Insurance and Other Determinants of Elderly Longevity in a Costa Rican Panel

Luis Rosero-Bixby, Universidad de Costa Rica
William H. Dow, University of North Carolina at Chapel Hill
Adriana Lacle

We followed 900 Costa Ricans aged 60+ in Coronado, a community at the outskirts of the capital city, from 1984 to 2001. This panel was drawn from the 1984 census. We made follow-up visits to the households in 1985 and 1986, tracked deaths in the national death registry database, and double-checked survival status in the voting registry in 1990 and 2002. We also contacted all the survivors by phone and household visits in 2002. Of the 876 individuals in the panel, 67% died by December 2001. We used Kaplan-Meier curves, life tables, and Cox regression to analyze the panel?s survival. Mortality resulted slightly higher than the Costa Rican average and similar to the US. Survival was substantially lower among not married men and individuals with no full autonomy at the beginning of the study. The effect of socio-economic status is weak. Insurance effects seem to be confounded by selection biases.