Self-rated and changes in self-rated health as predictors of mortality: First evidence from german panel data

Background: Studies from several countries have shown that self-rated health is an independent predictor of mortality. However, no empirical evidence exists for Germany so far. We investigate the effectiveness of (i) self-ratings of health by individuals and (ii) changes in self-rated health, as predictors of mortality for Germany. Methods: A sub-sample of 3,096 respondents, aged 50 years and over, is drawn from the annual collections of data of the German Socio-Economic Panel Study (GSOEP), between 1994 and 1996. Several probit models are estimated in order to analyse the impact of selfrated health and of changes in self-rated health on predictions of mortality. Results: We find that, while currently self-rated health is shown to be a valid predictor of mortality in Germany, adding previously self-rated health has no effect on explaining the probability of death. Furthermore, one-year changes in self-rated health do not have an additional impact on predicting mortality. Conclusion: Our results for Germany confirm international evidence. In addition, the assumption that self-rated health reflects trajectories and not only the current level of health can be neglected. This leads us to the conclusion that self-rated health reflects a static rather than a dynamic perspective of health. Therefore, when evaluating a population's state of health, it may be sufficient to rely on self-assessments of health at one point of time instead of using panel data.