Incident user cohorts for assessing medication cost-offsets

Area of Science:

• Health Economics
• Pharmacoeconomics
• Clinical Trial Design

Background:

• Assessing the cost-effectiveness of medications is crucial for healthcare management.
• Statin therapy in diabetic patients is a key area for cost-saving evaluations.
• Existing prevalent user designs may be subject to confounding biases.

Purpose of the Study:

• To develop and validate novel incident drug user designs for evaluating cost savings associated with statin use in diabetic individuals.
• To compare the robustness of different incident user designs against prevalent user designs.

Main Methods:

• A seven-step incident user design was implemented using a 5% random sample of Medicare beneficiaries (2006-2008).
• Designs included unadjusted analyses, difference-in-difference (DID) methods, and propensity score (PS)-matched DID with static and dynamic covariates.
• Analyses specifically addressed adherence strata and controlled for healthy adherer bias, with replication for ACE-inhibitor/ARB initiators.

Main Results:

• Statistically significant Medicare spending reductions were observed starting 7 months post-statin initiation and 13 months post-ACE-inhibitor/ARB initiation.
• However, these observed cost savings were primarily evident in adherent patient groups, particularly in propensity score-matched analyses controlling for healthy adherer bias.
• The study highlights the critical role of patient adherence in realizing medication-related cost offsets.

Conclusions:

• Incident drug user designs offer a more robust approach to assessing medication cost offsets compared to prevalent user designs.
• Effective cost-offset inference necessitates incorporating adjustments for confounding factors and conducting thorough sensitivity analyses.
• Adherence is a critical determinant of observed cost savings from pharmacologic interventions.