Skip to main content

Do patient decision aids improve patient adherence? Considering definitions and measures when interpreting systematic review findings

Date and Location




Monday 23 September 2013 - 10:30 - 12:00


Presenting author and contact person

Presenting author

Dawn Stacey

Contact person

Sophie Desroches
Abstract text
Background:Although adherence is an outcome in trials of patient decision aids (DAs), little is known about how it is defined or measured within these trials. Objectives:To assess the impact of DAs on patient adherence to the chosen treatment option, and to describe the definitions and measures for adherence used in these trials. Methods: A sub-analysis of randomized control trials included in the 2011 Cochrane Review of DAs for people facing health treatment or screening decisions was conducted. Two reviewers independently screened 86 trials for eligibility (e.g., measured adherence), extracted data (e.g. study characteristics, type of DA, results), and assessed risk of bias. Reviewers also extracted data on adherence definitions and measures, length of follow-up, and use of adherence as the primary outcome and/or for calculation of the sample size. Results: Eight eligible trials measured adherence to medications for menopause symptoms, atrial fibrillation, osteoporosis, depression, dyslipidemia, hypertension, and diabetes. Six trials compared DAs to usual care and two compared a detailed versus simple DAs. Each trial defined adherence differently. Three measured adherence to the chosen option, and five measured adherence to taking prescribed medication. All trials measured patient-reported adherence using three different instruments or their own question, and two also used pharmacy records. Follow-up was 2 to 36 months (median 6). Sample size was not calculated for adherence despite being a primary outcome in four trials. Although near perfect in both groups, one trial showed adherence to diabetes medications at 6 months based on pharmacy records was significantly better in the usual care group; while patient-reported adherence did not differ. The other seven trials reported no significant differences on adherence between groups. Conclusions: Adherence has been inconsistently defined and measured in trials of DAs. Therefore, the effectiveness of DAs for improving adherence to a chosen option remains unclear.