Impact of unplanned readmissions on lumbar surgery outcomes: a national study of 33,447 patients

Unplanned readmissions following lumbar spine surgery have immense clinical and financial implications. However, little is known regarding the impact of unplanned readmissions on patient-reported outcomes (PROs) following lumbar spine surgery. To evaluate the impact of unplanned readmissions, including specific readmission reasons, on patient reported outcomes 12 months after lumbar spine surgery. A retrospective cohort study of prospectively collected data was conducted using patients included in the lumbar module of the Quality and Outcomes Database (QOD), a national, multicenter spine registry. A total of 33,447 patients who underwent elective lumbar spine surgery for degenerative diseases were included. Mean age was 59.8 (SD=14.04), 53.6% were male, 89.5% were white, 45.9% were employed, and 47.5% had private insurance. Unplanned 90-day readmissions and 12-month patient-reported outcomes (PROs) including numeric rating scale (NRS) scores for back and leg pain, Oswestry Disability Index (ODI) scores, EuroQol-5 Dimension (EQ-5D) scores, and North American Spine Society (NASS) patient-satisfaction scores. The lumbar module of the QOD was queried for adults undergoing elective lumbar spine surgery for degenerative disease. Unplanned 90-day readmissions were classified into 4 groups: medical, surgical, pain-only, and no readmissions. Medical and surgical readmissions were further categorized into primary reason for readmission. 12-month PROs assessing patient back and leg pain (NRS), disability (ODI), quality of life (EQ-5D), and patient satisfaction were collected. Multivariable models predicting 12-month PROs were built controlling for covariates. A total of 31,430 patients (94%) had no unplanned readmission while 2,017 patients (6%) had an unplanned readmission within 90 days following lumbar surgery. Patients with readmissions had significantly worse 12-month PROs compared with those with no unplanned readmissions in covariate-adjusted models. Using Wald-df as a measure of predictor importance, surgical readmissions were associated with the worst 12-month outcomes, followed by pain-only, then medical readmissions. In separate covariate adjusted models, we found that readmissions for pain, SSI/wound dehiscence, and revisions were among the most important predictors of worse outcomes at 12-months. Unplanned 90-day readmissions were associated with worse pain, disability, quality of life, and greater dissatisfaction at 12-months, with surgical readmissions