Performance of ChatGPT on NASS Clinical Guidelines for the Diagnosis and Treatment of Low Back Pain: A Comparison Study

STUDY DESIGN

Comparative analysis.

OBJECTIVE

To evaluate Chat Generative Pre-trained Transformer (ChatGPT's) ability to predict appropriate clinical recommendations based on the most recent clinical guidelines for the diagnosis and treatment of low back pain.

BACKGROUND

Low back pain is a very common and often debilitating condition that affects many people globally. ChatGPT is an artificial intelligence model that may be able to generate recommendations for low back pain.

MATERIALS AND METHODS

Using the North American Spine Society Evidence-Based Clinical Guidelines as the gold standard, 82 clinical questions relating to low back pain were entered into ChatGPT (GPT-3.5) independently. For each question, we recorded ChatGPT's answer, then used a point-answer system-the point being the guideline recommendation and the answer being ChatGPT's response-and asked ChatGPT if the point was mentioned in the answer to assess for accuracy. This response accuracy was repeated with one caveat-a prior prompt is given in ChatGPT to answer as an experienced orthopedic surgeon-for each question by guideline category. A two-sample proportion z test was used to assess any differences between the preprompt and postprompt scenarios with alpha=0.05.

RESULTS

ChatGPT's response was accurate 65% (72% postprompt, P =0.41) for guidelines with clinical recommendations, 46% (58% postprompt, P =0.11) for guidelines with insufficient or conflicting data, and 49% (16% postprompt, P =0.003*) for guidelines with no adequate study to address the clinical question. For guidelines with insufficient or conflicting data, 44% (25% postprompt, P =0.01*) of ChatGPT responses wrongly suggested that sufficient evidence existed.

CONCLUSION

ChatGPT was able to produce a sufficient clinical guideline recommendation for low back pain, with overall improvements if initially prompted. However, it tended to wrongly suggest evidence and often failed to mention, especially postprompt, when there is not enough evidence to adequately give an accurate recommendation.

Anatomic TSA for B2 glenoids treated with asymmetric reaming has equivalent survivorship and patient outcomes compared to type A glenoids at mean 9 years follow-up

BACKGROUND

Walch B2 glenoids can present a challenge to shoulder arthroplasty surgeons. Short-term studies have demonstrated that corrective reaming to 10° of retroversion in anatomic total shoulder arthroplasty (aTSA) can result in good outcomes, however there is little data reporting the long-term outcomes in this cohort. B2 glenoids treated with high-side reaming present a theoretical risk of early glenoid component failure as one may ream into the subchondral bone. This study aimed to demonstrate that 1) B2 glenoids treated with corrective reaming have durable results and 2) offer similar results to Walch A1/2 in long-term follow-up.

METHODS

Patients who underwent aTSA by a single surgeon (E.L.F.) were identified from a shoulder arthroplasty registry. Inclusion criteria included Walch A1, A2 or B2 glenoid, a diagnosis of primary shoulder osteoarthritis, and a minimum radiographic and clinical follow-up of 5 years. 43 patients with B2 glenoids were compared to a cohort of 42 patients with A1 or A2 glenoids. Preoperative computed tomography (CT) and radiographs were utilized to assess deformity, glenoid version, and posterior subluxation of the humeral head. Postoperatively, patients were assessed with radiographs and patient-reported outcome measures including American Shoulder and Elbow Surgeons (ASES) score, Simple Shoulder Test (SST) score, and Visual Analog Scale (VAS).

RESULTS

85 shoulders (82 patients, 42 B2 and 43 A1/A2 glenoids) with an average follow-up of 9.4 years were included. In the B2 cohort, the average retroversion was 21.1° and posterior subluxation was 69.4% compared with 10.6°(p<0.001) and 57.5% (p<0.001) in the A1 or A2 cohort. The cohort demographics were similar except for male sex (B2 69.8% vs A1 or A2 37.2%, p=0.008). There was no difference between the cohorts in their improvement in ASES (p=0.807), SST (p=0.586), or VAS (p=0.930) scores. There was no difference in lateral humeral offset (p=0.889) or acromial humeral interval (p=0.468) between initial postoperative and final follow-up visits. Survivorship for B2 glenoids was 97.6%, 94.1% and 73.3% at 5, 10 and 15 years, respectively compared to 97.6%, 91.9% and 83.5% in type A glenoids. The revision rate was similar between the two groups (p=0.432). Lazarus score (p=0.682) and rates of humeral radiolucency (p=0.366) and humeral osteolysis (p=0.194) were similar between the two cohorts at final follow-up.

CONCLUSION

Asymmetric reaming of patients with B2 glenoids is a reliable method of glenoid preparation with excellent mid- to long-term clinical results, patient reported outcomes, and low revision rates similar to their A1 and A2 counterparts.

ChatGPT versus NASS clinical guidelines for degenerative spondylolisthesis: a comparative analysis

BACKGROUND CONTEXT

Clinical guidelines, developed in concordance with the literature, are often used to guide surgeons' clinical decision making. Recent advancements of large language models and artificial intelligence (AI) in the medical field come with exciting potential. OpenAI's generative AI model, known as ChatGPT, can quickly synthesize information and generate responses grounded in medical literature, which may prove to be a useful tool in clinical decision-making for spine care. The current literature has yet to investigate the ability of ChatGPT to assist clinical decision making with regard to degenerative spondylolisthesis.

PURPOSE

The study aimed to compare ChatGPT's concordance with the recommendations set forth by The North American Spine Society (NASS) Clinical Guideline for the Diagnosis and Treatment of Degenerative Spondylolisthesis and assess ChatGPT's accuracy within the context of the most recent literature.

METHODS

ChatGPT-3.5 and 4.0 was prompted with questions from the NASS Clinical Guideline for the Diagnosis and Treatment of Degenerative Spondylolisthesis and graded its recommendations as "concordant" or "nonconcordant" relative to those put forth by NASS. A response was considered "concordant" when ChatGPT generated a recommendation that accurately reproduced all major points made in the NASS recommendation. Any responses with a grading of "nonconcordant" were further stratified into two subcategories: "Insufficient" or "Over-conclusive," to provide further insight into grading rationale. Responses between GPT-3.5 and 4.0 were compared using Chi-squared tests.

RESULTS

ChatGPT-3.5 answered 13 of NASS's 28 total clinical questions in concordance with NASS's guidelines (46.4%). Categorical breakdown is as follows: Definitions and Natural History (1/1, 100%), Diagnosis and Imaging (1/4, 25%), Outcome Measures for Medical Intervention and Surgical Treatment (0/1, 0%), Medical and Interventional Treatment (4/6, 66.7%), Surgical Treatment (7/14, 50%), and Value of Spine Care (0/2, 0%). When NASS indicated there was sufficient evidence to offer a clear recommendation, ChatGPT-3.5 generated a concordant response 66.7% of the time (6/9). However, ChatGPT-3.5's concordance dropped to 36.8% when asked clinical questions that NASS did not provide a clear recommendation on (7/19). A further breakdown of ChatGPT-3.5's nonconcordance with the guidelines revealed that a vast majority of its inaccurate recommendations were due to them being "over-conclusive" (12/15, 80%), rather than "insufficient" (3/15, 20%). ChatGPT-4.0 answered 19 (67.9%) of the 28 total questions in concordance with NASS guidelines (P = 0.177). When NASS indicated there was sufficient evidence to offer a clear recommendation, ChatGPT-4.0 generated a concordant response 66.7% of the time (6/9). ChatGPT-4.0's concordance held up at 68.4% when asked clinical questions that NASS did not provide a clear recommendation on (13/19, P = 0.104).

CONCLUSIONS

This study sheds light on the duality of LLM applications within clinical settings: one of accuracy and utility in some contexts versus inaccuracy and risk in others. ChatGPT was concordant for most clinical questions NASS offered recommendations for. However, for questions NASS did not offer best practices, ChatGPT generated answers that were either too general or inconsistent with the literature, and even fabricated data/citations. Thus, clinicians should exercise extreme caution when attempting to consult ChatGPT for clinical recommendations, taking care to ensure its reliability within the context of recent literature.

Performance of a Large Language Model in the Generation of Clinical Guidelines for Antibiotic Prophylaxis in Spine Surgery

OBJECTIVE

Large language models, such as chat generative pre-trained transformer (ChatGPT), have great potential for streamlining medical processes and assisting physicians in clinical decision-making. This study aimed to assess the potential of ChatGPT's 2 models (GPT-3.5 and GPT-4.0) to support clinical decision-making by comparing its responses for antibiotic prophylaxis in spine surgery to accepted clinical guidelines.

METHODS

ChatGPT models were prompted with questions from the North American Spine Society (NASS) Evidence-based Clinical Guidelines for Multidisciplinary Spine Care for Antibiotic Prophylaxis in Spine Surgery (2013). Its responses were then compared and assessed for accuracy.

RESULTS

Of the 16 NASS guideline questions concerning antibiotic prophylaxis, 10 responses (62.5%) were accurate in ChatGPT's GPT-3.5 model and 13 (81%) were accurate in GPT-4.0. Twenty-five percent of GPT-3.5 answers were deemed as overly confident while 62.5% of GPT-4.0 answers directly used the NASS guideline as evidence for its response.

CONCLUSION

ChatGPT demonstrated an impressive ability to accurately answer clinical questions. GPT-3.5 model's performance was limited by its tendency to give overly confident responses and its inability to identify the most significant elements in its responses. GPT-4.0 model's responses had higher accuracy and cited the NASS guideline as direct evidence many times. While GPT-4.0 is still far from perfect, it has shown an exceptional ability to extract the most relevant research available compared to GPT-3.5. Thus, while ChatGPT has shown far-reaching potential, scrutiny should still be exercised regarding its clinical use at this time.

Surgical Treatment of Displaced Proximal Humerus Fractures Is Associated With Decreased 1-Year Mortality in Patients Aged 65 and Older: A Retrospective Study of Medicare Patients

BACKGROUND

Proximal humerus fracture (PHF) is a risk factor for 1-year mortality. This study aimed to determine if surgery is associated with lower mortality compared to nonoperative treatment following PHF in older patients.

METHODS

This retrospective cohort study used the Medicare Limited Data Set. Patients aged 65+ with a PHF diagnosis in 2017-2020 were included. Treatment was classified as nonoperative, open reduction internal fixation (ORIF), total shoulder arthroplasty (TSA), or hemiarthroplasty. Multivariable logistic regression models examined (a) predictors of treatment type and (b) the association of treatment type with 1-year mortality, adjusting for patient demographics, comorbidities, frailty, and fracture severity among other variables. A subgroup analysis examined how the relationship between treatment type and 1-year mortality varied based on fracture severity. Adjusted odds ratios (aOR) and 95% confidence intervals (CI) are reported.

RESULTS

In total, 49,072 patients were included (mean age = 76.6 years, 82.3% female). Most were treated nonoperatively (77.5%), 10.9% underwent ORIF, 10.6% underwent TSA, and 1.0% underwent hemiarthroplasty. Examples of factors associated with receipt of operative (versus nonoperative treatment) included worse fracture severity and lower frailty. The 1-year mortality rate after the initial PHF diagnosis was 11.0% for the nonoperative group, 4.0% for ORIF, 5.2% for TSA, and 6.0% for hemiarthroplasty. Compared to nonoperative treatment, ORIF (aOR 0.55; 95% CI [0.47, 0.64]; P<.001) and TSA (aOR 0.59; 95% CI [0.50, 0.68]; P<.001) were associated with decreased odds of 1-year mortality. In the subgroup analysis, ORIF and TSA were associated with a lower 1-year mortality risk for 2-part and 3-/4-part fractures.

CONCLUSIONS

Compared to nonoperative treatment, surgery (particularly TSA and ORIF) was associated with a decreased odds of 1-year mortality. This relationship remained significant for 2-part and 3-/4-part fractures after stratifying by fracture severity.

Mid-term outcomes following reverse total shoulder arthroplasty

PURPOSE

The utilization of reverse total shoulder arthroplasty now exceeds the incidence of anatomic shoulder arthroplasty. Previous mid-to-long-term studies on rTSA have reported a decrease in shoulder function as follow-up increased. The purpose of this study was to provide data on mid-term outcomes and implant survival in a series focusing on reverse total shoulder arthroplasty.

MATERIALS AND METHODS

Demographic information such as age at surgery, revision surgery status, BMI, and smoking status were recorded. The clinical endpoints measured in this study were range of motion scores (forward elevation, external rotation, internal rotation) and patient reported outcomes (VAS, ASES, SST). Radiographic variables captured included preoperative glenoid morphology, humeral lucency, and glenoid loosening.

RESULTS

Fifty-six shoulders were included in this study. The overall mean age at surgery was 72.5 ± 7.2 years with an average follow-up time of 6.8 ± 3.5 years. The mean BMI was 28.1 ± 5.5. All measurements of range of motion saw significant and sustained improvements. Overall, forward elevation improved from 82° preoperatively to 133° postoperatively (p < 0.01). External rotation improved from 23° preoperatively to 36° (p < 0.01), while internal rotation improved from L3 to L1 (p = 0.05). ASES scores improved from 31 preoperatively to 70 postoperatively (p < 0.01). SST scores improved from 2 preoperatively to 7 (p < 0.01). VAS pain index scores improved from 6 to 2 following surgery (p < 0.01). Postoperative scapular notching was seen in 18 patients at final follow-up. Glenoid loosening was seen in 3 shoulders. Humeral loosening was seen in 18 shoulders. Tuberosity resorption was seen in 8 shoulders. The 5 year survival estimate was 98%, and the 10 year survival estimate was 83%.

CONCLUSION

In this series, we found that rTSA provides mid-term improvements in range of motion in patients while reducing pain levels. When considered together, this demonstrates that most patients undergoing rTSA can have excellent use of their shoulder from age at surgery to end-of-life.

Cemented vs. press-fit humeral stems: a matched cohort analysis at a mean follow-up of 10 years

BACKGROUND

Although cementation of humeral stems has long been considered the gold standard for anatomic shoulder arthroplasty (aTSA), cementless, or press-fit, fixation offers a relatively cheaper and less demanding alternative, particularly in the setting of a revision procedure. However, this approach has been accompanied by concerns of implant loosening and high rates of radiolucency. In the present study, we performed a propensity-matched comparison of clinical and patient-reported outcomes between cemented and cementless fixation techniques for aTSA. We hypothesized that cemented fixation of the humeral component would have significantly better implant survival while providing comparable functional outcomes at final follow-up.

METHODS

This study was a retrospective comparison of 50 shoulders undergoing aTSA: 25 using cemented humeral fixation vs. 25 using press-fit humeral fixation. Patients in the 2 groups were propensity matched according to age, sex, and preoperative American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form (ASES) score. Primary outcome measures included range of motion (ROM) (forward elevation, external rotation, internal rotation), patient-reported outcomes (ASES, Simple Shoulder Test [SST], visual analog scale [VAS]), and implant survival.

RESULTS

At baseline, the 2 fixation groups were similar in regard to age, sex, body mass index, preoperative ASES score, and surgical indication. Mean follow-up was 11.7 ± 4.95 years in the cemented cohort and 9.13 ± 3.77 years in the press-fit cohort (P = .045). Both groups demonstrated significant improvements postoperatively in all included ROM and patient-reported outcomes. However, press-fit patients reported significantly better VAS, ASES, and SST scores. Mean VAS pain score was 1.1 ± 1.8 in press-fit patients and 3.2 ± 3.0 in cemented patients (P = .005). The mean ASES score was 87.7 ± 12.4 in press-fit patients and 69.5 ± 22.7 in cemented patients (P = .002). Lastly, the mean SST score was 9.8 ± 3.1 in press-fit patients and 7.7 ± 3.7 in cemented patients (P = .040). Both fixation techniques provided lasting implant survivorship with only a single revision operation in each of the cohorts.

CONCLUSION

Herein, we provide a propensity-matched, long-term comparison of patients receiving anatomic shoulder arthroplasty stratified according to humeral stem fixation technique. The results of this analysis illustrate that both types of humeral fixation techniques yield durable and significant improvements in shoulder function with similar rates of survival at 10 years of follow-up.

Superior migration of the humeral head does not significantly affect outcomes at an average of 11 years after total shoulder arthroplasty

BACKGROUND

Superior migration of the humeral head has been linked with rotator cuff dysfunction and glenoid loosening after total shoulder arthroplasty (TSA). We aimed to determine if superior migration was associated with poor shoulder function following anatomic TSA at long-term follow-up.

METHODS

In this retrospective cohort study, we reviewed patients undergoing TSA by a single surgeon at an urban, academic institution. To study the effect of superior migration on TSA outcomes, we stratified the cohort by ≥ and <7 mm of acromiohumeral interval (AHI) and compared range of motion and patient reported outcomes (PROs). Clinical variables included preoperative and postoperative forward elevation (FE), internal rotation, external rotation, visual analog scale, American Shoulder and Elbow Surgeons shoulder score, and Simple Shoulder Text score. Radiographic variables included immediate postoperative and long-term follow-up AHI, lateral humeral offset, and glenoid loosening scores.

RESULTS

After applying exclusion criteria, 121 TSAs were included. The mean age was 63.9 ± 9.5 years, and 66 surgeries (55%) were in male patients. The mean follow-up for our cohort was 11.2 years (range, 5-26 years). Nine shoulders underwent revision surgery. All range of motion and PROs improved significantly from preoperative to the most recent postoperative follow-up. The mean AHI immediately following surgery was 10.9 ± 4.1 mm, while the mean AHI at most recent follow-up was 8.4 ± 3.5 mm. Glenoid loosening was observed in 29 (23.8%) shoulders at the most recent follow-up appointment. Although AHI correlated weakly with FE (r = 0.252; P = .006), we did not observe a clear threshold of migration which led to degraded function. Importantly, glenoid loosening was not related to AHI at long-term follow-up (P = .631). None of FE, internal rotation, external rotation, visual analog scale, American Shoulder and Elbow Surgeons shoulder score, Simple Shoulder Text, or revisions were significantly different between patients with ≥ and <7 mm of AHI.

CONCLUSION

Our results suggest that anatomic TSA provides durable improvements to pain, function, and PROs despite changes to the AHI.

Did the COVID-19 Pandemic Coincide With an Increase in Osteonecrosis as Indication for Total Hip Arthroplasty in Older Patients?

BACKGROUND

Osteonecrosis of the femoral head is a common indication for total hip arthroplasty (THA). It is unclear to what extent the COVID-19 pandemic has impacted its incidence. Theoretically, the combination of microvascular thromboses and corticosteroid use in patients who have COVID-19 may increase the risk of osteonecrosis. We aimed to (1) assess recent osteonecrosis trends and (2) investigate if a history of COVID-19 diagnosis is associated with osteonecrosis.

METHODS

This retrospective cohort study utilized a large national database between 2016 and 2021. Osteonecrosis incidence in 2016 to 2019 was compared to 2020 to 2021. Secondly, utilizing a cohort from April 2020 through December 2021, we investigated whether a prior COVID-19 diagnosis was associated with osteonecrosis. For both comparisons, Chi-square tests were applied.

RESULTS

Among 1,127,796 THAs performed between 2016 and 2021, we found an osteonecrosis incidence of 1.6% (n = 5,812) in 2020 to 2021 compared to 1.4% (n = 10,974) in 2016 to 2019; P < .0001. Furthermore, using April 2020 to December 2021 data from 248,183 THAs, we found that osteonecrosis was more common among those who had a history of COVID-19 (3.9%; 130 of 3,313) compared to patients who had no COVID-19 history (3.0%; 7,266 of 244,870); P = .001).

CONCLUSION

Osteonecrosis incidence was higher in 2020 to 2021 compared to previous years and a previous COVID-19 diagnosis was associated with a greater likelihood of osteonecrosis. These findings suggest a role of the COVID-19 pandemic on an increased osteonecrosis incidence. Continued monitoring is necessary to fully understand the impact of the COVID-19 pandemic on THA care and outcomes.

Improved functional, radiographic and patient-reported outcomes at midterm follow-up for shoulder arthroplasty patients 75 years and older

Background

Current discussion centers around the appropriateness of shoulder arthroplasty in elderly patients, and whether anatomic and reverse total shoulder arthroplasty yield acceptable results in this population. The purpose of this study was to examine midterm outcomes in patients 75 years and older who underwent either procedure.

Methods

A retrospective review was performed on patients who underwent anatomic total shoulder arthroplasty (aTSA) or reverse total shoulder arthroplasty (rTSA) between 2000 and 2018. Inclusion criteria was patient age ≥75 years at time of surgery and ≥1 years postoperative follow-up. Primary outcomes were differences in Simple Shoulder Test (SST), American Shoulder and Elbow Surgeons (ASES) score, Visual Analog Scale (VAS) pain score, and range of motion. A secondary outcome was revision surgery.

Results

22 patients who underwent aTSA and 17 patients who underwent rTSA were included. Mean age at surgery was 79 and 80 in the aTSA and rTSA groups, respectively. Mean postoperative follow-up was 6.6 years across both cohorts. Anatomic TSA patients experienced improvements in VAS (7 preop vs. 1 postop; p < 0.001), ASES (16 vs. 75; p < 0.001), and SST (2 vs. 8; p < 0.001) scores. Reverse TSA patients also experienced improvements in VAS (7 vs. 1; p < 0.001), ASES (29 vs. 74; p < 0.001), and SST (1 vs. 7; p < 0.001) scores. Anatomic TSA patients experienced improved external rotation (17° vs. 53°; p < 0.001), forward elevation (108° vs. 155°; p < 0.001), and internal rotation (L5 vs. T10; p < 0.001). Reverse TSA patients experienced improved forward elevation (52° vs. 126°; p < 0.001). 21 aTSA patients (100%) and 16 rTSA patients (94%) experienced survival free from revision. 100% of aTSA and rTSA patients experienced survival free from loosening.

Conclusion

Both aTSA and rTSA can be performed in the elderly population with acceptable midterm outcomes, suggesting that implant survival and patient satisfaction have the potential to endure through the end of life.

Level of evidence

IV.

Thromboembolic prophylaxis in spine surgery: an analysis of ChatGPT recommendations

BACKGROUND CONTEXT

Venous thromboembolism is a negative outcome of elective spine surgery. However, the use of thromboembolic chemoprophylaxis in this patient population is controversial due to the possible increased risk of epidural hematoma. ChatGPT is an artificial intelligence model which may be able to generate recommendations for thromboembolic prophylaxis in spine surgery.

PURPOSE

To evaluate the accuracy of ChatGPT recommendations for thromboembolic prophylaxis in spine surgery.

STUDY DESIGN/SETTING

Comparative analysis.

PATIENT SAMPLE

None.

OUTCOME MEASURES

Accuracy, over-conclusiveness, supplemental, and incompleteness of ChatGPT responses compared to the North American Spine Society (NASS) clinical guidelines.

METHODS

ChatGPT was prompted with questions from the 2009 NASS clinical guidelines for antithrombotic therapies and evaluated for concordance with the clinical guidelines. ChatGPT-3.5 responses were obtained on March 5, 2023, and ChatGPT-4.0 responses were obtained on April 7, 2023. A ChatGPT response was classified as accurate if it did not contradict the clinical guideline. Three additional categories were created to further evaluate the ChatGPT responses in comparison to the NASS guidelines: over-conclusiveness, supplementary, and incompleteness. ChatGPT was classified as over-conclusive if it made a recommendation where the NASS guideline did not provide one. ChatGPT was classified as supplementary if it included additional relevant information not specified by the NASS guideline. ChatGPT was classified as incomplete if it failed to provide relevant information included in the NASS guideline.

RESULTS

Twelve clinical guidelines were evaluated in total. Compared to the NASS clinical guidelines, ChatGPT-3.5 was accurate in 4 (33%) of its responses while ChatGPT-4.0 was accurate in 11 (92%) responses. ChatGPT-3.5 was over-conclusive in 6 (50%) of its responses while ChatGPT-4.0 was over-conclusive in 1 (8%) response. ChatGPT-3.5 provided supplemental information in 8 (67%) of its responses, and ChatGPT-4.0 provided supplemental information in 11 (92%) responses. Four (33%) responses from ChatGPT-3.5 were incomplete, and 4 (33%) responses from ChatGPT-4.0 were incomplete.

CONCLUSIONS

ChatGPT was able to provide recommendations for thromboembolic prophylaxis with reasonable accuracy. ChatGPT-3.5 tended to cite nonexistent sources and was more likely to give specific recommendations while ChatGPT-4.0 was more conservative in its answers. As ChatGPT is continuously updated, further validation is needed before it can be used as a guideline for clinical practice.

Daily Pain Prediction Using Smartphone Speech Recordings of Patients With Spine Disease

BACKGROUND

Pain evaluation remains largely subjective in neurosurgical practice, but machine learning provides the potential for objective pain assessment tools.

OBJECTIVE

To predict daily pain levels using speech recordings from personal smartphones of a cohort of patients with diagnosed neurological spine disease.

METHODS

Patients with spine disease were enrolled through a general neurosurgical clinic with approval from the institutional ethics committee. At-home pain surveys and speech recordings were administered at regular intervals through the Beiwe smartphone application. Praat audio features were extracted from the speech recordings to be used as input to a K-nearest neighbors (KNN) machine learning model. The pain scores were transformed from a 0 to 10 scale to low and high pain for better discriminative capacity.

RESULTS

A total of 60 patients were enrolled, and 384 observations were used to train and test the prediction model. Using the KNN prediction model, an accuracy of 71% with a positive predictive value of 0.71 was achieved in classifying pain intensity into high and low. The model showed 0.71 precision for high pain and 0.70 precision for low pain. Recall of high pain was 0.74, and recall of low pain was 0.67. The overall F1 score was 0.73.

CONCLUSION

Our study uses a KNN to model the relationship between speech features and pain levels collected from personal smartphones of patients with spine disease. The proposed model is a stepping stone for the development of objective pain assessment in neurosurgery clinical practice.

Evaluating the effects of age on the long-term functional outcomes following anatomic total shoulder arthroplasty

BACKGROUND

In the past decade, the number of anatomic total shoulder arthroplasty (aTSA) procedures has steadily increased. Patients over 65 years of age comprise the vast majority of recipients, and outcomes have been well documented; however, patients are opting for definitive surgical treatment at younger ages.We aim to report on the effects of age on the long-term clinical outcomes following aTSA.

METHODS

Among the patients who underwent TSA, 119 shoulders were retrospectively analyzed. Preoperative and postoperative clinical outcome data were collected. Linear regression analysis (univariate and multivariate) was conducted to evaluate the associations of clinical outcomes with age. Kaplan-Meier curves and Cox regression analyses were performed to evaluate implant survival.

RESULTS

At final follow-up, patients of all ages undergoing aTSA experienced significant and sustained improvements in all primary outcome measures compared with preoperative values. Based on multivariate analysis, age at the time of surgery was a significant predictor of postoperative outcomes. Excellent implant survival was observed over the course of this study, and Cox regression survival analysis indicated age and sex to not be associated with an increased risk of implant failure.

CONCLUSIONS

When controlling for sex and follow-up duration, older age was associated with significantly better patient-reported outcome measures. Despite this difference, we noted no significant effects on range of motion or implant survival. Level of evidence: IV.

Does humeral fixation technique affect long-term outcomes of total shoulder arthroplasty?

BACKGROUND

For anatomic total arthroscopic repair, cementless humeral fixation has recently gained popularity. However, few studies have compared clinical, radiographic, and patient-reported outcomes between cemented and press-fit humeral fixation, and none have performed follow-up for longer than 5 years. In this study, we compared long-term postoperative outcomes in patients receiving a cemented versus press-fit humeral stem anatomic arthroscopic repair.

METHODS

This study retrospectively analyzed 169 shoulders that required primary anatomic total shoulder arthroplasty (aTSA). Shoulders were stratified by humeral stem fixation technique: cementation or press-fit. Data were collected pre- and postoperatively. Primary outcome measures included range of motion, patient reported outcomes, and radiographic measures.

RESULTS

One hundred thirty-eight cemented humeral stems and 31 press-fit stems were included. Significant improvements in range of motion were seen in all aTSA patients with no significant differences between final cemented and press-fit stems (forward elevation: P=0.12, external rotation: P=0.60, and internal rotation: P=0.77). Patient reported outcome metrics also exhibited sustained improvement through final follow-up. However, at final follow-up, the press-fit stem cohort had significantly better overall scores when compared to the cemented cohort (visual analog score: P=0.04, American Shoulder and Elbow Surgeon Score: P<0.01, Simple Shoulder Test score: P=0.03). Humeral radiolucency was noted in two cemented implants and one press-fit implant. No significant differences in implant survival were observed between the two cohorts (P=0.75).

CONCLUSIONS

In this series, we found that irrespective of humeral fixation technique, aTSA significantly improves shoulder function. However, within this cohort, press-fit stems provided significantly better outcomes than cemented stems in terms of patient reported outcome scores. Level of evidence: III.

Diabetes increases risk for readmission and infection after shoulder arthroplasty: A national readmissions study of 113,713 patients

Background

The recent increasing popularity of shoulder arthroplasty has been paralleled by a rise in prevalence of diabetes in the United States. We aimed to evaluate the impact of diabetes status on readmission and short-term complications among patients undergoing shoulder arthroplasty.

Methods

We analyzed the Healthcare Cost and Utilization Project National Readmissions Database (NRD) between the years 2016-2018. Patients were included in the study if they underwent anatomic total shoulder arthroplasty (aTSA) or reverse total shoulder arthroplasty (rTSA) according to ICD-10 procedure codes. Postoperative complications including surgical site/joint infection, dislocation, prosthetic complications, hardware-related complications, non-infectious wound complications, 30-day, and 90-day readmission were collected.

Results

A total of 113,713 shoulder arthroplasty patients were included. 23,749 (20.9%) had a diagnosis of diabetes and 89,964 (79.1%) did not. On multivariate analysis, a diagnosis of diabetes led to an increased risk of 30-day (OR: 1.24; 95% CI: [1.14, 1.34]; p < 0.001) and 90-day (OR: 1.18; 95% CI: [1.12, 1.25]; p < 0.001) readmission, surgical site/joint infection (OR: 1.21; 95% CI: [1.06, 1.38]; p = 0.005), respiratory complication (OR: 1.34; 95% CI: [1.09, 1.64]; p = 0.005), postoperative infection (OR: 1.22; 95% CI [1.07, 1.39]; p = 0.003), and deep vein thrombosis (OR: 1.38; 95% CI: [1.09, 1.74]; p = 0.007).

Conclusions

Our findings suggest that patients with diabetes may be at an increased risk of readmission, infection, respiratory complication, and deep vein thrombosis following shoulder arthroplasty. Shoulder surgeons should consider these potential adverse events when planning postoperative care for patients with diabetes.

Can a Novel Natural Language Processing Model and Artificial Intelligence Automatically Generate Billing Codes From Spine Surgical Operative Notes?

STUDY DESIGN

Retrospective cohort.

OBJECTIVE

Billing and coding-related administrative tasks are a major source of healthcare expenditure in the United States. We aim to show that a second-iteration Natural Language Processing (NLP) machine learning algorithm, XLNet, can automate the generation of CPT codes from operative notes in ACDF, PCDF, and CDA procedures.

METHODS

We collected 922 operative notes from patients who underwent ACDF, PCDF, or CDA from 2015 to 2020 and included CPT codes generated by the billing code department. We trained XLNet, a generalized autoregressive pretraining method, on this dataset and tested its performance by calculating AUROC and AUPRC.

RESULTS

The performance of the model approached human accuracy. Trial 1 (ACDF) achieved an AUROC of .82 (range: .48-.93), an AUPRC of .81 (range: .45-.97), and class-by-class accuracy of 77% (range: 34%-91%); trial 2 (PCDF) achieved an AUROC of .83 (.44-.94), an AUPRC of .70 (.45-.96), and class-by-class accuracy of 71% (42%-93%); trial 3 (ACDF and CDA) achieved an AUROC of .95 (.68-.99), an AUPRC of .91 (.56-.98), and class-by-class accuracy of 87% (63%-99%); trial 4 (ACDF, PCDF, CDA) achieved an AUROC of .95 (.76-.99), an AUPRC of .84 (.49-.99), and class-by-class accuracy of 88% (70%-99%).

CONCLUSIONS

We show that the XLNet model can be successfully applied to orthopedic surgeon's operative notes to generate CPT billing codes. As NLP models as a whole continue to improve, billing can be greatly augmented with artificial intelligence assisted generation of CPT billing codes which will help minimize error and promote standardization in the process.

Robust prediction of nonhome discharge following elective anterior cervical discectomy and fusion using explainable machine learning

PURPOSE

Predict nonhome discharge (NHD) following elective anterior cervical discectomy and fusion (ACDF) using an explainable machine learning model.

METHODS

2227 patients undergoing elective ACDF from 2008 to 2019 were identified from a single institutional database. A machine learning model was trained on preoperative variables, including demographics, comorbidity indices, and levels fused. The validation technique was repeated stratified K-Fold cross validation with the area under the receiver operating curve (AUROC) statistic as the performance metric. Shapley Additive Explanation (SHAP) values were calculated to provide further explainability regarding the model's decision making.

RESULTS

The preoperative model performed with an AUROC of 0.83 ± 0.05. SHAP scores revealed the most pertinent risk factors to be age, medicare insurance, and American Society of Anesthesiology (ASA) score. Interaction analysis demonstrated that female patients over 65 with greater fusion levels were more likely to undergo NHD. Likewise, ASA demonstrated positive interaction effects with female sex, levels fused and BMI.

CONCLUSION

We validated an explainable machine learning model for the prediction of NHD using common preoperative variables. Adding transparency is a key step towards clinical application because it demonstrates that our model's "thinking" aligns with clinical reasoning. Interactive analysis demonstrated that those of age over 65, female sex, higher ASA score, and greater fusion levels were more predisposed to NHD. Age and ASA score were similar in their predictive ability. Machine learning may be used to predict NHD, and can assist surgeons with patient counseling or early discharge planning.

A comparison of pegged and keeled glenoid clinical outcomes at long-term follow-up after total shoulder arthroplasty

Background

The two main glenoid types used in total shoulder arthroplasty (TSA) are the pegged and keeled glenoid designs. We aimed to determine if a pegged glenoid is superior to a keeled glenoid at long-term follow-up as measured by range of motion (ROM), patient reported outcomes (PROs), and radiographic glenoid loosening.

Methods

We retrospectively reviewed all patients undergoing TSA by a single surgeon at an urban, academic hospital. The cohort was stratified into two groups based on glenoid type - one group consisting of keeled implants and a second group consisting of pegged implants. For each patient, forward elevation (FE), internal rotation (IR), external rotation (ER), visual analog scale (VAS), American Shoulder and Elbow Surgeons (ASES) shoulder score, and simple shoulder test (SST) scores were collected preoperatively and at the most recent follow-up visit. Radiographic variables included acromiohumeral interval (AHI) and glenoid loosening.

Results

After applying exclusion criteria, 144 TSAs were included in our study. Of these, 42 (29.2%) had keeled glenoids and 102 (70.8%) had pegged glenoids. Patients with a pegged glenoid implant were older (67.4 vs. 60.7 years; p < 0.001) and had a shorter follow-up time (9.3 vs. 14.4 years; p < 0.001) than patients with a keeled glenoid implant. At the most recent follow-up visit, there were no significant differences among postoperative FE, ER, AHI, or PROs. However, pegged glenoid implants provided significantly more internal rotation (T11 vs. L1; p = 0.010) and were less likely to show evidence of radiographic glenoid loosening (16.7% vs. 42.9%; p=<0.001). Revision rates were not significantly different between the pegged and keeled groups (6.9% vs. 14.3%; p = 0.158).

Conclusion

Although a pegged design correlated with superior internal rotation and less radiographic glenoid loosening, both pegged and keeled glenoid designs offered favorable long-term clinical outcomes following TSA over the long-term.

Shoulder arthroplasty following solid organ transplant: A systematic review and meta-analysis

Introduction

The purpose of this study is to report a systematic review and meta-analysis of solid organ transplant (SOT) patients undergoing shoulder arthroplasty to compare functional and radiographic outcomes, demographics, and complications with non-transplant patients.

Methods

Studies were included if they examined patients undergoing shoulder arthroplasty in the setting of prior solid organ transplantation and included post operative range of motion, patient-reported outcomes, complications, or revisions. Studies were excluded if they were national database analyses or lacked clinical data. Pubmed, MEDLine, Scopus, and Web of Science were queried using relevant search terms in July 2022. Data was pooled, weighted, and a paired t-test and chi-square analysis was performed.

Results

There were 71 SOT and 159 non-SOT shoulders included in the study. The most common indication for surgery was avascular necrosis (n = 26) in the solid organ transplant group and osteoarthritis (n = 60) in the non-SOT group. Forward elevation, external rotation, ASES, and VAS pain scores improved significantly in both cohorts following surgery. There was no significant difference in age at surgery (p-value = 0.20), postoperative forward elevation (p-value = 0.08), postoperative external rotation (0.84), and postoperative ASES scores (p-value = 0.11) between the two cohorts. VAS pain scores were significantly lower in the SOT cohort (p-value<0.01). The risk of death was significantly higher in the SOT group (p-value<0.01). but the rate of overall complications (p = 0.47), surgical complication (p-value = 0.79), or revision surgery (p-value = 1.00) was not significantly different between the two cohorts.

Conclusion

Shoulder arthroplasty is a safe, effective option in patients following solid organ transplant. There is not an increased risk of adverse outcomes, and SOT patients had comparable range of motion and patient-reported outcomes when compared to their non-SOT peers.

Level of evidence

III.

A national analysis of the effect alcohol use disorder has on short-term complications and readmissions following total shoulder arthroplasty

Background

Alcohol use disorder has been associated with broad health consequences that may interfere with healing after total shoulder arthroplasty. The aim of this study was to explore the impact of alcohol use disorder on readmissions and complications following total shoulder arthroplasty.

Methods

We used data from the Healthcare Cost and Utilization Project National Readmissions Database (NRD) from 2016 to 2018. Patients were included based on International Classification of Diseases, 10th Revision (ICD-10) procedure codes for anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (rTSA). Patients with an alcohol use disorder (AUD) were identified using the ICD-10 diagnosis code F10.20. Demographics, complications, and 30-day and 90-day readmission were collected for all patients. A univariate logistic regression was performed to investigate AUD as a factor affecting readmission and complication rates. A multivariate logistic regression model was created to assess the impact of alcohol use disorder on complications and readmission while controlling for demographic factors.

Results

In total, 164,527 patients were included, and 503 (0.3%) patients had a prior diagnosis of AUD. Revision surgery was more common in patients with an alcohol use disorder (8.8% vs. 6.2%; p = 0.022). Postoperative infection (p = 0.026), dislocation (p = 0.025), liver complications (p < 0.01), and 90-day readmission (p < 0.01) were more common in patients with a diagnosed AUD. On multivariate analysis, patients with an AUD were found to be at increased odds for liver complications (OR: 46.8; 95% CI: [32.8, 66.8]; p < 0.01). Comparatively, mean age, length of stay, and over healthcare costs were also higher for patients with an AUD.

Conclusion

Patients with a diagnosis of AUD were more likely to suffer from shoulder dislocation, liver complications, and 90-day readmission, while also being younger and having longer hospital stays. Therefore, surgeons should take caution to anticipate and prevent complications and readmissions following total shoulder arthroplasty in patients with an AUD.

The Effect of Subsidence on Segmental and Global Lordosis at Long-term Follow-up After Anterior Cervical Discectomy and Fusion

OBJECTIVE

Subsidence following anterior cervical discectomy and fusion (ACDF) may lead to disruptions of cervical alignment and lordosis. The purpose of this study was to evaluate the effect of subsidence on segmental, regional, and global lordosis.

METHODS

This was a retrospective cohort study performed between 2016-2021 at a single institution. All measurements were performed using lateral cervical radiographs at the immediate postoperative period and at final follow-up greater than 6 months after surgery. Associations between subsidence and segmental lordosis, total fused lordosis, C2-7 lordosis, and cervical sagittal vertical alignment change were determined using Pearson correlation and multivariate logistic regression analyses.

RESULTS

One hundred thirty-one patients and 244 levels were included in the study. There were 41 one-level fusions, 67 two-level fusions, and 23 three-level fusions. The median follow-up time was 366 days (interquartile range, 239-566 days). Segmental subsidence was significantly negatively associated with segmental lordosis change in the Pearson (r = -0.154, p = 0.016) and multivariate analyses (beta = -3.78; 95% confidence interval, -7.15 to -0.42; p = 0.028) but no associations between segmental or total fused subsidence and any other measures of cervical alignment were observed.

CONCLUSION

We found that subsidence is associated with segmental lordosis loss 6 months following ACDF. Surgeons should minimize subsidence to prevent long-term clinical symptoms associated with poor cervical alignment.

Repertoires of SARS-CoV-2 epitopes targeted by antibodies vary according to severity of COVID-19

Antibodies to SARS-CoV-2 are central to recovery and immunity from COVID-19. However, the relationship between disease severity and the repertoire of antibodies against specific SARS-CoV-2 epitopes an individual develops following exposure remains incompletely understood. Here, we studied seroprevalence of antibodies to specific SARS-CoV-2 and other betacoronavirus antigens in a well-annotated, community sample of convalescent and never-infected individuals obtained in August 2020. One hundred and twenty-four participants were classified into five groups: previously exposed but without evidence of infection, having no known exposure or evidence of infection, seroconverted without symptoms, previously diagnosed with symptomatic COVID-19, and recovered after hospitalization with COVID-19. Prevalence of IgGs specific to the following antigens was compared between the five groups: recombinant SARS-CoV-2 and betacoronavirus spike and nucleocapsid protein domains, peptides from a tiled array of 22-mers corresponding to the entire spike and nucleocapsid proteins, and peptides corresponding to predicted immunogenic regions from other proteins of SARS-CoV-2. Antibody abundance generally correlated positively with severity of prior illness. A number of specific immunogenic peptides and some that may be associated with milder illness or protection from symptomatic infection were identified. No convincing association was observed between antibodies to Receptor Binding Domain(s) (RBDs) of less pathogenic betacoronaviruses HKU1 or OC43 and COVID-19 severity. However, apparent cross-reaction with SARS-CoV RBD was evident and some predominantly asymptomatic individuals had antibodies to both MERS-CoV and SARS-CoV RBDs. Findings from this pilot study may inform development of diagnostics, vaccines, and therapeutic antibodies, and provide insight into viral pathogenic mechanisms.

Effects of the COVID-19 pandemic on stroke response times: a systematic review and meta-analysis

OBJECTIVES

COVID-19 presents a risk for delays to stroke treatment. We examined how COVID-19 affected stroke response times.

METHODS

A literature search was conducted to identify articles covering stroke during COVID-19 that included time metrics data pre- and post-pandemic. For each outcome, pooled relative change from baseline and 95% CI were calculated using random-effects models. Heterogeneity was explored through subgroup analyses comparing comprehensive stroke centers (CSCs) to non-CSCs.

RESULTS

38 included studies reported on 6109 patients during COVID-19 and 14 637 patients during the pre-COVID period. Pooled increases of 20.9% (95% CI 5.8% to 36.1%) in last-known-well (LKW) to arrival times, 1.2% (-2.9% to 5.3%) in door-to-imaging (DTI), 0.8% (-2.9% to 4.5%) in door-to-needle (DTN), 2.8% (-5.0% to 10.6%) in door-to-groin (DTG), and 19.7% (11.1% to 28.2%) in door-to-reperfusion (DTR) times were observed during COVID-19. At CSCs, LKW increased by 24.0% (-0.3% to 48.2%), DTI increased by 1.6% (-3.0% to 6.1%), DTN increased by 3.6% (1.2% to 6.0%), DTG increased by 4.6% (-5.9% to 15.1%), and DTR increased by 21.2% (12.3% to 30.1%). At non-CSCs, LKW increased by 12.4% (-1.0% to 25.7%), DTI increased by 0.2% (-2.0% to 2.4%), DTN decreased by -4.6% (-11.9% to 2.7%), DTG decreased by -0.6% (-8.3% to 7.1%), and DTR increased by 0.5% (-31.0% to 32.0%). The increases during COVID-19 in LKW (p=0.01) and DTR (p=0.00) were statistically significant, as was the difference in DTN delays between CSCs and non-CSCs (p=0.04).

CONCLUSIONS

Factors during COVID-19 resulted in significantly delayed LKW and DTR, and mild delays in DTI, DTN, and DTG. CSCs experience more pronounced delays than non-CSCs.

Deep Learning for Adjacent Segment Disease at Preoperative MRI for Cervical Radiculopathy

Background Patients who undergo surgery for cervical radiculopathy are at risk for developing adjacent segment disease (ASD). Identifying patients who will develop ASD remains challenging for clinicians. Purpose To develop and validate a deep learning algorithm capable of predicting ASD by using only preoperative cervical MRI in patients undergoing single-level anterior cervical diskectomy and fusion (ACDF). Materials and Methods In this Health Insurance Portability and Accountability Act-compliant study, retrospective chart review was performed for 1244 patients undergoing single-level ACDF in two tertiary care centers. After application of inclusion and exclusion criteria, 344 patients were included, of whom 60% (n = 208) were used for training and 40% for validation (n = 43) and testing (n = 93). A deep learning-based prediction model with 48 convolutional layers was designed and trained by using preoperative T2-sagittal cervical MRI. To validate model performance, a neuroradiologist and neurosurgeon independently provided ASD predictions for the test set. Validation metrics included accuracy, areas under the curve, and F1 scores. The difference in proportion of wrongful predictions between the model and clinician was statistically tested by using the McNemar test. Results A total of 344 patients (median age, 48 years; interquartile range, 41-58 years; 182 women) were evaluated. The model predicted ASD on the 93 test images with an accuracy of 88 of 93 (95%; 95% CI: 90, 99), sensitivity of 12 of 15 (80%; 95% CI: 60, 100), and specificity of 76 of 78 (97%; 95% CI: 94, 100). The neuroradiologist and neurosurgeon provided predictions with lower accuracy (54 of 93; 58%; 95% CI: 48, 68), sensitivity (nine of 15; 60%; 95% CI: 35, 85), and specificity (45 of 78; 58%; 95% CI: 56, 77) compared with the algorithm. The McNemar test on the contingency table demonstrated that the proportion of wrongful predictions was significantly lower by the model (test statistic, 2.000; P < .001). Conclusion A deep learning algorithm that used only preoperative cervical T2-weighted MRI outperformed clinical experts at predicting adjacent segment disease in patients undergoing surgery for cervical radiculopathy. © RSNA, 2021 An earlier incorrect version appeared online. This article was corrected on September 22, 2021.