Sentiment Analysis of Online Patient-Written Reviews of Vascular Surgeons.

BACKGROUND: Online patient reviews influence a patient's choice of a vascular surgeon. The aim of this study is to examine underlying factors that contribute to positive and negative patient reviews by leveraging sentiment analysis and machine learning methods. METHODS: The Society of Vascular Surgeons publicly accessible member directory was queried and cross-referenced with a popular patient-maintained physician review website, healthgrades.com. Sentiment analysis and machine learning methods were used to analyze several parameters. Demographics (gender, age, and state of practice), star rating (of 5 stars), and written reviews were obtained for corresponding vascular surgeons. A sentiment analysis model was applied to patient-written reviews and validated against the star ratings. Student's t-test or one-way analysis of variance assessed demographic relationships with reviews. Word frequency assessments and multivariable logistic regression analyses were conducted to identify common and determinative components of written reviews. RESULTS: A total of 1,799 vascular surgeons had public profiles with reviews. Female gender of surgeon was associated with lower star ratings (male = 4.19, female = 3.95, P < 0.01) and average sentiment score (male = 0.50, female = 0.40, P < 0.01). Younger physician age was associated with higher star rating (P = 0.02) but not average sentiment score (P = 0.12). In the Best reviews, the most commonly used one-words were Care (N = 999), Caring (N = 767), and Kind (N = 479), while the most commonly used two-word pairs were Saved/Life (N = 189), Feel/Comfortable (N = 106), and Kind/Caring (N = 104). For the Worst reviews, the most commonly used one-words were Pain (N = 254) and Rude (N = 148), while the most commonly used two-word pairs were No/One (N = 27), Waste/Time (N = 25), and Severe/Pain (N = 18). In a multiple logistic regression, satisfactory reviews were associated with words such as Confident (odds ratio [OR] = 8.93), Pain-free (OR = 4.72), Listens (OR = 2.55), and Bedside Manner (OR = 1.70), while unsatisfactory reviews were associated with words such as Rude (OR = 0.01), Arrogant (OR = 0.09), Infection (OR = 0.20), and Wait (OR = 0.48). CONCLUSIONS: Female surgeons received significantly worse reviews and younger surgeons tended to receive better reviews. The positivity and negativity of reviews were largely related to words associated with the patient-doctor experience and pain. Vascular surgeons should focus on these 2 areas to improve patient experiences and their own reviews.

Comparison of machine learning techniques to predict unplanned readmission following total shoulder arthroplasty.

BACKGROUND: Machine learning (ML) techniques have been shown to successfully predict postoperative complications for high-volume orthopedic procedures such as hip and knee arthroplasty and to stratify patients for risk-adjusted bundled payments. The latter has not been done for more heterogeneous, lower-volume procedures such as total shoulder arthroplasty (TSA) with equally limited discussion around strategies to optimize the predictive ability of ML algorithms. The purpose of this study was to (1) assess which of 5 ML algorithms best predicts 30-day readmission, (2) test select ML strategies to optimize the algorithms, and (3) report on which patient variables contribute most to risk prediction in TSA across algorithms. METHODS: We identified 9043 patients in the American College of Surgeons National Surgical Quality Improvement Database who underwent primary TSA between 2011 and 2015. Predictors included demographics, comorbidities, laboratory data, and intraoperative variables. The outcome of interest was 30-day unplanned readmission. Five ML algorithms-support-vector machine (SVM), logistic regression, random forest (RF), an adaptive boosting algorithm, and neural network-were trained on the derivation cohort (2011-2014 TSA patients) to predict 30-day unplanned readmission rates. After training, weights for each respective model were fixed and the classifiers were evaluated on the 2015 TSA cohort to simulate a prospective evaluation. C-statistic and f1 scores were used to assess the performance of each classifier. After evaluation, features were removed independently to assess which features most affected classifier performance. RESULTS: The derivation and validation cohorts comprised 5857 and 3186 primary TSA patients, respectively, with similar demographics, comorbidities, and 30-day unplanned readmission rates (2.9% vs. 2.7%). Of the ML algorithms, SVM performed the worst with a c-statistic of 0.54 and an f1-score of 0.07, whereas the random-forest classifier performed the best with the highest c-statistic of 0.74 and an f1-score of 0.18. In addition, SVM was most sensitive to loss of single features, whereas the performance of RF did not dramatically decrease after loss of single features. Within the trained RF classifier, 5 variables achieved weights >0.5 in descending order: high bilirubin (>1.9 mg/dL), age >65, race, chronic obstructive pulmonary disease, and American Society of Anesthesiologists' scores ≥3. In our validation cohort, we observed a 2.7% readmission rate. From this cohort, using the RF classifier we were then able to identify 436 high-risk patients with a predicted risk score >0.6, of whom 36 were readmitted (readmission rate of 8.2%). CONCLUSION: Predictive analytics algorithms can achieve acceptable prediction of unplanned readmission for TSA with the RF classifier outperforming other common algorithms.

Thoracolumbar corpectomy/spondylectomy for spinal metastasis: a pooled analysis comparing the outcome of seven different surgical approaches.

OBJECTIVE: To compare surgical outcomes between seven different approaches for thoracolumbar corpectomy/spondylectomy in the setting of spinal metastasis. METHODS: A systematic review of literature was performed including articles on corpectomy for thoracolumbar spinal metastasis. Data were extracted and sorted by surgical approach: en bloc spondylectomy (group 1), transpedicular (group 2), costotransversectomy (group 3), mini-open retropleural/retroperitoneal (group 4a), lateral extracavitary approach (group 4b), open transthoracic/transretroperitoneal (group 5), and thoracoscopic (group 6). Comparison of demographics, blood loss, directly procedure related complications, operating time, and postoperative improvement of pain. RESULTS: A total of 63 articles were included comprising data of 774 patients with various primary tumor entities. Mean age was 51.8 years, 54% of patients were female, on average 1.46 levels were treated per patient, and mean follow-up was 1.59 years. The following statistically significant findings were observed: Blood loss was lowest for the mini-open retropleural/retroperitoneal (917 ml), thoracoscopic (1107 ml) and transthoracic approach (1172 ml) versus the posterior approach groups (1633-2261 ml); directly procedure related complications were lowest for mini-open retropleural/retroperitoneal and thoracoscopic approach (0% each) versus 7-15% in the other groups; operating time was lowest in mini-open retropleural/retroperitoneal approach (184 min) versus 300-588 min in the other groups. CONCLUSION: Less invasive approaches (mini-open retropleural/retroperitoneal and thoracoscopic) not only had superior outcome in terms of blood loss and operating time, but also were shown to be safe techniques in cancer patients with low rates of procedure-related complications. These slides can be retrieved under Electronic Supplementary Material.

High-content image informatics of the structural nuclear protein NuMA parses trajectories for stem/progenitor cell lineages and oncogenic transformation.

Stem and progenitor cells that exhibit significant regenerative potential and critical roles in cancer initiation and progression remain difficult to characterize. Cell fates are determined by reciprocal signaling between the cell microenvironment and the nucleus; hence parameters derived from nuclear remodeling are ideal candidates for stem/progenitor cell characterization. Here we applied high-content, single cell analysis of nuclear shape and organization to examine stem and progenitor cells destined to distinct differentiation endpoints, yet undistinguishable by conventional methods. Nuclear descriptors defined through image informatics classified mesenchymal stem cells poised to either adipogenic or osteogenic differentiation, and oligodendrocyte precursors isolated from different regions of the brain and destined to distinct astrocyte subtypes. Nuclear descriptors also revealed early changes in stem cells after chemical oncogenesis, allowing the identification of a class of cancer-mitigating biomaterials. To capture the metrology of nuclear changes, we developed a simple and quantitative "imaging-derived" parsing index, which reflects the dynamic evolution of the high-dimensional space of nuclear organizational features. A comparative analysis of parsing outcomes via either nuclear shape or textural metrics of the nuclear structural protein NuMA indicates the nuclear shape alone is a weak phenotypic predictor. In contrast, variations in the NuMA organization parsed emergent cell phenotypes and discerned emergent stages of stem cell transformation, supporting a prognosticating role for this protein in the outcomes of nuclear functions.

Antecubital Burn Resulting From Antenna Coupling: A Case Report.

CASE: We present the case of a 54-year-old man who underwent elective hip disarticulation complicated by third-degree burn of the left antecubital fossa requiring skin graft. After careful review, it was determined that "antenna coupling" as a result of electrosurgery was the likely cause. We present an experiment demonstrating this phenomenon. CONCLUSION: Antenna coupling is a real but rare cause of intraoperative burns not previously described in the orthopaedic literature. Care should be taken to avoid coiling or running bovie or other electrosurgical device cords with other metallic cords or corded devices.

Does an improvement in cord-level intraoperative neuromonitoring data lead to a reduced risk for postoperative neurologic deficit in spine deformity surgery?

PURPOSE: To determine if an improvement in cord-level intraoperative neuromonitoring (IONM) data following data loss results in a reduced risk for new postoperative motor deficit in pediatric and adult spinal deformity surgery. METHODS: A consecutive series of 1106 patients underwent spine surgery from 2015 to 2023 by a single surgeon. Cord alerts were defined by Somatosensory-Evoked Potentials (SSEP; warning criteria: 10% increase in latency or > 50% loss in amplitude) and Motor-Evoked Potentials (MEP; warning criteria: 75% loss in amplitude without return to acceptable limits after stimulation up 100 V above baseline level). Timing of IONM loss and recovery, interventions, and baseline/postoperative day 1 (POD1) lower extremity motor scores were analyzed. RESULTS: IONM Cord loss was noted in 4.8% (53/11,06) of patients and 34% (18/53) with cord alerts had a POD1 deficit compared to preoperative motor exam. MEP and SSEP loss attributed to 98.1% (52/53) and 39.6% (21/53) of cord alerts, respectively. Abnormal descending neurogenic-evoked potential (DNEP) was seen in 85.7% (12/14) and detected 91.7% (11/12) with POD1 deficit. Abnormal wake-up test (WUT) was seen in 38.5% (5/13) and detected 100% (5/5) with POD1 deficit. Most cord alerts occurred during a three-column osteotomy (N = 23/53, 43%); decompression (N = 12), compression (N = 7), exposure (N = 4), and rod placement (N = 14). Interventions were performed in all 53 patients with cord loss and included removing rods/less correction (N = 11), increasing mean arterial pressure alone (N = 10), and further decompression with three-column osteotomy (N = 9). After intervention, IONM data improved in 45(84.9%) patients (Full improvement: N = 28; Partial improvement: 17). For those with full and partial IONM improvement, the POD1 deficit was 10.7% (3/28) and 41.2% (7/17), respectively. For those without any IONM improvement (15.1%, 8/53), 100% (8/8) had a POD1 deficit, P < 0.001. CONCLUSION: A full or partial improvement in IONM data loss after intraoperative intervention was significantly associated with a lower risk for POD1 deficit with an absolute risk reduction of 89.3% and 58.8%, respectively. All patients without IONM improvement had a POD1 neurologic deficit.

Giant Cell Reparative Granuloma in the Thumb Distal Phalanx: A Case Report.

Giant cell reparative granulomas (GCRG) often affect the bones of the hands and the feet. Treatment of this lesion depends on the exact location and amount of localized bony destruction. Ours is the first case report to discuss the nuances of treating this lesion in the thumb distal phalanx. A 19-year-old male presented with lytic, destructive expansion of his left thumb distal phalanx; imaging was suggestive of an aneurysmal bone cyst. Open biopsy was interpreted as giant cell reparative granuloma. Curettage and bone grafting resulted in complete healing of the distal phalanx with an excellent range of motion and interphalangeal joint stability. GCRG is a rare, benign entity typically presenting as a lytic bone lesion. Despite the initial massive bony destruction, this lesion nevertheless healed with curettage and bone grafting with maintained flexor pollicis longus and extensor pollicis longus function, permitting excellent active motion postoperatively.

How Are Patients Reviewing Spine Surgeons Online? A Sentiment Analysis of Physician Review Website Written Comments.

STUDY DESIGN: A Sentiment Analysis of online reviews of spine surgeons. OBJECTIVES: Physician review websites have significant impact on a patient's provider selection. Written reviews are subjective, but sentiment analysis through machine learning can quantitatively analyze these reviews. This study analyzes online written reviews of spine surgeons and reports biases associated with demographic factors and trends in words utilized. METHODS: Online written and star-reviews of spine surgeons were obtained from healthgrades.com. A sentiment analysis package was used to analyze the written reviews. The relationship of demographic variables to these scores was analyzed with t-tests and word and bigram frequency analyses were performed. Additionally, a multiple regression analysis was performed on key terms. RESULTS: 8357 reviews of 480 surgeons were analyzed. There was a significant difference between the means of sentiment analysis scores and star scores for both gender and age. Younger, male surgeons were rated more highly on average (P < .01). Word frequency analysis indicated that behavioral factors and pain were the main contributing factors to both the best and worst reviewed surgeons. Additionally, several clinically relevant words, when included in a review, affected the odds of a positive review. CONCLUSIONS: The best reviews laud surgeons for their ability to manage pain and for exhibiting positive bedside manner. However, the worst reviews primarily focus on pain and its management, as exhibited by the frequency and multivariate analysis. Pain is a clear contributing factor to reviews, thus emphasizing the importance of establishing proper pain expectations prior to any intervention.

Scoliosis Surgery: A Social Media Analysis of Content, Tone, and Perspective.

INTRODUCTION: The purpose of this study was to analyze posts shared on social media sites, Twitter and Instagram, referencing scoliosis surgery for tone, content, and perspective of the posts. METHODS: Public Twitter and Instagram posts from November 2020 to April 2021 were isolated using the hashtag #ScoliosisSurgery or the words "scoliosis surgery." A total of 5,022 Instagram and 1,414 Twitter posts were collected, of which 500 of each were randomly selected to be analyzed by the authors for the variables previously listed. RESULTS: Of the Instagram posts, 91.8% were associated with an image, and 47.8% were postoperative. 96.9% of the posts had either a positive or neutral tone. 38% delivered a progress update, and 29.9% disseminated education or sought to provide awareness. 48.6% of the posts were from the perspective of the patient. Of the Twitter posts, 60.1% contained only words, and 37.8% were postoperative. 75% of the posts had either a negative or neutral tone. 38.4% described a personal story, and 19.3% provided a progress update. 42.3% of the posts were from the perspective of the patient. CONCLUSION: Patients reported a positive tone on Instagram, displaying their progress updates and demonstrating contentment with scoliosis surgery, and a negative tone on Twitter, showing discontentment toward inadequate access to surgery. Although both platforms were used to distribute information and provide awareness, only a small percentage of posts were from physicians and hospitals, indicating opportunities for surgeons to use social media to connect with patients.

How are Patients Describing You Online? A Natural Language Processing Driven Sentiment Analysis of Online Reviews on CSRS Surgeons.

STUDY DESIGN: A quantitative analysis of written, online reviews of Cervical Spine Research Society (CSRS) surgeons. OBJECTIVE: This study quantitatively analyzes the written reviews of members of the CSRS to report biases associated with demographic factors and frequently used words in reviews to help aid physician practices. SUMMARY OF BACKGROUND DATA: Physician review websites have influence on a patient's selection of a provider, but written reviews are subjective. Sentiment analysis of writing through artificial intelligence can quantify surgeon reviews to provide actionable feedback. METHODS: Online written and star-rating reviews of CSRS surgeons were obtained from healthgrades.com. A sentiment analysis package was used to obtain compound scores of each physician's reviews. The relationship between demographic variables and average sentiment score of written reviews were evaluated through t -tests. Positive and negative word and bigram frequency analysis was performed to indicate trends in the reviews' language. RESULTS: In all, 2239 CSRS surgeon's reviews were analyzed. Analysis showed a positive correlation between the sentiment scores and overall average star-rated reviews ( r2 =0.60, P <0.01). There was no difference in review sentiment by provider sex. However, the age of surgeons showed a significant difference as those <55 had more positive reviews (mean=+0.50) than surgeons >=55 (mean=+0.37) ( P <0.01). The most positive reviews focused both on pain and behavioral factors, whereas the most negative focused mainly on pain. Behavioral attributes increased the odds of receiving positive reviews while pain decreased them. CONCLUSION: The top-rated surgeons were described as considerate providers and effective at managing pain in their most frequently used words and bigrams. However, the worst-rated ones were mainly described as unable to relieve pain. Through quantitative analysis of physician reviews, pain is a clear factor contributing to both positive and negative reviews of surgeons, reinforcing the need for proper pain expectation management. LEVEL OF EVIDENCE: Level 4-retrospective case-control study.

Development of a machine learning algorithm to identify total and reverse shoulder arthroplasty implants from X-ray images.

Introduction: Demand for total shoulder arthroplasty (TSA) has risen significantly and is projected to continue growing. From 2012 to 2017, the incidence of reverse total shoulder arthroplasty (rTSA) rose from 7.3 cases per 100,000 to 19.3 per 100,000. Anatomical TSA saw a growth from 9.5 cases per 100,000 to 12.5 per 100,000. Failure to identify implants in a timely manner can increase operative time, cost and risk of complications. Several machine learning models have been developed to perform medical image analysis. However, they have not been widely applied in shoulder surgery. The authors developed a machine learning model to identify shoulder implant manufacturers and type from anterior-posterior X-ray images. Methods: The model deployed was a convolutional neural network (CNN), which has been widely used in computer vision tasks. 696 radiographs were obtained from a single institution. 70% were used to train the model, while evaluation was done on 30%. Results: On the evaluation set, the model performed with an overall accuracy of 93.9% with positive predictive value, sensitivity and F-1 scores of 94% across 10 different implant types (4 reverse, 6 anatomical). Average identification time was 0.110 s per implant. Conclusion: This proof of concept study demonstrates that machine learning can assist with preoperative planning and improve cost-efficiency in shoulder surgery.

Using Sentiment Analysis to Understand What Patients Are Saying About Hand Surgeons Online.

BACKGROUND: Physician review websites have influence on a patient's selection of a provider. Written reviews are subjective and difficult to quantitatively analyze. Sentiment analysis of writing can quantitatively assess surgeon reviews to provide actionable feedback for surgeons to improve practice. The objective of this study is to quantitatively analyze large subset of written reviews of hand surgeons using sentiment analysis and report unbiased trends in words used to describe the reviewed surgeons and biases associated with surgeon demographic factors. METHODS: Online written and star-rating reviews of hand surgeons were obtained from healthgrades.com and webmd.com. A sentiment analysis package was used to calculate compound scores of all reviews. Mann-Whitney U tests were performed to determine the relationship between demographic variables and average sentiment score of written reviews. Positive and negative word and word-pair frequency analysis was also performed. RESULTS: A total of 786 hand surgeons' reviews were analyzed. Analysis showed a significant relationship between the sentiment scores and overall average star-rated reviews (r2 = 0.604, P ≤ .01). There was no significant difference in review sentiment by provider sex; however, surgeons aged 50 years and younger had more positive reviews than older (P < .01). The most frequently used bigrams used to describe top-rated surgeons were associated with good bedside manner and efficient pain management, whereas those with the worst reviews are often characterized as rude and unable to relieve pain. CONCLUSIONS: This study provides insight into both demographic and behavioral factors contributing to positive reviews and reinforces the importance of pain expectation management.

Readmission and Associated Factors in Surgical Versus Non-Surgical Management of Spinal Epidural Abscess: A Nationwide Readmissions Database Analysis.

STUDY DESIGN: Retrospective cohort study. OBJECTIVES: Spinal epidural abscess (SEA) is a rare but potentially life-threatening infection treated with antimicrobials and, in most cases, immediate surgical decompression. Previous studies comparing medical and surgical management of SEA are low powered and limited to a single institution. As such, the present study compares readmission in surgical and non-surgical management using a large national dataset. METHODS: We identified all hospital admissions for SEA using the Nationwide Readmissions Database (NRD), which is the largest collection of hospital admissions data. Patients were grouped into surgically and non-surgically managed cohorts using ICD-10 coding and compared using information retrieved from the NRD such as demographics, comorbidities, length of stay and cost of admission. RESULTS: We identified 350 surgically managed and 350 non-surgically managed patients. The 90-day readmission rates for surgical and non-surgical management were 26.0% and 35.1%, respectively (P < .05). Expectedly, surgical management was associated with a significantly higher charge and length of stay at index hospital admission. Surgically managed patients had a significantly lower risk of readmission for osteomyelitis (P < .05). Finally, in patients with a low comorbidity burden, we observed a significantly lower 90-day readmission rate for surgically managed patients (surgical: 23.0%, non-surgical: 33.8%, P < .05). CONCLUSION: In patients with a low comorbidity burden, we observed a significantly lower readmission rate for surgically managed patients than non-surgically managed patients. The results of this study suggest a lower readmission rate as an advantage to surgical management of SEA and emphasize the importance of SEA as a not-to-miss diagnosis.

Previous Emergency Department Admission Is Associated With Increased 90-Day Readmission Following Cervical Spine Surgery: Evidenced Using Propensity Score Matching.

STUDY DESIGN: This was a retrospective case-control study. OBJECTIVE: The objective of this study was to evaluate whether prior emergency department admission was associated with an increased risk for 90-day readmission following elective cervical spinal fusion. SUMMARY OF BACKGROUND DATA: The incidence of cervical spine fusion reoperations has increased, necessitating the improvement of patient outcomes following surgery. Currently, there are no studies assessing the impact of emergency department visits before surgery on the risk of 90-day readmission following elective cervical spine surgery. This study aimed to fill this gap and identify a novel risk factor for readmission following elective cervical fusion. METHODS: The 2016-2018 Nationwide Readmissions Database was queried for patients aged 18 years and older who underwent an elective cervical fusion. Prior emergency admissions were defined using the variable HCUP_ED in the Nationwide Readmissions Database database. Univariate analysis of patient demographic details, comorbidities, discharge disposition, and perioperative complication was evaluated using a χ 2 test followed by multivariate logistic regression. RESULTS: In all, 2766 patients fit the inclusion criteria, and 18.62% of patients were readmitted within 90 days. Intraoperative complications, gastrointestinal complications, valvular, uncomplicated hypertension, peripheral vascular disorders, chronic obstructive pulmonary disease, cancer, and experiencing less than 3 Charlson comorbidities were identified as independent predictors of 90-day readmission. Patients with greater than 3 Charlson comorbidities (OR=0.04, 95% CI 0.01-0.12, P <0.001) and neurological complications (OR=0.29, 95% CI 0.10-0.86, P =0.026) had decreased odds for 90-day readmission. Importantly, previous emergency department visits within the calendar year before surgery were a new independent predictor of 90-day readmission (OR=9.74, 95% CI 6.86-13.83, P <0.001). CONCLUSIONS: A positive association exists between emergency department admission history and 90-day readmission following elective cervical fusion. Screening cervical fusion patients for this history and optimizing outcomes in those patients may reduce 90-day readmission rates.

Weekend Admission Increases Risk of Readmissions Following Elective Cervical Spinal Fusion.

OBJECTIVE: The "weekend effect" occurs when patients cared for during weekends versus weekdays experience worse outcomes. But reasons for this effect are unclear, especially amongst patients undergoing elective cervical spinal fusion (ECSF). Our aim was to analyze whether index weekend admission affects 30- and 90-day readmission rates post-ECSF. METHODS: All ECSF patients > 18 years were retrospectively identified from the 2016-2018 Healthcare Cost and Utilization Project Nationwide Readmissions Database (NRD), using unique patient linkage codes and International Classification of Diseases, Tenth Revision codes. Patient demographics, comorbidities, and outcomes were analyzed. Univariate logistic regression analyzed primary outcomes of 30- and 90-day readmission rates in weekday or weekend groups. Multivariate regression determined the impact of complications on readmission rates. RESULTS: Compared to the weekday group (n = 125,590), the weekend group (n = 1,026) held a higher percentage of Medicare/Medicaid insurance, incurred higher costs, had longer length of stay, and fewer routine home discharge (all p < 0.001). There was no difference in comorbidity burden between weekend versus weekday admissions, as measured by the Elixhauser Comorbidity Index (p = 0.527). Weekend admissions had higher 30-day (4.30% vs. 7.60%, p < 0.001) and 90-day (7.80% vs. 16.10%, p < 0.001) readmission rates, even after adjusting for sex, age, insurance status, and comorbidities. All-cause complication rates were higher for weekend admissions (8.62% vs. 12.7%, p < 0.001), specifically deep vein thrombosis, infection, neurological conditions, and pulmonary embolism. CONCLUSION: Index weekend admission increases 30- and 90-day readmission rates after ECSF. In patients undergoing ECSF on weekends, postoperative care for patients at risk for specific complications will allow for improved outcomes and health care utilization.

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.

Can Natural Language Processing and Artificial Intelligence Automate The Generation of Billing Codes From Operative Note Dictations?

STUDY DESIGN: Retrospective Cohort Study. OBJECTIVES: Using natural language processing (NLP) in combination with machine learning on standard operative notes may allow for efficient billing, maximization of collections, and minimization of coder error. This study was conducted as a pilot study to determine if a machine learning algorithm can accurately identify billing Current Procedural Terminology (CPT) codes on patient operative notes. METHODS: This was a retrospective analysis of operative notes from patients who underwent elective spine surgery by a single senior surgeon from 9/2015 to 1/2020. Algorithm performance was measured by performing receiver operating characteristic (ROC) analysis, calculating the area under the ROC curve (AUC) and the area under the precision-recall curve (AUPRC). A deep learning NLP algorithm and a Random Forest algorithm were both trained and tested on operative notes to predict CPT codes. CPT codes generated by the billing department were compared to those generated by our model. RESULTS: The random forest machine learning model had an AUC of .94 and an AUPRC of .85. The deep learning model had a final AUC of .72 and an AUPRC of .44. The random forest model had a weighted average, class-by-class accuracy of 87%. The LSTM deep learning model had a weighted average, class-by-class accuracy 0f 59%. CONCLUSIONS: Combining natural language processing with machine learning is a valid approach for automatic generation of CPT billing codes. The random forest machine learning model outperformed the LSTM deep learning model in this case. These models can be used by orthopedic or neurosurgery departments to allow for efficient billing.

What are patients saying about you online? A sentiment analysis of online written reviews on Scoliosis Research Society surgeons.

PURPOSE: Physician review websites have significant influence on a patient's selection of a provider, but written reviews are subjective. Sentiment analysis of writing through artificial intelligence can quantify surgeon reviews to provide actionable feedback. The objective of this study is to quantitatively analyze the written reviews of members of the Scoliosis Research Society (SRS) through sentiment analysis. METHODS: Online written reviews and star-rating reviews of SRS surgeons were obtained from healthgrades.com, and a sentiment analysis package was used to obtain compound scores of each physician's reviews. A t test and ANOVA was performed to determine the relationship between demographic variables and average sentiment score of written reviews. Positive and negative word and word-pair frequency analysis was performed to provide context to words used to describe surgeons. RESULTS: Seven hundred and twenty-one SRS surgeon's reviews were analyzed. Analysis showed a positive correlation between the sentiment scores and overall average star-rated reviews (r2 = 0.5, p < 0.01). There was no difference in review sentiment by provider gender. However, the age of surgeons showed a significant difference as younger surgeons, on average, had more positive reviews (p < 0.01). CONCLUSION: The most frequently used word pairs used to describe top-rated surgeons describe compassionate providers and efficiency in pain management. Conversely, those with the worst reviews are characterized as unable to relieve pain. Through quantitative analysis of physician reviews, pain is a clear factor contributing to both positive and negative reviews of surgeons, reinforcing the need to properly manage pain expectations. LEVEL OF EVIDENCE: IV.

What Are Patients Saying About Minimally Invasive Spine Surgeons Online: A Sentiment Analysis of 2,235 Physician Review Website Reviews.

Objective Physician review websites are becoming increasingly popular for patients to find and review healthcare providers. The goal of this study was to utilize quantitative analyses to understand trends in ratings and written comments on physician review websites for Society of Minimally Invasive Spine Surgery (SMISS) members. Methods This is a cross-sectional study. The reviews of SMISS surgeons were obtained from healthgrades.com, and sentiment analysis was used to obtain compound scores of each physicians' reviews. SMISS surgeons who were international or had fewer than three written reviews, often consisting of residents and fellows, were excluded. Inferential statistics were utilized, and word frequency analysis reported the phrases used to characterize reviews. Results One hundred sixty-nine surgeons met the inclusion criteria. 98.6% were males and the mean age was 51.7 years old. A total of 2,235 written reviews were analyzed. Younger surgeons were significantly more likely to receive higher star ratings (p<0.01). Positive behavioral characteristics, such as "kind" and "bedside manner," conferred significantly improved odds of receiving positive reviews (p<0.01). Ancillary "staff" conferred a 2x greater odds of receiving a positive review whereas a comment on "wait" times halved a surgeon's odds (p<0.01). Sentences describing pain drove down the odds of positive reviews whereas those describing pain relief produced greater odds of positive reviews (p<0.01). Conclusion Physicians who were younger, personable, provided sufficient pain relief, and who worked in favorable offices received the most positive reviews. This study informs SMISS members on the traits deemed important by patients who ultimately review surgeons online.

Trends in the Charges and Utilization of Computer-Assisted Navigation in Cervical and Thoracolumbar Spinal Surgery.

STUDY DESIGN: Retrospective national database study. PURPOSE: This study is conducted to assess the trends in the charges and usage of computer-assisted navigation in cervical and thoracolumbar spinal surgery. OVERVIEW OF LITERATURE: This study is the first of its kind to use a nationwide dataset to analyze trends of computer-assisted navigation in spinal surgery over a recent time period in terms of use in the field as well as the cost of the technology. METHODS: Relevant data from the National Readmission Database in 2015-2018 were analyzed, and the computer-assisted procedures of cervical and thoracolumbar spinal surgery were identified using International Classification of Diseases 9th and 10th revision codes. Patient demographics, surgical data, readmissions, and total charges were examined. Comorbidity burden was calculated using the Charlson and Elixhauser comorbidity index. Complication rates were determined on the basis of diagnosis codes. RESULTS: A total of 48,116 cervical cases and 27,093 thoracolumbar cases were identified using computer-assisted navigation. No major differences in sex, age, or comorbidities over time were found. The utilization of computer-assisted navigation for cervical and thoracolumbar spinal fusion cases increased from 2015 to 2018 and normalized to their respective years' total cases (Pearson correlation coefficient=0.756, p =0.049; Pearson correlation coefficient=0.9895, p =0.010). Total charges for cervical and thoracolumbar cases increased over time (Pearson correlation coefficient=0.758, p =0.242; Pearson correlation coefficient=0.766, p =0.234). CONCLUSIONS: The use of computer-assisted navigation in spinal surgery increased significantly from 2015 to 2018. The average cost grossly increased from 2015 to 2018, and it was higher than the average cost of nonnavigated spinal surgery. With the increased utilization and standardization of computer-assisted navigation in spinal surgeries, the cost of care of more patients might potentially increase. As a result, further studies should be conducted to determine whether the use of computer-assisted navigation is efficient in terms of cost and improvement of care.