The three-year mortality rate is 3.1%. This seems consistent with published mortality data for Puerto Rico. According to the United Nations, the crude (annual) death rate in Puerto Rico during 2000-2005 under “medium variant” mortality assumptions is 8.3 per 1,000 population.6 This implies a three-year mortality rate of approximately 2.49% (= 3 * 0.83%). Mortality of ASES beneficiaries may be somewhat higher than that of other residents of Puerto Rico.
We can compare the vintage distribution of drugs used by ASES beneficiaries to the vintage distribution of drugs used by all Americans and by American Medicaid beneficiaries using the 2000 Prescribed Medicines file of the Household Component of the Medical Expenditure Panel Survey (MEPS), which is conducted by the Agency for Healthcare Research and Quality. This comparison is shown in Figure 1. The fraction of ASES Rx’s that were approved after 1970 (63.2%) is similar to the fraction of U.S. Medicaid Rx’s that were approved after 1970 (61.7%). However, the fractions of ASES Rx’s that were approved after 1980 and 1990 (29.7% and 8.2%) are much smaller than the fractions of U.S. Medicaid Rx’s7 that were approved after 1980 and 1990 (48.4% and 25.5%). Use of older drugs in Puerto Rico’s Medicaid program may be partly attributable to the fact that, in Puerto Rico, the physician bears the costs of the drugs—the cost is deducted from the physician’s capitation payment. this
0LS estimates of the mortality model based on the entire population of ASES beneficiaries with pharmaceutical claims (N = 539,750) are shown in Table 1.8 The dependent variable is DIED: a dummy variable equal to one if the person died during 2000-2002, otherwise equal to zero.9 In addition to the three drug-vintage variables (P0ST70, P0ST80, and P0ST90), the model includes: three utilization variables (number of medical, pharmacy, and hospital claims), 15 diagnosis category variables, the index of the number of diseases (N_DISEASE), 9 region dummy variables, and 200 single-year-of-age-by-sex dummy variables (not shown to conserve space).
The coefficients of all three drug-vintage variables are negative and highly significant (p-value