Browsing by Subject "predictive model"

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • Thumbnail Image
    ItemOpen Access
    A model to predict risk of blood transfusion after gynecologic surgery.
    (Am J Obstet Gynecol, 2017-05) Stanhiser, Jamie; Chagin, Kevin; Jelovsek, J Eric
    BACKGROUND: A model that predicts a patient's risk of receiving a blood transfusion may facilitate selective preoperative testing and more efficient perioperative blood management utilization. OBJECTIVE: We sought to construct and validate a model that predicts a patient's risk of receiving a blood transfusion after gynecologic surgery. STUDY DESIGN: In all, 18,319 women who underwent gynecologic surgery at 10 institutions in a single health system by 116 surgeons from January 2010 through June 2014 were analyzed. The data set was split into a model training cohort of 12,219 surgeries performed from January 2010 through December 2012 and a separate validation cohort of 6100 surgeries performed from January 2013 through June 2014. In all, 47 candidate risk factors for transfusion were collected. Multiple logistic models were fit onto the training cohort to predict transfusion within 30 days of surgery. Variables were removed using stepwise backward reduction to find the best parsimonious model. Model discrimination was measured using the concordance index. The model was internally validated using 1000 bootstrapped samples and temporally validated by testing the model's performance in the validation cohort. Calibration and decision curves were plotted to inform clinicians about the accuracy of predicted probabilities and whether the model adds clinical benefit when making decisions. RESULTS: The transfusion rate in the training cohort was 2% (95% confidence interval, 1.72-2.22). The model had excellent discrimination and calibration during internal validation (bias-corrected concordance index, 0.906; 95% confidence interval, 0.890-0.928) and maintained accuracy during temporal validation using the separate validation cohort (concordance index, 0.915; 95% confidence interval, 0.872-0.954). Calibration curves demonstrated the model was accurate up to 40% then it began to overpredict risk. The model provides superior net benefit when clinical decision thresholds are between 0-50% predicted risk. CONCLUSION: This model accurately predicts a patient's risk of transfusion after gynecologic surgery facilitating selective preoperative testing and more efficient perioperative blood management utilization.
  • Thumbnail Image
    ItemOpen Access
    Calibration of a comprehensive predictive model for the development of proximal junctional kyphosis and failure in adult spinal deformity patients with consideration of contemporary goals and techniques.
    (Journal of neurosurgery. Spine, 2023-06) Tretiakov, Peter S; Lafage, Renaud; Smith, Justin S; Line, Breton G; Diebo, Bassel G; Daniels, Alan H; Gum, Jeffrey; Protopsaltis, Themistocles; Hamilton, D Kojo; Soroceanu, Alex; Scheer, Justin K; Eastlack, Robert K; Mundis, Gregory; Nunley, Pierce D; Klineberg, Eric O; Kebaish, Khaled; Lewis, Stephen; Lenke, Lawrence; Hostin, Richard; Gupta, Munish C; Ames, Christopher P; Hart, Robert A; Burton, Douglas; Shaffrey, Christopher I; Schwab, Frank; Bess, Shay; Kim, Han Jo; Lafage, Virginie; Passias, Peter G

    Objective

    The objective of this study was to calibrate an updated predictive model incorporating novel clinical, radiographic, and prophylactic measures to assess the risk of proximal junctional kyphosis (PJK) and failure (PJF).

    Methods

    Operative patients with adult spinal deformity (ASD) and baseline and 2-year postoperative data were included. PJK was defined as ≥ 10° in sagittal Cobb angle between the inferior uppermost instrumented vertebra (UIV) endplate and superior endplate of the UIV + 2 vertebrae. PJF was radiographically defined as a proximal junctional sagittal Cobb angle ≥ 15° with the presence of structural failure and/or mechanical instability, or PJK with reoperation. Backstep conditional binary supervised learning models assessed baseline demographic, clinical, and surgical information to predict the occurrence of PJK and PJF. Internal cross validation of the model was performed via a 70%/30% cohort split. Conditional inference tree analysis determined thresholds at an alpha level of 0.05.

    Results

    Seven hundred seventy-nine patients with ASD (mean 59.87 ± 14.24 years, 78% female, mean BMI 27.78 ± 6.02 kg/m2, mean Charlson Comorbidity Index 1.74 ± 1.71) were included. PJK developed in 50.2% of patients, and 10.5% developed PJF by their last recorded visit. The six most significant demographic, radiographic, surgical, and postoperative predictors of PJK/PJF were baseline age ≥ 74 years, baseline sagittal age-adjusted score (SAAS) T1 pelvic angle modifier > 1, baseline SAAS pelvic tilt modifier > 0, levels fused > 10, nonuse of prophylaxis measures, and 6-week SAAS pelvic incidence minus lumbar lordosis modifier > 1 (all p < 0.015). Overall, the model was deemed significant (p < 0.001), and internally validated receiver operating characteristic analysis returned an area under the curve of 0.923, indicating robust model fit.

    Conclusions

    PJK and PJF remain critical concerns in ASD surgery, and efforts to reduce the occurrence of PJK and PJF have resulted in the development of novel prophylactic techniques and enhanced clinical and radiographic selection criteria. This study demonstrates a validated model incorporating such techniques that may allow for the prediction of clinically significant PJK and PJF, and thus assist in optimizing patient selection, enhancing intraoperative decision making, and reducing postoperative complications in ASD surgery.
  • Thumbnail Image
    ItemOpen Access
    Predicting development of severe clinically relevant distal junctional kyphosis following adult cervical deformity surgery, with further distinction from mild asymptomatic episodes.
    (Journal of neurosurgery. Spine, 2021-12) Passias, Peter G; Naessig, Sara; Kummer, Nicholas; Passfall, Lara; Lafage, Renaud; Lafage, Virginie; Line, Breton; Diebo, Bassel G; Protopsaltis, Themistocles; Kim, Han Jo; Eastlack, Robert; Soroceanu, Alex; Klineberg, Eric O; Hart, Robert A; Burton, Douglas; Bess, Shay; Schwab, Frank; Shaffrey, Christopher I; Smith, Justin S; Ames, Christopher P
    OBJECTIVE:This retrospective cohort study aimed to develop a formal predictive model distinguishing between symptomatic and asymptomatic distal junctional kyphosis (DJK). In this study the authors identified a DJK rate of 32.2%. Predictive models were created that can be used with high reliability to help distinguish between severe symptomatic DJK and mild asymptomatic DJK through the use of surgical factors, radiographic parameters, and patient variables. METHODS:Patients with cervical deformity (CD) were stratified into asymptomatic and symptomatic DJK groups. Symptomatic: 1) DJK angle (DJKA) > 10° and either reoperation due to DJK or > 1 new-onset neurological sequela related to DJK; or 2) either a DJKA > 20° or ∆DJKA > 20°. Asymptomatic: ∆DJK > 10° in the absence of neurological sequelae. Stepwise logistic regressions were used to identify factors associated with these types of DJK. Decision tree analysis established cutoffs. RESULTS:A total of 99 patients with CD were included, with 32.2% developing DJK (34.3% asymptomatic, 65.7% symptomatic). A total of 37.5% of asymptomatic patients received a reoperation versus 62.5% symptomatic patients. Multivariate analysis identified independent baseline factors for developing symptomatic DJK as follows: pelvic incidence (OR 1.02); preoperative cervical flexibility (OR 1.04); and combined approach (OR 6.2). Having abnormal hyperkyphosis in the thoracic spine, more so than abnormal cervical lordosis, was a factor for developing symptomatic disease when analyzed against asymptomatic patients (OR 1.2). Predictive modeling identified factors that were predictive of symptomatic versus no DJK, as follows: myelopathy (modified Japanese Orthopaedic Association score 12-14); combined approach; uppermost instrumented vertebra C3 or C4; preoperative hypermobility; and > 7 levels fused (area under the curve 0.89). A predictive model for symptomatic versus asymptomatic disease (area under the curve 0.85) included being frail, T1 slope minus cervical lordosis > 20°, and a pelvic incidence > 46.3°. Controlling for baseline deformity and disability, symptomatic patients had a greater cervical sagittal vertical axis (4-8 cm: 47.6% vs 27%) and were more malaligned according to their Scoliosis Research Society sagittal vertical axis measurement (OR 0.1) than patients without DJK at 1 year (all p < 0.05). Despite their symptomatology and higher reoperation rate, outcomes equilibrated in the symptomatic cohort at 1 year following revision. CONCLUSIONS:Overall, 32.2% of patients with CD suffered from DJK. Symptomatic DJK can be predicted with high reliability. It can be further distinguished from asymptomatic occurrences by taking into account pelvic incidence and baseline cervicothoracic deformity severity.