Evaluating Recall Periods for Patient-Reported Outcome Measures: A Systematic Review of Quantitative Methods

The current guidance for selection of recall periods recommends considering the design of the study, nature of the condition, patient's burden and ability to recall, and intent of the outcome measure. Empirical study of the accuracy of recall periods is recommended; however, there is not consensus on how to quantitatively evaluate the consistency of results from patient-reported outcome measures (PROMs) with different recall periods. We conducted a systematic review to describe quantitative methods for evaluating results obtained from PROMs with differing recall periods to lay the groundwork for establishing consensus.

We searched MEDLINE, Embase, Scopus, and American Psychological Association PsycINFO for studies where participants are given the same health-related measure (eg, quality of life, well-being, functioning, and pain) with differing recall periods.

A total of 7174 abstracts were screened. The 30 included studies reflected a wide range of domains, including pain, fatigue, and sexual behavior and function. The recall periods ranged from momentary to 6 months. The analytic approaches varied, including different methods for assessing relative agreement, absolute agreement, and for assessing combined relative and absolute agreement.

We found variability in how PROM recall periods were evaluated, suggesting an opportunity for greater consensus on methodological approach. As a starting point, we provide recommendations for which methods are preferred for which contexts.