they work in the health care industry or in a health care–related industry and are familiar with their
organization’s use of health data and digital technologies for research, diagnosis, and/or treatment
decisions.

Size of Organization Seniority Industry Sectors Job Functions Regions

31% 37% 30% 16% 63%

10,000 or more Executive Hospital/practice Professional North America
employees management/ (health care/legal)
board members
14% 17%
27% Consulting 13% Europe
1,000–9,999 29% General
employees Senior management
management 10% 10%
Integrated health Asia Pacific
20% care delivery 13%
100–999 15% system Admin
employees Middle 6%
management Latin America
9% 11%
22% Academic/ Consulting
Fewer than 100 19% research/ 4%
employees Other grades medical training All other functions Middle East/Africa
less than 9% each.
All other sectors
less than 9% each.

Figures may not add up to 100% due to rounding.

ABOUT US

Harvard Business Review Analytic Services is an independent commercial research

unit within Harvard Business Review Group, conducting research and comparative
analysis on important management challenges and emerging business opportunities.
Seeking to provide business intelligence and peer-group insight, each report is
published based on the findings of original quantitative and/or qualitative research
and analysis. Quantitative surveys are conducted with the HBR Advisory Council,
HBR’s global research panel, and qualitative research is conducted with senior
business executives and subject matter experts from within and beyond the Harvard
Business Review author community. Email us at [email protected].

hbr.org/hbr-analytic-services

Copyright © 2023 Harvard Business School Publishing. MCCRE39940723