Adherence to osteoporosis pharmacotherapy is underestimated using days supply values in electronic pharmacy claims data Academic Article uri icon

  •  
  • Overview
  •  
  • Research
  •  
  • Identity
  •  
  • Additional Document Info
  •  
  • View All
  •  

abstract

  • PURPOSE: Days supply (prescription duration) values are commonly used to estimate drug exposure and quantify adherence to therapy, yet accuracy is not routinely assessed, and potential inaccurate reporting has been previously identified. We examined the impact of cleaning days supply values on the measurement of adherence to oral bisphosphonates. METHODS: We identified new users of oral bisphosphonates among Ontario seniors (April 2001-March 2011). Days supply values were examined by dose, and we identified misclassification by comparing observed values to dose-specific expected values. Days supply values not matching expected values were cleaned using dose-specific algorithms. One-year adherence to therapy was defined using measures of compliance (mean proportion of days covered [PDC], and categorized into high [PDC ≥ 80%], medium [50% < PDC < 80%], low [PDC ≤ 50%]) and persistence (30-day permissible gap). Estimates were compared using the observed and cleaned days supply values, stratified by site of patient residence (community or long-term care [LTC]). RESULTS: We identified 337 729 (5% LTC) eligible new users. Among LTC patients, adherence estimates increased significantly following data cleaning: mean PDC (59 to 83%), proportion with high compliance (47 to 76%), and proportion persisting with therapy (62 to 78%). Modest increases were identified among community-dwelling patients following data cleaning (mean PDC, 71 to 74%; high compliance, 54 to 58%; and persistence, 56 to 61%). CONCLUSIONS: Data cleaning to correct for exposure misclassification can influence estimates of adherence with oral bisphosphonate therapy, particularly in LTC. Results highlight the importance of developing data cleaning strategies to correct for exposure misclassification and improve transparency in pharmacoepidemiologic studies.

publication date

  • January 2015

has subject area