216
Has pharmaceutical innovation increased medication adherence?
In chronic diseases, medication non-adherence is associated with increased hospital admissions, worse health outcomes and higher overall health care costs. However, patients often do not meet appropriate adherence levels. Besides other factors (e.g. patient and health system characteristics), the level of adherence that can be explained by the drug administered is unknown. Pharmaceutical innovation has been shown to improve health in terms of life expectancy and better functional status. In this study, we hypothesize that drug vintage – i.e. the degree of technological progress embodied in a drug agent – contributes to improved medication adherence.
Methods
To test this hypothesis, we estimate models of adherence to medications for chronic medical conditions using drug-by-patient-level data. We combine Truven Health MarketScan claims data during the period 1996-2013 of more than 50 million employees with Food and Drug Administration (FDA) data on drug approvals. We use the Medication Possession Ratio (MPR), i.e. the proportion of time a patient has access to medication as our measure of adherence. The explanatory variable of primary interest is the drug’s vintage (initial FDA approval year). Moreover, we consider a drug’s priority status assigned by the FDA, the proportion of drugs prescribed as extended or delayed release formulations and dose frequency as additional potential determinants of adherence. We use two complementary approaches to analyze the effect of pharmaceutical innovation on adherence; both approaches control for unobserved patient heterogeneity. The first approach is based on aggregate longitudinal data on a large number of drug classes. The second approach is based on patient-level data on drug utilization. We further account for patient characteristics, co-payment structure and different modes of drug administration (number of refills, mail-order prescriptions).
Results
We find that poor medication adherence is common. Mean MPR was 54.85% (s.d. 37.36%). Variation in MPR across diagnoses was moderate (highest in diabetes MPR=59.75%, lowest in gout disease MPR=45.81%. In contrast, we observed substantial variation in the vintage of drugs used within the diagnostic categories. The panel regression results suggest that use of newer drugs positively influences adherence levels (1% marginal increase by year of FDA registration, p<0.05) after controlling for the level of co-payments. At patient level, we find that besides drug vintage (log odds: 0.02, p<0.0001), medication adherence is positively influenced by drugs having assigned priority status (log odds: 0.001, p<0.0001), a higher proportion of drugs with extended or delayed release (log odds: 0.835, p<0.0001) and lower dose frequency (log odds: -0.172, p<0.0001).
Conclusions
Newer drugs contribute to increasing medication adherence in chronic conditions, even after controlling for the level of co-payments and other factors that are related to ease of drug administration. The differential effect of newer drugs may be due to greater efficacy or fewer adverse effects of newer drugs. Greater adherence to newer drugs is one important reason why patients using newer drugs live longer and have fewer hospital admissions/work-loss days. Physicians and pharmacy benefit managers should be aware of this.