Editorial Commentary: Knowledge is Power: A Primer for Machine Learning.

Machine learning (ML) has become an increasingly common statistical methodology in medical research. In recent years, ML techniques have been used with greater frequency to evaluate orthopaedic data. ML allows for the creation of adaptive predictive models that can be applied to clinical patient outcomes. However, ML models for predicting clinical or safety outcomes may be made available online so that physicians may apply these models to their patients to make predictions. If the algorithms have not been externally validated, then the models are not likely to generalize, and their predictions will suffer from inaccuracy. This is especially important to bear in mind because the recent increase in ML papers in the medical literature includes publications with fundamental flaws.

Outcomes of reverse shoulder arthroplasty following failed superior capsular reconstruction.

BACKGROUND: Superior capsular reconstruction (SCR) can be used for massive irreparable rotator cuff tears in the absence of significant degenerative changes; however, those who fail an SCR may require reverse shoulder arthroplasty (RSA). The effect of a previously performed SCR on outcomes following RSA remains unknown. METHODS: Subjects who underwent RSA from May 2015 to January 2021 at 2 separate institutions were retrospectively identified through prospectively collected databases. Patients who underwent RSA after failed SCR were matched to those who underwent RSA after failed rotator cuff repair (RCR) based on the number of previous ipsilateral shoulder procedures (n = 1, 2, ≥3) and secondarily by age within 5 years. American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form (ASES) score, Single Assessment Numeric Evaluation (SANE), visual analog scale (VAS) for pain, and Western Ontario Osteoarthritis of the Shoulder index (WOOS) scores were compared between groups. The minimal clinically important difference (MCID), substantial clinical benefit (SCB), and patient acceptable symptomatic state (PASS) thresholds were calculated to determine clinically relevant differences between groups. RESULTS: Forty-five patients were included (32 RSA following RCR, 13 following SCR). There were more smokers (P = .001) and worker's compensation cases (P = .034) in the SCR group. The RCR cohort was older (P = .007) and had a greater incidence of mental health (P > .999) and somatic disorders (P = .698), although these did not reach statistical significance. The mean follow-up for the RCR and SCR groups were 24.2 ± 23.3 and 20.4 ± 14.9 months following RSA, respectively (P = .913). The time from index RCR or SCR to RSA were 94.4 ± 22.2 and 89.2 ± 5.3 months, respectively (P = .003). Pre- and postoperative range of motion were similar between groups, as was the overall change in forward flexion (P = .879), abduction (P = .971), and external rotation (P = .968) following RSA. The RCR group had lower postoperative VAS pain (P = .009), higher SANE (P = .015), higher ASES (P = .008), and higher WOOS (P = .018) scores. The percentage achieving the MCID (P = .676) and SCB (P > .999) were similar; however, 56.7% of the RCR group met the SANE PASS threshold compared with 0.0% in the SCR group (P = .005). There were no differences in postoperative complications (P = .698) or revision rates (P = .308) following RSA between cohorts. CONCLUSION: When matched for number of previous procedures to the ipsilateral extremity and age, patients who underwent RSA following failed SCR had worse clinical outcome scores than their RSA following failed RCR counterparts. No patient in the SCR group met the SANE PASS threshold, whereas more than half of the RCR group did.

ORIF versus arthroplasty for open proximal humerus fractures: Nationwide Inpatient Sample data between 1998 and 2013.

BACKGROUND: Limited data exists in analyzing open reduction and internal fixation (ORIF) and arthroplasty in the management of open proximal humerus fractures. We analyzed differences in hospital course between these procedures, patient demographics, complication rate, length of stay, hospital charges, and mortality rate. MATERIALS AND METHODS: This is a retrospective review of the Nationwide Inpatient Sample database. ICD-9 codes identified patients hospitalized for open proximal humerus fractures from 1998 to 2013 who underwent ORIF or shoulder arthroplasty (hemi-, total, or reverse). Demographics and in-hospital complications were compared. Logistic regression controlling for age, gender, and Deyo index tested the impact of ORIF vs ARTH on any complications. RESULTS: Seven hundred thirty patients were included (ORIF, n = 662 vs ARTH, n = 68). ORIF patients were younger (p < 0.001), more likely to be males (p < 0.001), and had a lower Deyo score (p = 0.012). Both groups had comparable complication rates (21.4% vs 18.0%, p = 0.535), lengths of stay (7.86 days vs 7.44 days, p = 0.833), hospital charges ($76,998 vs $64,133, p = 0.360), and mortality rates (0.2% vs 0%, p = 0.761). Type of surgery was not a predictor of any complications (OR = 0.67 [95% CI 0.33-1.35], p = 0.266), extended length of stay (OR = 1.01 [95% CI 0.58-1.78], p = 0.967), or high hospital charges (OR = 1.39 [95% CI 0.68-2.86], p = 0.366). CONCLUSION: We revealed no differences in hospital course between ORIF and arthroplasty for management of open proximal humerus fractures. Although differences in demographics existed, no differences in complication rates, length of stay, hospital charges and mortality rates were noted. Future studies can evaluate the long-term outcomes of these procedures. LEVEL OF EVIDENCE: Level III.

Improving Applicant Satisfaction in Orthopaedic Surgery Residency Matching: The Role of Preference Signaling.

Background: The orthopaedic surgery residency match is becoming increasingly more competitive with a disproportionate number of applicants to positions. As the residency application process has become more competitive, applicants have resorted to applying broadly to improve their chance of a successful match. Preference signaling was implemented for orthopaedic surgery for the 2022-2023 match cycle which allowed applicants to "signal" 30 programs of their choosing. Methods: The purpose of this study was to assess the impact of preference signaling on orthopaedic surgery applicant experiences and outcomes in the 2023 residency application cycle and match. An anonymous electronically based survey study developed using Research Electronic Data Capture (REDCap) was send to 895 applicants to a single orthopedic residency program. 148 applicants filled out some portion of the survey for a 16.5% response rate. Results: 51% of applicants applied to 61-100 programs. Applicants received more interview offers from programs they signaled compared to programs they did not signal. 50% of applicants responded that the number of allotted signals was "just right", with more applicants responding that the number of signals allotted was "too many" rather than "too few". 62% of applicants agreed that signaling increased his/her chances of receiving an interview offer at a signaled program, 66% were satisfied with the results of the match, and 50% thought signaling had a positive impact on the application process. Conclusion: Overall, preference signaling was well received by applicants and may help to connect applicants with residency programs they are specifically interested in. Level of Evidence: III.

Total Hip Arthroplasty Learning Curves Based on Approach for New Fellowship-trained Surgeons.

INTRODUCTION: There is concern that the direct anterior (DA) approach is associated with a steeper learning curve than the posterolateral (PL) approach in total hip arthroplasty (THA). The purpose of this study was to investigate if the learning curve is similar for newly trained arthroplasty fellowship-trained surgeons using the DA and PL approaches. METHODS: The first 100 primary THA cases of six DA and PL fellowship-trained arthroplasty surgeons were identified and divided into 50 case cohorts. Demographics, indications for surgery, and 90-day Hip Society standardized complications were collected. Variables were analyzed using independent sample t tests, chi-square tests, or Fisher exact tests. RESULTS: In total, there were 600 patients, and no notable differences were observed in revision surgeries, surgical complications, and total complications between the DA and PL groups. Both groups had lower revision surgery rates, surgical complications, and total complications during their second 50 cases. Higher revision surgeries, and surgical and total complication rates were observed among all surgeons during the first 50 cases. DISCUSSION AND CONCLUSION: No differences were observed in the learning curve when comparing the DA and PL approach. With proper training, early-career surgeons can safely perform THA with similar complication rates regardless of the approach.

Substantial clinical benefit values demonstrate a high degree of variability when stratified by time and geographic region.

Background: A Substantial Clinical Benefit (SCB) value is the amount of change in a patient-reported outcome measure required for a patient to feel they significantly improved from an intervention. Previously published SCB values are often cited by researchers when publishing outcomes data. Where these SCB values are set can have a large impact on the conclusions drawn from a study citing them. As such, the goal of this study was to determine the generalizability of SCB values for a procedure when stratified by time from surgery and geographic region. Methods: A nationwide outcomes database was utilized to obtain preoperative, one-year, and two-year postoperative outcome measurements for patients who underwent anatomic total shoulder arthroplasty (TSA) or reverse TSA. The data were divided into three geographic regions: the South, the Midwest, and the West. An East region was not included due to its limited number of patients. SCB values were calculated for four outcomes measures: Single Assessment Numeric Evaluation score, American Shoulder Elbow Surgeons score, Visual Analog Scale, and Western Ontario Osteoarthritis of the Shoulder score. SCB values were calculated for each region, for each procedure, and at both one and two years postoperatively. To determine the variability of potential SCBs within each region, simulated datasets were created to determine a distribution of possible calculated SCBs. Results: A total of 380 anatomic TSA patients and 543 reverse TSA patients were included for analysis. There was a high degree of variability of SCB values when stratified by procedure, time, and region. While some simulated datasets did produce homogenous SCB distributions among regions, some outcome measures demonstrated a large heterogeneity in distribution among regions, with concomitant large distributions of values within individual regions. Conclusions: There is notable heterogeneity of SCB values when stratified by region or time. The current method of citing previously published SCB values for determining the efficacy of an intervention may be inappropriate. It is likely that this variability holds true in other areas of orthopedics.

Readability assessment of patient educational materials for shoulder arthroplasty from top academic orthopedic institutions.

BACKGROUND: Previous studies have demonstrated that online patient educational materials are written at reading levels too advanced for the average patient. The average American reads at the eighth-grade reading level. To date, the readability of online educational material of academic centers for shoulder arthroplasty has not been analyzed. METHODS: Online patient educational materials from the top 25 orthopedic institutions, as ranked by U.S. News & World Report, were assessed utilizing the following readability assessments: Flesch-Kincaid (FK), Flesch Reading Ease, Gunning Fog Index, Coleman-Liau Index, Simple Measure of the Gobbledygook Index, Automated Readability Index, FORCAST, and the New Dale and Chall Readability. All of these scores, with the exception of the Flesch Reading Ease, provide an output indicating reading difficulty based on grade level. Correlations between academic institutional ranking and FK scores were evaluated using a Spearman regression. Lastly, additional factors including geographical location, private versus public institution, and use of concomitant multi-media modalities that may impact institutional readability scores (as determined by FK) were evaluated. RESULTS: Only 16.0% of the top 25 institutions included online material at or below the eighth-grade reading level. Moreover, half of the online resources evaluated (those with FK score ≥9.3) were not at a suitable reading level for more than two-thirds of the general United States population (∼70%). Overall, the composite mean scores were 9.5 ± 2.1 for FK, 52.8 ± 9. for 4 Flesch Reading Ease, 12.2 ± 2.4 for Gunning Fog, 11.6 ± 1.8 for Coleman-Liau, 12.3 ± 1.7 for Simple Measure of the Gobbledygook Index, 9.6 ± 2.6 for Automated Readability, 11.1 ± 0.6 for FORCAST, and 5.9 ± 0.6 for New Dale and Chall. There was no correlation between institutional ranking and FK scores (ρ = -0.15; P = .946). Geographical location, private versus public institution, and use of concomitant multi-media modalities were not significantly associated with readability. CONCLUSION: Shoulder arthroplasty online patient educational material at top-ranked orthopedic institutions have poor readability and are likely not suitable for the majority of patients in the United States.

Efficacy of Arthroscopic Shavers for the Retrieval and Processing of Connective Tissue Progenitor Cells from Subacromial Bursal Tissue.

The purpose of this study is to determine if arthroscopic shavers can effectively collect and process connective tissue progenitor (CTP) cells from subacromial bursal tissue for utilization in rotator cuff repair augmentation. Subacromial bursal tissue was collected and processed using two arthroscopic shavers, Shaver A and Shaver B, in 10 patients undergoing arthroscopic rotator cuff repair. Each shaver was used in a random order for the same patient. Tissue samples underwent testing for cellular proliferation, cellular concentration, number of colony-forming units (CFU), live/dead assay, fluorescence-activated cells sorting (FACS) analysis, cytokine analysis, and growth factor analysis. Shaver A produced more CFUs compared to Shaver B (210.3 vs. 125.9; p < 0.001). At 3 weeks, cells collected via Shaver A had greater cellular proliferation (0.35 vs. 0.51; p < 0.001) as well as more viable cells (214,773 vs. 132,356 cells/gram; p < 0.001). Tissue collected with Shaver B had greater amounts of the cytokines MMP-1 (3741 vs. 5500 pg/mL; p < 0.001), MMP-3 (1131 vs. 1871 pg/mL; p < 0.001), and MMP-13 (179 vs. 401 pg/mL; p < 0.001), while those collected with Shaver A had greater vascular endothelial growth factor (VEGF) (47.8 vs. 9.0 pg/mL; p < 0.05). Arthroscopic shavers are capable of harvesting and processing CTP cells from subacromial bursal tissue. Different shavers may produce different yields of viable CTP cells.

Restoration of Global Sagittal Alignment After Surgical Correction of Cervical Hyperlordosis in a Patient with Emery-Dreifuss Muscular Dystrophy: A Case Report.

CASE: We report a rare cervical hyperlordotic deformity in a 19-year-old woman with Emery-Dreifuss muscular dystrophy and concomitant scoliosis. After standard posterolateral instrumentation and fusion of C2-T1 and extensive soft-tissue release, her neck pain improved and unassisted maintenance of cervical alignment and horizontal gaze were preserved through an 8-year follow-up. More importantly, she exhibited reciprocal correction of compensatory global sagittal malalignment, including lumbar lordosis. CONCLUSIONS: This case highlights the importance of full-spine analysis for all patients with spinal deformity to identify and differentiate primary driver(s) of deformity from compensatory mechanisms to individualize treatment toward what truly drives the patient's disability.

Adolescent Idiopathic Scoliosis Care in an Underserved Inner-City Population: Screening, Bracing, and Patient- and Parent-Reported Outcomes.

STUDY DESIGN: Retrospective review of a prospectively collected database. OBJECTIVES: This preliminary investigation sought to identify the quality of care adolescent idiopathic scoliosis (AIS) patients from our large, underserved community had received before presenting at this institution's clinic. SUMMARY OF BACKGROUND DATA: AIS affects 1% to 4% of children between ages 10 and 16. Barriers to health care for patients in underserved populations have not been well studied. METHODS: Patients who visited a single surgeon's clinic for primary AIS between June 2016 and January 2017 were enrolled. Patients had 36-inch full-spine radiographs and completed a survey of demographics, prior AIS care received (screening, bracing, etc), socioeconomic parameters, and patient-reported outcomes (PROs; Scoliosis Research Society [SRS]-30 Questionnaire and Body Image Disturbance Questionnaire [BIDQ]). Parametric and nonparametric analyses were used and percentages and mean/median values were reported. RESULTS: 47 patients (age: 15 ± 3 years; 82.7% female) were included. Overall, 25.5% of patients reported a family history of scoliosis, and 42.6% had no prior knowledge of scoliosis. Per Scoliosis Research Society (SRS) recommendations, 15 patients required observation (main Cobb angle: <25°), 22 patients were eligible for bracing (25°-45°), and 10 patients were surgical candidates (>45°). In addition, 21.3% of all patients were never screened for scoliosis; of these, 50% had a main scoliosis curve >25°. Seventy percent of surgical candidates never wore a brace, and 59.3% of screened patients who were eligible for bracing were not braced at initial presentation. Patients who were left unbraced when eligible exhibited worse BIDQ scores (1.7 vs. 1.4, p < .05). CONCLUSIONS: One of five children in our population was never screened for scoliosis, and nearly three of five children did not receive optimal care as recommended by SRS. AIS patients in our inner-city populations are potentially at risk of continuing to experience a significant disadvantage in health care access. LEVEL OF EVIDENCE: Level IV case series.

Noncontact sports participation in adolescent idiopathic scoliosis: effects on parent-reported and patient-reported outcomes.

Comparing risks against benefits of adolescent idiopathic scoliosis (AIS) patients participating in sports represents a controversial topic in the literature. Previous studies have reported sports participation as a possible risk factor for AIS development, while others describe its functional benefits for AIS athletes. The objective of this study was to determine if sports participation had an impact on pain, function, mental status, and self-perception of deformity in patients and their parents. Patients had full spine radiographs and completed baseline surveys of demographics, socioeconomics, and patient-reported outcomes (PRO): Scoliosis Research Society (SRS)-30, Body Image Disturbance Questionnaire, and Spinal Appearance Questionnaire (SAQ: Children and Parent). Patients were grouped by their participation (sports) or nonparticipation (no-sports) in noncontact sports. Demographics, radiographic parameters, and PRO were compared using parametric/nonparametric tests with means/medians reported. Linear regression models identified significant predictors of PRO. Forty-nine patients were included (sports: n=29, no-sports: n=20). Both groups had comparable age, sex, BMI, bracing status, and history of physical therapy (all P>0.05). Sports and no-sports also had similar coronal deformity (major Cobb: 31.1° vs. 31.5°). Sagittal alignment profiles (pelvic incidence, pelvic incidence minus lumbar lordosis, thoracic kyphosis, and sagittal vertical axis) were similar between groups (all P>0.05). Sports had better SRS-30 (Function, Self-image, and Total) scores, SAQ-Child Expectations, and SAQ-Parent Total Scores (P<0.05). Regression models revealed major Cobb angle (ß coefficient: -0.312) and sports participation (ß coefficient: 0.422) as significant predictors of SRS-30 Function score (R=0.434, P<0.05). Our data show that for AIS patients with statistically similar bracing status and coronal and sagittal deformities, patients who participated in sports were more likely to have improved functionality, self-image, expectations, and parental perception of deformity. Further investigation is warranted to acquire a comprehensive understanding of the relationship between AIS and patient participation in sports. Maintaining moderate levels of physical activity and participating in safe sports may benefit treatment outcomes. Level of Evidence III - Retrospective Comparative Study.

The Risks of Hepatitis C in Association With Cervical Spinal Surgery: Analysis of Radiculopathy and Myelopathy Patients.

STUDY DESIGN: Retrospective review. OBJECTIVE: To investigate rates of in-hospital postsurgical complications among hepatitis C-infected patients after cervical spinal surgery in comparison with uninfected patients and determine independent risk factors. SUMMARY OF BACKGROUND DATA: Studying hepatitis C virus (HCV) as a possible risk factor for cervical spine postoperative complications is prudent, given the high prevalence of cervical spondylosis and HCV in older patients. Spine literature is limited with respect to the impact of chronic HCV upon complications after surgery. MATERIALS AND METHODS: Patients who underwent cervical spine surgery for cervical radiculopathy (CR) or cervical myelopathy (CM) from 2005 to 2013 were retrospectively reviewed using the Nationwide Inpatient Sample database. Patients were divided into CR and CM groups, with comparative subgroup analysis of HCV and no-HCV patients. Univariate analysis compared demographics and complications. Binary logistic stepwise regression modeling identified any independent outcome predictors (covariates: age, sex, Deyo score, and surgical approach). RESULTS: In total, 227,310 patients (HCV: n=2542; no-HCV: n=224,764) were included. From 2005 to 2013, HCV infection prevalence among all cervical spinal fusion cases increased from 0.8% to 1.2%. HCV patients were more likely to be African American or Hispanic and have Medicare and/or Medicaid (all P<0.001). Overall complication rates among HCV patients with CR or CM increased, specifically related to device (CR: 3.1% vs. 1.9%; CM: 2.9% vs. 1.3%), hematoma/seroma (CR: 1.1% vs. 0.4%; CM: 1.8% vs. 0.8%), and sepsis (CR: 0.4% vs. 0.1%; CM: 1.1% vs. 0.5%) (all P≤0.001). Among CR and CM patients, HCV significantly predicted increased complication rates [odds ratio (OR): 1.268; OR: 1.194], hospital stay (OR: 1.738; OR: 1.861), and hospital charges (OR: 1.516; OR: 1.732; all P≤0.044). CONCLUSIONS: HCV patients undergoing cervical spinal surgery were found to have increased risks of postoperative complications and increased risk associated with surgical approach. These findings should augment preoperative risk stratification and counseling for HCV patients and their spine surgeons. LEVEL OF EVIDENCE: Level III.