Published in the American Journal of Epidemiology, June 2009
Researchers found that self-rated national health studies are too inconsistent to be used to measure trends in overall population health. The study, Are Americans feeling less healthy? The puzzle of trends in self-rated health, evaluated four national surveys from 1971 to 2007, which covered more than 900,000 people aged 18 and older. The work was done in collaboration with scientists at the Harvard University Initiative for Global Health.
Research findings
The most commonly used survey item in global measures of self-rated health asks people to characterize their health as “excellent, very good, good, fair, or poor.” The study authors found that while responses to this question are good predictors of health care utilization and even mortality at the individual level, they are too inconsistent to be used to measure trends in overall population health.
Researchers also noted striking inconsistencies in trends. Whereas data from one survey demonstrated that Americans were increasingly likely to report “fair” or “poor” health over the last decade, those from another survey indicate the opposite trend. Subgroup analyses revealed that the greatest inconsistencies were among young respondents, Hispanics, and those without a high school education. Trends in “fair” or “poor” ratings were more inconsistent than trends in “excellent” ratings. This finding has potentially profound implications for analyses of socioeconomic disparities in health that rely on self-rated health responses.
Analytical approach
Researchers conducted a comparative analysis of responses to a common self-rated health question in four national surveys from 1971 to 2007: the National Health and Nutrition Examination Survey, Behavioral Risk Factor Surveillance System, National Health Interview Survey, and Current Population Survey.
Sample weights were applied in each data set to account for unequal probabilities of selection, nonresponse, and noncoverage. For each survey year, confidence intervals were computed around age-standardized proportions using different response categories. Logistic regression models were used to assess trends over the last decade overall and by sex, age, race, and education, as well as to assess the probability of reporting “excellent” health.
Research objective
Discrepancies in responses to the same item in different nationally representative surveys raise questions about the validity of inferences about population health based on self-rated health. With federal and state policymakers looking for new ways to deliver high-quality health care to people who are uninsured or underinsured and international groups (such as the World Health Organization and the G8) discussing fund allocation for health aid to developing countries, such discrepancies can have a profound impact on policy decisions. This research is part of ongoing work by IHME to assess health system performance and to develop analytic tools to harness the value of available data.
Recommendations for future work
Currently, the US Centers for Disease Control and Prevention, World Health Organization, and the European Commission all recommend using a simple question that asks people to rate their own current health as “excellent, very good, good, fair, or poor” as a reliable way to track progress over time in population health.
The researchers suggest that the widespread use of this single, global self-report on health in research studies and policy analyses needs to be reconsidered. They also suggest that more time be spent evaluating the strength of surveys that use detailed questions and more symptom-specific questions. More work is urgently needed on robust and comparable approaches to tracking population health.
Citation: Salomon JA, Nordhagen S, Oza S, Murray CJL. Are Americans feeling less healthy? The puzzle of trends in self-rated health. American Journal of Epidemiology. 2009; 170(3):343–351.
