Health and Functional Limitations Predict Depression Scores in the Health and Retirement Study

Results straight from MARS

Abstract

Objective: To examine the effects of chronic health conditions and functional status limitations on depression scores in a large representative sample of Americans. Method: The data included 27,461 respondents ages 50 to 90 who completed up to eight test occasions from the Health and Retirement Study. Multivariate adaptive regression splines (MARS) modeling was applied. Possible covariates of depression included arthritis, lung disease, back pain, diabetes, heart disease, high blood pressure, cancer, 28 pairwise combinations of the aforementioned conditions, ADL functional limitations, age, education and being female, being white, and being Hispanic. Results: The best fitting model had a GRSq of 0.18 (comparable to R2 ) and included 12 of 42 covariates. Depression score was predicted by: 1) ADL limitations, 2) education, 3) back pain, 4) lung disease, 5) being female, 6) being Hispanic, 7) heart disease, 8) being white, 9) high blood pressure plus stroke, 10) age, 11) back pain plus arthritis, and 12) back pain plus diabetes. Conclusions: Functional limitations was the strongest predictor of depression; reporting one limitation increased depression scores by nearly double the increase associated with two or more limitations. Back pain and lung disease were the strongest chronic disease predictors of depression; both are associated with considerable discomfort.