Clinical Epidemiology

BackgroundPatients undergoing surgery for traumatic spinal fractures face a substantially elevated risk of postoperative lower extremity deep vein thrombosis (DVT). While generic risk assessment tools exist, a purpose-built model integrating spine-specific and readily available preoperative predictors is lacking. This study aimed to develop and internally validate a novel predictive model for this specific complication. MethodsThis retrospective cohort study analyzed data from 1,676 patients who underwent surgery for traumatic spinal fractures at a single center. All patients received standardized DVT surveillance. The cohort was randomly split into training (70%) and testing (30%) sets. Univariate and multivariable logistic regression with stepwise selection were used to identify independent predictors from 29 candidate variables. Model performance was evaluated by its discriminative ability (area under the curve, AUC), calibration (calibration curves and Hosmer-Lemeshow test), and clinical utility (decision curve analysis, DCA). A nomogram was constructed for clinical use. ResultsThe incidence of postoperative DVT was 14.26% (239/1,676). Six independent preoperative predictors were identified: prolonged bed rest > 72 hours (adjusted odds ratio [aOR] = 5.208), pre-existing lower extremity vascular disease (aOR = 2.938), elevated D-dimer (aOR = 1.582), elevated fibrinogen (aOR = 1.434), severe neurological impairment (ASIA grade A/B), and advanced age (aOR = 1.019). The model demonstrated robust discrimination (AUC: 0.891 training, 0.885 testing) and excellent calibration (Hosmer-Lemeshow p > 0.7), with high sensitivity (90.5- 91.2%) and moderate specificity (74.3-74.5%). Decision curve analysis confirmed its clinical utility across a wide range of threshold probabilities. ConclusionWe developed and validated a parsimonious and clinically practical prediction model for postoperative DVT in traumatic spinal fracture patients. This tool, which leverages six preoperatively accessible variables, facilitates individualized risk stratification and could guide the implementation of targeted prophylactic strategies to improve patient outcomes.

ObjectiveTo compare the predictive accuracy of the Risk Analysis Index (RAI) versus the 5-factor Modified Frailty Index (mFI-5) for postoperative outcomes in olecranon fracture open reduction internal fixation (ORIF). MethodsThis retrospective study analyzed 3,987 patients from the ACS-NSQIP database who underwent olecranon ORIF between 2015-2020. Outcomes included 30-day mortality, non-home discharge (NHD), complications, readmission, and extended length of stay. Predictive accuracy was assessed using area under ROC curves (AUROC). ResultsRAI demonstrated superior predictive accuracy for NHD (AUROC: 0.81 vs 0.68, p<0.001), major complications (AUROC: 0.72 vs 0.65, p=0.05), and reoperation (AUROC: 0.63 vs 0.57, p=0.03) compared to mFI-5. Severely frail patients identified by RAI showed significantly increased odds for NHD (OR: 4.78, p=0.005), extended length of stay (OR: 2.83, p=0.008), and major complications (OR: 9.23, p=0.03). No significant differences were found between indices for mortality, minor complications, or readmission rates. ConclusionThe RAI demonstrates superior discriminatory accuracy compared to mFI-5 for predicting adverse outcomes after olecranon ORIF, particularly for NHD and major complications. Implementation of RAI in preoperative assessment may improve risk stratification and resource allocation for olecranon fracture patients.

BackgroundPostoperative new-onset deep vein thrombosis (PNO-DVT) of the lower extremities represents a prevalent and serious clinical complication following pelvic fractures, which substantially impedes patient rehabilitation and diminishes quality of life. Conventional risk assessment methodologies exhibit inherent limitations, rendering them inadequate for precise prediction and individualized management of DVT. In recent years, machine learning techniques have demonstrated significant advantages in data analysis, emerging as promising tools for predicting postoperative DVT risk. This study sought to investigate the predictive efficacy of machine learning models for the development of new-onset lower extremity deep vein thrombosis following pelvic fracture surgery. MethodsData from 745 patients who underwent pelvic fracture surgery at our hospital between January 2016 and December 2019 were collected. The analysis encompassed demographic information, general patient data, preoperative laboratory test results, surgical details, and scoring systems. Initially, the data were analyzed using univariate logistic regression, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression to identify 12 independent risk factors, including age, HDL-C, and ApoB. Subsequently, the dataset was partitioned into a training set and a test set at a 7:3 ratio. Six models were employed for analysis, including logistic regression, support vector machine (SVM), random forest, XGBoost, LightGBM, and AdaBoost. ConclusionComparative analysis of the six machine learning models revealed XGBoost exhibited the highest performance (AUC: 0.8633), followed by LightGBM (0.8349), random forest (0.8055), logistic regression (0.7503), SVM (0.7505), and AdaBoost (0.8179). Model sensitivity ranged from 0.3684 to 0.8421, and accuracy ranged from 0.6502 to 0.9238.

PurposeCompared with the general population, people with intellectual disabilities (ID) have higher incidences of major osteoporotic fracture (MOF) and hip fracture (HF), and osteoporosis develops at a younger age. The rate of HF in those aged 50 years and over is two and four times higher than that in women and men, respectively, without ID. It is essential to identify people with ID at risk of such fractures so that a targeted fracture prevention strategy can be designed. However, current fracture prediction models are derived from the general population and may underestimate risk in the ID population. MethodsPrediction models (IDFracture) for the 10-year risk of HF and MOF were developed and validated in populations of people with ID aged 30-79 years. Models were developed in the CPRD GOLD database and temporally validated in the Aurum database. The predictors included those in current fracture prediction models and ID-specific predictors such as Down syndrome. All the predictors were included in the Cox regression models. Bootstrapping was used to adjust for overfitting. ResultsThe development cohort included 38,665 people with IDs, 1045 with MOFs and 360 with HFs within 10 years. The external validation cohort included 76,385 people, 2420 MOFs and 1001 HFs. Discrimination, as judged by the C statistic, was good: MOF 0.775, HF 0.839. The calibration was also good but tended to overpredict at the highest predicted risks. ConclusionIDFracture has potential as a screening tool in clinical practice to identify people with ID who are at increased risk of MOF and HF. Mini AbstractO_ST_ABSBrief rationaleC_ST_ABSPeople with intellectual disability (ID) have a relatively high incidence of fracture, so current risk prediction models are not appropriate. Main resultA new prediction model for people with ID showed good calibration and discrimination in external validation. Significance of paperThis is the first such prediction model developed for people with ID.

Introduction/AimsMultiple screening tools exist for identifying patients at risk of extended stay following lower limb arthroplasty. Use of these models at other hospital sites requires verification of appropriate data coverage and evidence of validity in a new population. The aim of this study was to adapt and assess 1) data compatibility, 2) discrimination, and 3) calibration of three published models for identifying patients at risk of an extended (5+ day) stay, or those likely to stay for the target 3 or fewer days following lower limb arthroplasty. MethodsRetrospective study, utilising a randomly selected (N=200 of a total 331 available in the electronic medical record) cohort of lower-limb Total Joint Arthroplasty (TJA) patients, to externally validate an adaptation of predictive tools and regression models published by three independent groups: Winemaker et al (2015)1, Oldmeadow et al (2003)2 and Gabriel et al (2018)3. Electronic medical records of a single, medium-sized public hospital were accessed to extract data required for the models and respective predictive tools, and model characteristics (included predictors, data coding, sample sizes) were modified according to the available data. ResultsThe study cohort comprised 200 patients (60% female) at a median 70yrs of age (IQR 62-75). Approximately 58% received total knee arthroplasty (TKA) and 42% underwent total hip arthroplasty (THA). The two prediction tools and three regression models all required modifications due to data items being unavailable in the electronic records. A modification of the RAPT tool applied to 176 eligible patients resulted in sensitivity of 85.71% (95%CI 71.46-94.57) and poor specificity 32.09% (24.29-40.70), with 68% of short-stay patients classified in the high risk group. Adaptation of the second tool to 85 eligible patients resulted in unreliable estimates of sensitivity due to limited data. The three adapted regression models performed similarly well with regard to discrimination when used to predict patients staying for 5 days or longer (concordance index: Winemaker et al:, 0.79, n=198; Oldmeadow et al: 0.79, n=176), or those staying 3 days or less (Gabriel et al: 0.70, n=199). Estimates of calibration suggested the models were relatively well calibrated (spiegelhalter Z -0.01-0.29, p>0.05), although calibration plots indicated some variation remained unaccounted for, particularly with patients considered at intermediate risk. ConclusionThe three resulting regression models performed adequately in terms of discrimination and calibration for identification of patients at risk of an extended stay. However, comparison with published models was hampered by systemic issues with data compatibility. Further evaluation of such models in a specific hospital setting should incorporate improvements in data collection, and establish key thresholds for use in targeting resources to patients in need of greater support.

ObjectivesMortality risk scores, such as SOFA, qSOFA, and CURB-65, are quick, effective tools for communicating a patients prognosis and guiding therapeutic decisions. Most use simple calculations that can be performed by hand. While several COVID-19 specific risk scores exist, they lack the ease of use of these simpler scores. The objectives of this study were (1) to design, validate, and calibrate a simple, easy-to-use mortality risk score for COVID-19 patients and (2) to recalibrate SOFA, qSOFA, and CURB-65 in a hospitalized COVID-19 population. DesignRetrospective cohort study incorporating demographic, clinical, laboratory, and admissions data from electronic health records. SettingMulti-hospital health system in New York City. Five hospitals were included: one quaternary care facility, one tertiary care facility, and three community hospitals. ParticipantsPatients (n=4840) with laboratory-confirmed SARS-CoV2 infection who were admitted between March 1 and April 28, 2020. Main outcome measuresGrays K-sample test for the cumulative incidence of a competing risk was used to assess and rank 48 different variables associations with mortality. Candidate variables were added to the composite score using DeLongs test to evaluate their effect on predictive performance (AUC) of in-hospital mortality. Final AUCs for the new score, SOFA, qSOFA, and CURB-65 were assessed on an independent test set. ResultsOf 48 variables investigated, 36 (75%) displayed significant (p<0.05 by Grays test) associations with mortality. The variables selected for the final score were (1) oxygen support level, (2) troponin, (3) blood urea nitrogen, (4) lymphocyte percentage, (5) Glasgow Coma Score, and (6) age. The new score, COBALT, outperforms SOFA, qSOFA, and CURB-65 at predicting mortality in this COVID-19 population: AUCs for initial, maximum, and mean COBALT scores were 0.81, 0.91, and 0.92, compared to 0.77, 0.87, and 0.87 for SOFA. We provide COVID-19 specific mortality estimates at all score levels for COBALT, SOFA, qSOFA, and CURB-65. ConclusionsThe COBALT score provides a simple way to estimate mortality risk in hospitalized COVID-19 patients with superior performance to SOFA and other scores currently in widespread use. Evaluation of SOFA, qSOFA, and CURB-65 in this population highlights the importance of recalibrating mortality risk scores when they are used under novel conditions, such as the COVID-19 pandemic. This studys approach to score design could also be applied in other contexts to create simple, practical and high-performing mortality risk scores. Trial registrationNA Funding sourceThe authors declare that there was no external funding provided. Summary boxO_ST_ABSWhat is already known on this topicC_ST_ABSO_LIMortality risk scores are widely used in clinical settings to facilitate communication with patients and families, guide goals of care discussions, and optimize resource allocation. C_LIO_LIAlthough popular mortality risk scores like SOFA, qSOFA, and CURB-65 are routinely used in COVID-19 populations, they were originally calibrated in different contexts and their true performance among hospitalized COVID-19 patients is unknown. C_LIO_LISeveral dedicated COVID-19 mortality risk scores have been created during the 2019-2020 pandemic, but all use complicated formulae or machine learning algorithms and are difficult or impossible to calculate by hand, limiting their applicability at the bedside. C_LI What this study addsO_LIWe describe a data-driven, simple, and hand-calculable COVID-specific mortality risk score (COBALT) that has superior performance to SOFA, qSOFA, and CURB-65 in a hospitalized COVID-19 patient population. C_LIO_LIWe provide COVID-specific mortality estimates for SOFA, qSOFA, and CURB-65 using data from 4840 patients in a large and diverse New York City multihospital health system. C_LI

BackgroundVTE is a serious postoperative complication after surgery with resultant higher morbidity and mortality. Despite years of experience with current risk models, rates continue to be high and more information is needed on individual patient risk in the prophylaxis era. Research QuestionsCan we assess the individualized risk of postoperative venous thromboembolism (VTE) for broad categories of surgery? MethodsThis study was performed using data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) Database. Patient data (n=2,875,190) from 2015-2017 were used for study analysis. Eight predictors were selected for the model: age, preoperative platelet count[&ge;]450 (x109/L), disseminated cancer, corticosteroid use, serum albumin [&le;]2.5 g/dL, preoperative sepsis, hospital length of stay and surgery type. The second model included 7 predictors without hospital length of stay. A predictive model was trained using ACS-NSQIP data from 2015-2016 (n=1,859,227) and tested using data from 2017 (n= 1,015,963). Primary outcomes are postoperative 30-day VTE, including deep vein thrombosis (DVT) and/or pulmonary embolism (PE). ResultsVTE occurred in 23,249 patients (0.81%) and 49.9% of VTE occurred after discharge from index hospitalization. The risk prediction model had high AUC (area under the receiver operating characteristic curve) for postoperative VTE of 0.78 (training cohort) and 0.78 (test cohort). InterpretationThis clinical prediction model is a validated, practical and easy-to-use tool to identify surgical patients at the highest risk of postoperative VTE and provide an individualized assessment of risk based on clinical factors and type of surgery. This prediction model may be used as a tool to assess individualized risk of postoperative VTE and promote broader discussion and awareness of the VTE risk during the perioperative period.

ObjectiveCompare the effect of posterior surgical approach (PA) vs direct anterior approach (DAA) on speed of recovery from pain and dysfunction and on intrathecal oxytocin analgesia after total hip arthroplasty (THA). Study designNested cohort within a randomized, controlled, double-blind trial SettingHospital SubjectsIndividuals aged 31 to 80 years undergoing total hip arthroplasty (THA) MethodsIn this secondary analysis of a randomized controlled trial, the association between surgical approach and number of daily steps was assessed, and whether this was modified by receipt of intrathecal oxytocin. Data were collected from accelerometers and daily patient diaries in the first eight weeks postoperatively. Outcomes were analyzed using generalized linear regression models. ResultsNinety patients underwent THA, of which 35 (38.9%) received a PA. Patients were predominantly female (57.8%) with a mean age of 60.6 (standard deviation [SD] 9.3) years. On postoperative day one patients who received a PA with placebo took more steps (mean difference [MD] 53.72, 95% CI: -1717.86, 1825.31) than patients who underwent a DAA. Trajectories were significantly modified by whether they received oxytocin, in which patients who underwent PA with oxytocin took more steps than patients who underwent DAA (p<0.001). DiscussionFurther studies are needed to understand mechanisms underlying oxytocins interaction with surgical approach and guide considerations for recovery after THA.

BackgroundPhysician specialists (PS) are often the type of healthcare provider initially contacted by an individual with low back pain (LBP). LBP clinical practice guidelines (CPG) recommend a stepped approach to management with an emphasis on first-line non-pharmaceutical and non-interventional services. ObjectiveExamine the association between the incorporation of CPG recommended first-line services, exposure to second- and third-line services and total episode cost for individuals with non-surgical LBP initially contacting a PS. DesignRetrospective observational study with identical design to previous study focused on primary care physicians. Setting/PatientsNational sample of individuals with non-surgical LBP occurring in 2017-2019. MeasurementsIndependent variables were initial contact with a PS, and the timing of incorporation of five types of first-line services. Dependent measures included exposure to thirteen types of health care services and total episode cost. Results91,096 individuals were associated with 98,992 episodes of non-surgical LBP. 36.2% of the 33,277 PS initially contacted for an episode of LBP incorporated any first-line service at any time during an episode. A first-line service was provided in 24.0% of episodes with active care (19.5% of episodes), manual therapy (13.7%) and chiropractic manipulative therapy (6.5%) the most common. 7.3% of non-surgical LBP episodes included a first-line service within seven days of initial contact with a PS. These episodes were associated with a reduction in the use of prescription skeletal muscle relaxants (risk ratio (RR) 0.88) and opioids (RR 0.55), spinal injections (RR 0.84), and CT scans (RR 0.71), with no impact on the use of prescription NSAIDs, radiography, or MRI scans. First-line services were associated with an increase in total episode cost at any time of incorporation with chiropractic manipulation associated with the lowest cost increase. Younger individuals from zip codes with higher adjusted gross income were more likely to receive a first-line service in the first seven days of an episode. LimitationsAs a retrospective observational analysis of associations there are numerous potential confounders and limitations. ConclusionsFor individuals with non-surgical LBP PS provide second- or third-line services more frequently and earlier than CPG recommended first-line services. There is an opportunity to improve concordance with LBP CPGs for individuals with LBP initially contacting a PS.

PurposeThe objective was to investigate the consistency in cumulative revision rates for a selection of total hip arthroplasty cups and stems across national/regional hip arthroplasty registries worldwide. MethodsTen cups and 10 stems from total hip systems were randomly selected. Two frequently used implants across registries were added, totalling 11 cups and 11 stems. Cumulative revision rates (CRR) and 95%CIs were extracted from the latest annual registry reports using these implants. CRRs were pooled for each cup or stem, and differences between cup-stem combinations and between registries were investigated. ResultsCRRs were available for 10 cups and 8 stems from 8 registries, totalling 552,148 cups and 727,447 stems. Follow-up was 1-20 years. Five-year CRRs pooled on all cups was 2.9% (95%CI 2.3 to 3.6) and on all stems 3.0% (95%CI 2.4 to 3.8). Homogenous (consistent) CRRs with respect to both, associated implant and country, were observed for 2 cups and 3 stems. Significant differences in CRR were identified in 1 cup by associated implant only, in 1 cup by registry only, and in 2 cups and 4 stems for both. Sparse data prevented evaluation of 4 cups and 1 stem. ConclusionRegistries annual reports provide a large amount of publicly available information on CRRs of specific implants. These CRRs can be synthesized to improve the assessment of implant performance over time. Our CRR analysis represents a promising approach to detect implants with a consistent low- or high-risk pattern across registries.

Background Graft versus host disease (GVHD) remains a major determinant of morbidity and mortality following allogeneic hematopoietic stem cell transplantation (allo HSCT). Existing GVHD prediction models demonstrate modest discrimination and limited generalizability, and calibration drift across external populations is rarely characterized despite its essential role in the clinical interpretability of predicted probabilities. Objectives To develop and externally validate an explainable machine learning framework for predicting acute and chronic GVHD and associated overall survival in patients with acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), and myelodysplastic syndromes (MDS) undergoing allo HSCT, and to systematically characterize calibration across heterogeneous external validation cohorts to inform deployment requirements. Study Design The model was developed on three publicly available registry-derived datasets (N = 2,509) and externally validated across six independent cohorts (N = 14,788) comprising adult and pediatric allo HSCT recipients, including a regional Middle Eastern cohort (UAE and Jordan). A standardized preprocessing pipeline harmonized heterogeneous datasets. Gradient boosting models (CatBoost) were used for binary GVHD prediction; exploratory overall survival analysis used a Cox proportional hazards model with predicted acute GVHD risk as a covariate. Discrimination (AUROC with bootstrap 95% CI), calibration (logistic recalibration intercept and slope with analytical 95% CI), and feature importance (SHapley Additive exPlanations, SHAP) were assessed in training out-of-fold and all external cohorts. Results In internal validation, AUROC was 0.63 (95% CI 0.61-0.65) for acute GVHD and 0.72 (95% CI 0.70-0.74) for chronic GVHD. External validation demonstrated AUROC ranges of 0.51-0.57 (acute) and 0.54-0.64 (chronic), with consistent performance across disease subgroups despite substantial heterogeneity in transplant practices and feature availability. In exploratory survival analysis, the acute-GVHD-informed Cox model achieved a training-cohort C-index of 0.679 (95% CI 0.658-0.697); external C-indices ranged from 0.47-0.53. Calibration analysis identified systematic external risk overestimation (negative calibration intercept in 10 of 11 evaluable external cohort-target combinations) with heterogeneous slope drift requiring cohort-specific recalibration. Key predictors included recipient age, graft source, conditioning intensity, GVHD prophylaxis, and HLA match ratio. Conclusions An explainable, externally validated GVHD prediction framework was developed using heterogeneous registry-derived datasets, with systematic characterization of calibration drift across multiple external cohorts, an analysis rarely reported in prior GVHD prediction literature. Predictive performance was modest for acute GVHD and moderate for chronic GVHD, constrained by missing immunobiological variables and incomplete HLA characterization. Per-cohort recalibration is required before clinical deployment, with prospective validation and benchmarking against established GVHD risk scores identified as priority next steps.

BackgroundHip fracture remains a leading cause of morbidity and mortality among older adults in the United States. The aim of this systematic and meta-analytical review is to synthesize available evidence on predictors of one-year mortality following hip fracture among older adults, guided by a socio-ecological framework. MethodsWe searched PubMed, Embase, Web of Science, CINAHL, and Scopus for U.S.-based studies published between 2010 and 2025 reporting one-year mortality after hip fracture. Studies were included if they evaluated predictors of mortality across pre-injury, perioperative, or post-discharge phases. Data were extracted on study design, population characteristics, mortality outcomes, and risk factors. Predictors examined in [&ge;]3 studies were pooled using random-effects meta-analysis, and narrative synthesis was conducted for predictors with limited data. Methodological quality was assessed using the Joanna Briggs Institute checklist. ResultsTwenty-eight studies (n = 835,226) met inclusion criteria. Pooled one-year mortality was 21.8%, ranging from 7.1% to 54.4%. Advancing age and male sex were consistent non-modifiable risk factors. Comorbidity burden, including congestive heart failure, chronic kidney disease, myocardial infarction, and dementia, and measures of frailty and functional impairment were among the strongest predictors, often doubling mortality odds. Perioperative factors such as higher injury severity and delayed surgery, and post-discharge factors including hospital readmission, missed follow-up visits, and postoperative complications, were also associated with increased mortality. ConclusionOne-year hip fracture-related mortality remains high and stems from multifactorial causes. A multi-level, systems-oriented approach may be necessary to meaningfully reduce long-term mortality in this growing and vulnerable population.

Aim: The global population of older adults is growing, and older age is linked to higher bleeding risk. Although guidelines discourage aspirin for primary prevention in healthy older adults due to bleeding harms outweighing benefits, many continue taking it without a clear indication. It remains unclear whether all older adults face uniform aspirin-related bleeding risk or if certain subgroups are more vulnerable. Methods: We analyzed data from 19,114 ASPREE trial participants to develop machine learning models using 116 baseline variables. Random forest (RF) and random survival forest (RSF) models predicted 5-year bleeding risk, and participants were stratified into low, intermediate, and high-risk groups based on the 20th and 80th percentiles of predicted risk. We assessed heterogeneity of treatment effect (HTE) by testing treatment-by-risk group interactions on the relative scale using Fine-Gray models, and on the absolute scale using observed 5-year cumulative incidence rates. Results: Over a median follow-up of 4.7 years, 626 major bleeding events occurred. The RF model had moderate discrimination (AUC = 0.65, 95% CI: 0.63-0.67) and good calibration (Brier = 0.032, 95% CI: 0.029-0.034). Statistically significant HTE was observed on the relative scale, with the greatest relative increase in bleeding risk seen in the low-risk group (subdistribution hazard ratio = 2.26, 95% CI: 1.27-4.01). On the absolute scale, low-risk participants experienced higher bleeding with aspirin (absolute risk difference (ARD) = 1.17%, 95% CI: 0.37-1.95), but heterogeneity in ARDs was not statistically significant (Cochran's Q p > 0.45). Similar findings were observed when using the RSF model. Conclusion: Participants at lowest baseline bleeding risk experienced the greatest relative increase in bleeding risk with aspirin therapy. We found statistically significant heterogeneity in treatment effects on the relative but not absolute scale. These findings support an individualized, risk-based approach to aspirin therapy decision-making in older adults.

Background Lumbar fusion and decompression procedures are widely used for degenerative spine conditions but are associated with substantial health care costs and variable outcomes. Orthobiologic treatments, including platelet rich plasma (PRP) and bone marrow aspirate concentrate (BMAC), have emerged as less invasive options for select patients who meet surgical criteria. However, concerns remain that orthobiologic care may delay rather than avert surgery, potentially increasing downstream utilization and costs. Comparative evidence on real world utilization and costs is limited. Methods We conducted a retrospective, observational study using linked commercial insurance claims and a national orthobiologic treatment registry. Adults with lumbar degenerative disc disease (DDD) who met criteria for lumbar fusion or laminectomy, foraminotomy, discectomy, and facetectomy (LFDF) procedures, and who received PRP injection (with or without BMAC) or surgery between 2016 and 2023 were included. Two comparisons were evaluated: PRP versus lumbar fusion and PRP versus lumbar decompression procedures. Propensity score matching was used to balance cohorts on demographic characteristics, comorbidities, spine related diagnoses, prior health care use, and severity proxies. Outcomes included spine-related health care resource use and aggregate costs at 12 and 24 months, with exploratory analyses at 36 and 48 months. Costs were estimated using multiple approaches, including Medicare based estimates and commercial payer methods. Results After matching, 133 patients receiving PRP were compared with 2,560 patients undergoing fusion, and 198 patients receiving PRP were compared with 3,960 patients undergoing LFDF. Rates of subsequent spine surgery following PRP were low and below cell suppression thresholds through 24 months, with similar findings in exploratory longer-term analyses. Compared with surgical cohorts, patients receiving PRP had lower rates of postoperative imaging, home health services, and outpatient visits, with no consistent differences in opioid use, magnetic resonance imaging, or physical therapy. At 12 and 24 months, mean aggregate costs were significantly higher for fusion and LFDF cohorts across most costing methods. Cost differences were largest for fusion comparisons and were driven primarily by index procedure costs and higher reoperation and imaging rates in surgical cohorts. Findings were generally consistent across sensitivity and exploratory analyses. Conclusions Among select patients with degenerative spine conditions who meet surgical criteria, PRP was associated with lower health care utilization and substantially lower costs compared with lumbar fusion or LFDF, without evidence of increased progression to surgery. These findings support consideration of orthobiologic options for appropriately selected patients when surgery is not the only viable treatment option. Limitations include selection bias, absence of patient reported outcomes, and claims-based severity measures.

ObjectivesTo examine temporal trends in incidence of arthroscopic subacromial decompression (ASAD) surgery internationally during conduct and after publication of placebo controlled trials finding no evidence of meaningful benefit of ASAD for shoulder impingement. DesignObservational study of incidence rates. SettingLarge routinely collected datasets were used: outpatient data from Belgium and UK, and insurance claims and outpatient data from US. UK data were from Clinical Practice Research Datalink and Belgium and US data were from IQVIA. US and UK data spanned 2005 - 2019 and Belgium data 2011 - 2019. ParticipantsPatients were eligible for inclusion in the study if they had at least one visit recorded in the database in a given year and cases were defined as patients undergoing ASAD for the first time in their records in a given year. Outcome measuresWe calculated incidence of ASAD over time, overall and stratified by age and sex. Characteristics of patients undergoing ASAD were also assessed over time. ResultsUK incidence has fallen since a peak of 4.7 per 10,000 person years in 2011 (when the CSAW trial began) to 1.8 in 2019. US incidence shows no clear pattern and remains consistently higher than the UK, at 11.5 per 100,000 person years in 2019. Changes in incidence patterns were similar across different age groups and sexes. The number of cases in Belgium was too small for meaningful conclusions. ConclusionsWe found ASAD rates have fallen in the UK during conduct and after publication of two large surgical RCTs from the UK and Finland that questioned the effectiveness of ASAD for shoulder impingement. A similar impact on clinical practice has not been seen in US. Further work to understand the barriers or concerns preventing international uptake of high quality evidence into clinical practice is needed. Strengths and limitations of this studyO_LIThis is the most comprehensive study of ASAD incidence we are aware of. Routinely collected datasets were used to assess proportions of the patients undergoing this procedure in several countries C_LIO_LIStandardised case definitions were used across databases to increase comparability of findings C_LIO_LITemporal changes in database coverage and quality of reporting can influence findings. The observed variation in ASAD incidence may not be entirely attributable to changes in ASAD surgery rates. C_LI

ObjectiveWe simplified and evaluated the Kaiser Permanente inpatient risk adjustment methodology (KP method) to predict in-hospital mortality, using open-source tools to measure comorbidity and diagnosis groups, and removing troponin, which is difficult to standardize across clinical assays. Study Design and SettingRetrospective cohort study of adult general medical inpatients at 7 hospitals in Ontario, Canada. ResultsIn 206,155 unique hospitalizations with 6.9% in-hospital mortality, the simplified KP method accurately predicted the risk of mortality. Bias-corrected c-statistics were 0.874 (95%CI 0.872-0.877) with troponin and 0.873 (95%CI 0.871-0.876) without troponin, and calibration was excellent for both approaches. Discrimination and calibration were similar with and without troponin for patients with heart failure and acute myocardial infarction. The Laboratory-based Acute Physiology Score (LAPS, a component of the KP method) predicted inpatient mortality on its own with and without troponin with bias-corrected c-statistics of 0.687 (95%CI 0.682-0.692) and 0.680 (95%CI 0.675-0.685), respectively. LAPS was well calibrated, except at very high scores. ConclusionA simplification of the KP method accurately predicted in-hospital mortality risk in an external general medicine cohort. Without troponin, and using common open-source tools, the KP method can be implemented for risk adjustment in a wider range of settings.

BackgroundLiver transplantation (LT) is the definitive treatment for end-stage liver disease, but remains complicated by thrombotic events such as hepatic artery and portal vein thrombosis. This network meta-analysis compares the efficacy and safety of different antithrombotic strategies post-LT, as well as dosing and duration to inform individualized, evidence-based management. MethodsMedline and Embase were searched to 22 November 2024 for RCTs and cohort studies reporting outcomes of postoperative antithrombotic methods in liver transplant patients. Random-effects frequentist network meta-analysis pooled odds ratios (OR), 95% confidence interval (CI) values, and P-score to rank the antithrombotic medications. ResultsAmong 18 studies involving 8,856 patients, aspirin (OR: 0.30, 95%CI: 0.18 - 0.51, p=0.04) and UFH (OR: 0.31, 95%CI: 0.12 - 0.84, p<0.01) use were associated with lowest overall thrombotic risk relative to the control group. Aspirin was ranked the highest for HAT prevention (P-score = 0.88) and had no significant bleeding risk. Low molecular weight heparin (LMWH) was ranked highest for DVT prevention (P-score = 0.67) but also second for worst associated bleeding risk (P-score = 0.43) after VKA (OR: 3.53, 95%CI: 1.86 - 6.71, p<0.01, P-score = 0.01). LMWH also ranked worse than control in prevention of PVT. DOACs were ranked first for the lowest associated bleeding risk and third for reduction in overall thrombotic risk. None of the antithrombotic medications showed any significant association with overall patient mortality. ConclusionsAspirin remains a mainstay of arterial thromboprophylaxis. DOACs appear promising for thromboprophylaxis post-LT, with the lowest bleeding risk. Targeted rather than routine prophylaxis, guided by individual risk profiles, likely maximises post-LT outcomes.

AimsDefinitive and successful treatment of end stage ankle arthritis is either Total Ankle Replacement (TAR) or Ankle Fusion (AF). Both options place patients on an irreversible pathway that risks harm from further surgery. AF may predispose patients to subsequent hindfoot joint fusion and TAR is associated with high rates of complex revision surgery. The aim is to improve decision making by investigating the risks of further surgery, adjacent joint surgery and rare but serious complications of AF versus TAR. MethodsAn England population cohort study was performed using the Hospital Episode Statistics database, linked to ONS mortality data (19982023). The primary outcome was Kaplan Meier curve analysis of revision surgery free survival of TAR versus AF. Secondary outcome measures were the rates of adjacent joint/hindfoot fusion, any further reintervention to the ankle, perioperative mortality, 90 day complications, and serious adverse events. Results10,335 TAR and 30,704 AF were analysed. The AF revision rate was significantly lower than TAR at all time points including; 5 years (2% vs 6.1%, RR 0.12; 95% CI 0.10 to 0.16), 10 years (2.5% vs 10.2%, RR 0.12; 95% CI 0.08 to 0.18) and 20 years (3.1% vs 13.55%, RR 0.12; 95% CI 0.01 to 0.23). There was no significant difference in 25 year risk of adjacent joint fusion following AF (8.64%, 95% CI 7.79% to 9.58%) versus TAR (6.82%; 95% CI 5.36% to 8.66%). TAR was associated with higher risks of intra operative fracture (0.42% vs 0.10%, RR = 4.35) and reintervention for wound infection (0.26% vs 0.15%, RR 1.74) but fewer pulmonary emboli (0.23% vs 0.58%, RR = 0.40). ConclusionBoth TAR and AF are safe definitive treatments of ankle arthritis with low perioperative risk. TAR is associated with a significantly higher rate of further revision surgery than AF. AF does not predispose patients to hindfoot fusion surgery.

BackgroundPersistent postsurgical pain and opioid use after primary total hip and knee arthroplasty (THA and TKA) have major consequences for the patient and for society. High-dose perioperative treatment with glucocorticoids reduces inflammation and acute pain, both of which are associated with persistent postsurgical pain. We therefore hypothesise that routine treatment with glucocorticoids reduces the number of patients with persistent opioid use. ObjectiveTo determine if perioperative glucocorticoids for primary THA or TKA surgery, relative to no glucocorticoids, decreases the number of patients taking opioids in the period from 3 to 12 months after surgery. DesignTarget trial emulation trial with data from Danish national registries. SettingAll departments of orthopaedic surgery in Denmark, from 1 January 2010 to 31 December 2020. ParticipantsPatients with primary osteoarthritis undergoing primary THA or TKA, excluding presurgical users of glucocorticoids or insulin because these patients do not always receive the intervention. InterventionA single high-dose glucocorticoids ([&ge;]125 mg methylprednisolone or [&ge;]24 mg dexamethasone) after induction of anaesthesia. ComparatorNo glucocorticoids during surgery. AllocationPatients operated at departments where treatment with high-dose glucocorticoids was standard of care at the time of surgery constitute the treatment arm, while patients operated at departments where high-dose glucocorticoids was not used serve as controls. Thus, all patients will be analysed according to their allocation, regardless of whether they received the treatment or not. Main outcome measuresThe primary outcome is number of persistent opioid users, defined as patients who redeem a prescription within at least two of the last three quarters during the first postsurgical year. The primary safety outcome is number of days alive and out of hospital within 90 days after surgery. ExpectationsThese results will provide important evidence for or against the use of perioperative glucocorticoids in total hip and knee arthroplasty.

BackgroundPre-procedural risk prediction of 30-day all-cause mortality after percutaneous coronary intervention (PCI) aids in clinical decision-making and benchmarking hospital performance. This study aimed to identify pre-procedural factors to predict the risk of 30-day all-cause mortality post-PCI using machine learning (ML) approaches. MethodsThe study analysed 93,055 consecutive PCI procedures. Boruta feature selection method was used to identify key predictive variables. Seven ML algorithms were employed for models development and validation. Model performance was assessed using standard metrics for validation dataset. SHapley Additive exPlanations (SHAP) method was used to explain leading predictive variables. ResultsAmong the seven ML algorithms, the Extreme Gradient Booster (XGB) had the better performance across most metrics, such as accuracy (86.7%), root mean square error (36.5%), specificity (82.5%), precision (54.0%), F1 score (52.7%), and Brier score (13.3%). The XGB model also demonstrated strong discriminatory power, achieving a receiver operating characteristics-area under the curve (ROC-AUC) of 85.5% (95% CI: 83.5%-87.4%). The XGB model identified left ventricular ejection fraction (LVEF), acute coronary syndrome (ACS), estimated glomerular filtration rate (eGFR), age, and complex lesion as the five leading factors associated with 30-day mortality post-PCI. Other factors, in order, were cardiogenic shock, body mass index (BMI), intubated out-of-hospital cardiac arrest (OHCA), lesion location, mechanical ventricular support, gender, and peripheral vascular disease (PVD). ConclusionThe XGB algorithm was identified as the best predictive model for 30-day all-cause mortality post-PCI. It is essential to underscore the need for further validation of the model with external data to ensure its applicability to other populations. WHAT IS ALREADY KNOWN ON THIS TOPICO_LIrisk-adjustment model for an Australian percutaneous coronary intervention (PCI) patient population was previously developed to predict 30-day mortality post-PCI using traditional regression model. C_LIO_LIknowledge, patient characteristics, and clinical practices evolve over time, requiring frequent model updates to reflect new evidence, guidelines, and interventions C_LI WHAT THIS STUDY ADDSO_LIA machine learning (ML)-based preprocedural risk prediction model for 30-day mortality post-PCI was developed. The Extreme Gradient Booster (XGB) model was identified as the top performer in predicting 30-day all-cause mortality post-PCI. The model selected left ventricular ejection fraction, acute coronary syndrome, estimated glomerular filtration rate, age, and complex lesion as the top influential factors. C_LI HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYO_LIRisk prediction models aid clinical decision-making, enhance patient counselling, improve care quality, inform healthcare policies, and advance research. C_LI