Associations between statin adherence level, health care costs, and utilization.

BACKGROUND Evidence of the associations between statin adherence level, health care costs, and utilization is still limited. It is not clear whether better clinical outcomes derived from increasing statin adherence levels can be translated into cost savings and lower health care utilization. OBJECTIVES To evaluate the associations between statin adherence level, health care costs, hospital admission, and emergency room (ER) visits after statin therapy is taken for 1 year. METHODS A retrospective cohort study was performed to examine whether higher statin adherence level, measured as medication possession ratio (MPR), is associated with lower health care costs and hospital admission rate and with fewer ER visits. The study sample consisted of adult patients aged 18-64 years on an index date with continuous enrollment 12 months prior to and 12 months after the index date (the first fill date of a statin between January 1, 2009, and December 31, 2010). Study subjects also needed to have a minimum of 2 ICD-9-CM diagnoses for hyperlipidemia or diabetes in the pre-index date period. Main data sources were medical and prescription (Rx) claims, as well as enrollment files provided by a health benefit program and a medical carrier of state government and public school employees in a midwestern state. Study subjects were stratified into 8 groups based on statin MPR level: less than  40%, 40%-59%, 60%-69%, 70%-79%, 80%-84%, 85%-89%, 90%-95%, and 96%-100%. Total medical and Rx costs, as well as all-cause hospital admission rates and ER visits in a year after the index date, were computed based on medical and Rx claims. A separate breakout of statin costs, part of total Rx costs, was also computed. Generalized linear models (GLMs) were developed to test the hypothesis that higher statin adherence levels are associated with lower health care costs and utilization. RESULTS A total of 10,312 subjects met the criteria and were selected. The average statin MPR in a year after the index date was 71.95%. Mean total costs (medical + Rx) in a year after the index date were $6,064.36. There were significant variations in Rx costs and total health care costs as well as ER visits among the 8 patient groups stratified using the statin MPR level. A GLM model showed that all the ratios of health care costs among groups with statin MPR from 40%-59%, 60%-69%, 80%-84%, 85%-89%, 90%-95%, and 96%-100% were larger than 1 and statistically significant compared with the reference group with statin MPR less than 40%, suggesting those groups had higher health care costs than the reference group with the lowest statin MPR level. Based on a logistic regression model of hospital utilization for this study population, all the odds ratios of all-cause hospitalization among the groups with higher statin MPR were not statistically significant, suggesting that the likelihood of hospitalization for patients with higher statin MPR was not statistically lower than that of the reference group with statin MPR less than 40%. After controlling for all other covariates, another GLM model based on the Poisson distribution and log link function showed that ratios of ER visits among groups with statin MPR from 60%-69%, 80%-84%, 85%-89%, 90%-95%, and 96%-100% were smaller than 1 and statistically significant, suggesting the groups had fewer ER visits than the reference group with statin MPR less than  40%. The patient group with statin MPR from 96%-100% was estimated to have the lowest number of ER visits. CONCLUSIONS Our study results show that much higher statin adherence levels are related to fewer ER visits after statin treatment is taken for a year among beneficiaries; however, the study is inconclusive whether higher statin adherence levels are associated with lower overall health care costs in a year after statin therapy is taken. Further research is needed to evaluate the associations between statin adherence level, the cost of cardiovascular care alone, and utilization over a longer period.