Comparison of hospital charge prediction models for gastric cancer patients: neural network vs. decision tree models

Implementation of Machine Learning to Predict Cost of Care Associated with Ambulatory Single-Level Lumbar Decompression

BACKGROUND

With the emergence of the concept of value-based care, efficient resource allocation has become an increasingly prominent factor in surgical decision-making. Validated machine learning (ML) models for cost prediction in outpatient spine surgery are limited. As such, we developed and internally validated a supervised ML algorithm to reliably identify cost drivers associated with ambulatory single-level lumbar decompression surgery.

METHODS

A retrospective review of the New York State Ambulatory Surgical Database was performed to identify patients who underwent single-level lumbar decompression from 2014 to 2015. Patients with a length of stay of >0 were excluded. Using pre- and intraoperative parameters (features) derived from the New York State Ambulatory Surgical Database, an optimal supervised ML model was ultimately developed and internally validated after 5 candidate models were rigorously tested, trained, and compared for predictive performance related to total charges. The best performing model was then evaluated by testing its performance on identifying relationships between features of interest and cost prediction. Finally, the best performing algorithm was entered into an open-access web application.

RESULTS

A total of 8402 patients were included. The gradient-boosted ensemble model demonstrated the best performance assessed via internal validation. Major cost drivers included anesthesia type, operating room time, race, patient income and insurance status, community type, worker's compensation status, and comorbidity index.

CONCLUSIONS

The gradient-boosted ensemble model predicted total charges and associated cost drivers associated with ambulatory single-level lumbar decompression using a large, statewide database with excellent performance. External validation of this algorithm in future studies may guide practical application of this clinical tool.

Identifying modifiable and nonmodifiable cost drivers of ambulatory rotator cuff repair: a machine learning analysis

INTRODUCTION

Implementing novel tools that identify contributors to the cost of orthopedic procedures can help hospitals maximize efficiency, minimize waste, improve surgical decision-making, and practice value-based care. The purpose of this study was to develop and internally validate a machine learning algorithm to identify key drivers of total charges after ambulatory arthroscopic rotator cuff repair and compare its performance with a state-of-the-art statistical learning model.

METHODS

A retrospective review of the New York State Ambulatory Surgery and Services Database was performed to identify patients who underwent elective outpatient rotator cuff repair (RCR) from 2015 to 2016. Initial models were constructed using patient characteristics (age, gender, insurance status, patient income, Elixhauser Comorbidity Index) as well as intraoperative variables (concomitant procedures and services, operative time). These were subsequently entered into 5 separate machine learning algorithms and a generalized additive model using natural splines. Global variable importance and partial dependence curves were constructed to identify the greatest contributors to cost.

RESULTS

A total of 33,976 patients undergoing ambulatory RCR were included. Median total charges after ambulatory RCR were $16,017 (interquartile range: $11,009-$22,510). The ensemble model outperformed the generalized additive model and demonstrated the best performance on internal validation (root mean squared error: $7112, 95% confidence interval: 7036-7188; logarithmic root mean squared error: 0.354, 95% confidence interval: 0.336-0.373, R²: 0.53), and identified major drivers of total charges after RCR as increasing operating room time, patient income level, number of anchors used, use of local infiltration anesthesia/peripheral nerve blocks, non-White race/ethnicity, and concurrent distal clavicle excision. The model was integrated into a web-based open-access application capable of providing individual predictions and explanations on a case-by-case basis.

CONCLUSION

This study developed an ensemble supervised machine learning algorithm that outperformed a sophisticated statistical learning model in predicting total charges after ambulatory RCR. Important contributors to total charges included operating room time, duration of care, number of anchors used, type of anesthesia, concomitant distal clavicle excision, community characteristics, and patient demographic factors. Generation of a patient-specific payment schedule based on the Agency for Healthcare Research and Quality risk of mortality highlighted the financial risk assumed by physicians in flat episodic reimbursement schedules given variable patient comorbidities and the importance of an accurate prediction algorithm to appropriately reward high-value care at low costs.

Machine learning models predict total charges and drivers of cost for transcatheter aortic valve replacement

Background

Given the increasing healthcare costs, there is an interest in developing machine learning (ML) prediction models for estimating hospitalization charges. We use ML algorithms to predict hospitalization charges for patients undergoing transfemoral transcatheter aortic valve replacement (TF-TAVR) utilizing the National Inpatient Sample (NIS) database.

Methods

Patients who underwent TF-TAVR from 2012 to 2016 were included in the study. The primary outcome was total hospitalization charges. Study dataset was divided into 80% training and 20% testing sets. We used following ML regression algorithms: random forest, gradient boosting, k-nearest neighbors (KNN), multi-layer perceptron and linear regression. ML algorithms were built for for 3 stages: Stage 1, including variables that were known pre-procedurally (prior to TF-TAVR); Stage 2, including variables that were known post-procedurally; Stage 3, including length of stay (LOS) in addition to the stage 2 variables.

Results

A total of 18,793 hospitalization for TF-TAVR were analyzed. The mean and median adjusted hospitalization charges were $220,725.2 ($137,675.1) and $187,212.0 ($137,971.0-264,824.8) respectively. Random forest regression algorithm outperformed other ML algorithms at all stages with higher R² score and lower mean absolute error (MAE), root mean squared area (RMSE) and root mean squared logarithmic error (RMSLE) (Stage 1: MAE 79,979.11, R² 0.157; Stage 2: MAE 76,200.09, R² 0.256; Stage 3: MAE 69,350.09, R² 0.453). LOS was the most important predictor of hospitalization charges.

Conclusions

We built ML algorithms that predict hospitalization charges with good accuracy in patients undergoing TF-TAVR at different stages of hospitalization and that can be used by healthcare providers to better understand the drivers of charges.

Duration of Care and Operative Time Are the Primary Drivers of Total Charges After Ambulatory Hip Arthroscopy: A Machine Learning Analysis

PURPOSE

To develop a machine learning algorithm to predict total charges after ambulatory hip arthroscopy and create a risk-adjusted payment model based on patient comorbidities.

METHODS

A retrospective review of the New York State Ambulatory Surgery and Services database was performed to identify patients who underwent elective hip arthroscopy between 2015 and 2016. Features included in initial models consisted of patient characteristics, medical comorbidities, and procedure-specific variables. Models were generated to predict total charges using 5 algorithms. Model performance was assessed by the root-mean-square error, root-mean-square logarithmic error, and coefficient of determination. Global variable importance and partial dependence curves were constructed to show the impact of each input feature on total charges. For performance benchmarking, the best candidate model was compared with a multivariate linear regression using the same input features.

RESULTS

A total of 5,121 patients were included. The median cost after hip arthroscopy was $19,720 (interquartile range, $12,399-$26,439). The gradient-boosted ensemble model showed the best performance (root-mean-square error, $3,800 [95% confidence interval, $3,700-$3,900]; logarithmic root-mean-square error, 0.249 [95% confidence interval, 0.24-0.26]; R² = 0.73). Major cost drivers included total hours in facility less than 12 or more than 15, longer procedure time, performance of a labral repair, age younger than 30 years, Elixhauser Comorbidity Index (ECI) of 1 or greater, African American race, residence in extreme urban and rural areas, and higher household and neighborhood income.

CONCLUSIONS

The gradient-boosted ensemble model effectively predicted total charges after hip arthroscopy. Few modifiable variables were identified other than anesthesia type; nonmodifiable drivers of total charges included duration of care less than 12 hours or more than 15 hours, operating room time more than 100 minutes, age younger than 30 years, performance of a labral repair, and ECI greater than 0. Stratification of patients based on the ECI highlighted the increased financial risk borne by physicians via flat reimbursement schedules given variable degrees of comorbidities.

LEVEL OF EVIDENCE

Level III, retrospective cohort study.

Prediction of total hospital expenses of patients undergoing breast cancer surgery in Shanghai, China by comparing three models

BACKGROUND

Breast cancer imposes a considerable burden on both the health care system and society, and becomes increasingly severe among women in China. To reduce the economic burden of this disease is crucial for patients undergoing the breast cancer surgery, hospital managers, and medical insurance providers. However, few studies have evidenced the prediction of the total hospital expenses (THE) for breast cancer surgery. The aim of the study is to predict THE for breast cancer surgery and identify the main influencing factors.

METHODS

Data were retrieved from the first page of medical records of 3699 patients undergoing breast cancer surgery in one tertiary hospital from 2017 to 2018. Multiple liner regression (MLR), artificial neural networks (ANNs), and classification and regression tree (CART) were constructed and compared.

RESULTS

The dataset from 3699 patients were randomly divided into training and test sets at a 70:30 ratio (2599 and 1100 records, respectively). The average total hospital expenses were 12520.54 ± 7844.88 ¥ (US$ 1929.20 ± 1208.11). MLR results revealed six factors to be significantly associated with THE: age, LOS, type of disease, having medical insurance, minimally invasive surgery, and receiving general anesthesia. After comparing three models, ANNs was the best model to predict THEs in patients undergoing breast cancer surgery, and its strong predictive performance was also validated.

CONCLUSIONS

To reduce the THEs, more attention should be paid to related factors of LOS, major and minimally invasive surgeries, and general anesthesia for these patient groups undergoing breast cancer surgery. This may reduce the information asymmetry between doctors and patients and provide more reliable cost, practical inpatient medical consumption standards and reimbursement standards reference for patients, hospital managers, and medical insurance providers ,respectively.

Status quo and future prospects of artificial neural network from the perspective of gastroenterologists

Artificial neural networks (ANNs) are one of the primary types of artificial intelligence and have been rapidly developed and used in many fields. In recent years, there has been a sharp increase in research concerning ANNs in gastrointestinal (GI) diseases. This state-of-the-art technique exhibits excellent performance in diagnosis, prognostic prediction, and treatment. Competitions between ANNs and GI experts suggest that efficiency and accuracy might be compatible in virtue of technique advancements. However, the shortcomings of ANNs are not negligible and may induce alterations in many aspects of medical practice. In this review, we introduce basic knowledge about ANNs and summarize the current achievements of ANNs in GI diseases from the perspective of gastroenterologists. Existing limitations and future directions are also proposed to optimize ANN’s clinical potential. In consideration of barriers to interdisciplinary knowledge, sophisticated concepts are discussed using plain words and metaphors to make this review more easily understood by medical practitioners and the general public.

Artificial intelligence in gastric cancer: a systematic review

OBJECTIVE

This study aims to systematically review the application of artificial intelligence (AI) techniques in gastric cancer and to discuss the potential limitations and future directions of AI in gastric cancer.

METHODS

A systematic review was performed that follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Pubmed, EMBASE, the Web of Science, and the Cochrane Library were used to search for gastric cancer publications with an emphasis on AI that were published up to June 2020. The terms "artificial intelligence" and "gastric cancer" were used to search for the publications.

RESULTS

A total of 64 articles were included in this review. In gastric cancer, AI is mainly used for molecular bio-information analysis, endoscopic detection for Helicobacter pylori infection, chronic atrophic gastritis, early gastric cancer, invasion depth, and pathology recognition. AI may also be used to establish predictive models for evaluating lymph node metastasis, response to drug treatments, and prognosis. In addition, AI can be used for surgical training, skill assessment, and surgery guidance.

CONCLUSIONS

In the foreseeable future, AI applications can play an important role in gastric cancer management in the era of precision medicine.

Prediction of medical expenditures of diagnosed diabetics and the assessment of its related factors using a random forest model, MEPS 2000-2015

OBJECTIVE

To predict the medical expenditures of individual diabetics and assess the related factors of it.

DESIGN AND SETTING

Cross-sectional study.

SETTING AND PARTICIPANTS

Data were collected from the US household component of the medical expenditure panel survey, 2000-2015.

MAIN OUTCOME MEASURE

Random forest (RF) model was performed with the programs of randomForest in R software. Spearman correlation coefficients (rs), mean absolute error (MAE) and mean-related error (MRE) was computed to assess the prediction of all the models.

RESULTS

Total medical expenditure was increased from $105 Billion in 2000 to $318 Billion in 2015. rs, MAE and MRE between the predicted and actual values of medical expenditures in RF model were 0.644, $0.363 and 0.043%. Top one factor in prediction was being treated by the insulin, followed by type of insurance, employment status, age and economical level. The latter four variables had no impact in predicting of medical expenditure by being treated by the insulin. Further, after the sub-analysis of gender and age-groups, the evaluating indicators of prediction were almost identical to each other. Top five variables of total medical expenditure among male were same as those among all the diabetics. Expenses for doctor visits, hospital stay and drugs were also predicted with RF model well. Treatment with insulin was the top one factor of total medical expenditure among female, 18-, 25- and 65-age-groups. Additionally, it indicated that RF model was little superior to traditional regression model.

CONCLUSIONS

RF model could be used in prediction of medical expenditure of diabetics and assessment of its related factors well.

Machine learning ensemble models predict total charges and drivers of cost for transsphenoidal surgery for pituitary tumor

OBJECTIVE

Efficient allocation of resources in the healthcare system enables providers to care for more and needier patients. Identifying drivers of total charges for transsphenoidal surgery (TSS) for pituitary tumors, which are poorly understood, represents an opportunity for neurosurgeons to reduce waste and provide higher-quality care for their patients. In this study the authors used a large, national database to build machine learning (ML) ensembles that directly predict total charges in this patient population. They then interrogated the ensembles to identify variables that predict high charges.

METHODS

The authors created a training data set of 15,487 patients who underwent TSS between 2002 and 2011 and were registered in the National Inpatient Sample. Thirty-two ML algorithms were trained to predict total charges from 71 collected variables, and the most predictive algorithms combined to form an ensemble model. The model was internally and externally validated to demonstrate generalizability. Permutation importance and partial dependence analyses were performed to identify the strongest drivers of total charges. Given the overwhelming influence of length of stay (LOS), a second ensemble excluding LOS as a predictor was built to identify additional drivers of total charges.

RESULTS

An ensemble model comprising 3 gradient boosted tree classifiers best predicted total charges (root mean square logarithmic error = 0.446; 95% CI 0.439-0.453; holdout = 0.455). LOS was by far the strongest predictor of total charges, increasing total predicted charges by approximately $5000 per day.In the absence of LOS, the strongest predictors of total charges were admission type, hospital region, race, any postoperative complication, and hospital ownership type.

CONCLUSIONS

ML ensembles predict total charges for TSS with good fidelity. The authors identified extended LOS, nonelective admission type, non-Southern hospital region, minority race, postoperative complication, and private investor hospital ownership as drivers of total charges and potential targets for cost-lowering interventions.

Predicting the Survival of Gastric Cancer Patients Using Artificial and Bayesian Neural Networks

Introduction and purpose: In recent years the use of neural networks without any premises for investigation of prognosis in analyzing survival data has increased. Artificial neural networks (ANN) use small processors with a continuous network to solve problems inspired by the human brain. Bayesian neural networks (BNN) constitute a neural-based approach to modeling and non-linearization of complex issues using special algorithms and statistical methods. Gastric cancer incidence is the first and third ranking for men and women in Iran, respectively. The aim of the present study was to assess the value of an artificial neural network and a Bayesian neural network for modeling and predicting of probability of gastric cancer patient death. Materials and Methods: In this study, we used information on 339 patients aged from 20 to 90 years old with positive gastric cancer, referred to Afzalipoor and Shahid Bahonar Hospitals in Kerman City from 2001 to 2015. The three layers perceptron neural network (ANN) and the Bayesian neural network (BNN) were used for predicting the probability of mortality using the available data. To investigate differences between the models, sensitivity, specificity, accuracy and the area under receiver operating characteristic curves (AUROCs) were generated. Results: In this study, the sensitivity and specificity of the artificial neural network and Bayesian neural network models were 0.882, 0.903 and 0.954, 0.909, respectively. Prediction accuracy and the area under curve ROC for the two models were 0.891, 0.944 and 0.935, 0.961. The age at diagnosis of gastric cancer was most important for predicting survival, followed by tumor grade, morphology, gender, smoking history, opium consumption, receiving chemotherapy, presence of metastasis, tumor stage, receiving radiotherapy, and being resident in a village. Conclusion: The findings of the present study indicated that the Bayesian neural network is preferable to an artificial neural network for predicting survival of gastric cancer patients in Iran.

Evaluating the High Risk Groups for Suicide: A Comparison of Logistic Regression, Support Vector Machine, Decision Tree and Artificial Neural Network

BACKGROUND

We aimed to assess the high-risk group for suicide using different classification methods includinglogistic regression (LR), decision tree (DT), artificial neural network (ANN), and support vector machine (SVM).

METHODS

We used the dataset of a study conducted to predict risk factors of completed suicide in Hamadan Province, the west of Iran, in 2010. To evaluate the high-risk groups for suicide, LR, SVM, DT and ANN were performed. The applied methods were compared using sensitivity, specificity, positive predicted value, negative predicted value, accuracy and the area under curve. Cochran-Q test was implied to check differences in proportion among methods. To assess the association between the observed and predicted values, Ø coefficient, contingency coefficient, and Kendall tau-b were calculated.

RESULTS

Gender, age, and job were the most important risk factors for fatal suicide attempts in common for four methods. SVM method showed the highest accuracy 0.68 and 0.67 for training and testing sample, respectively. However, this method resulted in the highest specificity (0.67 for training and 0.68 for testing sample) and the highest sensitivity for training sample (0.85), but the lowest sensitivity for the testing sample (0.53). Cochran-Q test resulted in differences between proportions in different methods (P<0.001). The association of SVM predictions and observed values, Ø coefficient, contingency coefficient, and Kendall tau-b were 0.239, 0.232 and 0.239, respectively.

CONCLUSION

SVM had the best performance to classify fatal suicide attempts comparing to DT, LR and ANN.

Complementarity of Clinician Judgment and Evidence Based Models in Medical Decision Making: Antecedents, Prospects, and Challenges

Early accounts of the development of modern medicine suggest that the clinical skills, scientific competence, and doctors' judgment were the main impetus for treatment decision, diagnosis, prognosis, therapy assessment, and medical progress. Yet, clinician judgment has its own critics and is sometimes harshly described as notoriously fallacious and an irrational and unfathomable black box with little transparency. With the rise of contemporary medical research, the reputation of clinician judgment has undergone significant reformation in the last century as its fallacious aspects are increasingly emphasized relative to the evidence based options. Within the last decade, however, medical forecasting literature has seen tremendous change and new understanding is emerging on best ways of sharing medical information to complement the evidence based medicine practices. This review revisits and highlights the core debate on clinical judgments and its interrelations with evidence based medicine. It outlines the key empirical results of clinician judgments relative to evidence based models and identifies its key strengths and prospects, the key limitations and conditions for the effective use of clinician judgment, and the extent to which it can be optimized and professionalized for medical use.

Intelligent clinical decision support systems based on SNOMED CT

The decision support systems that have been developed to assist physicians in the diagnostic process often are based on static data which may be out of date. We present a comprehensive analysis of artificial intelligent methods which could be applied to documents encoded by SNOMED CT. By mining information directly from SNOMED CT encoded documents, a decision support system could contain timely updated diagnostic information, which is of significant value in fast changing situations such as minimally understood emerging diseases and epidemics. Through a high level comparison of many AI methods it is found that a TAN-Bayesian method could be the most suitable to apply to SNOMED CT data.