Price difference as a predictor of the selection between brand name and generic statins in Japan.

OBJECTIVES: This study aimed to explore the predictors of the selection between brand name drug (BR) and generic drug (GE) and to clarify the quantitative relationship about selection.
METHODS: We identified "incident users" who dispensed statins between April 2008 and June 2011 in commercially databases consisted of dispensing claims databases (DCD) of out-of-hospital pharmacies and hospital claims databases (HCD) of in-house pharmacies in Japan. Predictors of the selection between BR and GE, including price difference (PD), the price of BR, their interaction and percent change of the price of GE relative to BR were explored by logistic regression using DCD and HCD separately.
RESULTS: We extracted records of 670 patients who have opportunity for selection both BR and GE. Logistic regression analysis demonstrated that PD, the price of BR, interaction between them, and prescriber affiliation were factors significantly associated with the selection in the DCD; logit (p)=9.735-0.251×PD-0.071×the price of BR+0.002×PD×the price of BR-1.816×affiliation+0.220×gender-0.008×age+0.038×monthly medical fee. PD was inversely proportional to BR choice in DCD and lead to the opposite result in HCD. Numerical simulation of selection revealed that the quantitative relationships heavily depend on situations.
CONCLUSIONS: PD and the price of BR are predictors of the selection between BR and GE interactively in out-of-hospital pharmacies, but not in in-house pharmacies of medical facilities. Results may support policies which increase the power of out-of-hospital pharmacies for selection.