Purpose The benefit of early detection of asymptomatic disease declines with remaining life expectancy. It makes no sense to subject patients to invasive tests and procedures if they will die of an unrelated cause before the disease becomes clinically manifest. This issue is particularly relevant to cancer screening, since there is often a long “lead time” between tumor onset and presentation of symptoms. Cross-sectional studies that estimate the relationship between patient characteristics and screening may mischaracterize the decision process relating life expectancy to screening because many people never undergo screening in the first place. Instead, we want to know whether patients who regularly receive screens discontinue screening in response to a large decrease in life expectancy. We estimate the impact of hospitalization on continuation of prostate cancer screening. We expect that men who receive a negative health shock (e.g. hospitalization) will be less likely to continue regular screening, but persistence in screening decisions and the fact that hospitalization leads to more contact with the health system may mitigate this effect.

Method Using the SEER-Medicare 5% control file, we identified men age 65+ who received a prostate specific antigen (PSA) test in 1997 and were not diagnosed with prostate cancer previously. We classified men in this sample based on whether or not they were hospitalized in the 6 month window following the 1997 index PSA test. Our follow-up period was from 6 months after the initial PSA test until December 31, 2004. We used a Cox model to estimate the impact of being hospitalized on time to receipt of a 2nd PSA test, controlling for age, race, and comorbidities.

Results 28,297 (88%) of the sample of 32,039 men had a 2nd PSA test. Of the remaining 3,742, 2,447 died without receiving a 2nd PSA and 1,279 were alive at the end of 2004. The odds ratio from the Cox survival model for hospitalization was 0.74 (p<0.01), implying that men who were hospitalized were less likely to receive a 2nd PSA test.

Conclusions: New information about life expectancy – as measured by whether or not a person is hospitalized – influences prostate cancer screening decisions. Patients and physicians do not blindly persist in past screening behavior. However, we are unable to determine whether screening decisions are optimal.