Abstract
Aims: We examined 2006 total self-reported and database-calculated out-of-pocket (OOP) costs and prescription drug behavior under Medicare Part D to assess the association between cost measures and behavior change.
Methods: We collected information on 2006 total self-reported OOP costs and cost responses through a stratified random sample survey of Part D members of a large integrated delivery system (IDS) (n=1040, response rate=74.9%). Behavior changes included reduced adherence (e.g. skipping pills), cost-coping (e.g. switching to cheaper drugs), and financial burden (e.g. borrowing money). We also calculated total 2006 OOP costs using the IDS’ pharmacy database. We used logistic regression to examine the association of self-reported and database OOP costs with behavior change, adjusting for gender, education, marital status, race/ethnicity, self-reported health, income, assistance with medications, and number of prescriptions in 2006. All analyses were weighted for sampling proportions.
Results: Sixty-seven percent of patients reported costs = $500, 16% reported $500–$1000, 10% reported $1000–$2000, and 8% reported = $2000. Fifty-one percent of respondents reported OOP costs that agreed with their database OOP cost category, 18% reported lower costs, 32% reported higher costs. Patients taking 5 different medications and those with higher database OOP costs were less likely to report OOP costs in the same category as the database calculation. Overall, 36% of respondents reported 1 behavior change due to drug costs. Patients with self-reported OOP costs >$500 had higher odds of reporting behavior change compared to those with costs <$500 (OR=3.4 [1.8–6.2] for $500–$1000, OR=3.6 [1.8–7.5] for $1000–$2000, OR=3.3 [1.7–6.5] for >$2000) after adjustment for database costs and covariates. Database OOP costs were not significantly associated with behavior change after adjustment for self-reported costs. Patients with income below 200% of federal poverty level and those with 5 prescriptions had higher odds of behavior change.
Conclusions: Less than half of patients reported OOP costs in the same category as our database calculation; over 1-in-4 reported higher costs. Higher self-reported costs, but not database costs, were associated with behavior changes due to costs, and patients with lower incomes or more prescriptions may also be more likely to reduce adherence, switch to cheaper drugs, or experience financial burden due to costs.
- Received September 11, 2008.
