Cancer Mortality Feasibility Study

Crespo, Carlos

Abstract

Little is known about the risk factors for cancer in Puerto Rican men. Puerto Ricans are the second largest group of Hispanics in the United States and have higher cancer mortality rates than all Hispanics combined. The age-adjusted cancer mortality in Puerto Rico is lower than in the United States; however, the age-adjusted stomach, oral cavity and prostate cancer mortality are higher in Puerto Rico than in the United States. Moreover, as cancer mortality rates in the United States have started to decline, the cancer mortality rates in Puerto Rico continue to increase. There is only one longitudinal study with a large enough sample of aging Puerto Ricans that can improve our understanding of the association between lifestyles and cancer mortality. The Puerto Rico Heart Health Program (PRHHP) began in 1965 and included a random sample of 9,824 men, 35-79 years of age at baseline, to study risk factors for coronary heart disease incidence and mortality. Between 1965-1978 four examinations collected information on diet, physical activity, cardiovascular end-points, skin color and others. Currently, we have 12-year mortality follow-up data on this cohort. Only nine of the original sample of 9,824 men were lost to follow-up over this 12-year period that ended in 1980. Cause of death in the PRHHP were established from autopsy reports, hospital and physician records, and information from family and next of kin. After 12-year of follow-up, there were 382 cancer deaths out of 1737 total deaths. We will select a sample of these deceased subjects to study the feasibility of ascertaining cancer mortality using the following sources of data: (1) the National Death Index Plus, (2) the Puerto Rico Demographic Registry; (3) the Puerto Rico Cancer Registry, and (4) the Puerto Rico Heart Health Program. Sensitivity, specificity, and percent agreement measurements will be obtained for these approaches using the PRHHP data as the standard. We have examined the probability of death through 1998 in this cohort, and expect 4,354 total deaths and 962 cancer deaths. Our results in establishing the approach, or combination of approaches with the highest net sensitivity can be used to ascertain multiple cause of death in a future study. Based on the results from this pilot research, we will propose a larger investigation to ascertain cause of death in the entire cohort and to study the association between diet, lifestyle and cancer mortality.