Dataset Information

Development and Validation of a Prediction Model for Pain and Functional Outcomes After Lumbar Spine Surgery.

ABSTRACT:

Importance

Functional impairment and pain are common indications for the initiation of lumbar spine surgery, but information about expected improvement in these patient-reported outcome (PRO) domains is not readily available to most patients and clinicians considering this type of surgery.

Objective

To assess population-level PRO response after lumbar spine surgery, and develop/validate a prediction tool for PRO improvement.

Design, setting, and participants

This statewide multicenter cohort was based at 15 Washington state hospitals representing approximately 75% of the state's spine fusion procedures. The Spine Surgical Care and Outcomes Assessment Program and the survey center at the Comparative Effectiveness Translational Network prospectively collected clinical and PRO data from adult candidates for lumbar surgery, preoperatively and postoperatively, between 2012 and 2016. Prediction models were derived for PRO improvement 1 year after lumbar fusion surgeries on a random sample of 85% of the data and were validated in the remaining 15%. Surgical candidates from 2012 through 2015 were included; follow-up surveying continued until December 31, 2016, and data analysis was completed from July 2016 to April 2017.

Main outcomes and measures

Functional improvement, defined as a reduction in Oswestry Disability Index score of 15 points or more; and back pain and leg pain improvement, defined a reduction in Numeric Rating Scale score of 2 points or more.

Results

A total of 1965 adult lumbar surgical candidates (mean [SD] age, 61.3 [12.5] years; 944 [59.6%] female) completed baseline surveys before surgery and at least 1 postoperative follow-up survey within 3 years. Of these, 1583 (80.6%) underwent elective lumbar fusion procedures; 1223 (77.3%) had stenosis, and 1033 (65.3%) had spondylolisthesis. Twelve-month follow-up participation rates for each outcome were between 66% and 70%. Improvements were reported in function, back pain, and leg pain at 12 months by 306 of 528 surgical patients (58.0%), 616 of 899 patients (68.5%), and 355 of 464 patients (76.5%), respectively, whose baseline scores indicated moderate to severe symptoms. Among nonoperative patients, 35 (43.8%), 47 (53.4%), and 53 (63.9%) reported improvements in function, back pain, and leg pain, respectively. Demographic and clinical characteristics included in the final prediction models were age, sex, race, insurance status, American Society of Anesthesiologists score, smoking status, diagnoses, prior surgery, prescription opioid use, asthma, and baseline PRO scores. The models had good predictive performance in the validation cohort (concordance statistic, 0.66-0.79) and were incorporated into a patient-facing, web-based interactive tool (https://becertain.shinyapps.io/lumbar_fusion_calculator).

Conclusions and relevance

The PRO response prediction tool, informed by population-level data, explained most of the variability in pain reduction and functional improvement after surgery. Giving patients accurate information about their likelihood of outcomes may be a helpful component in surgery decision making.

SUBMITTER: Khor S

PROVIDER: S-EPMC5875305 | biostudies-literature | 2018 Jul

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development and Validation of a Prediction Model for Pain and Functional Outcomes After Lumbar Spine Surgery.

Khor Sara S Lavallee Danielle D Cizik Amy M AM Bellabarba Carlo C Chapman Jens R JR Howe Christopher R CR Lu Dawei D Mohit A Alex AA Oskouian Rod J RJ Roh Jeffrey R JR Shonnard Neal N Dagal Armagan A Flum David R DR

JAMA surgery 20180701 7

<h4>Importance</h4>Functional impairment and pain are common indications for the initiation of lumbar spine surgery, but information about expected improvement in these patient-reported outcome (PRO) domains is not readily available to most patients and clinicians considering this type of surgery.<h4>Objective</h4>To assess population-level PRO response after lumbar spine surgery, and develop/validate a prediction tool for PRO improvement.<h4>Design, setting, and participants</h4>This statewide ...[more]

PMID: 29516096

Similar Datasets

Project description:Study designRetrospective database analysis.ObjectiveTo assess the effect glycemic control has on perioperative morbidity and mortality in patients undergoing elective degenerative lumbar spine surgery.Summary of background dataDiabetes mellitus (DM) is a prevalent disease of glucose dysregulation that has been demonstrated to increase morbidity and mortality after spine surgery. However, there is limited understanding of whether glycemic control influences surgical outcomes in patients with DM undergoing lumbar spine procedures for degenerative conditions.MethodsThe Nationwide Inpatient Sample was analyzed from 2002 to 2011. Hospitalizations were isolated on the basis of International Classification of Diseases, Ninth Revision, Clinical Modification, procedural codes for lumbar spine surgery and diagnoses codes for degenerative conditions of the lumbar spine. Patients were then classified into 3 cohorts: controlled diabetic, uncontrolled diabetic, and nondiabetic. Patient demographic data, acute complications, and hospitalization outcomes were determined for each cohort.ResultsA total of 403,629 (15.7%) controlled diabetic patients and 19,421 (0.75%) uncontrolled diabetic patients underwent degenerative lumbar spine surgery from 2002 to 2011. Relative to nondiabetic patients, uncontrolled diabetic patients had significantly increased odds of cardiac complications, deep venous thrombosis, and postoperative shock; in addition, uncontrolled diabetic patients also had an increased mean length of stay (approximately, 2.5 d), greater costs (1.3-fold), and a greater risk of inpatient mortality (odds ratio=2.6, 95% confidence interval=1.5-4.8, P<0.0009). Controlled diabetic patients also had increased risk of acute complications and inpatient mortality when compared with nondiabetic patients, but not nearly to the same magnitude as uncontrolled diabetic patients.ConclusionSuboptimal glycemic control in diabetic patients undergoing degenerative lumbar spine surgery leads to increased risk of acute complications and poor outcomes. Patients with uncontrolled DM, or poor glucose control, may benefit from improving glycemic control prior to surgery.Level of evidence3.

Project description:BackgroundEstimating surgical case duration accurately is an important operating room efficiency metric. Current predictive techniques in spine surgery include less sophisticated approaches such as classical multivariable statistical models. Machine learning approaches have been used to predict outcomes such as length of stay and time returning to normal work, but have not been focused on case duration.ObjectiveThe primary objective of this 4-year, single-academic-center, retrospective study was to use an ensemble learning approach that may improve the accuracy of scheduled case duration for spine surgery. The primary outcome measure was case duration.MethodsWe compared machine learning models using surgical and patient features to our institutional method, which used historic averages and surgeon adjustments as needed. We implemented multivariable linear regression, random forest, bagging, and XGBoost (Extreme Gradient Boosting) and calculated the average R2, root-mean-square error (RMSE), explained variance, and mean absolute error (MAE) using k-fold cross-validation. We then used the SHAP (Shapley Additive Explanations) explainer model to determine feature importance.ResultsA total of 3189 patients who underwent spine surgery were included. The institution's current method of predicting case times has a very poor coefficient of determination with actual times (R2=0.213). On k-fold cross-validation, the linear regression model had an explained variance score of 0.345, an R2 of 0.34, an RMSE of 162.84 minutes, and an MAE of 127.22 minutes. Among all models, the XGBoost regressor performed the best with an explained variance score of 0.778, an R2 of 0.770, an RMSE of 92.95 minutes, and an MAE of 44.31 minutes. Based on SHAP analysis of the XGBoost regression, body mass index, spinal fusions, surgical procedure, and number of spine levels involved were the features with the most impact on the model.ConclusionsUsing ensemble learning-based predictive models, specifically XGBoost regression, can improve the accuracy of the estimation of spine surgery times.

Project description:Length-of-hospital stay (LOS) after lumbar spine surgery (LSS) can be affected by many factors. However, few studies have evaluated predictors of LOS, and all have used limited number of variables as predictors.The purpose of the study was to identify presurgical, surgical, and postsurgical predictors of LOS after LSS.Retrospective review of consecutive patients who had LSS at the University of Kansas Hospital from October 2008 to April, 2012.Five hundred ninety-three patients underwent LSS consisting of laminotomy, laminectomy, or arthrodesis.LOS. Multiple presurgical, surgical, and postsurgical variables were extracted from the patients' medical records and considered as possible predictors (independent variables) of LOS.Potential predictors that were significantly correlated with LOS were used as indicators to construct three latent factors presurgical, surgical, and postsurgical, which were in turn used to predict LOS in a structural equation model.The average LOS was 4.01±2.73 days. The presurgical factor was indicated by age (61.97±14.49 years), previous level of function (60.5% were totally independent), previous hemoglobin level (13.70±1.36 mg/dL), and use of assistive devices (60% were assistive device users). The surgical factor was indicated by severity of illness (50.2% had minor disease severity), presence of complications (1.9%), and stay in an intensive care unit (4.0%). The postsurgical factor was indicated by postsurgical walking distance (166.43±175.75 ft), level of assistance during walking (5.18±0.81 out of 7 points), balance scores (6.18±1.82 out of 10 points), and bed mobility and transfer dependency scores (9.81±1.99 out of 14 points). These three latent factors explained 47% of variation in LOS.Postsurgical factors predicted the highest variation in LOS in comparison with presurgical and surgical factors and should be taken into consideration for discharge planning. Postsurgical factors are related to the patient's function, modifiable with rehabilitation, and can be improved to shorten LOS. Inclusion of more reliable and standardized presurgical variables could improve the predictability of the model.

Dataset Information

Development and Validation of a Prediction Model for Pain and Functional Outcomes After Lumbar Spine Surgery.

Importance

Objective

Design, setting, and participants

Main outcomes and measures

Results

Conclusions and relevance

Publications

Development and Validation of a Prediction Model for Pain and Functional Outcomes After Lumbar Spine Surgery.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets