PRM51 - Application of survival analysis to adult pneumococcal vaccination rate in the United States

Abstract

Using observational databases to calculate vaccination rate is challenging because of limited continuous enrollment and loss to follow-up. The objective of this study was to apply the concept of survival analysis in estimating rate of pneumococcal vaccination among adults with the high-risk medical conditions included in the ACIP (Advisory Committee on Immunization Practices) recommendations. This was a retrospective cohort study using a large administrative claims database. The study cohort was 19-64 years old adults with newly diagnosed high-risk conditions (HIV/AIDS, diabetes, asplenia, chronic lung disease, chronic renal disease, chronic heart disease, organ transplant, cochlear implant, cancer, chronic liver disease and alcoholism) during 2007 to 2010 and had at least three years of continuous enrollment. Subjects were followed from the initial diagnosis date to the end of enrollment or 2011. Cox proportional hazard model was used to analyze the association of high-risk medical conditions with pneumococcal vaccination after controlling potential confounders. Survival curves were estimated from the Cox model. A total of 946,898 eligible subjects were followed for a total of 2,358,563 person-years, of which, 71,298 subjects received pneumococcal vaccine. The overall pneumococcal vaccination rate was 3.023/100 person years (95%CI=3.001, 3.045). Pneumococcal vaccination rate was the highest in HIV/AIDS patients (16.499/100 person years, 95%CI=(15.374,17.683)) followed by diabetes patients (4.520/100 person years, 95%CI=(4.466,4.574)) and was the lowest in alcoholism patients (1.042/100 person years, 95%CI=(0.953,1.137)). After controlling potential confounders, high-risk conditions had a significant association with pneumococcal vaccination. Compared to patients with alcoholism, HIV/AIDS patients were 11 times more likely to receive pneumococcal vaccination (HR=11.964, 95%CI=(10.696,13.382), p<0.0001). Pneumococcal vaccination rate was more accurately analyzed by applying survival analysis because a larger and more representative sample was constructed and loss to follow-up was taken into account. Study suggests pneumococcal vaccination rate was low in adults with newly diagnosed high-risk conditions.