Reoperations after lumbar disc surgery - A population-based study of regional and interspecialty variations

Study Design. A follow-up study using nationwide administrative databases. Objectives. To explore rates of reoperation after lumbar disc surgery and t heir regional and interspecialty variations. Summary of Background Data. In many Western countries, rates of lumbar disc surgery display significant geographic variations suggesting varying treat ment criteria among operating surgeons. Few population-based studies have e xplored the risk of reoperation after disc surgery, and regional or intersp ecialty variations in the reoperations are unknown. Methods. Patients who underwent lumbar spine surgery from January 1, 1987 t hrough December 31, 1995, were identified in the Finnish Hospital Discharge Register. Data an the patients' initial disc operations, subsequent operat ions, and cause-of-death records were linked using personal identification codes. The Kaplan-Meier method and proportional hazard model were used to a nalyze risks of reoperation after initial surgery, according to hospital ca tchment area rates of disc surgery and for neurosurgical and orthopedic pat ients of university hospitals. Results. 12.3% of 25,359 surgical patients with herniated lumbar discs unde rwent subsequent lumbar operations corresponding to the cumulative risk of 18.9% in the 9-year follow-up. Reoperation rates increased during the study period with the recent patient cohorts exhibiting risks. The reoperation r isk showed a systematic geographic variation: the higher the regional disc surgery rate, the higher the reoperation risk. Overall, neurosurgical patie nts had a higher reoperation risk than orthopedic patients (relative risk [ RR]: 1.57, 95% confidence interval [Cl]: 1.17-2.10), but this was not a uni form finding. Conclusions. The reoperation risk after disc surgery increased during the s tudy period and was higher in hospital catchment areas with higher overall discectomy rates. The reoperation risks varied among the university hospita ls but tended to be higher for neurosurgical rather than for orthopedic pat ients.