Reliability of retrospective clinical data to evaluate the effectiveness of lumbar fusion in chronic low back pain.

STUDY DESIGN: Patients in whom a posterior spinal fusion instrumentation had been performed to treat low back pain were asked to recall their preoperative clinical status by retrospectively filling out the same 3 self-evaluation scales they had completed before surgery in a prospective fashion.

OBJECTIVES: To evaluate the impact of recollection error and compare outcomes using retrospective versus prospective methodologies among a cohort of patients treated with posterior spinal fusion instrumentation.

SUMMARY OF BACKGROUND DATA: Literature on spine surgery from 1990 to 2000 shows a greater increase in retrospective studies as compared to randomized controlled trials and other prospective studies. Cross-sectional studies evaluate therapeutic effectiveness by comparing the current condition with the recalled (retrospectively recorded) pretreatment condition. There are no studies analyzing the characteristics of recalled data in a cohort of patients with chronic low back pain treated with posterior spinal fusion instrumentation.

METHODS: The preoperative clinical status of 58 patients, 33 women and 25 men, with a mean age of 48.3 years (22-84 years) was assessed prospectively with 3 self evaluation questionnaires and retrospectively at a mean of 37.5 months (2-58 months) after surgery using the same questionnaires. The Wilcoxon test was used to compare prospective and retrospective preoperative data and to compare prospective outcomes with outcomes determined from cross-sectional data. Agreement between prospective and retrospective measures was estimated with intraclass correlation coefficients for absolute agreement and consistency.

RESULTS: Comparisons between prospective and recalled data showed significant differences, demonstrating a worse preoperative situation when using retrospective data. Assessment of treatment effectiveness showed that cross-sectional evaluation significantly improved the real surgical outcome. Both absolute agreement and consistency intraclass correlation coefficients showed poor agreement between prospective and cross-sectional data, revealing no systematic bias. Follow-up, age, and gender did not modify agreement and cross-sectional overestimation.

CONCLUSIONS: Relying on a patient's recall of preoperative clinical status is not an accurate method to evaluate surgical outcome after posterior spinal fusion instrumentation. Cross-sectional studies may overestimate the effectiveness of surgery.