Evaluating the fairness and accuracy of machine learning-based predictions of clinical outcomes after anatomic and reverse total shoulder arthroplasty.

BACKGROUND: Machine learning (ML)-based clinical decision support tools (CDSTs) make personalized predictions for different treatments; by comparing predictions of multiple treatments, these tools can be used to optimize decision making for a particular patient. However, CDST prediction accuracy varies for different patients and also for different treatment options. If these differences are sufficiently large and consistent for a particular subcohort of patients, then that bias may result in those patients not receiving a particular treatment. Such level of bias would deem the CDST "unfair." The purpose of this study is to evaluate the "fairness" of ML CDST-based clinical outcomes predictions after anatomic (aTSA) and reverse total shoulder arthroplasty (rTSA) for patients of different demographic attributes. METHODS: Clinical data from 8280 shoulder arthroplasty patients with 19,249 postoperative visits was used to evaluate the prediction fairness and accuracy associated with the following patient demographic attributes: ethnicity, sex, and age at the time of surgery. Performance of clinical outcome and range of motion regression predictions were quantified by the mean absolute error (MAE) and performance of minimal clinically important difference (MCID) and substantial clinical benefit classification predictions were quantified by accuracy, sensitivity, and the F1 score. Fairness of classification predictions leveraged the "four-fifths" legal guideline from the US Equal Employment Opportunity Commission and fairness of regression predictions leveraged established MCID thresholds associated with each outcome measure. RESULTS: For both aTSA and rTSA clinical outcome predictions, only minor differences in MAE were observed between patients of different ethnicity, sex, and age. Evaluation of prediction fairness demonstrated that 0 of 486 MCID (0%) and only 3 of 486 substantial clinical benefit (0.6%) classification predictions were outside the 20% fairness boundary and only 14 of 972 (1.4%) regression predictions were outside of the MCID fairness boundary. Hispanic and Black patients were more likely to have ML predictions out of fairness tolerance for aTSA and rTSA. Additionally, patients <60 years old were more likely to have ML predictions out of fairness tolerance for rTSA. No disparate predictions were identified for sex and no disparate regression predictions were observed for forward elevation, internal rotation score, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form score, or global shoulder function. CONCLUSION: The ML algorithms analyzed in this study accurately predict clinical outcomes after aTSA and rTSA for patients of different ethnicity, sex, and age, where only 1.4% of regression predictions and only 0.3% of classification predictions were out of fairness tolerance using the proposed fairness evaluation method and acceptance criteria. Future work is required to externally validate these ML algorithms to ensure they are equally accurate for all legally protected patient groups.

Using machine learning to predict internal rotation after anatomic and reverse total shoulder arthroplasty.

BACKGROUND: Improvement in internal rotation (IR) after anatomic (aTSA) and reverse (rTSA) total shoulder arthroplasty is difficult to predict, with rTSA patients experiencing greater variability and more limited IR improvements than aTSA patients. The purpose of this study is to quantify and compare the IR score for aTSA and rTSA patients and create supervised machine learning that predicts IR after aTSA and rTSA at multiple postoperative time points. METHODS: Clinical data from 2270 aTSA and 4198 rTSA patients were analyzed using 3 supervised machine learning techniques to create predictive models for internal rotation as measured by the IR score at 6 postoperative time points. Predictions were performed using the full input feature set and 2 minimal input feature sets. The mean absolute error (MAE) quantified the difference between actual and predicted IR scores for each model at each time point. The predictive accuracy of the XGBoost algorithm was also quantified by its ability to distinguish which patients would achieve clinical improvement greater than the minimal clinically important difference (MCID) and substantial clinical benefit (SCB) patient satisfaction thresholds for IR score at 2-3 years after surgery. RESULTS: rTSA patients had significantly lower mean IR scores and significantly less mean IR score improvement than aTSA patients at each postoperative time point. Both aTSA and rTSA patients experienced significant improvements in their ability to perform activities of daily living (ADLs); however, aTSA patients were significantly more likely to perform these ADLs. Using a minimal feature set of preoperative inputs, our machine learning algorithms had equivalent accuracy when predicting IR score for both aTSA (0.92-1.18 MAE) and rTSA (1.03-1.25 MAE) from 3 months to >5 years after surgery. Furthermore, these predictive algorithms identified with 90% accuracy for aTSA and 85% accuracy for rTSA which patients will achieve MCID IR score improvement and predicted with 85% accuracy for aTSA patients and 77% accuracy for rTSA which patients will achieve SCB IR score improvement at 2-3 years after surgery. DISCUSSION: Our machine learning study demonstrates that active internal rotation can be accurately predicted after aTSA and rTSA at multiple postoperative time points using a minimal feature set of preoperative inputs. These predictive algorithms accurately identified which patients will, and will not, achieve clinical improvement in IR score that exceeds the MCID and SCB patient satisfaction thresholds.

Risk factors for complications and revision surgery after anatomic and reverse total shoulder arthroplasty.

BACKGROUND: Complications and revisions following anatomic (aTSA) and reverse (rTSA) total shoulder arthroplasty have deleterious effects on patient function and satisfaction. The purpose of this study is to evaluate patient-specific, implant-specific and technique-specific risk factors for intraoperative complications, postoperative complications, and the occurrence of revisions after aTSA and rTSA. METHODS: A total of 2964 aTSA and 5616 rTSA patients were enrolled in an international database of primary shoulder arthroplasty. Intra- and postoperative complications, as well as revisions, were reported and evaluated. Multivariate analyses were performed to quantify the risk factors associated with complications and revisions. RESULTS: aTSA patients had a significantly higher complication rate (P = .0026) and a significantly higher revision rate (P < .0001) than rTSA patients, but aTSA patients also had a significantly longer average follow-up (P < .0001) than rTSA patients. No difference (P = .2712) in the intraoperative complication rate was observed between aTSA and rTSA patients. Regarding intraoperative complications, female sex (odds ratio [OR] 2.0, 95% confidence interval [CI] 1.17-3.68) and previous shoulder surgery (OR 2.9, 95% CI 1.73-4.90) were identified as significant risk factors. In regard to postoperative complications, younger age (OR 0.987, 95% CI 0.977-0.996), diagnosis of rheumatoid arthritis (OR 1.76, 95% 1.12-2.65), and previous shoulder surgery (OR 1.42, 95% CI 1.16-1.72) were noted to be risks factors. Finally, in regard to revision surgery, younger age (OR 0.964, 95% CI 0.933-0.998), more glenoid retroversion (OR 1.03, 95% CI 1.001-1.058), larger humeral stem size (OR 1.09, 95% CI 1.01-1.19), larger humeral liner thickness or offset (OR 1.50, 95% CI 1.18-1.96), larger glenosphere diameter (OR 1.16, 95% CI 1.07-1.26), and more intraoperative blood loss (OR 1.002, 95% CI 1.001-1.004) were noted to be risk factors. CONCLUSIONS: Studying the impact of numerous patient- and implant-specific risk factors and determining their impact on complications and revision shoulder arthroplasty can assist surgeons in counseling patients and guide patient expectations following aTSA or rTSA. Care should be taken in patients with a history of previous shoulder surgery, who are at increased risk of both intra- and postoperative complications.

Glenoid component loosening associated with increased critical shoulder angle at midterm follow-up.

BACKGROUND: Glenoid component loosening is a common failure mode of total shoulder arthroplasty (TSA). A larger critical shoulder angle (CSA) may cause superior glenoid component loading and more rapid component loosening. The purpose of this study was to define the relationship between the CSA and glenoid component loosening in midterm follow-up after TSA. METHODS: We conducted a retrospective study of 61 primary TSAs for osteoarthritis with an average follow-up of 5.0 ± 2.2 years without surgical revision. Standard true anteroposterior radiographs postoperatively and at longest follow-up were graded in a blinded and repetitive nature for pegged glenoid radiolucent lines and measured for the CSA. An "at-risk" glenoid was defined as grade 3 or higher lucency. RESULTS: The average CSA was 32° ± 5°, median midterm lucency grade was 2 (range, 0-5), and median progression of lucency grade was 1 (range, -1 to 4). At midterm follow-up, 20% of TSAs were grade 3 or higher mean glenoid lucency, with an average CSA of 36°. There was a statistically significant correlation between CSA and both glenoid lucency grade (odds ratio, 1.20 per degree CSA) and progression of lucency grade (odds ratio, 1.24). An increase in CSA of 10° was associated with a 6.2-fold increased odds of having an at-risk glenoid. CONCLUSION: This study identifies the CSA as a risk factor for glenoid component loosening after TSA. Our findings suggest that the CSA may be a modifiable factor during surgery to improve glenoid component outcomes.

Risk factors associated with blood transfusion after shoulder arthroplasty.

BACKGROUND: Closed-suction drainage has been studied extensively in hip and knee arthroplasty literature. However, little is known about outcomes in patients treated with drainage after shoulder arthroplasty, particularly relative to transfusion requirements. METHODS: All primary total and reverse total shoulder arthroplasties (TSAs and RSAs) performed at a single institution during a 5-year period were retrospectively reviewed. Data collected included patient demographic information, estimated blood loss (EBL), drain output, length of drain use, changes in hemoglobin (Hgb) level postoperatively, transfusions, and complications. A multivariable regression analysis was performed to identify independent risk factors for transfusion. RESULTS: There were no differences in surgery duration, EBL, or complications between TSA and RSA patients (P > .05). Patients undergoing RSA were older (74.0 vs. 68.4 years; P < .001) and had lower preoperative and postoperative Hgb levels (P < .001) compared with TSA patients. Reverse arthroplasty was also associated with longer hospital stays (2.8 vs. 2.2 days; P < .001), longer drain durations (1.6 vs. 1.2 days; P < .001), increased total wound drainage (209 vs. 168 m; P = .006), and higher transfusion rates (11.7% vs. 3.1%; P = .002). Independent risk factors for transfusion included low preoperative Hgb levels in both TSA (P = .024) and RSA (P = .002) and higher EBL in TSA (P = .031). CONCLUSION: Low preoperative Hgb level is an independent risk factor for requiring blood transfusion after TSA and RSA. Increased wound drainage was not a risk factor for transfusion, and the 40-mL increase in wound drainage found in RSA is of questionable clinical significance.

Reliability of a field-based drop vertical jump screening test for ACL injury risk assessment.

OBJECTIVES: There is an epidemic of anterior cruciate ligament (ACL) injuries in youth athletes. Poor neuromuscular control is an easily modifiable risk factor for ACL injury, and can be screened for by observing dynamic knee valgus on landing in a drop vertical jump test. This study aims to validate a simple, clinically useful population-based screening test to identify at-risk athletes prior to participation in organized sports. We hypothesized that both physicians and allied health professionals would be accurate in subjectively assessing injury risk in real-time field and office conditions without motion analysis data and would be in agreement with each other. METHODS: We evaluated the inter-rater reliability of risk assessment by various observer groups, including physicians and allied health professionals, commonly involved in the care of youth athletes. Fifteen athletes age 11-17 were filmed performing a drop vertical jump test. These videos were viewed by 242 observers including orthopaedic surgeons, orthopaedic residents/fellows, coaches, athletic trainers (ATCs), and physical therapists (PTs), with the observer asked to subjectively estimate the risk level of each jumper. Objective injury risk was calculated using normalized knee separation distance (measured using Dartfish, Alpharetta, GA), based on previously published studies. Risk assessments by observers were compared to each other to determine inter-rater reliability, and to the objectively calculated risk level to determine sensitivity and specificity. Seventy one observers repeated the test at a minimum of 6 weeks later to determine intra-rater reliability. RESULTS: Between groups, the inter-rater reliability was high, κ = 0.92 (95% CI 0.829-0.969, p < 0.05), indicating that no single group gave better (or worse) assessments, including comparisons between physicians and allied health professionals. With a screening cutoff isolated to subjects identified by observers as "high risk", the sensitivity was 63.06% and specificity 82.81%. Reducing the screening cutoff to also include jumpers identified as "medium risk" increased sensitivity to 95.04% and decreased the specificity to 46.07%. Intra-rater reliability was moderate, κ = 0.55 (95% CI 0.49-0.61, p < 0.05), indicating that individual observers made reproducible risk assessments. CONCLUSIONS: This study supports the use of a simple, field-based observational drop vertical jump screening test to identify athletes at risk for ACL injury. Our study shows good inter- and intra-rater reliability and high sensitivity and suggests that screening can be performed without significant training by physicians as well as allied health professionals, including: coaches, athletic trainers and physical therapists. Identification of these high-risk athletes may play a role in enrollment in appropriate preventative neuromuscular training programs, which have been shown to decrease the incidence of ACL injuries in this population.

Markedly elevated intra-articular white cell count caused by gout alone.

Joint pain accompanied by erythema, swelling, and decreased range of motion is concerning for septic arthritis and typically warrants joint aspiration. The synovial fluid white blood cell count plays a central role in the decision-making process regarding these patients. Traditional teaching holds that a cell count greater than 50,000 white blood cells/µL is likely caused by infection and therefore warrants either operative intervention or serial aspiration. This report describes 2 patients with extremely high synovial fluid white blood cell counts in the absence of infection. Case 1 involved a 59-year-old man who presented to the emergency department with sudden onset of atraumatic left elbow pain and was found to have a white blood cell count of 168,500 white blood cells/µL on joint aspiration and innumerable monosodium urate crystals. The patient ultimately improved with treatment with oral prednisone, avoiding operative intervention. Case 2 involved a 69-year-old man who presented to the emergency department with acute onset of atraumatic left knee pain. On arthrocentesis, the patient had a cell count of 500,000 white blood cells/µL and was therefore taken to the operating room for arthroscopic irrigation and debridement. Final analysis of the synovial fluid showed monosodium urate crystals and negative culture findings. These cases illustrate the highest synovial fluid white blood cell count reported in patients with gout and highlight the potential difficulty in differentiating between acute gout and septic arthritis in the setting of markedly elevated white blood cell count.

Factors affecting fixation of the glenoid component of a reverse total shoulder prothesis.

The semiconstrained design of the reverse arthroplasty allows loads from the humerus to challenge the fixation of the glenoid component to the scapula. We examined some of the factors affecting the quality of glenoid screw fixation, including the density of the material into which the screws are placed, the purchase of individual screws, and the direction of loading in relation to screw placement. Loads were applied by the humeral component to glenoid components with different conditions of fixation. The load to failure for each set of conditions was measured and compared statistically. Load to failure was less when the glenoid component was fixed to material of lesser density. Each screw contributed to the quality of fixation; the screw nearest the point of load application made the largest contribution. Load to failure was less when the load was colinear with a line through the nonlocking holes in the base plate compared to colinear with a line through the locking holes. In performing a reverse total shoulder, surgeons should emphasize secure intraosseous placement of the fixation screws in the best quality bone available. The placement of the inferior screw appears to be the most critical.