Aims. This study identifies early risk factors for symptomatic nonunion of displaced midshaft fractures of the clavicle that aid identification of an at risk group who may benefit from surgery. . Methods . We performed a retrospective study of 88 patients aged between 16 and 60 years that were managed non-operatively. . Results . The rate of symptomatic nonunion requiring surgery was 14% (n = 13). Smoking (odds ratio (OR) 40.76, 95% confidence intervals (CI) 1.38 to 120.30) and the six week Disabilities of the Arm Shoulder and Hand (DASH) score (OR 1.11, 95% CI 1.01 to 1.22, for each point increase) were independent predictors of nonunion. A six week DASH score of 35 or more was identified as a threshold value to predict nonunion using receiver operating characteristic curve analysis. Smoking and the threshold value in the DASH and were additive risk factors for nonunion, when neither were present the risk of nonunion was 2%, if one or the other were present the nonunion rate was between 17% to 20%, and if both were present the rate increased to 44%. Discussion. Patients with either of these risk factors, which include approximately half of all patients sustaining displaced midshaft fractures of the clavicle, are at an increased risk of developing a symptomatic non-union. Take home message: Smoking and failure of functional return at six weeks are significant predictors of nonunion of the midshaft of the clavicle. Such patients warrant further investigation as to whether they would benefit from early surgical fixation in order to avoid the morbidity of a nonunion. Cite this article: Bone Joint J 2016;98-B:125–30

Aims. Machine learning (ML), a branch of artificial intelligence that uses algorithms to learn from data and make predictions, offers a pathway towards more personalized and tailored surgical treatments. This approach is particularly relevant to prevalent joint diseases such as osteoarthritis (OA). In contrast to end-stage disease, where joint arthroplasty provides excellent results, early stages of OA currently lack effective therapies to halt or reverse progression. Accurate prediction of OA progression is crucial if timely interventions are to be developed, to enhance patient care and optimize the design of clinical trials. Methods. A systematic review was conducted in accordance with PRISMA guidelines. We searched MEDLINE and Embase on 5 May 2024 for studies utilizing ML to predict OA progression. Titles and abstracts were independently screened, followed by full-text reviews for studies that met the eligibility criteria. Key information was extracted and synthesized for analysis, including types of data (such as clinical, radiological, or biochemical), definitions of OA progression, ML algorithms, validation methods, and outcome measures. Results. Out of 1,160 studies initially identified, 39 were included. Most studies (85%) were published between 2020 and 2024, with 82% using publicly available datasets, primarily the Osteoarthritis Initiative. ML methods were predominantly supervised, with significant variability in the definitions of OA progression: most studies focused on structural changes (59%), while fewer addressed pain progression or both. Deep learning was used in 44% of studies, while automated ML was used in 5%. There was a lack of standardization in evaluation metrics and limited external validation. Interpretability was explored in 54% of studies, primarily using SHapley Additive exPlanations. Conclusion. Our systematic review demonstrates the feasibility of ML models in predicting OA progression, but also uncovers critical limitations that currently restrict their clinical applicability. Future priorities should include diversifying data sources, standardizing outcome measures, enforcing rigorous validation, and integrating more sophisticated algorithms. This paradigm shift from predictive modelling to actionable clinical tools has the potential to transform patient care and disease management in orthopaedic practice. Cite this article: Bone Joint J 2024;106-B(11):1216–1222

Aims. Hip arthroscopy has gained prominence as a primary surgical intervention for symptomatic femoroacetabular impingement (FAI). This study aimed to identify radiological features, and their combinations, that predict the outcome of hip arthroscopy for FAI. Methods. A prognostic cross-sectional cohort study was conducted involving patients from a single centre who underwent hip arthroscopy between January 2013 and April 2021. Radiological metrics measured on conventional radiographs and magnetic resonance arthrography were systematically assessed. The study analyzed the relationship between these metrics and complication rates, revision rates, and patient-reported outcomes. Results. Out of 810 identified hip arthroscopies, 359 hips were included in the study. Radiological risk factors associated with unsatisfactory outcomes after cam resection included a dysplastic posterior wall, Tönnis grade 2 or higher, and over-correction of the α angle. The presence of acetabular retroversion and dysplasia were also significant predictors for worse surgical outcomes. Notably, over-correction of both cam and pincer deformities resulted in poorer outcomes than under-correction. Conclusion. We recommend caution in performing hip arthroscopy in patients who have three positive acetabular retroversion signs. Acetabular dysplasia with a lateral centre-edge angle of less than 20° should not be treated with isolated hip arthroscopy. Acetabular rim-trimming should be avoided in patients with borderline dysplasia, and care should be taken to avoid over-correction of a cam deformity and/or pincer deformity. Cite this article: Bone Joint J 2024;106-B(8):775–782

Aims. The preoperative grading of chondrosarcomas of bone that accurately predicts surgical management is difficult for surgeons, radiologists, and pathologists. There are often discrepancies in grade between the initial biopsy and the final histology. Recent advances in the use of imaging methods have shown promise in the ability to predict the final grade. The most important clinical distinction is between grade 1 chondrosarcomas, which are amenable to curettage, and resection-grade chondrosarcomas (grade 2 and 3) which require en bloc resection. The aim of this study was to evaluate the use of a Radiological Aggressiveness Score (RAS) to predict the grade of primary chondrosarcomas in long bones and thus to guide management. Methods. A total of 113 patients with a primary chondrosarcoma of a long bone presenting between January 2001 and December 2021 were identified on retrospective review of a single oncology centre’s prospectively collected database. The nine-parameter RAS included variables from radiographs and MRI scans. The best cut-off of parameters to predict the final grade of chondrosarcoma after resection was determined using a receiver operating characteristic curve (ROC), and this was correlated with the biopsy grade. Results. A RAS of ≥ four parameters was 97.9% sensitive and 90.5% specific in predicting resection-grade chondrosarcoma based on a ROC cut-off derived using the Youden index. Cronbach’s α of 0.897 was derived as the interclass correlation for scoring the lesions by four blinded reviewers who were surgeons. Concordance between resection-grade lesions predicted from the RAS and ROC cut-off with the final grade after resection was 96.46%. Concordance between the biopsy grade and the final grade was 63.8%. However, when the patients were analyzed based on surgical management, the initial biopsy was able to differentiate low-grade from resection-grade chondrosarcomas in 82.9% of biopsies. Conclusion. These findings suggest that the RAS is an accurate method for guiding the surgical management of patients with these tumours, particularly when the initial biopsy results are discordant with the clinical presentation. Cite this article: Bone Joint J 2023;105-B(7):808–814

Several studies have reported the rate of post-operative mortality after the surgical treatment of a fracture of the hip, but few data are available regarding the delayed morbidity. In this prospective study, we identified 568 patients who underwent surgery for a fracture of the hip and who were followed for one year. Multivariate analysis was carried out to identify possible predictors of mortality and morbidity. The 30-day, four-month and one-year rates of mortality were 4.3%, 11.4%, and 18.8%, respectively. General complications and pre-operative comorbidities represented the basic predictors of mortality at any time interval (p < 0.01). In-hospital, four-month and one-year general complications occurred in 29.4%, 18.6% and 6.7% of patients, respectively. After adjusting for confounding variables, comorbidities and poor cognitive status determined the likelihood of early and delayed general complications, respectively (p < 0.001). Operative delay was the main predictor of the length of hospital stay (p < 0.001) and was directly related to in-hospital (p = 0.017) and four-month complications (p = 0.008). Cite this article: Bone Joint J 2015;97-B:383–90

Acute Haematogenous Osteomyelitis (AHO) remains a cause of severe illness among children. Contemporary research aims to identify predictors of acute and chronic complications. Trends in C-reactive protein (CRP) following treatment initiation may predict disease course. We have sought to identify factors associated with acute and chronic complications in the New Zealand population. A retrospective review of all patients <16 years with presumed AHO presenting to a tertiary referral centre between 2008–2018 was performed. Multivariate was analysis used to identify factors associated with an acute or chronic complication. An “acute” complication was defined as need for two or more surgical procedures, hospital stay longer than 14-days, or recurrence despite IV antibiotics. A “chronic” complication was defined as growth or limb length discrepancy, avascular necrosis, chronic osteomyelitis, pathological fracture, frozen joint or dislocation. 151 cases met inclusion criteria. The median age was 8 years (69.5% male). Within this cohort, 53 (34%) experienced an acute complication and 18 (12%) a chronic complication. Regression analysis showed that contiguous disease, delayed presentation, and failure to reduce CRP by 50% at day 4/5 predicted an acutely complicated disease course. Chronic complication was predicted by need for surgical management and failed CRP reduction by 50% at day 4/5. We conclude that CRP trends over 96 hours following commencement of treatment differentiate patients with AHO likely to experience severe disease

Aims. The aim of this study was to assess the ability of morphological spinal parameters to predict the outcome of bracing in patients with adolescent idiopathic scoliosis (AIS) and to establish a novel supine correction index (SCI) for guiding bracing treatment. Methods. Patients with AIS to be treated by bracing were prospectively recruited between December 2016 and 2018, and were followed until brace removal. In all, 207 patients with a mean age at recruitment of 12.8 years (SD 1.2) were enrolled. Cobb angles, supine flexibility, and the rate of in-brace correction were measured and used to predict curve progression at the end of follow-up. The SCI was defined as the ratio between correction rate and flexibility. Receiver operating characteristic (ROC) curve analysis was carried out to assess the optimal thresholds for flexibility, correction rate, and SCI in predicting a higher risk of progression, defined by a change in Cobb angle of ≥ 5° or the need for surgery. Results. The baseline Cobb angles were similar (p = 0.374) in patients whose curves progressed (32.7° (SD 10.7)) and in those whose curves remained stable (31.4° (SD 6.1)). High supine flexibility (odds ratio (OR) 0.947 (95% CI 0.910 to 0.984); p = 0.006) and correction rate (OR 0.926 (95% CI 0.890 to 0.964); p < 0.001) predicted a lower incidence of progression after adjusting for Cobb angle, Risser sign, curve type, menarche status, distal radius and ulna grading, and brace compliance. ROC curve analysis identified a cut-off of 18.1% for flexibility (sensitivity 0.682, specificity 0.704) and a cut-off of 28.8% for correction rate (sensitivity 0.773, specificity 0.691) in predicting a lower risk of curve progression. A SCI of greater than 1.21 predicted a lower risk of progression (OR 0.4 (95% CI 0.251 to 0.955); sensitivity 0.583, specificity 0.591; p = 0.036). Conclusion. A higher supine flexibility (18.1%) and correction rate (28.8%), and a SCI of greater than 1.21 predicted a lower risk of progression. These novel parameters can be used as a guide to optimize the outcome of bracing. Cite this article: Bone Joint J 2022;104-B(4):495–503

Aims. The aim of this study was to evaluate the reliability and validity of a patient-specific algorithm which we developed for predicting changes in sagittal pelvic tilt after total hip arthroplasty (THA). Methods. This retrospective study included 143 patients who underwent 171 THAs between April 2019 and October 2020 and had full-body lateral radiographs preoperatively and at one year postoperatively. We measured the pelvic incidence (PI), the sagittal vertical axis (SVA), pelvic tilt, sacral slope (SS), lumbar lordosis (LL), and thoracic kyphosis to classify patients into types A, B1, B2, B3, and C. The change of pelvic tilt was predicted according to the normal range of SVA (0 mm to 50 mm) for types A, B1, B2, and B3, and based on the absolute value of one-third of the PI-LL mismatch for type C patients. The reliability of the classification of the patients and the prediction of the change of pelvic tilt were assessed using kappa values and intraclass correlation coefficients (ICCs), respectively. Validity was assessed using the overall mean error and mean absolute error (MAE) for the prediction of the change of pelvic tilt. Results. The kappa values were 0.927 (95% confidence interval (CI) 0.861 to 0.992) and 0.945 (95% CI 0.903 to 0.988) for the inter- and intraobserver reliabilities, respectively, and the ICCs ranged from 0.919 to 0.997. The overall mean error and MAE for the prediction of the change of pelvic tilt were -0.3° (SD 3.6°) and 2.8° (SD 2.4°), respectively. The overall absolute change of pelvic tilt was 5.0° (SD 4.1°). Pre- and postoperative values and changes in pelvic tilt, SVA, SS, and LL varied significantly among the five types of patient. Conclusion. We found that the proposed algorithm was reliable and valid for predicting the standing pelvic tilt after THA. Cite this article: Bone Joint J 2024;106-B(1):19–27

Aims. The aim of this study was to investigate the association between fracture displacement and survivorship of the native hip joint without conversion to a total hip arthroplasty (THA), and to determine predictors for conversion to THA in patients treated nonoperatively for acetabular fractures. Methods. A multicentre cross-sectional study was performed in 170 patients who were treated nonoperatively for an acetabular fracture in three level 1 trauma centres. Using the post-injury diagnostic CT scan, the maximum gap and step-off values in the weightbearing dome were digitally measured by two trauma surgeons. Native hip survival was reported using Kaplan-Meier curves. Predictors for conversion to THA were determined using Cox regression analysis. Results. Of 170 patients, 22 (13%) subsequently received a THA. Native hip survival in patients with a step-off ≤ 2 mm, > 2 to 4 mm, or > 4 mm differed at five-year follow-up (respectively: 94% vs 70% vs 74%). Native hip survival in patients with a gap ≤ 2 mm, > 2 to 4 mm, or > 4 mm differed at five-year follow-up (respectively: 100% vs 84% vs 78%). Step-off displacement > 2 mm (> 2 to 4 mm hazard ratio (HR) 4.9, > 4 mm HR 5.6) and age > 60 years (HR 2.9) were independent predictors for conversion to THA at follow-up. Conclusion. Patients with minimally displaced acetabular fractures who opt for nonoperative fracture treatment may be informed that fracture displacement (e.g. gap and step-off) up to 2 mm, as measured on CT images, results in limited risk on conversion to THA. Step-off ≥ 2 mm and age > 60 years are predictors for conversion to THA and can be helpful in the shared decision-making process. Cite this article: Bone Joint J 2023;105-B(9):1020–1029

Aims. We aimed to develop a gene signature that predicts the occurrence of postmenopausal osteoporosis (PMOP) by studying its genetic mechanism. Methods. Five datasets were obtained from the Gene Expression Omnibus database. Unsupervised consensus cluster analysis was used to determine new PMOP subtypes. To determine the central genes and the core modules related to PMOP, the weighted gene co-expression network analysis (WCGNA) was applied. Gene Ontology enrichment analysis was used to explore the biological processes underlying key genes. Logistic regression univariate analysis was used to screen for statistically significant variables. Two algorithms were used to select important PMOP-related genes. A logistic regression model was used to construct the PMOP-related gene profile. The receiver operating characteristic area under the curve, Harrell’s concordance index, a calibration chart, and decision curve analysis were used to characterize PMOP-related genes. Then, quantitative real-time polymerase chain reaction (qRT-PCR) was used to verify the expression of the PMOP-related genes in the gene signature. Results. We identified three PMOP-related subtypes and four core modules. The muscle system process, muscle contraction, and actin filament-based movement were more active in the hub genes. We obtained five feature genes related to PMOP. Our analysis verified that the gene signature had good predictive power and applicability. The outcomes of the GSE56815 cohort were found to be consistent with the results of the earlier studies. qRT-PCR results showed that RAB2A and FYCO1 were amplified in clinical samples. Conclusion. The PMOP-related gene signature we developed and verified can accurately predict the risk of PMOP in patients. These results can elucidate the molecular mechanism of RAB2A and FYCO1 underlying PMOP, and yield new and improved treatment strategies, ultimately helping PMOP monitoring. Cite this article: Bone Joint Res 2022;11(8):548–560

Osteosynthesis aims to maintain fracture reduction until bone healing occurs, which is not achieved in case of mechanical fixation failure. One form of failure is plastic plate bending due to overloading, occurring in up to 17% of midshaft fracture cases and often necessitating reoperation. This study aimed to replicate in-vivo conditions in a cadaveric experiment and to validate a finite element (FE) simulation to predict plastic plate bending. Six cadaveric bones were used to replicate an established ovine tibial osteotomy model with locking plates in-vitro with two implant materials (titanium, steel) and three fracture gap sizes (30, 60, 80 mm). The constructs were tested monotonically until plastic plate deformation under axial compression. Specimen-specific FE models were created from CT images. Implant material properties were determined using uniaxial tensile testing of dog bone shaped samples. The experimental tests were replicated in the simulations. Stiffness, yield, and maximum loads were compared between the experiment and FE models. Implant material properties (Young's modulus and yield stress) for steel and titanium were 184 GPa and 875 MPa, and 105 GPa and 761 MPa, respectively. Yield and maximum loads of constructs ranged between 469–491 N and 652–683 N, and 759–995 N and 1252–1600 N for steel and titanium fixations, respectively. FE models accurately and quantitatively correctly predicted experimental results for stiffness (R2=0.96), yield (R2=0.97), and ultimate load (R2=0.97). FE simulations accurately predicted plastic plate bending in osteosynthesis constructs. Construct behavior was predominantly driven by the implant itself, highlighting the importance of modelling correct material properties of metal. The validated FE models could predict subject-specific load bearing capacity of osteosyntheses in vivo in preclinical or clinical studies. Acknowledgements: This study was supported by the AO Foundation via the AOTRAUMA Network (Grant No.: AR2021_03)

Aim. Recurrence of bone and joint infection, despite appropriate therapy, is well recognised and stimulates ongoing interest in identifying host factors that predict infection recurrence. Clinical prediction models exist for those treated with DAIR, but to date no models with a low risk of bias predict orthopaedic infection recurrence for people with surgically excised infection and removed metalwork. The aims of this study were to construct and internally validate a risk prediction model for infection recurrence at 12 months, and to identify factors that predict recurrence. Predictive factors must be easy to check in pre-operative assessment and relevant across patient groups. Methods. Four prospectively collected datasets including 1173 participants treated in European centres between 2003 and 2021, followed up to 12 months after surgery for orthopaedic infections, were included in logistic regression modelling [1–3]. The definition of infection recurrence was identical and ascertained separately from baseline factors in three contributing cohorts. Eight predictive factors were investigated following a priori sample size calculation: age, gender, BMI, ASA score, the number of prior operations, immunosuppressive medication, glycosylated haemoglobin (HbA1c), and smoking. Missing data, including systematically missing predictors, were imputed using Multiple Imputation by Chained Equations. Weekly alcohol intake was not included in modelling due to low inter-observer reliability (mean reported intake 12 units per week, 95% CI for mean inter-rater error −16.0 to +15.4 units per week). Results. Participants were 64% male, with a median age of 60 years (range 18–95). 86% of participants had lower limb orthopaedic infections. 732 participants were treated for osteomyelitis, including FRI, and 432 for PJI. 16% of participants experienced treatment failure by 12 months. The full prediction model had moderate apparent discrimination: AUROC (C statistic) 0.67, Brier score 0.13, and reasonable apparent calibration. Of the predictors of interest, associations with failure were seen with prior operations at the same anatomical site (odds ratio for failure 1.51 for each additional prior surgery; 95% CI 1.02 to 2.22, p=0.06), and the current use of immunosuppressive medications (odds ratio for failure 2.94; 95% CI 0.89 to 9.77, p=0.08). Conclusions. This association between number of prior surgeries and treatment failure supports the urgent need to streamline referral pathways for people with orthopaedic infection to specialist multidisciplinary units

Aims. Transfusion after primary total hip arthroplasty (THA) has become rare, and identification of causative factors allows preventive measures. The aim of this study was to determine patient-specific factors that increase the risk of needing a blood transfusion. Methods. All patients who underwent elective THA were analyzed retrospectively in this single-centre study from 2020 to 2021. A total of 2,892 patients were included. Transfusion-related parameters were evaluated. A multiple logistic regression was performed to determine whether age, BMI, American Society of Anesthesiologists (ASA) grade, sex, or preoperative haemoglobin (Hb) could predict the need for transfusion within the examined patient population. Results. The overall transfusion rate was 1.2%. Compared to the group of patients without blood transfusion, the transfused group was on average older (aged 73.8 years (SD 9.7) vs 68.6 years (SD 10.1); p = 0.020) and was mostly female (p = 0.003), but showed no significant differences in terms of BMI (28.3 kg/m. 2. (SD 5.9) vs 28.7 kg/m. 2. (SD 5.2); p = 0.720) or ASA grade (2.2 (SD 0.5) vs 2.1 (SD 0.4); p = 0.378). The regression model identified a cutoff Hb level of < 7.6 mmol/l (< 12.2 g/dl), aged > 73 years, and a BMI of 35.4 kg/m² or higher as the three most reliable predictors associated with postoperative transfusion in THA. Conclusion. The possibility of transfusion is predictable based on preoperatively available parameters. The proposed thresholds for preoperative Hb level, age, and BMI can help identify patients and take preventive measures if necessary. Cite this article: Bone Jt Open 2024;5(7):560–564

Aims. Precise implant positioning, tailored to individual spinopelvic biomechanics and phenotype, is paramount for stability in total hip arthroplasty (THA). Despite a few studies on instability prediction, there is a notable gap in research utilizing artificial intelligence (AI). The objective of our pilot study was to evaluate the feasibility of developing an AI algorithm tailored to individual spinopelvic mechanics and patient phenotype for predicting impingement. Methods. This international, multicentre prospective cohort study across two centres encompassed 157 adults undergoing primary robotic arm-assisted THA. Impingement during specific flexion and extension stances was identified using the virtual range of motion (ROM) tool of the robotic software. The primary AI model, the Light Gradient-Boosting Machine (LGBM), used tabular data to predict impingement presence, direction (flexion or extension), and type. A secondary model integrating tabular data with plain anteroposterior pelvis radiographs was evaluated to assess for any potential enhancement in prediction accuracy. Results. We identified nine predictors from an analysis of baseline spinopelvic characteristics and surgical planning parameters. Using fivefold cross-validation, the LGBM achieved 70.2% impingement prediction accuracy. With impingement data, the LGBM estimated direction with 85% accuracy, while the support vector machine (SVM) determined impingement type with 72.9% accuracy. After integrating imaging data with a multilayer perceptron (tabular) and a convolutional neural network (radiograph), the LGBM’s prediction was 68.1%. Both combined and LGBM-only had similar impingement direction prediction rates (around 84.5%). Conclusion. This study is a pioneering effort in leveraging AI for impingement prediction in THA, utilizing a comprehensive, real-world clinical dataset. Our machine-learning algorithm demonstrated promising accuracy in predicting impingement, its type, and direction. While the addition of imaging data to our deep-learning algorithm did not boost accuracy, the potential for refined annotations, such as landmark markings, offers avenues for future enhancement. Prior to clinical integration, external validation and larger-scale testing of this algorithm are essential. Cite this article: Bone Jt Open 2024;5(8):671–680

External validation of machine learning predictive models is achieved through evaluation of model performance on different groups of patients than were used for algorithm development. This important step is uncommonly performed, inhibiting clinical translation of newly developed models. Recently, machine learning was used to develop a tool that can quantify revision risk for a patient undergoing primary anterior cruciate ligament (ACL) reconstruction (https://swastvedt.shinyapps.io/calculator_rev/). The source of data included nearly 25,000 patients with primary ACL reconstruction recorded in the Norwegian Knee Ligament Register (NKLR). The result was a well-calibrated tool capable of predicting revision risk one, two, and five years after primary ACL reconstruction with moderate accuracy. The purpose of this study was to determine the external validity of the NKLR model by assessing algorithm performance when applied to patients from the Danish Knee Ligament Registry (DKLR). The primary outcome measure of the NKLR model was probability of revision ACL reconstruction within 1, 2, and/or 5 years. For the index study, 24 total predictor variables in the NKLR were included and the models eliminated variables which did not significantly improve prediction ability - without sacrificing accuracy. The result was a well calibrated algorithm developed using the Cox Lasso model that only required five variables (out of the original 24) for outcome prediction. For this external validation study, all DKLR patients with complete data for the five variables required for NKLR prediction were included. The five variables were: graft choice, femur fixation device, Knee Injury and Osteoarthritis Outcome Score (KOOS) Quality of Life subscale score at surgery, years from injury to surgery, and age at surgery. Predicted revision probabilities were calculated for all DKLR patients. The model performance was assessed using the same metrics as the NKLR study: concordance and calibration. In total, 10,922 DKLR patients were included for analysis. Average follow-up time or time-to-revision was 8.4 (±4.3) years and overall revision rate was 6.9%. Surgical technique trends (i.e., graft choice and fixation devices) and injury characteristics (i.e., concomitant meniscus and cartilage pathology) were dissimilar between registries. The model produced similar concordance when applied to the DKLR population compared to the original NKLR test data (DKLR: 0.68; NKLR: 0.68-0.69). Calibration was poorer for the DKLR population at one and five years post primary surgery but similar to the NKLR at two years. The NKLR machine learning algorithm demonstrated similar performance when applied to patients from the DKLR, suggesting that it is valid for application outside of the initial patient population. This represents the first machine learning model for predicting revision ACL reconstruction that has been externally validated. Clinicians can use this in-clinic calculator to estimate revision risk at a patient specific level when discussing outcome expectations pre-operatively. While encouraging, it should be noted that the performance of the model on patients undergoing ACL reconstruction outside of Scandinavia remains unknown

Aims. The aims of this study were to validate the minimal clinically important difference (MCID) and patient-acceptable symptom state (PASS) thresholds for Western Ontario Shoulder Instability Index (WOSI), Rowe score, American Shoulder and Elbow Surgeons (ASES), and visual analogue scale (VAS) scores following arthroscopic Bankart repair, and to identify preoperative threshold values of these scores that could predict the achievement of MCID and PASS. Methods. A retrospective review was conducted on 131 consecutive patients with anterior shoulder instability who underwent arthroscopic Bankart repair between January 2020 and January 2023. Inclusion criteria required at least one episode of shoulder instability and a minimum follow-up period of 12 months. Preoperative and one-year postoperative scores were assessed. MCID and PASS were estimated using distribution-based and anchor-based methods, respectively. Receiver operating characteristic curve analysis determined preoperative patient-reported outcome measure thresholds predictive of achieving MCID and PASS. Results. MCID thresholds were determined as 169.6, 6.8, 7.2, and 1.1 for WOSI, Rowe, ASES, and VAS, respectively. PASS thresholds were calculated as ≤ 480, ≥ 80, ≥ 87, and ≤ 1 for WOSI, Rowe, ASES, and VAS, respectively. Preoperative thresholds of ≥ 760 (WOSI) and ≤ 50 (Rowe) predicted achieving MCID for WOSI score (p < 0.001). Preoperative thresholds of ≤ 60 (ASES) and ≥ 2 (VAS) predicted achieving MCID for VAS score (p < 0.001). A preoperative threshold of ≥ 40 (Rowe) predicted achieving PASS for Rowe score (p = 0.005). Preoperative thresholds of ≥ 50 (ASES; p = 0.002) and ≤ 2 (VAS; p < 0.001) predicted achieving PASS for the ASES score. Preoperative thresholds of ≥ 43 (ASES; p = 0.046) and ≤ 4 (VAS; p = 0.024) predicted achieving PASS for the VAS. Conclusion. This study defined MCID and PASS values for WOSI, Rowe, ASES, and VAS scores in patients undergoing arthroscopic Bankart repair. Higher preoperative functional scores may reduce the likelihood of achieving MCID but increase the likelihood of achieving the PASS. These findings provide valuable guidance for surgeons to counsel patients realistically regarding their expectations. Cite this article: Bone Joint J 2024;106-B(10):1118–1124

Single level discectomy (SLD) is one of the most commonly performed spinal surgery procedures. Two key drivers of their cost-of-care are duration of surgery (DOS) and postoperative length of stay (LOS). Therefore, the ability to preoperatively predict SLD DOS and LOS has substantial implications for both hospital and healthcare system finances, scheduling and resource allocation. As such, the goal of this study was to predict DOS and LOS for SLD using machine learning models (MLMs) constructed on preoperative factors using a large North American database. The American College of Surgeons (ACS) National Surgical and Quality Improvement (NSQIP) database was queried for SLD procedures from 2014-2019. The dataset was split in a 60/20/20 ratio of training/validation/testing based on year. Various MLMs (traditional regression models, tree-based models, and multilayer perceptron neural networks) were used and evaluated according to 1) mean squared error (MSE), 2) buffer accuracy (the number of times the predicted target was within a predesignated buffer), and 3) classification accuracy (the number of times the correct class was predicted by the models). To ensure real world applicability, the results of the models were compared to a mean regressor model. A total of 11,525 patients were included in this study. During validation, the neural network model (NNM) had the best MSEs for DOS (0.99) and LOS (0.67). During testing, the NNM had the best MSEs for DOS (0.89) and LOS (0.65). The NNM yielded the best 30-minute buffer accuracy for DOS (70.9%) and ≤120 min, >120 min classification accuracy (86.8%). The NNM had the best 1-day buffer accuracy for LOS (84.5%) and ≤2 days, >2 days classification accuracy (94.6%). All models were more accurate than the mean regressors for both DOS and LOS predictions. We successfully demonstrated that MLMs can be used to accurately predict the DOS and LOS of SLD based on preoperative factors. This big-data application has significant practical implications with respect to surgical scheduling and inpatient bedflow, as well as major implications for both private and publicly funded healthcare systems. Incorporating this artificial intelligence technique in real-time hospital operations would be enhanced by including institution-specific operational factors such as surgical team and operating room workflow

Aims. The aim of the study was to apply 3D measurements for fracture displacement in minimally to moderately displaced acetabular fractures treated nonoperatively, and to evaluate whether this measurement can be used to estimate the likelihood of conversion to total hip arthroplasty (THA) at follow-up. Methods. A multicentre, cross-sectional study was performed on 144 patients who were treated nonoperatively for an acetabular fracture in four level 1 trauma centres between January 2000 and December 2020. For each patient, fracture displacement was measured on CT-based 3D models. The 3D gap area represents fracture displacement (mm. 2. ) between all fracture fragments. A receiver operating characteristic curve was generated to determine a 3D gap area threshold representing the optimal sensitivity and specificity to predict conversion to THA. Native hip survival was reported using Kaplan-Meier curves. Predictors of conversion to THA were determined using Cox regression analysis. Results. Of 144 patients, 18 (12%) received a THA. The median 3D gap area of the complete study cohort was 692 mm. 2. (IQR 216 to 1,400). Native hip survival in patients with a gap area threshold ≤ 1,171 or > 1,171 mm. 2. differed at one-year (98% vs 85%), two-year (96% vs 77%), and five-year follow-up (92% vs 73%). 3D gap area > 1,171 mm. 2. (hazard ratio (HR) 4.3; 95% CI 1.7 to 11.0) and the presence of osteoarthritis grade 2 or higher (HR 3.1; 95% CI 1.2 to 7.8) independently predicted the likelihood of conversion to THA during follow-up. Conclusion. For nonoperatively treated acetabular fractures, introducing 3D gap area assessment to measure fracture displacement offers a promising approach for assessing the risk of conversion to THA. With the introduction of (semi-)automatic segmentation and measurement techniques, or their implementation into commercially available software, this 3D gap area measurement can serve as an addition to the Letournel classification and an alternative to traditional 2D gap and step-off measurements, improving accuracy, reproducibility, and applicability in clinical decision-making. Cite this article: Bone Joint J 2025;107-B(2):204–212

Aims. Periprosthetic fractures (PPFs) around the knee are challenging injuries. This study aims to describe the characteristics of knee PPFs and the impact of patient demographics, fracture types, and management modalities on in-hospital mortality. Methods. Using a multicentre study design, independent of registry data, we included adult patients sustaining a PPF around a knee arthroplasty between 1 January 2010 and 31 December 2019. Univariate, then multivariable, logistic regression analyses were performed to study the impact of patient, fracture, and treatment on mortality. Results. Out of a total of 1,667 patients in the PPF study database, 420 patients were included. The in-hospital mortality rate was 6.4%. Multivariable analyses suggested that American Society of Anesthesiologists (ASA) grade, history of peripheral vascular disease (PVD), history of rheumatic disease, fracture around a loose implant, and cerebrovascular accident (CVA) during hospital stay were each independently associated with mortality. Each point increase in ASA grade independently correlated with a four-fold greater mortality risk (odds ratio (OR) 4.1 (95% confidence interval (CI) 1.19 to 14.06); p = 0.026). Patients with PVD have a nine-fold increase in mortality risk (OR 9.1 (95% CI 1.25 to 66.47); p = 0.030) and patients with rheumatic disease have a 6.8-fold increase in mortality risk (OR 6.8 (95% CI 1.32 to 34.68); p = 0.022). Patients with a fracture around a loose implant (Unified Classification System (UCS) B2) have a 20-fold increase in mortality, compared to UCS A1 (OR 20.9 (95% CI 1.61 to 271.38); p = 0.020). Mode of management was not a significant predictor of mortality. Patients managed with revision arthroplasty had a significantly longer length of stay (median 16 days; p = 0.029) and higher rates of return to theatre, compared to patients treated nonoperatively or with fixation. Conclusion. The mortality rate in PPFs around the knee is similar to that for native distal femur and neck of femur fragility fractures. Patients with certain modifiable risk factors should be optimized. A national PPF database and standardized management guidelines are currently required to understand these complex injuries and to improve patient outcomes. Cite this article: Bone Joint J 2024;106-B(2):158–165

Excessive resident duty hours (RDH) are a recognized issue with implications for physician well-being and patient safety. A major component of the RDH concern is on-call duty. While considerable work has been done to reduce resident call workload, there is a paucity of research in optimizing resident call scheduling. Call coverage is scheduled manually rather than demand-based, which generally leads to over-scheduling to prevent a service gap. Machine learning (ML) has been widely applied in other industries to prevent such issues of a supply-demand mismatch. However, the healthcare field has been slow to adopt these innovations. As such, the aim of this study was to use ML models to 1) predict demand on orthopaedic surgery residents at a level I trauma centre and 2) identify variables key to demand prediction. Daily surgical handover emails over an eight year (2012-2019) period at a level I trauma centre were collected. The following data was used to calculate demand: spine call coverage, date, and number of operating rooms (ORs), traumas, admissions and consults completed. Various ML models (linear, tree-based and neural networks) were trained to predict the workload, with their results compared to the current scheduling approach. Quality of models was determined by using the area under the receiver operator curve (AUC) and accuracy of the predictions. The top ten most important variables were extracted from the most successful model. During training, the model with the highest AUC and accuracy was the multivariate adaptive regression splines (MARS) model, with an AUC of 0.78±0.03 and accuracy of 71.7%±3.1%. During testing, the model with the highest AUC and accuracy was the neural network model, with an AUC of 0.81 and accuracy of 73.7%. All models were better than the current approach, which had an AUC of 0.50 and accuracy of 50.1%. Key variables used by the neural network model were (descending order): spine call duty, year, weekday/weekend, month, and day of the week. This was the first study attempting to use ML to predict the service demand on orthopaedic surgery residents at a major level I trauma centre. Multiple ML models were shown to be more appropriate and accurate at predicting the demand on surgical residents as compared to the current scheduling approach. Future work should look to incorporate predictive models with optimization strategies to match scheduling with demand in order to improve resident well being and patient care