The Oswestry Disability Index and 12-Item Short Form Health Survey Physical Component Scores Are Not Affected by Recall Bias in Posterior Lumbar Spine Surgery Patients: A Prospective Study Using Data From Fitness Trackers.

Abstract

To assess the effects of recall bias on prospectively collected patient-reported outcome (PRO) measures after lumbar laminectomy by analyzing correlations between PROs and step counts in measurement windows preceding PRO measurement. Responses to the 12-item Short Form Health Survey (SF-12) and the Oswestry Disability Index (ODI) were collected postoperatively from 22 patients. Accelerometers recorded daily step counts. Median and maximum step counts were calculated for windows (1 day, 3 days, 1 week, and 2 weeks) preceding PRO measurement. Spearman rank correlation coefficients between PROs and step counts were calculated. Median and maximum step counts from 1- and 2-week windows more consistently correlated with SF-12 Physical Component Scores scores than 1- and 3-day windows over the postoperative period. Median steps from 1-, 2-week, and 3-day windows correlated more with ODI scores than the 1-day window. Maximum steps from 1- and 2-week windows correlated more than 1- and 3-day windows. PROs had higher concordance with step counts from the 1 week and 2 weeks before PRO measurement than the 1 day and 3 days prior. We therefore conclude that SF-12 and ODI are not markedly affected by recall bias because scores are not skewed by the events of the days immediately preceding measurement.