Do Lumbar Decompression and Fusion Patients Recall Their Preoperative Status?: A Cohort Study of Recall Bias in Patient-Reported Outcomes.

STUDY DESIGN: Prospective cohort study. OBJECTIVE: To characterize the accuracy of patient recollection of preoperative symptoms after lumbar spine surgery. SUMMARY OF BACKGROUND DATA: Although patient-reported outcomes have become important in the evaluation of spine surgery patients, the accuracy of patient recall remains unknown. METHODS: Patients undergoing lumbar decompression with or without fusion were enrolled. Back and leg Numeric Pain Scores and Oswestry Disability Indices were recorded preoperatively. Patients were asked to recall their preoperative status at a minimum of 1 year after surgery. Actual and recalled scores were compared using paired t tests and relations were quantified using Pearson correlation coefficients. Multivariable linear regression was used to identify factors that affected recollection. RESULTS: Sixty-two patients with a mean age of 66.1 years were included. Compared to their preoperative scores, patients showed significant improvement in back pain (mean difference [MD] = -3.2, 95% CI -4.0 to -2.4), leg pain (MD -3.3, 95% CI -4.3 to -2.2), and disability (MD -25.0%, 95% CI -28.7 to -19.6). Patient recollection of preoperative status was significantly more severe than actual for back pain (MD +2.3, 95% CI 1.5-3.2), leg pain (MD +1.8, 95% CI 0.9-2.7), and disability (MD +9.6%, 95% CI 5.6-14.0). No significant correlation between actual and recalled scores with regards to back (r = 0.18) or leg (r = 0.24) pain and only moderate correlation with disability (r = 0.44) were seen. This was maintained across age, sex, and time between date of surgery and recollection. More than 40% of patients switched their predominant symptom from back pain to leg pain or leg pain to back pain on recall. CONCLUSION: Relying on patient recollection does not provide an accurate measure of preoperative status after lumbar spine surgery. Recall bias indicates the importance of obtaining true baseline scores and patient-reported outcomes prospectively and not retrospectively. LEVEL OF EVIDENCE: 2.