Self-Rated Health as a Measure of Healthy Ageing

Healthy ageing includes dimensions that are difficult to reduce to diagnoses or laboratory values. Self-rated health provides one compact way to ask how a person evaluates their health as a whole. Its value lies not in identifying a specific mechanism, but in summarizing information across bodily, psychological, and functional domains from the respondent's perspective. [1] [3]

Who This Is Useful For

This page is useful for readers interpreting cohort studies, population surveys, and healthy-ageing research in which a single general-health question is used as an outcome, predictor, or adjustment variable. It is especially relevant when comparing self-rated health with mortality, disability, multimorbidity, or more detailed measures of physical and mental health. [4] [5] [6]

What Self-Rated Health Measures

A common item asks, in some form, how a respondent would rate their health in general. Responses are usually ordered from a favourable category such as “excellent” or “very good” to an unfavourable category such as “poor” or “very poor.” The item is global: it does not tell the respondent which diseases, symptoms, time period, or domain should receive the most weight. [1] [2]

The resulting answer is better understood as an evaluation than as a direct inventory. Conceptual work describes the respondent as drawing on information about the body and mind, interpreting that information through prior experience and knowledge, and then selecting a response relative to personal and social standards. The same observed impairment may therefore contribute differently to two people's ratings. [1] [7]

Why It Is Relevant to Healthy Ageing

Healthy ageing is multidimensional, and a person can live with diagnosed disease while retaining substantial function or experience poor health without one measurement explaining the whole picture. Studies of the self-rated-health construct find that physical symptoms, longstanding illness, mental health, tiredness, mobility, and medication use can all contribute to the rating. No single measured component accounts for all of its variation. [3]

This breadth gives self-rated health a complementary role. It can add the respondent's integrated perspective to disease counts, physical-performance tests, or clinical assessments, but it cannot show which underlying domain produced a particular response. [1] [9]

Common Measurement Choices

Similar-looking self-rated-health variables are not automatically interchangeable. Studies differ in wording, time frame, comparison standard, response labels, and analysis. These choices should be checked before prevalence estimates or changes over time are compared. [2] [8]

Association With Mortality

Self-rated health has repeatedly predicted subsequent mortality in prospective population studies. A meta-analysis of 22 community cohorts found a graded association: compared with excellent health, adjusted mortality risk increased across good, fair, and poor ratings. The association remained in analyses accounting for comorbidity, function, cognition, and depression. [4]

In an individual-participant meta-analysis of eight cohorts containing 424,791 adults aged 60 years or older, fair and poor ratings were associated with higher all-cause mortality than ratings of at least good health. Adjustment for sociodemographic characteristics, lifestyle, body measures, and medical history explained only part of the association. [6]

The pattern is robust but not universal. A systematic review focused on older adults found that 20 of 26 included studies reported an association, while six did not. Differences between populations, follow-up periods, covariate sets, and measurement methods mean that self-rated health should be interpreted as a probabilistic predictor rather than a fixed risk category. [5]

Association With Functional Decline

Mortality is not the only relevant outcome. In a population-based study of 7,527 community-dwelling adults aged 70 years or older, less favourable self-assessments of general and physical health predicted later functional decline as well as mortality over six years. Functional change was assessed using basic and instrumental activities of daily living. [10]

This association makes the measure relevant to healthspan research, but it does not establish that the rating itself causes loss of function. Symptoms, underlying disease, psychological state, and early functional changes may contribute both to how health is rated and to what happens later. [1] [3] [10]

Why One Question Can Predict Later Outcomes

Several explanations can operate together. Respondents may know about symptoms, recent changes, diagnoses, family history, tolerance of effort, and functional difficulties that are only partly captured by a study's covariates. They may also integrate the severity and combination of problems rather than merely count them. Self-rated health can therefore retain predictive information after statistical adjustment without being a hidden objective test. [1] [4]

Reporting processes also matter. Mood, expectations, cultural standards, and the comparison group used by the respondent can influence the category selected. Prediction does not prove that every respondent applies the scale in the same way, nor that the unmeasured part of the rating is wholly biological. [1] [7]

Change Over Time

A single assessment describes one evaluation at one time. Repeated assessments can distinguish stable, improving, and worsening patterns, but observed change can reflect both change in health and change in expectations or reference standards. Studies that alter the question wording or response categories between waves introduce an additional source of measurement difference. [1] [2] [7]

For longitudinal interpretation, preserving the item wording, response order, response labels, and survey context helps make within-person and population trends more interpretable. Even with a stable instrument, a change in rating does not identify which physical, mental, or functional component changed. [2] [3]

Comparability Across People and Populations

Response categories depend on thresholds: one person may label a given level of difficulty “good,” while another may call it “fair.” A multi-country study using anchoring vignettes found differences in health expectations across age groups and countries, illustrating how identical labels can represent different implicit standards. [7]

These issues do not make self-rated health unusable. They mean that comparisons across languages, cultures, age groups, or surveys require attention to translation, question form, response scales, and possible differences in reporting behaviour. The measure is often most informative when the same instrument is used consistently within a defined population. [2] [7] [8]

Evidence Quality and Interpretation

Evidence is strong that self-rated health is associated with later mortality at the population level. This conclusion is supported by decades of prospective cohorts, conventional meta-analysis, an individual-participant meta-analysis, and a recent systematic review focused on older adults. [4] [5] [6] [11]

Evidence is less definitive about exactly what portion of that association comes from unmeasured disease, embodied sensations, function, psychological factors, health behaviour, or reporting style. Most outcome evidence is observational, residual confounding is possible, and findings for mortality are not uniform across every older population. [1] [3] [5]

What This Does Not Mean

• It does not mean self-rated health diagnoses a disease or identifies a biological mechanism. [1]
• It does not mean a poor rating makes decline inevitable; associations describe differences in group-level probability. [5] [6]
• It does not mean self-rated and clinically measured health are interchangeable; each can contain information not represented by the other. [9]
• It does not mean identical response labels have identical meanings across languages, cultures, ages, or question forms. [2] [7] [8]
• It does not mean a change in rating reveals which health domain has improved or worsened. [1] [3]

Practical Interpretation Examples

• If a study reports “poor self-rated health”: check the exact question, response scale, and category cut-off before comparing its prevalence with another study. [2] [8]
• If a rating predicts mortality after adjustment: interpret it as retained prognostic information, not proof that the subjective judgment causes mortality or measures one hidden disease. [1] [4]
• If two people have similar diagnoses but different ratings: symptoms, function, mental health, recent change, expectations, and personal weighting may contribute to the difference. [1] [3]
• If ratings improve across survey waves: consider both genuine health change and possible changes in wording, response options, expectations, or sample composition. [2] [7]

Summary

Self-rated health is a concise global evaluation that complements more specific measures of disease, function, and mental health. Less favourable ratings consistently identify groups with higher risks of mortality and functional decline, but the item remains subjective, multidimensional, and sensitive to how questions and response scales are framed. It is most useful in healthy-ageing research when treated as a broad, context-dependent indicator rather than a diagnosis or direct measure of biological age. [1] [2] [4] [10]

References

1. Jylhä, M. (2009). What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social Science & Medicine. https://pubmed.ncbi.nlm.nih.gov/19520474/
2. Cullati, S., Bochatay, N., Rossier, C., et al. (2020). Does the single-item self-rated health measure the same thing across different wordings? Construct validity study. Quality of Life Research. https://pmc.ncbi.nlm.nih.gov/articles/PMC7434800/
3. Singh-Manoux, A., Martikainen, P., Ferrie, J., et al. (2006). What does self rated health measure? Results from the British Whitehall II and French Gazel cohort studies. Journal of Epidemiology & Community Health. https://pmc.ncbi.nlm.nih.gov/articles/PMC2566175/
4. DeSalvo, K. B., Bloser, N., Reynolds, K., et al. (2006). Mortality prediction with a single general self-rated health question: a meta-analysis. Journal of General Internal Medicine. https://pmc.ncbi.nlm.nih.gov/articles/PMC1828094/
5. Dramé, M., Cantegrit, E., & Godaert, L. (2023). Self-rated health as a predictor of mortality in older adults: a systematic review. International Journal of Environmental Research and Public Health. https://pmc.ncbi.nlm.nih.gov/articles/PMC10001164/
6. Bamia, C., Orfanos, P., Juerges, H., et al. (2017). Self-rated health and all-cause and cause-specific mortality of older adults: individual data meta-analysis of prospective cohort studies in the CHANCES Consortium. Maturitas. https://pubmed.ncbi.nlm.nih.gov/28778331/
7. Salomon, J. A., Tandon, A., & Murray, C. J. L. (2004). Comparability of self rated health: cross sectional multi-country survey using anchoring vignettes. BMJ. https://pmc.ncbi.nlm.nih.gov/articles/PMC324453/
8. Eriksson, I., Undén, A.-L., & Elofsson, S. (2001). Self-rated health. Comparisons between three different measures. Results from a population study. International Journal of Epidemiology. https://pubmed.ncbi.nlm.nih.gov/11369738/
9. Giltay, E. J., Vollaard, A. M., & Kromhout, D. (2012). Self-rated health and physician-rated health as independent predictors of mortality in elderly men. Age and Ageing. https://pubmed.ncbi.nlm.nih.gov/22180414/
10. Lee, Y. (2000). The predictive value of self assessed general, physical, and mental health on functional decline and mortality in older adults. Journal of Epidemiology & Community Health. https://pmc.ncbi.nlm.nih.gov/articles/PMC1731623/
11. Idler, E. L., & Benyamini, Y. (1997). Self-rated health and mortality: a review of twenty-seven community studies. Journal of Health and Social Behavior. https://pubmed.ncbi.nlm.nih.gov/9097506/