Using resource use logs to reduce the amount of missing data in economic evaluations alongside trials

Economic evaluations alongside randomized controlled trials that collect data using patient-completed questionnaires are prone to missing data. Our objective was to determine whether giving patients a resource use log (RUL) at baseline would improve the odds of completing questions in a follow-up resource use questionnaire (RUQ) and to identify patients' views on RUL's usefulness and acceptability.

The RUL study was a randomized controlled trial and qualitative study nested within a larger randomized controlled trial (the Arthroplasty Pain Experience Study trial). Eighty-five patients were randomized at baseline to receive or not receive an RUL. At 3-month follow-up, all participants received a postal RUQ. We created dummy variables for 13 resource use categories indicating whether complete information had been given for each category. We compared the completion rates between arms by using descriptive statistics and logistic regression. We explored patients' experience of using the RUL by interviewing a different subsample of Arthroplasty Pain Experience Study patients (n = 24) at 2- to 4-week follow-up.

At 3 months, 74 of the 85 (87% in each arm) patients returned the RUQ. Patients in the RUL arm were 3.5 times more likely to complete the National Health Service community-based services category (P = 0.08). The RUL was positively received by patients and was generally seen as a useful memory aid.

The RUL is a useful and acceptable tool in reducing the amount of missing data for some types of resource use.