TY - JOUR
T1 - Patient-specific prediction model for clinical and quality-of-life outcomes after lumbar spine surgery
AU - Lubelski, Daniel
AU - Feghali, James
AU - Nowacki, Amy S.
AU - Alentado, Vincent J.
AU - Planchard, Ryan
AU - Abdullah, Kalil G.
AU - Sciubba, Daniel M.
AU - Steinmetz, Michael P.
AU - Benzel, Edward C.
AU - Mroz, Thomas E.
N1 - Publisher Copyright:
© 2021 American Association of Neurological Surgeons. All rights reserved.
PY - 2021/4
Y1 - 2021/4
N2 - OBJECTIVE Patient demographics, comorbidities, and baseline quality of life (QOL) are major contributors to postoperative outcomes. The frequency and cost of lumbar spine surgery has been increasing, with controversy revolving around optimal management strategies and outcome predictors. The goal of this study was to generate predictive nomograms and a clinical calculator for postoperative clinical and QOL outcomes following lumbar spine surgery for degenerative disease. METHODS Patients undergoing lumbar spine surgery for degenerative disease at a single tertiary care institution between June 2009 and December 2012 were retrospectively reviewed. Nomograms and an online calculator were modeled based on patient demographics, comorbidities, presenting symptoms and duration of symptoms, indication for surgery, type and levels of surgery, and baseline preoperative QOL scores. Outcomes included postoperative emergency department (ED) visit or readmission within 30 days, reoperation within 90 days, and 1-year changes in the EuroQOL-5D (EQ-5D) score. Bootstrapping was used for internal validation. RESULTS A total of 2996 lumbar surgeries were identified. Thirty-day ED visits were seen in 7%, 30-day readmission in 12%, 90-day reoperation in 3%, and improvement in EQ-5D at 1 year that exceeded the minimum clinically important difference in 56%. Concordance indices for the models predicting ED visits, readmission, reoperation, and dichotomous 1-year improvement in EQ-5D were 0.63, 0.66, 0.73, and 0.84, respectively. Important predictors of clinical outcomes included age, body mass index, Charlson Comorbidity Index, indication for surgery, preoperative duration of symptoms, and the type (and number of levels) of surgery. A web-based calculator was created, which can be accessed here: https://riskcalc.org/PatientsEligibleForLumbarSpineSurgery/. CONCLUSIONS The prediction tools derived from this study constitute important adjuncts to clinical decision-making that can offer patients undergoing lumbar spine surgery realistic and personalized expectations of postoperative outcome. They may also aid physicians in surgical planning, referrals, and counseling to ultimately lead to improved patient experience and outcomes.
AB - OBJECTIVE Patient demographics, comorbidities, and baseline quality of life (QOL) are major contributors to postoperative outcomes. The frequency and cost of lumbar spine surgery has been increasing, with controversy revolving around optimal management strategies and outcome predictors. The goal of this study was to generate predictive nomograms and a clinical calculator for postoperative clinical and QOL outcomes following lumbar spine surgery for degenerative disease. METHODS Patients undergoing lumbar spine surgery for degenerative disease at a single tertiary care institution between June 2009 and December 2012 were retrospectively reviewed. Nomograms and an online calculator were modeled based on patient demographics, comorbidities, presenting symptoms and duration of symptoms, indication for surgery, type and levels of surgery, and baseline preoperative QOL scores. Outcomes included postoperative emergency department (ED) visit or readmission within 30 days, reoperation within 90 days, and 1-year changes in the EuroQOL-5D (EQ-5D) score. Bootstrapping was used for internal validation. RESULTS A total of 2996 lumbar surgeries were identified. Thirty-day ED visits were seen in 7%, 30-day readmission in 12%, 90-day reoperation in 3%, and improvement in EQ-5D at 1 year that exceeded the minimum clinically important difference in 56%. Concordance indices for the models predicting ED visits, readmission, reoperation, and dichotomous 1-year improvement in EQ-5D were 0.63, 0.66, 0.73, and 0.84, respectively. Important predictors of clinical outcomes included age, body mass index, Charlson Comorbidity Index, indication for surgery, preoperative duration of symptoms, and the type (and number of levels) of surgery. A web-based calculator was created, which can be accessed here: https://riskcalc.org/PatientsEligibleForLumbarSpineSurgery/. CONCLUSIONS The prediction tools derived from this study constitute important adjuncts to clinical decision-making that can offer patients undergoing lumbar spine surgery realistic and personalized expectations of postoperative outcome. They may also aid physicians in surgical planning, referrals, and counseling to ultimately lead to improved patient experience and outcomes.
KW - Calculator
KW - Lumbar
KW - Nomogram
KW - Outcome
KW - Prediction
KW - Quality of life
KW - Spine surgery
UR - http://www.scopus.com/inward/record.url?scp=85103471235&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85103471235&partnerID=8YFLogxK
U2 - 10.3171/2020.8.SPINE20577
DO - 10.3171/2020.8.SPINE20577
M3 - Article
C2 - 33528964
AN - SCOPUS:85103471235
SN - 1547-5654
VL - 34
SP - 580
EP - 588
JO - Journal of Neurosurgery: Spine
JF - Journal of Neurosurgery: Spine
IS - 4
ER -