Linking measurement error to minimal important change of patient-reported outcomes

Abstract

Objective: To provide a practical example showing how to relate measurement error to minimal important change (MIC) to estimate sample sizes required for detecting important changes in research, and to determine the usefulness of patient-reported outcomes (PROs) in daily clinical practice.

Study design and setting: The standard error of measurement (SEM) can be calculated as a parameter of measurement error. MIC can be determined by taking the mean change in score on the questionnaire in the subgroup of patients that reported to be "a little better" or "better" on a global rating scale minus the mean change in score in the subgroup of patients that reported to be "not changed." Based on SEM and MIC, formulas are presented that can be used to estimate sample sizes and to determine whether the questionnaire can be used to monitor changes in individual patients in daily clinical practice.

Results: A fully worked-out example is presented, using Western Ontario and McMaster University Osteoarthritis Index data from a cohort of patients undergoing hip or knee replacement.

Conclusion: We strongly recommend investigators to present the SEM and MIC of PROs and use these to estimate sample sizes for research and to determine the usefulness of these outcomes in daily clinical practice.