Optimizing statin treatment decisions for diabetes patients in the presence of uncertain future adherence

Jennifer E. Mason, Darin A. England, Brian T. Denton, Steven A. Smith, Murat Kurt, Nilay D. Shah

Research output: Contribution to journalArticlepeer-review

34 Scopus citations


Background. Statins are an important part of the treatment plan for patients with type 2 diabetes. However, patients who are prescribed statins often take less than the prescribed amount or stop taking the drug altogether. This suboptimal adherence may decrease the benefit of statin initiation. Objective. To estimate the influence of adherence on the optimal timing of statin initiation for patients with type 2 diabetes. Method. The authors use a Markov decision process (MDP) model to optimize the treatment decision for patients with type 2 diabetes. Their model incorporates a Markov model linking adherence to treatment effectiveness and long-term health outcomes. They determine the optimal time of statin initiation that minimizes expected costs and maximizes expected quality-adjusted life years (QALYs). Results. In the long run, approximately 25% of patients remain highly adherent to statins. Based on the MDP model, generic statins lower costs in men and result in a small increase in costs in women relative to no treatment. Patients are able to noticeably increase their expected QALYs by 0.5 to 2 years depending on the level of adherence. Conclusions. Adherence-improving interventions can increase expected QALYs by as much as 1.5 years. Given suboptimal adherence to statins, it is optimal to delay the start time for statins; however, changing the start time alone does not lead to significant changes in costs or QALYs.

Original languageEnglish (US)
Pages (from-to)154-166
Number of pages13
JournalMedical Decision Making
Issue number1
StatePublished - Jan 2012
Externally publishedYes

Bibliographical note

Funding Information:
This project was funded in part by the Agency for Healthcare Research and Quality (AHRQ) under grant 1R21HS017628 and by the National Science Foundation under grant CMMI-0844511.


  • Markov decision process
  • adherence
  • diabetes
  • statins


Dive into the research topics of 'Optimizing statin treatment decisions for diabetes patients in the presence of uncertain future adherence'. Together they form a unique fingerprint.

Cite this