Generalizability is the extent to which the results found in one study can be used to describe people who were not followed in that particular study. Generalizability also is called external validity.
A study is generalizable to a larger population of people if the sample of people examined in the study is representative of that larger population. In population health management, if we create policies based on findings from particular studies, it is essential that these findings are, in fact, generalizable to the population that will be affected by the policy.
Two essential criteria must be met before using findings from population health management evaluations. The studies must be 1) valid and 2) generalizable. Population health management studies are conducted on a select group(s) with certain characteristics. Hence, the findings are unique to that group. Under certain circumstances, the results may be applicable to other groups, either with or without adjustment.
Example 1: A research study that follows upper-class Caucasian youth in the United States may lack generalizability; the findings of the study may not accurately describe youth in other areas of the world or youth in the United States who live in different circumstances or are members of different racial or ethnic groups.
Example 2: An insurance company approves payment for a myocardial infarction treatment that has been found to be effective in men. If the results of this study are not generalizable to women, then the insurance company will lose money by paying for a treatment that is ineffective in women.
Example 3: A specific population health management related example is the generalizability of a randomized control study, in which patients are first solicited for participation in the study and then randomized into intervention and control groups. The results of such a randomized control study address efficacy of the intervention but are not generalizable back to the population from which the participants were drawn, because the participants are self-selected and therefore not representative.
Hennekens, C., and Buring, J. Epidemiology in Medicine. Boston: Little, Brown & Company (1987).
A set of genetic markers present in a given patient. Specific genomic and molecular (proteomic) markers or combinations of markers may indicate a patient’s risk of developing certain diseases or how a patient will respond to certain drugs. The number of markers that are clinically useful is rapidly expanding, providing greater opportunities to tailor prevention and treatment to an individual patient, often called “personalized medicine.”
Nash et al. (2010). Population health.
Offit, K. (2008). Genomic profiles for disease risk. JAMA, 299(11), 353-355. http://jama.ama-assn.org/content/299/11/1353.full