A full report on the workshop has been published As the Department of Health and Human Services

by the Office of the Assistant Secretary for Health moves to focus attention on and increase its efforts
and will be widely distributed through State maternal in disease prevention and health promotion, it is
and child health and crippled children's services appropriate to consider an alternative approach to
directors as well as through voluntary agencies. classification that is more closely linked to the prac-
tical considerations that govern proper application of
I am confident that the workshop's eventual out- preventive interventions.
come will be better health care for a greater number
-and a greater diversity-of children with dis- In the old scheme, the distinction between primary
abilities. The Department of Health and Human and secondary prevention depends on our identifica-
Services has a strong commitment to improve serv- tion of the biologic origin of disease. While the
ices to disabled children and their families. We will biologic origin of acute infections and injuries may
be using a variety of techniques to continue the be clear-cut, the same is not true of the chronic
momentum developed at the workshop, and I will diseases that now constitute our major causes of
report back to you as we make progress toward disability and death. Does myocardial infarction
achieving its goals. begin with the first pain, or with the first arterial
wall lesions which may have developed in youth
C. Everett Koop, MD, ScD (3)? Does cancer stem from the initiation event, or
Surgeon General only from the occurrence of effective promotion
(4)? As more is learned about multifactorial chronic
diseases with long periods of latency, the concept of
An Operational Classification biologic origin of disease becomes progressively
of Disease Prevention more diffuse. We also become entrapped by seman-
tic distinctions that have more historical than ra-
Three decades have elapsed since a working group tional scientific justification. Consider the three
under the Commission on Chronic Illness proposed common clinical situations of asymptomatic but
the classification of disease prevention into the abnormal elevations of blood sugar, blood pressure,
categories primary and secondary (1). An additional and serum cholesterol. They are logically identical
term, "tertiary prevention," has gained currency in that none produces discomfort or disability, each
since, and the classification is now ubiquitous in text- has serious diagnostic significance for future clinical
books of epidemiology and preventive medicine. events, and each is susceptible to intervention. Yet
These classes are summarily defined as primary- we commonly call diabetes and hypertension dis-
practiced prior to the biologic origin of disease; eases, but refer to hypercholesterolemia as a risk
secondary-practiced after the disease can be recog- factor. Dietary management of hypercholesterolemia
nized, but before it has caused suffering and disabil- is often called "primary prevention of heart disease,"
ity; and tertiary-practiced after suffering or disabil- but prescription of a diet for diabetes or a drug for
ity have been experienced, in order to prevent further hypertension is viewed as treatment, or possibly sec-
deterioration. This classification stems from an era ondary prevention.
when biomedical research was almost exclusively
the province of the laboratory scientist, and concepts A second disadvantage of the 1952 scheme, in
of health and disease were principally mechanistic. our opinion, is that the terms "primary" and "sec-
In recent years, the growth and success of epidemio- ondary" suggest an ordinal value. Although it was
logic research on chronic disease have introduced a not the intention of the Commission to suggest that
large body of nonmechanistic scientific knowledge primary is preferable, and secondary is second rate,
germane to disease prevention. We are conversant this impression may develop, particularly among
with statistical associations between risk factors and lay persons who may have responsibility for impor-
clinical events and have accepted a battery of cri- tant decisions that bear on preventive programs.
teria for judging whether or not the association rep- Careful quantitative analysis of benefits, costs, risks,
resents causation (2). The primary-secondary clas- and effectiveness frequently reveals that a preventive
sification is attractive and simple, but it does not intervention is best applied only to a high-risk group,
serve to distinguish between preventive interventions the evidence of high risk being a finding that can be
which have different epidemiologic justifications and related to the biologic origin of disease. Though
require different strategies for optimal utilization. "secondary," this may well be the optimal preven-
March-April 1963, Vol. 96, No. 2 107
tive strategy. This unintended side effect of the utilized among selected groups depend primarily on
primary-secondary classification can be avoided by the motivation of the individual who is being pro-
a choice of descriptive terms that differentiate quali- tected either to carry out or to seek out the preven-
tatively, but do not suggest a priority ranking. tive measure, so that public education, in this case
directed at the high-risk group, is an essential aspect
The classification that we propose would restrict of the strategy for optimal public health practice.
the use of the term "preventive" to measures, ac-
tions, or interventions that are practiced by or on
persons who are not, at the time, suffering from any The third class, of preventive measures, which
discomfort or disability due to the disease or condi- we propose to term indicated, encompasses those
tion being prevented. This distinction would serve that are advisable only for persons who, on exami-
to eliminate most of what is now encompassed in the nation, are found to manifest a risk factor, condition,
old category "tertiary." Persons (patients) who feel or abnormality that identifies them, individually, as
current discomfort or disability are motivated to being at sufficiently high risk to require the preven-
seek medical aid--both for cure and for prevention tive intervention. The majority of these measures
of further progression of the disease-by forces have been called secondary under the classical
which do not apply to the asymptomatic individual. scheme, since in most cases the observable indication
Provisions for their care do not have to emphasize is related to the biologic origin of disease. Preventive
persuading them to seek or accept the care that they measures that fall into this category, not surprisingly,
need, nor does one have to estimate the probability are usually not totally benign or minimal in cost.
that they will fall victim to the disease under con- If they were, the balance in the cost-benefit analysis
sideration; the presence of the disease is already might favor their wider application, including seg-
established. ments of the population at lower risk of disease, and
they would tend to move into the selective or uni-
Preventive measures-those which should be ap- versal classes. The identification of persons for
plied to persons not motivated by current suffering- whom indicated preventive measures are advisable
can be operationally classified on the basis of the is the objective of screening programs, and the cost,
population groups among which they are optimally risks, availability, and effectiveness of the preventive
used. The most generally applicable type, which we measure must be carefully weighed before a decision
shall call universal, is a measure that is desirable to initiate screening is made. Preventive interven-
for everybody. In this category fall all those meas- tions in this class include control of hypertension,
ures which can be advocated confidently for the dietary measures to reduce hypercholesterolemia,
general public and which, in many cases, can be antituberculous drugs for recent skin test converters,
applied without professional advice or assistance. the use of uricosuric drugs by persons with asympto-
Maintenance of an adequate diet, dental hygiene, matic hyperuricemia, and frequent, careful reexami-
use of seatbelts in automobiles, smoking cessation, nation of persons from whom a basal cell skin cancer
and many forms of immunization would fall clearly or a colonic polyp has been removed. Indicated pre-
into the universal category for which benefits out- vention is most commonly applied in the clinical
weigh costs and risks for everyone. setting, as the indication is ordinarily one discovered
through medical examination or laboratory testing,
There are many measures, however, in which the and many of the preventive measures require profes-
balance of benefits against risk and cost is such that sional advice or assistance for optimal results. The
the procedure can be recommended only when the word "treatment" is often used in connection with
individual is a member of a subgroup of the popu- these efforts but, on consideration, it will be apparent
lation distinguished by age, sex, occupation, or that the "treatment" of an asymptomatic, clinically
other obvious characteristic whose risk of becoming demonstrable abnormality is justified only if it will
ill is above average. These measures we shall call result in the prevention of some later, anticipated
selective. Examples would be active rabies immuni- symptoms or disability. It is also evident that a dif-
zation for veterinarians, annual influenza immuniza- ferent approach to securing patient compliance is
tion for the elderly, use of safety goggles by machin- required with indicated prevention than with treat-
ists, and the avoidance of alcohol and many drugs ment which is immediately therapeutic. We believe,
by pregnant women. As with universal preventive therefore, that the distinction between indicated pre-
measures, most of these actions which should be vention and treatment can reasonably be made and
108 Public Health Reports
that the distinction is useful in considering how best
to utilize these measures. LETTER TO THE EDITOR
In summary, we propose to define prevention as "Asbestos in Colorado Schools"-Erratum
measures adopted by or practiced on persons not
currently feeling the effects of a disease, intended It has come to my attention that there is an error in
to decrease the risk that that disease will afflict them my article "Asbestos in Colorado Schools," which
in the future. Prevention is classified into three appeared in the July-August 1982 issue of Public
levels on the basis of the population for whom the Health Reports (pages 325-331). The error is on
measure is advisable on cost-benefit analysis. Uni- page 330 in the conclusion. In the first sentence,
versal measures are recommended for essentially third line, it stated that ". . . 380 [public schools] (38
everyone. Selective measures are advisable for pop- percent) may have asbestos sprayed on ceilings."
ulation subgroups distinguished by age, sex, occupa- That should read ". . . 480 (38 percent) may have
tion, or other evident characteristics, but who, on asbestos sprayed on ceilings."
individual examination, are perfectly well. Indicated
measures are those that should be applied only in Cindy A. Baldwin, Industrial Hygiene Consultant,
the presence of a demonstrable condition that iden- State of Iowa Bureau of Labor, Des Moines.
tifies the individual as being at higher than average
risk for the future development of a disease. Preven-
tive measures in each of these classes share many
features that determine optimal strategies for pro-
moting acceptance by those persons in the popula-
tion for whom they will do the greatest good in
preserving health.
Robert S. Gordon, Jr., MD, MHS
Special Assistant to the Director
National Institutes of Health

References ..................................
1. Commission on Chronic Illness: Chronic illness in the
United States, vol. 1. Published for the Commonwealth
Fund by Harvard University Press, Cambridge, Mass.,
1957.
2. Smoking and health. Report of the Advisory Committee
to the Surgeon General of the Public Health Service.
PHS Publication No. 1103. U.S. Department of Health,
Education, and Welfare, Washington, D.C., 1964.
3. Enos, W. F., and Beyer, J.: Coronary disease among
U.S. soldiers killed in action in Korea. JAMA 152:
1090-1093 (1953).
4. Carcinogenesis - a comprehensive survey, vol. 2.
Mechanisms of tumor promotion and cocarcinogenesis,
edited by T. J. Slaga, A. Sivak, and R. K. Boutwell.
Raven Press, New York, 1978.

March-April 193, Vol. 9S, No. 2 109