National incidences and predictors of inefficiencies in perioperative care

A machine learning prediction model for total shoulder arthroplasty procedure duration: an evaluation of surgeon, patient, and shoulder-specific factors

BACKGROUND

Operating room efficiency is of paramount importance for scheduling, cost efficiency, and to allow for the high operating volume required to address the growing demand for arthroplasty. The purpose of this study was to develop a machine learning predictive model for total shoulder arthroplasty (TSA) procedure duration and to identify factors which are predictive of a prolonged procedure.

METHODS

A retrospective review was undertaken of all TSA between 2013 and 2021 in a large academic institution. Patient, surgeon, anesthetic, and shoulder-specific factors were assessed. The duration of time in the operating room was recorded and compared to the human scheduler and electronic health record predicted procedure duration. Two gradient-boosted decision tree regression models were created with both training and validation datasets. The mean squared logarithmic error was chosen as the loss function. The first model (M1) considered patient, surgeon, and anesthetic factors, while the second model (M2) considered shoulder anatomy and pathology specific factors in addition.

RESULTS

Human schedulers' predicted 64.1% of cases accurately, with 26.7% underpredicted and 9.2% overpredicted. M1 successfully predicted 79.7% of cases, with 6.9% underpredicted and 13.4% overpredicted. M2 successfully predicted 82.5% of cases with 8.8% underpredicted and 8.8% overpredicted. M2 was significantly more accurate in predicting anatomic total shoulder arthroplasty compared to reverse (rTSA) (90.6% vs. 78.1%, P < .001).The feature with the greatest impact on the shoulder-specific model's prediction was the historical median procedure duration; followed by the electronic health record prediction, surgeon prediction, patient age, and a traumatic indication. Factors which were associated with underpredicting procedure duration included younger age, traumatic indication, male sex, greater body mass index, and a B2 glenoid.

CONCLUSION

Machine learning predictive models outperformed traditional scheduling, with a model incorporating general and shoulder-specific data providing the most accurate prediction of TSA procedure duration. Integration of modeling has the potential to optimize theater utilization and improve efficiency.

Utilization of Machine Learning Models to More Accurately Predict Case Duration in Primary Total Joint Arthroplasty

BACKGROUND

Accurate operative scheduling is essential for the appropriation of operating room esources. We sought to implement a machine learning model to predict primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) case time.

METHODS

A total of 10,590 THAs and 12,179 TKAs between July 2017 and December 2022 were retrospectively identified. Cases were chronologically divided into training, validation, and test sets. The test set cohort included 1,588 TKAs and 1,204 THAs. There were four ML algorithms developed: linear ridge regression (LR), random forest, XGBoost, and explainable boosting machine. Each model's case time estimate was compared to the scheduled estimate measured in 15-minute "wait" time blocks ("underbooking") and "excess" time blocks ("overbooking"). Surgical case time was recorded, and SHAP values were assigned to patient characteristics, surgical information, and the patient's medical condition to understand feature importance.

RESULTS

The most predictive model input was "median previous 30 procedure case times." The XGBoost model outperformed the other models in predicting both TKA and THA case times. The model reduced TKA 'excess time blocks' by 85 blocks (P < 0.001) and 'wait time blocks' by 96 blocks (P < 0.001). The model did not significantly reduce 'excess time blocks' in THA (P = 0.89) but did significantly reduce 'wait time blocks' by 134 blocks (P < 0.001). In total, the model improved TKA operative booking by 181 blocks (2,715 minutes) and THA operative booking by 138 blocks (2,070 minutes).

CONCLUSIONS

Machine learning outperformed a traditional method of scheduling total joint arthroplasty cases. The median time of the prior 30 surgical cases was the most influential on scheduling case time accuracy. As ML models improve, surgeons should consider ML utilization in case scheduling; however, prior 30 surgical cases may serve as an adequate alternative.

Improving case duration accuracy of orthopedic surgery using bidirectional encoder representations from Transformers (BERT) on Radiology Reports

PURPOSE

A major source of inefficiency in the operating room is the mismatch between scheduled versus actual surgical time. The purpose of this study was to demonstrate a proof-of-concept study for predicting case duration by applying natural language processing (NLP) and machine learning that interpret radiology reports for patients undergoing radius fracture repair.

METHODS

Logistic regression, random forest, and feedforward neural networks were tested without NLP and with bag-of-words. Another NLP method tested used feedforward neural networks and Bidirectional Encoder Representations from Transformers specifically pre-trained on clinical notes (ClinicalBERT). A total of 201 cases were included. The data were split into 70% training and 30% test sets. The average root mean squared error (RMSE) were calculated (and 95% confidence interval [CI]) from 10-fold cross-validation on the training set. The models were then tested on the test set to determine proportion of times surgical cases would have scheduled accurately if ClinicalBERT was implemented versus historic averages.

RESULTS

The average RMSE was lowest using feedforward neural networks using outputs from ClinicalBERT (25.6 min, 95% CI: 21.5-29.7), which was significantly (P < 0.001) lower than the baseline model (39.3 min, 95% CI: 30.9-47.7). Using the feedforward neural network and ClinicalBERT on the test set, the percentage of accurately predicted cases, which was defined by the actual surgical duration within 15% of the predicted surgical duration, increased from 26.8 to 58.9% (P < 0.001).

CONCLUSION

This proof-of-concept study demonstrated the successful application of NLP and machine leaning to extract features from unstructured clinical data resulting in improved prediction accuracy for surgical case duration.

Patient-specific Predictors of Surgical Delay in a Large Tertiary-care Hospital Operating Room

PURPOSE

The purpose of this study was to describe patient-specific factors predictive of surgical delay in elective surgical cases.

DESIGN

Retrospective cohort study.

METHODS

Data were extracted retrospectively from the electronic health record of 32,818 patients who underwent surgery at a large academic hospital in Los Angeles between May 2012 and April 2017. Following bivariate analysis of patient-specific factors and surgical delay, statistically significant predictors were entered into a logistic regression model to determine the most significant predictors of surgical delay.

FINDINGS

Predictors of delay included having monitored anesthesia care (odds ratio [OR], 1.28; 95% confidence intervals [CI], 1.20-1.36), American Society of Anesthesiologist class 3 or above (OR, 1.21; 95% CI, 1.15-1.28), African American race (OR, 1.25; 95% CI, 1.12-1.39), renal failure (OR, 1.20; 95% CI, 1.09-1.32), steroid medication (OR, 1.13; 95% CI, 1.04-1.23) and Medicaid (OR,1.18; 95%CI, 1.09-1.30) or medicare insurance (OR, 1.14; 95% CI, 1.07-1.21). Six surgical specialties also increased the odds of delay. Obesity and cardiovascular anesthesia decreased the odds of delay.

CONCLUSIONS

Certain patient-specific factors including type of insurance, health status, and race were associated with surgical delay. Whereas monitored anesthesia care anesthesia was predictive of a delay, cardiovascular anesthesia reduced the odds of delay. Additionally, obese patients were less likely to experience a delay. While the electronic health record provided a large amount of detailed information, barriers existed to accessing meaningful data.

Contributing Factors to Operating Room Delays Identified from an Electronic Health Record: A Retrospective Study

The operating room (OR) is considered a major cost center and revenue generator for hospitals. Multiple factors contribute to OR delays and impact patient safety, patient satisfaction scores, and hospital financial performance. Reducing OR delays allows better utilization of OR resources and staffing and improves patient satisfaction while decreasing operating costs. Accurate scheduling can be the basis to achieve these goals. The objective of this initial study was to identify factors not normally documented in the electronic health record (EHR) that may contribute to or be indicators of OR delays. Materials and Methods. A retrospective data analysis was performed analyzing 67,812 OR cases from 12 surgical specialties at a small university medical center from 2010 through the first quarter of 2017. Data from the hospital's EHR were exported and subjected to statistical analysis using Statistical Analysis System (SAS) software (SAS Institute, Cary, NC). Results. Statistical analysis of the extracted EHR data revealed factors that were associated with OR delays including, surgical specialty, preoperative assessment testing, patient body mass index, American Society of Anesthesiologists (ASA) physical status classification, daily procedure count, and calendar year. Conclusions. Delays hurt OR efficiency on many levels. Identifying those factors may reduce delays and better accommodate the needs of surgeons, staff, and patients thereby leading to improved patient's outcomes and patient satisfaction. Reducing delays can decrease operating costs and improve the financial position of the operating theater as well as that of the hospital. Anesthesiology teams can play a key role in identifying factors that cause delays and implementing mitigating efficiencies.

Predicting surgical operative time in primary total knee arthroplasty utilizing machine learning models

BACKGROUND

Prolonged surgical operative time is associated with postoperative adverse outcomes following total knee arthroplasty (TKA). Increasing operating room efficiency necessitates the accurate prediction of surgical operative time for each patient. One potential way to increase the accuracy of predictions is to use advanced predictive analytics, such as machine learning. The aim of this study is to use machine learning to develop an accurate predictive model for surgical operative time for patients undergoing primary total knee arthroplasty.

METHODS

A retrospective chart review of electronic medical records was conducted to identify patients who underwent primary total knee arthroplasty at a tertiary referral center. Three machine learning algorithms were developed to predict surgical operative time and were assessed by discrimination, calibration and decision curve analysis. Specifically, we used: (1) Artificial Neural Networks (ANNs), (2) Random Forest (RF), and (3) K-Nearest Neighbor (KNN).

RESULTS

We analyzed the surgical operative time for 10,021 consecutive patients who underwent primary total knee arthroplasty. The neural network model achieved the best performance across discrimination (AUC = 0.82), calibration and decision curve analysis for predicting surgical operative time. Based on this algorithm, younger age (< 45 years), tranexamic acid non-usage, and a high BMI (> 40 kg/m²) were the strongest predictors associated with surgical operative time.

CONCLUSIONS

This study shows excellent performance of machine learning models for predicting surgical operative time in primary total knee arthroplasty. The accurate estimation of surgical duration is important in enhancing OR efficiency and identifying patients at risk for prolonged surgical operative time.

LEVEL OF EVIDENCE

Level III, case control retrospective analysis.

Preventing Surgical Delay and Cancellation with Patient-Centered Interventions

Delay and cancellation can significantly impact cost and outcomes among surgical patients. While the causes of delay and cancellation are not fully enumerated, possible reasons include delivery-related causes such as facility, equipment, and provider availability as well as patient-related issues such as readiness and health status. Despite limited research explaining patient-related causes, there are many studies that evaluate patient-centered interventions to decrease delay and cancellation. This article highlights patient-centered interventions including preoperative clinics, preoperative screening, and focused education that have been shown to reduce delay and cancellation. This information provides perianesthesia nurses and advanced practice nurses ideas to maximize their roles in improving efficiency by prevention of delay and cancellation. This article should also stimulate additional research to help better understand the causes and the role of the nurse in the implementation of evidence-based practice projects that use patient-centered interventions.

Case-mix adjusted postanaesthesia care unit length of stay and business intelligence dashboards for feedback to anaesthetists

BACKGROUND

Despite advances in business intelligence software and evidence that feedback to doctors can improve outcomes, objective feedback regarding patient outcomes for individual anaesthetists is hampered by lack of useful benchmarks. We aimed to address this issue by producing case-mix and risk-adjusted postanaesthesia care unit (PACU) length of stay (LOS) benchmarks for integration into modern reporting tools.

METHODS

We extended existing hospital information systems to calculate predicted PACU LOS using a neural network trained on patient age, surgery duration, sex, operating specialty, urgency, weekday, and insurance status (n=100 511). We then calculated the difference between observed mean and predicted PACU LOS for individual doctors, and compared the results with and without case-mix adjustment. We report practical implications of using visual analytics dashboards displaying the difference between observed and predicted PACU LOS to provide feedback to anaesthetic doctors.

RESULTS

The neural network accounted for over half of observed variation in individual doctors' mean PACU LOS (mean predicted and mean actual LOS Spearman's r²=0.57). Account for case-mix reduced apparent spread, with 80% of individual doctors falling in a band of 4.3 min after case-mix adjusting, compared with a range of 24 min without adjustment. Case-mix adjusting also identified different individual doctors as outliers (Weighted Cohen's kappa [κ]=0.27). Finally, we demonstrated that we were able to integrate the adjusted metrics into routine reporting tools.

CONCLUSION

With caution, case-mix adjustment of anaesthetic outcome measures such as PACU LOS potentially provides a useful continuous quality improvement tool. Unadjusted outcome measures are imprecise at best and misleading at worst.

A preoperative predictive model for prolonged post-anaesthesia care unit stay after outpatient surgeries

STUDY OBJECTIVE

To create a preoperative predictive model for prolonged post-anaesthesia care unit (PACU) stay for outpatient surgery and compare with an existing (University of California-San Diego, UCSD) model.

DESIGN

Retrospective observational study.

SETTING

Post-anaesthesia care unit. Patients: Outpatient surgical patients discharged on the same day in a large academic institution. Preoperative data were collected. The study period was three months in 2016. Measurements: Prolonged PACU stay defined as a length of stay longer than the third quartile. We utilized multivariate regression analyses and bootstrapping statistical techniques to create a predictive model for prolonged PACU stay. Main results: Four strong predictors for prolonged PACU stay: general anaesthesia, obstructive sleep apnoea, surgical specialty and scheduled case duration. Our model had an excellent discrimination performance and a good calibration.

CONCLUSION

We developed a predictive model for prolonged PACU stay in our institution. This model is different from the UCSD model probably secondary to local and regional differences in outpatient surgery practice. Therefore, individual practice study outcomes may not apply to other practices without careful consideration of these differences.

Reducing Preventable Surgical Cancellations: Improving the Preoperative Anesthesia Interview Process

PURPOSE

Thorough and accurate preoperative anesthesia interviews may help improve perioperative efficiency by reducing unnecessary preoperative testing and preventable surgical cancellations, both of which create financial burdens. Standardized anesthesia preoperative interview guidelines and online educational modules for registered nurses (RNs) conducting preoperative interviews may improve this process.

DESIGN

Predesign and postdesign, retrospective chart review.

METHODS

Online educational modules and standardized preoperative anesthesia interview guidelines were developed for RNs conducting preoperative interviews. A retrospective chart review compared preoperative anesthesia interview record completion rates, compliance with laboratory testing, and the total number of preventable surgical cancellations.

FINDINGS

Documentation of preoperative anesthesia interview records increased, whereas unnecessary preoperative testing decreased, neither with statistical significance. Preventable cancellation rates decreased significantly from 34.3% to 20% (P < .5) contributing to a 3-month postimplementation cost savings of approximately $30,000.

CONCLUSIONS

A standardized preoperative anesthesia interview guideline and online anesthesia educational modules for RNs conducting preoperative anesthesia interviews improved preoperative record completion rates, reduced unnecessary laboratory testing, and averted surgical cancellations.

Machine Learning Can Improve Estimation of Surgical Case Duration: A Pilot Study

Operating room (OR) utilization is a significant determinant of hospital profitability. One aspect of this is surgical scheduling, which depends on accurate predictions of case duration. This has been done historically by either the surgeon based on personal experience, or by an electronic health record (EHR) based on averaged historical means for case duration. Here, we compare the predicted case duration (pCD) accuracy of a novel machine-learning algorithm over a 3-month period. A proprietary machine learning algorithm was applied utilizing operating room factors such as patient demographic data, pre-surgical milestones, and hospital logistics and compared to that of a conventional EHR. Actual case duration and pCD (Leap Rail vs EHR) was obtained at one institution over the span of 3 months. Actual case duration was defined as time between patient entry into an OR and time of exit. pCD was defined as case time allotted by either Leap Rail or EHR. Cases where Leap Rail was unable to generate a pCD were excluded. A total of 1059 surgical cases were performed during the study period, with 990 cases being eligible for the study. Over all sub-specialties, Leap Rail showed a 7 min improvement in absolute difference between pCD and actual case duration when compared to conventional EHR (p < 0.0001). In aggregate, the Leap Rail method resulted in a 70% reduction in overall scheduling inaccuracy. Machine-learning algorithms are a promising method of increasing pCD accuracy and represent one means of improving OR planning and efficiency.

The impact of health literacy on shared decision making before elective surgery: a propensity matched case control analysis

BACKGROUND

Poor health literacy affects over 90 million Americans. The primary aim of the study was to evaluate a possible association between health literacy and decision conflict in surgical patients.

METHODS

Patients undergoing a diverse number of elective surgeries were enrolled in the study. Health literacy was measured using the Newest Vital Sign instrument and decision conflict using the low literacy version of the Decision Conflict Scale.

RESULTS

200 patients undergoing elective surgeries were included in the study. Patients who had greater health literacy scores had lower decision conflict scores, Spearman's rho = - 0.43, P < 0.001. Following propensity-score matching to account for potential covariates, the median (IQR) decision conflict score was 20 (0 to 40) for patients with poor health literacy compared to 0 (0 to 5) for patients with adequate literacy, P < 0.001.

CONCLUSIONS

Poor health literacy is associated with greater decision conflict in patients undergoing elective surgical procedures. Strategies should be implemented to minimize decision conflict in poor health literacy patients undergoing elective surgical procedures.

The Association of Anesthesia Clinical Directors (AACD) Glossary of Times Used for Scheduling and Monitoring of Diagnostic and Therapeutic Procedures

The Glossary of Times Used for Scheduling and Monitoring of Diagnostic and Therapeutic Procedures also known as the Procedural Times Glossary (PTG) was originally developed with the support of the Association of Anesthesia Clinical Directors (AACD). The goal was to establish standardized terms to measure and assess the performance of operating room and procedural areas. By incorporating standardized concepts of efficiency and utilization, the PTG codified operating room metrics and facilitated benchmarking and quality improvement initiatives. In the last three decades, these concepts have also served as the basis for research in operating room management, including incorporating frameworks from diverse fields. The metrics in the PTG are divided into four categories: (1) Procedural Times; (2) Procedural and Scheduling Definitions and Time Periods; (3) Utilization and Efficiency Indices; and (4) Patient Categories. We describe each of the categories and corresponding metrics. The PTG provides the fundamental building blocks for managing operating and non-operating room suites. We hope that reintroducing these important time markers will help facilitate the reporting of standardized metrics.

Operating Room Efficiency before and after Entrance in a Benchmarking Program for Surgical Process Data

Operating room (OR) efficiency continues to be a high priority for hospitals. In this context the concept of benchmarking has gained increasing importance as a means to improve OR performance. The aim of this study was to investigate whether and how participation in a benchmarking and reporting program for surgical process data was associated with a change in OR efficiency, measured through raw utilization, turnover times, and first-case tardiness. The main analysis is based on panel data from 202 surgical departments in German hospitals, which were derived from the largest database for surgical process data in Germany. Panel regression modelling was applied. Results revealed no clear and univocal trend of participation in a benchmarking and reporting program for surgical process data. The largest trend was observed for first-case tardiness. In contrast to expectations, turnover times showed a generally increasing trend during participation. For raw utilization no clear and statistically significant trend could be evidenced. Subgroup analyses revealed differences in effects across different hospital types and department specialties. Participation in a benchmarking and reporting program and thus the availability of reliable, timely and detailed analysis tools to support the OR management seemed to be correlated especially with an increase in the timeliness of staff members regarding first-case starts. The increasing trend in turnover time revealed the absence of effective strategies to improve this aspect of OR efficiency in German hospitals and could have meaningful consequences for the medium- and long-run capacity planning in the OR.