Development of a predictive model for a machine learning–derived shoulder arthroplasty clinical outcome score

BackgroundWe use machine learning to create predictive models from preoperative data to predict the Shoulder Arthroplasty Smart (SAS) score, the American Shoulder and Elbow Surgeons (ASES) score, and the Constant score at multiple postoperative time points and compare the accuracy of each algorithm for anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (rTSA).MethodsClinical data from 2270 patients who underwent aTSA and 4198 patients who underwent rTSA were analyzed using 3 supervised machine learning techniques to create predictive models for the SAS, ASES, and Constant scores at 6 different postoperative time points using a full input feature set and the 2 different minimal feature sets. Mean absolute errors (MAEs) quantified the difference between actual and predicted outcome scores for each model at each postoperative time point. The performance of each model was also quantified by its ability to predict improvement greater than the minimal clinically important difference (MCID) and the substantial clinical benefit (SCB) patient satisfaction thresholds for each outcome measure at 2-3 years after surgery.ResultsAll 3 machine learning techniques were more accurate at predicting aTSA and rTSA outcomes using the SAS score (aTSA: ±7.41 MAE; rTSA: ±7.79 MAE), followed by the Constant score (aTSA: ±8.32 MAE; rTSA: ±8.30 MAE) and finally the ASES score (aTSA: ±10.86 MAE; rTSA: ±10.60 MAE). These prediction accuracy trends were maintained across the 3 different model input categories for each of the SAS, ASES, and Constant models at each postoperative time point. For patients who underwent aTSA, the XGBoost predictive models achieved 94%-97% accuracy in MCID with an area under the receiver operating curve (AUROC) between 0.90-0.97 and 89%-94% accuracy in SCB with an AUROC between 0.89-0.92 for the 3 clinical scores using the full feature set of inputs. For patients who underwent rTSA, the XGBoost predictive models achieved 95%-99% accuracy in MCID with an AUROC between 0.88-0.96 and 88%-92% accuracy in SCB with an AUROC between 0.81-0.89 for the 3 clinical scores using the full feature set of inputs.DiscussionOur study demonstrated that the SAS score predictions are more accurate than the ASES and Constant predictions for multiple supervised machine learning techniques, despite requiring fewer input data for the SAS model. In addition, we predicted which patients will and will not achieve clinical improvement that exceeds the MCID and SCB thresholds for each score; this highly accurate predictive capability effectively risk-stratifies patients for a variety of outcome measures using only preoperative data.Level of evidenceLevel III; Retrospective Comparative Study     

Defining maximal outcome improvement thresholds for patient satisfaction after reverse total shoulder arthroplasty: a minimum two-year follow-up study

BackgroundPrevious research has investigated the percentage of maximal outcome improvement (MOI) for the American Shoulder and Elbow Surgeons (ASES) and Simple Shoulder Test scores after reverse total shoulder arthroplasty (rTSA); however, few other outcome scores have been validated. The purpose of this study was to establish thresholds of MOI percentage for the Single Assessment Numeric Evaluation (SANE) and subjective Constant-Murley (Constant) scores associated with patient satisfaction at 2 years following rTSA.MethodsA retrospective review of institutional registry data for all patients who underwent primary rTSA between 09/2016 and 02/2018 was performed. All patients completed the ASES, SANE, and Constant outcome measures preoperatively and at a minimum of 2 years postoperatively. Changes were assessed with Wilcoxon signed-rank tests. MOI percentage was defined as the score improvement from baseline divided by the highest possible improvement from baseline based on the maximum value of the outcome score used. Receiver operating characteristic (ROC) analyses were used to determine optimal MOI percentage thresholds associated with postoperative satisfaction. Stepwise multivariate logistic regression was performed to identify variables associated with achieving the MOI.ResultsA total of 105 patients were included in the final analysis, with a mean (±standard deviation) age of 70.0 ± 7.5 years and BMI of 28.9 ± 5.6 kg/m². The majority of patients were satisfied postoperatively (81.0%). Statistically significant increases were observed from baseline to 2 years postoperatively for the outcome measures assessed (P < .001 for all). ROC analysis demonstrated that the optimal MOI percentage threshold for achieving satisfaction was 66.9% for ASES ([Area under curve, AUC]: 0.87), 71.0% for SANE (AUC: 0.89), and 37.5% for Constant (AUC: 0.85). The linear regression model fit between the proportion of satisfied patients and the calculated MOI percentages was moderate for SANE (r² = 0.313, P < .001) and weak for Constant (r² = 0.228, P < .001). Multivariate logistic regression demonstrated that comorbid diabetes was associated with a decreased likelihood of MOI achievement for SANE ([Odds Ratio, OR]: 0.66, P = .020) and Constant (OR: 0.72, P = .027). Preoperative diagnosis of rotator cuff arthropathy (OR: 0.84, P = 0.049) was associated with a decreased likelihood of MOI achievement for Constant. No factors were significantly associated with MOI achievement for ASES.ConclusionAchieving 66.9%, 71.0%, and 37.5% of the MOI for the ASES, SANE, and Constant scores is associated with postoperative satisfaction following rTSA. Comorbid diabetes and diagnosis of rotator cuff arthropathy were significantly associated with a decreased likelihood of MOI achievement.Level of evidenceLevel III; Retrospective Case Series (Prognostic)     

Midterm outcomes of revision shoulder arthroplasty

BackgroundSince the introduction of shoulder arthroplasty, the indications have been expanding. Because of the increasing number of arthroplasty procedures, revision surgeries are also inevitable. The purpose of our study is to delineate a large number of revision shoulder arthroplasty cases treated in different ways, including antibiotic spacer placement (ABX), hemiarthroplasty (HA), anatomic total shoulder arthroplasty (aTSA), and reverse total shoulder arthroplasty (rTSA), and to analyze the relationship between preoperative factors and clinical outcomes.MethodsWe reviewed our institution's records of revision shoulder arthroplasty between January 1, 2000, and October 1, 2017. Preoperative information included age at the time of surgery, gender, body mass index, and infection status. Pre- and postoperatively, we gathered 6 patient-reported clinical outcomes and 3 range-of-motion parameters (elevation, abduction, and external rotation). Postoperative complications were also assessed. Then, we examined the differences between the pre- and postoperative outcomes. As a secondary analysis, we performed multivariable regression analysis on the same outcomes, accounting for age at the time of surgery, infection status, and previous surgery type.ResultsAmong the 341 revisions performed, 138 cases met inclusion criteria of at least a 2-year follow-up with pre- and postoperative functional outcome scores. The majority of our revision procedures were to a rTSA (92 cases, 67%), followed by aTSA (28 cases, 20%), and ABX/HA (18 cases, 13%). The mean age at the time of our index surgery was 66 years old. In aTSA and rTSA, all the postoperative outcomes (ie, ASES, Constant, UCLA, SST, SPADI, SF-12 scores, and 3 AROMs) were significantly improved beyond the minimal clinically important difference (MCID) except SF-12 scores in aTSA (P = .25) and active external rotation in rTSA (P = .73). None of the ABX/HA's postoperative outcomes achieve significant improvement or MCID. Multivariable regression analysis showed that older age at the time of surgery was significantly associated with better outcomes in 3 of the 6 patient-reported outcomes (ASES, SST, and SPADI; P = .023, .023, and .028, respectively).ConclusionRevision aTSA and rTSA showed statistically and clinically significant improvement postoperatively. ABX and HA did not achieve meaningful postoperative improvement. Overall, patients getting revision shoulder arthroplasty at an older age had better patient-reported outcomes.Level of evidenceLevel III; Retrospective Cohort Study     

Using inertial measurement units to quantify shoulder elevation after reverse total shoulder arthroplasty: a pilot study comparing goniometric measures captured clinically to inertial measures captured ‘in-the-wild’

BackgroundReverse total shoulder arthroplasty (rTSA) is utilized for a variety of indications, but most commonly for patients with rotator cuff arthropathy. This procedure reduces pain, improves satisfaction, and increases clinically measured range of motion (ROM). However, traditional clinical ROM measurements captured via goniometer may not accurately represent ‘real-world’ utilization of ROM. In contrast, inertial measurement units (IMUs) are useful for establishing ROM outside the clinical setting. We sought to measure ‘real-world’ ROM after rTSA using IMUs.MethodsA previously validated IMU-based method for continuously capturing shoulder elevation was used to assess 10 individuals receiving rTSA (1M, 82 ± 5 years) and compared to a previously captured 10 healthy individuals (4M, 69 ± 20 years) without shoulder dysfunction. Control subject data were previously collected over 1 week of continuous use. Patients undergoing rTSA donned sensors for 1 week pre-rTSA, 6 weeks at 3 months post-rTSA following clearance to perform active-independent ROM, and 1 week at 1 year and 2 years post-rTSA. Shoulder elevation was computed continuously each day. Daily continuous elevation was broken into 5° angle ‘bins’ (eg, 0-5°, 5-10°, etc.) and converted to percentage of the total day. IMU-based outcome measures were ROM binned percent (as described previously) and maximum/average elevation each week. Clinical goniometric ROM and patient-reported outcome measures were also captured.ResultsNo differences existed between patient and healthy control demographics. While patients showed improvement in American Shoulder and Elbow Surgeon (ASES) score, pain score, and goniometric ROM, IMU-based average and maximum elevation were equal between control subjects and patients both pre- and post-rTSA. The percent of time spent above 90° was equal between cohorts pre-rTSA, rose significantly at 3 months post-rTSA, and returned to preoperative levels thereafter.DiscussionAlthough pain, satisfaction, and ROM measured clinically may improve following rTSA, real-world utilization of improved ROM was not seen herein. Improvements during the acute rehabilitation phase may be transient, indicating longer or more specific rehabilitation protocols are necessary to see chronic improvements in post-rTSA movement patterns.     

Expectations affect outcome and satisfaction after anatomic total shoulder arthroplasty

BackgroundThere has been limited study of the relationship between expectations and outcome of anatomic total shoulder arthroplasty (aTSA). The purpose of this study was to investigate associations between preoperative expectations and the outcome of aTSA.MethodsThis was a retrospective review of 176 patients with advanced glenohumeral osteoarthritis who underwent aTSA. Baseline preoperative expectations, as well as postoperative met expectations and satisfaction with outcome at 2-year follow-up were determined using questions from the Musculoskeletal Outcomes Data Evaluation and Management System questionnaire. Patient reported outcomes measures (PROMs) were assessed at baseline and at 2-year follow-up with the Disabilities of arm, shoulder and hand score (DASH), Simple Shoulder Test, and Visual Analog Scale shoulder pain.ResultsPreoperative expectations were high (mean 4.57 ± 0.50 [max = 5]). Improvements from baseline to 2-year follow-up in PROMs were significant and clinically relevant. At 2-year follow-up the outcome met the expectations of a high percentage of patients (mean met expectations 4.63 ± 0.63 [max = 5]). There were statistically significant correlations between the preoperative mean expectations and the follow-up PROMs (DASH P < .001, rho = 0.30). Better PROMs at 2-year follow-up were significantly associated with better mean met expectations responses (DASH P< .001, rho = 0.51; Simple Shoulder Test P < .001, rho = 0.43; Visual Analog Scale shoulder pain P= 0.003, rho = 0.24). There was a statistically significant correlation between mean met expectations and satisfaction (p < .001, rho = 0.34). The strengths of a number of the statistically significant correlations were very weak or weak.ConclusionPatients with glenohumeral osteoarthritis who elect to have aTSA have high expectations of the outcome and the outcomes achieved are highly likely to meet their expectations. Although there are statistically significant associations and correlations between preoperative expectations and PROMs, PROMs and met expectations, and met expectation and satisfaction with the outcome, the strengths of the correlations are relatively weak. Further study of the factors that relate to and affect expectations is required to better understand these relationships and to appropriately set expectations to optimize patient satisfaction.Level of evidenceLevel IV; Retrospective Review     

The association of comorbidity indices and minimal clinically important difference in a population undergoing shoulder arthroplasty

BackgroundShoulder arthroplasty (SA) procedures are increasingly common. The Charlson and Elixhauser indices are ICD-10 based measures used in large databases to describe the patient case mix in terms of secondary medical conditions. There is a paucity of data on the relationship between these indices and patient-reported outcome measures (PROMs) after shoulder arthroplasty.MethodsPatients undergoing SA from 2016-2018 were identified in the electronic medical records. Charlson and Elixhauser comorbidities were used to calculate comorbidity scores according to established algorithms (eg Elixhauser-Walraven). Patient shoulder-specific (American Shoulder and Elbow Surgeons (ASES) score and Shoulder Activity Scale (SAS)) and general health scales (SF-12 Mental Component Score (MCS) and Physical Component Score (PCS) and Patient-Reported Outcomes Measurement Information System-Pain Interference (PROMIS-PI)) PROMs were obtained from our institution’s shoulder registry. Linear regression models adjusting for age and sex evaluated associations between comorbidity scores and PROMs. Receiver operating characteristic (ROC) curves determined optimal cutoffs, maximizing sensitivity and specificity to identify patients likely to fail to meet minimal clinically important difference (MCID) values.ResultsA total of 1817 SA procedures were identified. Higher Charlson and Elixhauser-Walraven scores were significantly associated with lower baseline SAS and SF-12 PCS. Patients with higher Charlson scores had lower baseline ASES (P = .003) and lower baseline (P = .0002) and 2-year (P = .02) SF-12 MCS. No significant associations were found for PROMIS-PI with either index. The Charlson score better predicted the failure to meet MCID for SF-12 PCS with an AUC of 0.64, compared to 0.55 for Elixhauser-Walraven. Conversely, Elixhauser-Walraven better predicted PROMIS-PI with an AUC of 0.66, compared to 0.53 for Charlson. Moderate AUCs were observed for the remaining PROMs, ranging from 0.57-0.64, with little difference between index scores.ConclusionHigher Charlson and Elixhauser-Walraven scores were associated with lower baseline scores on most PROMs. Generally, the Charlson score performed better than the Elixhauser-Walraven score in predicting worse outcomes at 2 years. Comorbidity indices may be useful as a decision aid to provide appropriate expectations of outcomes for patients undergoing SA.     

MRI-Based Shoulder Osteoarthritis Severity score (SOAS) is predictive of functional improvement following total shoulder arthroplasty

BackgroundTotal shoulder arthroplasty (TSA) is an increasingly common treatment for end-stage glenohumeral osteoarthritis. Current established radiographic measures and classification systems do not predict patient-reported outcomes from TSA. We hypothesized that the MRI-based Shoulder Osteoarthritis Severity (SOAS) Score would correlate with subjective improvement following TSA.MethodsPatients undergoing TSA with preoperative shoulder MRIs and pre- and postoperative ASES scores with minimum 2-year follow-up were included from a prospectively collected institutional shoulder arthroplasty database. SOAS scores, which is measured from 0 to 100 with an increasing score reflecting greater global degenerative changes, were assessed by two independent reviewers, and Samilson-Prieto grade and Walch classification were scored by one reviewer. Average SOAS scores were correlated with demographic factors and pre-, post-, and change (Δ) in ASES scores. Statistical analysis was performed with STATA with Pearson's correlation, one-way ANOVA, and ROC analysis, with significance defined by p <.05.Results30 patients (age 63 ± 10 years, 14 females, 16 males) who underwent primary anatomic TSA were included. The intraclass correlation coefficient (ICC) for total SOAS scores calculated by reviewers was 0.91. SOAS score correlated significantly with ΔASES (r = 0.61, p = .0003) and preoperative ASES (r = -0.37, p = .042), with greater MRI-based degenerative change associated with greater improvement after TSA and lower preoperative ASES score. No significant relationship was found between either Samilson-Prieto or Walch classification and SOAS or ASES scores. No significant relationship was found between SOAS scores and age, sex, or BMI. Using an MCID of 21 as previously reported, an ROC curve was generated and found to have an AUC of 0.96. A SOAS score cut-point of 36.25 was found to maximize sensitivity and specificity in predicting reaching MCID.ConclusionWe observed a significant positive correlation between the MRI-based SOAS score and functional improvement following TSA measured using change in ASES scores, indicating that patients with more advanced degenerative changes on MRI had greater improvement after shoulder replacement surgery. We found that the correlation strength was highest when comparing total SOAS score to ΔASES as opposed to any individual sub-component of the SOAS score. The MRI-based SOAS score for shoulder osteoarthritis may be a valuable tool for predicting patient outcomes following TSA.Level of evidenceLevel III; Retrospective Cohort Comparison; Prognosis Study     

Outcome analysis following open rotator cuff repair. Early effectiveness validated using four different shoulder assessment scales

Evaluation of upper extremity function after reconstructive surgery is increasingly important both to predict outcome and for the control of cost-effectiveness. Three validated, self-administered shoulder questionnaires were applied prospectively in 23 otherwise healthy patients with rotator cuff deficiency and correlated to the Constant-Murley Shoulder Score and a visual analogue scale for satisfaction. Seven women and 16 men with combined tears of supraspinatus and infraspinatus (mean age 55.3 ± 10.5 years, r/l: 14/9, follow-up 57.8 ± 15.7 weeks) were gathered prospectively and evaluated pre- and postoperatively with the American Shoulder and Elbow Surgeons (ASES) Shoulder Index, the Simple Shoulder Test (SST) and the Disabilities of the Arm, Shoulder and Hand Module (DASH questionnaire). Additionally, a visual analogue scale for satisfaction was employed. All four scores and the visual analogue scale revealed improvement at a statistically significant level (P < 0.01) after surgery. All questionnaires showed a significant correlation with the Constant-Murley Shoulder Score (ASES: r = 0.871, P < 0.01; DASH: r = –0.758, P < 0.01, SST: r = 0.494, P < 0.05, Pearson’s correlation coefficient). Taken together, all questionnaires were easy to apply, and reliable evaluation of shoulder function was possible with significant correlation to the Constant-Murley Shoulder Score postoperatively. The SST was easy to apply, and compound outcome analysis was possible with the ASES Shoulder Index and DASH questionnaire. The DASH scale was the most complex evaluation instrument. The Constant-Murley Shoulder Score comprises a physical examination, which is advantageous but restricts the application to the office. For postoperative assessment without the patient having to return to the clinic, the ASES Shoulder Index is preferred because of its good correlation to the Constant-Murley Shoulder Score (r = 0.871) and the visual analogue scale for satisfaction (r = 0.762). Received: 4 October 1999     

Outcomeanalyse nach offener Rekonstruktion von Rotatorenmanschettenrupturen Eine vergleichende Beurteilung neuer Bewertungsverfahren

Outcome evaluation is becoming increasingly important for reconstructive surgery in musculoskeletal diseases. In addition to established shoulder-scores, new outcome-scores are being developed to cover all effects caused by a disease or intervention. Three validated, self-administered shoulder questionnaires were applied prospectively in 23 otherwise healthy patients suffering from rotator cuff deficiency. These were correlated to the Constant-Murley Shoulder Score and to a visual analogue scale for satisfaction. 7 women and 16 men with combined tears of supraspinatus and infraspinatus (mean age 55.3 +/- 10.5, r/l: 14/9, FU 57.8 +/- 15.7 weeks) were gathered prospectively and evaluated pre- and postoperatively with the American Shoulder and Elbow Surgeons Shoulder Index (ASES), the Simple Shoulder Test (SST) and the Disabilities of the Arm, Shoulder and Hand Module (DASH questionnaire). In addition, a visual analogue scale for satisfaction was employed. All four scores as well as the visual analogue scale revealed improvement at a statistically significant level (paired, two-tailed t-test, P < 0.01) after surgery. All questionnaires had a significant correlation with the Constant-Murley Shoulder Score (Pearson's correlation coefficient: ASES: r = 0.871, P < 0.01; DASH: r = -0.758, P = < 0.01; SST: r = 0.494, P < 0.05). All were easy to apply and provided a reliable, postoperative evaluation of shoulder function. The SST was easy to apply, however compound outcome analysis was only possible with the ASES Shoulder Index and the DASH questionnaire. The DASH scale was the most complex evaluation instrument. The Constant-Murley Shoulder Score comprises a physical examination which is advantageous but must be carried out in the clinic. For postoperative assessment, without the patient having to return to the clinic, the ASES Shoulder Index is preferred as it correlates well with the Constant-Murley Shoulder Score (r = 0.871) and the visual analogue scale for satisfaction (r = 0.762).     

Prospective observational study of anatomic and reverse total shoulder arthroplasty using a single implant system with long-term follow-up

BackgroundBoth anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (rTSA) are the standard of care for various end-stage degenerative conditions of the glenohumeral joint. The purpose of this prospective study was to compare clinical outcomes of aTSA and rTSA using the same platform total shoulder arthroplasty system at a minimum follow-up of 8 years.MethodsAn international multicenter data registry was used to identify patients undergoing primary aTSA or rTSA with a minimum of 8-year follow-up. Patients were scored preoperatively and at latest follow-up using 6 outcome and 4 range of motion metrics. Patients graded global shoulder function on a 0-10 scale. Patient satisfaction was classified as much better, better, unchanged, or worse. Pain was graded using the Visual Analog Scale. A student’s 2-tailed paired t-test, Chi-squared test, or Wilcoxon rank-sum test were used when appropriate, where P < .05 indicates significant differences.ResultsA total of 364 aTSA patients and 278 rTSA patients were included. rTSA had greater mean age (aTSA 65.1, rTSA 71.2, P < .0001) and shorter mean follow-up (aTSA 115.2 months, rTSA 109.7 months; P = .0002). Postoperatively, all patients demonstrated significant improvements in pain and function. Preoperatively, aTSA patients had greater active abduction and forward elevation, global function, Constant Score, and the University of California, Los Angeles score. At latest follow-up, aTSA had greater active abduction, forward elevation, external rotation, and Simple Shoulder Test scores. There were no postoperative differences in patient satisfaction, shoulder function, pain, or any other outcome measure between the groups. aTSA patients had a greater revision rate (aTSA 5.8%, rTSA 1.8%; P = .0114), with no difference in complication rates. Humeral radiolucent lines were present in 20.9% of aTSA patients and 37.1% of rTSA patients (P = .0011). Glenoid radiolucent lines were present in 48.0% of aTSA patients (average score of 1.5). Scapular notching was present in 14.5% of rTSA patients (average grade of 0.26).ConclusionThis study is one of the largest to date that compares aTSA and rTSA with a minimum follow-up of 8 years. Significant improvements are seen in both aTSA and rTSA. For both groups, the complication rate was lower than previously reported in the literature. At the latest follow-up, pain relief and global function were significantly improved with high patient satisfaction in both groups, with no statistical differences found. This study provides physicians with information to help guide them when advising patients as to what they can expect at a minimum of 8 years after aTSA and rTSA.     

Clinical outcomes of upper extremity ambulators post-reverse total shoulder arthroplasty

BackgroundPatients requiring upper extremity assist devices such as canes, walkers, and wheelchairs have shown an increased incidence of rotator cuff tears (RCTs), distal clavicle osteolysis, and osteoarthritis. The frequent combination of glenohumeral arthritis and rotator cuff arthropathy makes reverse total shoulder arthroplasty (rTSA) an attractive option in treating the shoulder pathology in these patients. The objective of this study is to determine the effectiveness of the rTSA procedure for treatment of shoulder pathology in the upper extremity ambulator population, and what role preoperative functional status plays on outcomes.MethodsThe study involved a retrospective analysis of 30 upper extremity ambulator patients (ages 48-88, average = 72.54 years) who underwent a rTSA procedure performed by the senior author (S.G.) in Milwaukee, WI, USA. Of these 30, 5 patients received bilateral rTSAs, for a total of 35 procedures. Average follow-up for these patients was 32 months, (range 24-84 months). The following preoperative and postoperative characteristics were recorded for each patient: range of motion, Visual Analog Scale scores, American Shoulder and Elbow Surgeons scores, Constant Murley scores, and Simple Shoulder Test scores. Data were analyzed using ANCOVA analysis.ResultsAverage Visual Analog Scale scores improved from 6.40 to 2.08 (gain = 4.362; P< .0001). Average forward flexion improved from 70.32° to 127.74° (gain = 57.42°; P < .0001), while average abduction improved from 41.79° to 110.18° (gain = 68.69°; P = .0002). Average Murley scores improved from 25.68 to 63.32 (gain = 37.64; P = .030). American Shoulder and Elbow Surgeons and Simple Shoulder Test scores were not shown to be statistically significant, though patients showed a trend toward gains in these metrics. A patient's preoperative status was highly correlated with their expected gain in function postoperatively, with lower baseline preoperative function correlating to larger gains postoperatively. Patients showed statistically significant gains in the metrics listed above, regardless of preoperative diabetic status, opioid use, or the type of ambulatory assistive device used.ConclusionThis study shows that rTSA reliably leads to significant postoperative improvements in range of motion, function, and pain relief in the upper extremity ambulator population. Patients with lower preoperative scores showed overall higher gains in both pain and function postoperatively, ultimately leading to similar postoperative outcomes to those who started at a high baseline function. These gains were irrespective of a patient's other preoperative characteristics, including diagnosis of diabetes, preoperative opioid use, and the type of upper extremity assist device used.Level of evidenceLevel III.     

Use of machine learning to assess the predictive value of 3 commonly used clinical measures to quantify outcomes after total shoulder arthroplasty

BackgroundAn important psychometric parameter of validity that is rarely assessed is predictive value. In this study we utilize machine learning to analyze the predictive value of 3 commonly used clinical measures to assess 2-year outcomes after total shoulder arthroplasty (TSA).MethodsXGBoost was used to analyze data from 2790 TSA patients and create predictive algorithms for the American Shoulder and Elbow Surgeons (ASES), Constant, and the University of California Los Angeles (UCLA) scores and also quantify the most meaningful predictive features utilized by these measures and for all questions comprising each measure to rank and compare their value to predict 2-year outcomes after TSA.ResultsOur results demonstrate that the ASES, Constant, and UCLA measures rarely considered the most-predictive features relevant to 2-year TSA outcomes and that each outcome measure was composed of questions with different distributions of predictive value. Specifically, the questions composing the UCLA score were of greater predictive value than the Constant questions, and the questions composing the Constant score were of greater predictive value than the ASES questions. We also found the preoperative Shoulder Pain and Disability Index (SPADI) score to be of greater predictive value than the preoperative ASES, Constant, and UCLA scores. Finally, we identified the types of preoperative input questions that were most-predictive (subjective self-assessments of pain and objective measurements of active range of motion and strength) and also those that were least-predictive of 2-year TSA outcomes (subjective task-specific activities of daily living questions).DiscussionMachine learning can quantify the predictive value of the ASES, Constant, and UCLA scores after TSA. Future work should utilize this and related techniques to construct a more efficient and effective clinical outcome measure that incorporates subjective and objective input questions to better account for the preoperative factors that influence postoperative outcomes after TSA.Level of EvidenceLevel III; Retrospective Comparative Study     

Arthroscopic repair of type II SLAP lesions: Clinical and anatomic follow-up

Aims:The aim was to evaluate the clinical and anatomic outcome of arthroscopic repair of type II SLAP lesions.Results:At a mean follow-up of 54-month, the mean American Shoulder and Elbow Surgeons Shoulder Index (ASES) scores improved from 52.1 preoperatively to 86.1 postoperatively (P < 0.0001) and the Simple Shoulder Test (SST) scores from 7.7 to 10.6 (P < 0.0002). Twenty-two out of the 25 patients (88%) stated that they would have surgery again. Of the 21 patients who had postoperative magnetic resonance imaging arthrographys (MRAs), 9 patients (43%) demonstrated dye tracking between the labrum bone interface suggestive of a recurrent tear and 12 patients (57%) had a completely intact repair. There was no significant difference in ASES, SST, and patient satisfaction scores in patients with recurrent or intact repairs.Conclusions:Arthroscopic repair of type II SLAP lesions demonstrated improvements in clinical outcomes. However, MRA imaging demonstrated 43% of patients with recurrent tears. MRA results do not necessarily correlate with clinical outcome.     

Clinical and patient-reported outcomes of reverse shoulder arthroplasty in patients receiving Social Security Disability Insurance or workers’ compensation

BackgroundEvidence suggests that reverse shoulder arthroplasty (RSA) patients receiving workers’ compensation (WC) have worse patient-reported outcomes (PROs) than those not receiving WC. It is unknown whether Social Security Disability Insurance (SSDI) recipients also have worse outcomes of RSA. Our goals were to (1) compare PROs and range of motion (ROM) after RSA according to whether patients were receiving SSDI, WC, or neither form of assistance, and (2) identify factors associated with poor PROs.MethodsFrom a US institutional database of 454 patients who underwent RSA from January 2009 through December 2016, we identified 19 SSDI recipients and 25 WC recipients. From the same database, we created a control group of 81 patients not receiving SSDI or WC, matched by demographic variables. Between groups, we compared age, sex, operative arm dominance, preoperative diagnosis, number of previous shoulder surgeries, primary or revision arthroplasty, and Charlson Comorbidity Index value. Patients were evaluated preoperatively and at a minimum of 2 years postoperatively by physical examination, including range of motion, visual analog scale (VAS) for pain, American Shoulder and Elbow Surgeons (ASES) score, L'Insalata score, Simple Shoulder Test (SST), and Western Ontario Osteoarthritis of the Shoulder (WOOS) score. Significance was set at P< .05. Clinical relevance of improvements in forward flexion and abduction, VAS for pain, ASES, and SST were evaluated using established minimal clinically important difference (MCID) values.ResultsBoth the SSDI and WC groups experienced statistically significant improvements from preoperatively to postoperatively; these improvements exceeded MCID thresholds for forward flexion and abduction, VAS for pain, ASES (except in WC group), and SST score (all, P< .05). These outcomes were not significantly different between the SSDI and WC groups. Compared with the control group, both the SSDI and WC groups had statistically significantly worse outcomes for these same measures. The only factor associated with poor clinical outcomes was having undergone ≥2 previous surgical procedures on the same shoulder for which RSA was performed in both SSDI (odds ratio = 2.4, 95% confidence interval: 1.0-5.4) and WC (odds ratio = 1.6, 95% confidence interval: 1.1-4.5) groups.ConclusionAmong RSA patients, SSDI recipients did not have worse clinical outcomes than WC recipients. Having undergone ≥2 previous procedures on the same shoulder was associated with poor outcomes in both groups. These findings should be considered when determining the appropriateness of RSA for SSDI and WC recipients.Level of evidenceLevel III; Retrospective Cohort Comparison; Treatment Study     

The effect of an early mobilization rehabilitation protocol on outcomes after reverse shoulder arthroplasty

BackgroundGiven the continued growth of reverse shoulder arthroplasty (RSA), it is important to optimize factors that contribute to successful outcomes. Rehabilitation after RSA is critical to achieve successful outcomes including patient function and satisfaction; however, the ideal rehabilitation program has not been established. The purpose of this study was to evaluate the effect of early mobilization (EM) compared with delayed mobilization (DM) on outcomes and function after RSA.MethodsA retrospective study of prospectively collected data was performed comparing 67 patients who underwent RSA in two groups: the EM group began a rehabilitation program immediately after surgery, and the DM group began after 4-6 weeks postoperatively. Preoperative and postoperative Constant shoulder score, American Shoulder and Elbow Surgeons (ASES) score, Penn Shoulder Score (PSS), shoulder satisfaction value as well as demographics, opioid usage, range of motion, 90-day complications, reoperation, and readmission rates were collected. Statistical analysis was performed between each group’s outcomes and for the change (delta) of outcomes from baseline to each postoperative time point.ResultsAt 6 weeks postoperatively, the EM group showed significantly higher scores in ASES function, PSS, active forward elevation, passive forward elevation, active external rotation, and passive external rotation. There was significantly lower narcotic usage in the EM group at 6 weeks. At 3 months postoperatively, the EM group showed a lower numerical rating score for pain and higher active forward elevation. At the last follow-up, there was a lower numerical rating score and greater degree for all ranges of motion for the EM group than that for the DM group. A greater change from baseline to 6 weeks postoperatively was seen for the EM group in shoulder satisfaction value, ASES function, PSS, active forward elevation, passive forward elevation, and internal rotation. From baseline to 3 months postoperatively, ASES pain, PSS, active forward elevation, and passive forward elevation improved significantly in the EM group. At the final follow-up, active forward elevation increased greater for the EM group. There were 3 complications (2 hematomas and 1 infection) in the DM group and none in the EM group.ConclusionThe EM rehabilitation protocol after RSA achieved lower opioid usage at 6 weeks, lower pain scores at each time point, and better range of motion during the final follow-up than DM protocol. There were no complications reported with the EM protocol, suggesting it is a safe alternative for postoperative recovery. Orthopedic surgeons should consider the EM rehabilitation protocol after RSA.     

Correlation of patient-reported outcome measurement information system-upper extremity with American Shoulder and Elbow Surgeons scores in shoulder arthroplasty

BackgroundMultiple outcome measure surveys are available to assess preoperative and postoperative outcomes for a variety of orthopedic procedures such as shoulder arthroplasty. Although legacy instruments such as American Shoulder and Elbow Surgeons (ASES) remain popular, there remains no singular gold standard survey instrument for shoulder arthroplasty patients, and alternative instruments have been developed to better capture clinical outcomes. The goal of this study is to compare the efficacy of Patient-Reported Outcome Measurement Information System-Upper Extremity (PROMIS-UE) using computer adaptive technology with ASES scores both preoperatively and postoperatively in shoulder arthroplasty patients. Our hypothesis is that there would be a strong correlation between PROMIS-UE and ASES scores both preoperatively and postoperatively in total shoulder arthroplasty patients.MethodsPatients who underwent total shoulder arthroplasty and agreed to complete baseline and 1-year follow-up of ASES and PROMIS-UE scores were included. Patients also completed PROMIS-Physical Function (PROMIS-PF) as a third point comparison. Responses to these instruments were statistically analyzed and compared using Pearson correlation coefficients. Floor and ceiling effects were then calculated.ResultsNinety patients were included in this study, all of whom completed the PROMIS-UE, PROMIS-PF, and ASES surveys both preoperatively and 12 months postoperatively. The mean age in this cohort was 68.9 years (standard deviation [SD] 8.4 years, range 39-89). The mean preoperative and postoperative PROMIS-UE scores were 26.6 (SD 6.7, range 14.7-44.6) and 41.8 (SD 10.3, range 20.2-56.4) respectively. Mean preoperative and postoperative PROMIS-PF computer adaptive technology scores were 35.7 (SD 9.54, range 34.2-64) and 44.3 (SD 9.22, range 23.5-73.3) respectively. The mean ASES score was 37.5 preoperatively (SD 18.0, range 5-99.5) and 77.5 postoperatively (SD 20.5, range 25.0-100.0). PROMIS-UE demonstrated a moderate correlation with ASES preoperatively and a strong correlation postoperatively (r = 0.52, confidence interval 0.27-0.60; r = 0.70, confidence interval 0.55-0.78, respectively). PROMIS-UE demonstrated a minor floor effect preoperatively (7.8%) but significant ceiling effect postoperatively (24.4%) and ASES demonstrated a mild ceiling effect at final follow-up (8.9%). There were otherwise no other floor or ceiling effects at all other time points across each survey.ConclusionPROMIS-UE correlates well with ASES at both baseline and 1-year postoperation for patients undergoing total shoulder replacement. In addition, the change seen between both baseline and 1 year outcome scores for both scoring systems also correlates strongly, suggesting that PROMIS-UE may be a suitable alternative to ASES for this patient population.     

Is frozen shoulder completely resolved at 2 years after the onset of disease?

BackgroundAlthough numerous studies have suggested that frozen shoulder (FS) is a self-limiting disease with most patients recovering within 2 years, its long-term outcome is still controversial. The aims of this study were to evaluate the clinical outcomes after conservative treatment for FS and to determine the predictors of its clinical outcome.MethodsThis study included 234 shoulders of 215 patients who received conservative treatment for FS. The mean follow-up period was 41.8 months (range 27–117 months). Initial evaluation included demographics, detailed medical history, and clinical assessments of shoulder status. Questionnaires, which included the Visual Analogue Scale (VAS) pain score, American Shoulder and Elbow Surgeons (ASES) score, Subjective Shoulder Value (SSV) and satisfaction grading for the current shoulder status were assessed at the final follow-up.ResultsThe mean VAS pain score, ASES score, and SSV significantly improved from 6.7, 37.0, and 40.1% at the time of initial evaluation to 1.5, 87.6, and 85.0% at the final follow-up evaluation (all p < 0.001). According to satisfaction grading, the shoulder status at the final follow-up was very satisfied in 101 shoulders (43.2%), satisfied in 68 (29.1%), fair in 37 (15.8%), unsatisfied in 20 (8.5%), and very unsatisfied in 8 (3.4%). Univariate analysis revealed that gender, diabetes, simultaneous bilateral involvement, overall bilateral involvement and duration of symptoms were associated with clinical outcomes at the final follow-up. Multivariate analysis revealed that duration of symptoms (p = 0.002) was an independent risk factor for unsuccessful outcome.ConclusionsAt the mean follow-up period of 41.8 months, 72.3% of patients revealed subjective satisfaction for the current shoulder status. Duration of symptoms was an independent risk factor for poor prognosis.     

Reverse shoulder arthroplasty for proximal humeral fractures: better clinical midterm outcome after primary reverse arthroplasty versus secondary reverse arthroplasty after failed ORIF in the elderly

BackgroundJoint replacement surgery as a treatment for complex proximal humeral fractures is an established option, especially in the elderly. In light of the increased attention to reverse total shoulder arthroplasty (rTSA), this study has analyzed the outcomes of patients with primary reverse arthroplasty and after secondary reverse arthroplasty for failed osteosynthesis.MethodsWe retrospectively reviewed 57 patients with an average age of 76 years (min. 55; max. 94; SD 7) from 2010 and 2015 who underwent primary rTSA (30) and secondary rTSA after the failure of plate osteosynthesis (27) after proximal humeral fractures. The functional outcome of the operated shoulder was evaluated by clinical scores (Constant-Score, ASES, DASH and Oxford), range of motion (RoM), pain and activity level.ResultsPrimary rTSA had a significantly better functional outcome, mean-follow-up 37.3 months, measured by Constant-Score (57.13 vs 45.78 points; p= .015) compared to secondary RTSA, mean follow-up 42.1 months. A significantly better active abduction (P= .002), forward flexion (P = .003) and internal rotation (P = .037) was observed in the primary rTSA group, especially in the follow-up > 35 months.ConclusionReverse shoulder arthroplasty is an effective treatment for proximal humeral fractures as primary or revision surgery. The reliable clinical outcome especially in the follow-up to 40 months after primary reverse arthroplasty may suggest to prefer rTSA for complex humeral fractures in the elderly.Level of evidenceLevel III; Retrospective comparative study     

Is there a threshold of preoperative function that predicts failure to achieve clinically significant outcomes after total and reverse shoulder arthroplasty?

BackgroundThe purpose of this study was to determine if there is a threshold of preoperative function that is predictive of postoperative outcomes and the likelihood of achieving clinically significant outcomes following shoulder arthroplasty (SA).MethodsWe retrospectively identified patients who underwent a primary SA at our institution. Patients with preoperative and postoperative American Shoulder and Elbow Surgeons scores (ASES) were included in our analysis. A receiver operating characteristic (ROC) analysis was utilized to reach a preoperative ASES threshold correlated with achievement of the following clinically significant outcomes: minimal clinically important difference (MCID), substantial clinical benefit (SCB), and patient acceptable symptom state. This analysis was performed for our entire SA cohort and subanalyzed for total shoulder arthroplasty (TSA) and reverse shoulder arthroplasty (RSA). Fischer exact tests were used to analyze categorical data while continuous data were analyzed using t-tests. For the ROC, area under the curve (AUC) was calculated, along with bootstrap 95% confidence intervals and P values, with <0.05 as significant.ResultsA total of 516 patients were included (164 TSA and 352 RSA). ROC analysis yielded a preoperative ASES above 54 as predictive of failure to achieve MCID for all SAs (AUC, 0.77; P < .001), above 49 for TSA (AUC 0.74, P < .001), and above 56 for RSA (AUC 0.79, P < .001). Patients with preoperative ASES scores above 54 were significantly less likely to achieve MCID (odds ratio 5.1, P < .001) and SCB (odds ratio 7.2, P < .001); however, they had higher postoperative ASES scores (84 vs. 78, P < .001). A preoperative ASES score of 73 corresponded to a 50% chance of achieving MCID. ROC analysis also yielded a preoperative ASES score above 51 as predictive of failure to achieve SCB for all SA, TSA, and RSA (AUC: 0.79, 0.78, and 0.80, respectively, all P < .001). A preop ASES score of 52 corresponded to a 50% chance of achieving SCB.ConclusionPreoperative ASES scores above 49-56 are predictive of failure to achieve MCID and SCB following TSA and RSA. Although patients above these preoperative thresholds achieve higher absolute ASES scores at the final follow-up, they experienced less relative improvement from baseline. This will help surgeons counsel individual patients about appropriate expectations after arthroplasty.     

Is distal peripheral neuropathy common after shoulder arthroplasty?

BackgroundNeurologic injury after shoulder arthroplasty has been reported in previous outcome studies. However, the incidence and development of distal peripheral neuropathy (DPN) after shoulder arthroplasty requiring surgical intervention has not been previously described.HypothesisThe authors report on the incidence of DPN requiring surgical intervention following shoulder arthroplasty.MethodsA retrospective review was conducted of a prospectively collected shoulder arthroplasty registry at a single institution from a single surgeon from April 2006 to April 2017. Patients were included in the series if they had primary surgical intervention for ipsilateral or contralateral peripheral neuropathy following primary shoulder arthroplasty. Patients were excluded if they had known peripheral neuropathy or cervical radiculopathy prior to surgery. DPN was defined for the study as symptoms or diagnostic testing consistent with cubital tunnel or carpal tunnel syndrome. Demographics, patient reported outcome measures, and preoperative and postoperative shoulder range of motion were collected and analyzed with paired t-tests and multivariate regression models.ResultsOne thousand three hundred eighty-seven total shoulder arthroplasties were performed in this period. During the study period, 16 patients (1.2%) underwent surgery for ipsilateral DPN while 6 patients (0.4%) underwent surgery for contralateral DPN. ASES scores, SANE scores, and shoulder flexion improved significantly from pre- to postoperatively for both groups. There was no significant difference in postoperative scores between groups. Finally, the multivariate linear analysis did not yield any statistically significant regression equations for ASES, SANE VAS, AFE, and AER.ConclusionShoulder arthroplasty resulted in increased ipsilateral distal peripheral neuropathy. Emergence of DPN symptoms following arthroplasty is multifactorial and may be related to a form of the double crush syndrome.Level of evidenceLevel IV; Case series