Leveraging Machine Learning to Predict 30-Day Hospital Readmission after Cardiac Surgery

Abstract Objective To provide focused evaluation of predictive modeling of electronic medical record (EMR) data to predict 30 day hospital readmission. Design Systematic review. Data source Ovid Medline, Ovid Embase, CINAHL, Web of Science, and Scopus from January 2015 to January 2019. Eligibility criteria for selecting studies All studies of predictive models for 28 day or 30 day hospital readmission that used EMR data. Outcome measures Characteristics of included studies, methods of prediction, predictive features, and performance of predictive models. Results Of 4442 citations reviewed, 41 studies met the inclusion criteria. Seventeen models predicted risk of readmission for all patients and 24 developed predictions for patient specific populations, with 13 of those being developed for patients with heart conditions. Except for two studies from the UK and Israel, all were from the US. The total sample size for each model ranged between 349 and 1 195 640. Twenty five models used a split sample validation technique. Seventeen of 41 studies reported C statistics of 0.75 or greater. Fifteen models used calibration techniques to further refine the model. Using EMR data enabled final predictive models to use a wide variety of clinical measures such as laboratory results and vital signs; however, use of socioeconomic features or functional status was rare. Using natural language processing, three models were able to extract relevant psychosocial features, which substantially improved their predictions. Twenty six studies used logistic or Cox regression models, and the rest used machine learning methods. No statistically significant difference (difference 0.03, 95% confidence interval −0.0 to 0.07) was found between average C statistics of models developed using regression methods (0.71, 0.68 to 0.73) and machine learning (0.74, 0.71 to 0.77). Conclusions On average, prediction models using EMR data have better predictive performance than those using administrative data. However, this improvement remains modest. Most of the studies examined lacked inclusion of socioeconomic features, failed to calibrate the models, neglected to conduct rigorous diagnostic testing, and did not discuss clinical impact.

Download Full-text

Predicting all-cause 90-day hospital readmission for dental patients using machine learning methods

BDJ Open ◽

10.1038/s41405-021-00057-6 ◽

2021 ◽

Vol 7 (1) ◽

Author(s):

Wei Li ◽

Martin S. Lipsky ◽

Eric S. Hon ◽

Weicong Su ◽

Sharon Su ◽

...

Keyword(s):

Machine Learning ◽

Hospital Readmission ◽

Health Care Quality ◽

Prediction Models ◽

Characteristic Curve ◽

Model Performance ◽

Care Quality ◽

Support Vector ◽

Day Hospital ◽

Dental Patients

Abstract Introduction Hospital readmission rates are an indicator of the health care quality provided by hospitals. Applying machine learning (ML) to a hospital readmission database offers the potential to identify patients at the highest risk for readmission. However, few studies applied ML methods to predict hospital readmission. This study sought to assess ML as a tool to develop prediction models for all-cause 90-day hospital readmission for dental patients. Methods Using the 2013 Nationwide Readmissions Database (NRD), the study identified 9260 cases for all-cause 90-day index admission for dental patients. Five ML classification algorithms including decision tree, logistic regression, support vector machine, k-nearest neighbors, and artificial neural network (ANN) were implemented to build predictive models. The model performance was estimated and compared by using area under the receiver operating characteristic curve (AUC), and accuracy, sensitivity, specificity, and precision. Results Hospital readmission within 90 days occurred in 1746 cases (18.9%). Total charges, number of diagnosis, age, number of chronic conditions, length of hospital stays, number of procedures, primary expected payer, and severity of illness emerged as the top eight important features in all-cause 90-day hospital readmission. All models had similar performance with ANN (AUC = 0.743) slightly outperforming the rest. Conclusion This study demonstrates a potential annual saving of over $500 million if all of the 90-day readmission cases could be prevented for 21 states represented in the NRD. Among the methods used, the prediction model built by ANN exhibited the best performance. Further testing using ANN and other methods can help to assess important readmission risk factors and to target interventions to those at the greatest risk.

Download Full-text

Machine Learning Approach to Predict Risk of 90-Day Hospital Readmissions in Patients With Atrial Fibrillation: Implications for Quality Improvement in Healthcare

Health Services Research and Managerial Epidemiology ◽

10.1177/2333392820961887 ◽

2020 ◽

Vol 7 ◽

pp. 233339282096188

Author(s):

Man Hung ◽

Eric S. Hon ◽

Evelyn Lauren ◽

Julie Xu ◽

Gary Judd ◽

...

Keyword(s):

Machine Learning ◽

Atrial Fibrillation ◽

Support Vector Machine ◽

Catheter Ablation ◽

Hospital Readmission ◽

Hospital Readmissions ◽

Support Vector ◽

Day Hospital ◽

Learning Methods ◽

Machine Learning Methods

Background: Atrial fibrillation (AF) in the elderly population is projected to increase over the next several decades. Catheter ablation shows promise as a treatment option and is becoming increasingly available. We examined 90-day hospital readmission for AF patients undergoing catheter ablation and utilized machine learning methods to explore the risk factors associated with these readmission trends. Methods: Data from the 2013 Nationwide Readmissions Database on AF cases were used to predict 90-day readmissions for AF with catheter ablation. Multiple machine learning methods such as k-Nearest Neighbors, Decision Tree, and Support Vector Machine were employed to determine variable importance and build risk prediction models. Accuracy, precision, sensitivity, specificity, and area under the curve were compared for each model. Results: The 90-day hospital readmission rate was 17.6%; the average age of the patients was 64.9 years; 62.9% of patients were male. Important variables in predicting 90-day hospital readmissions in patients with AF undergoing catheter ablation included the age of the patient, number of diagnoses on the patient’s record, and the total number of discharges from a hospital. The k-Nearest Neighbor had the best performance with a prediction accuracy of 85%. This was closely followed by Decision Tree, but Support Vector Machine was less ideal. Conclusions: Machine learning methods can produce accurate models in predicting hospital readmissions for patients with AF. The likelihood of readmission to the hospital increases as the patient age, total number of hospital discharges, and total number of patient diagnoses increase. Findings from this study can inform quality improvement in healthcare and in achieving patient-centered care.

Download Full-text

A Machine Learning-Based Predictive Model for 30-Day Hospital Readmission Prediction for COPD Patients

2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC) ◽

10.1109/smc42975.2020.9283295 ◽

2020 ◽

Author(s):

Vijay Kumar Verma ◽

Wen-Yen Lin

Keyword(s):

Machine Learning ◽

Predictive Model ◽

Hospital Readmission ◽

Day Hospital ◽

Copd Patients

Download Full-text

Commentary: Right Ventricular dysfunction after cardiac surgery: Machine Learning to teach us what we already know?

Journal of Thoracic and Cardiovascular Surgery ◽

10.1016/j.jtcvs.2020.11.126 ◽

2020 ◽

Author(s):

Edgar Aranda-Michel ◽

Ibrahim Sultan

Keyword(s):

Machine Learning ◽

Cardiac Surgery ◽

Right Ventricular ◽

Ventricular Dysfunction ◽

Right Ventricular Dysfunction ◽

Learning To Teach

Download Full-text

Comparison of Inpatient Standard-of-Care to Outpatient Oritavancin Therapy for Patients With Acute Uncomplicated Cellulitis

Journal of Pharmacy Practice ◽

10.1177/08971900211021258 ◽

2021 ◽

pp. 089719002110212

Author(s):

Brandy Williams ◽

Justin Muklewicz ◽

Taylor D. Steuber ◽

April Williams ◽

Jonathan Edwards

Keyword(s):

Hospital Readmission ◽

Average Length ◽

Hospital Readmissions ◽

Standard Of Care ◽

Economic Outcomes ◽

Day Hospital ◽

Outpatient Parenteral Antimicrobial Therapy ◽

Average Length Of Stay ◽

Observational Cohort ◽

The Difference

Background: Shifting inpatient antibiotic treatment to outpatient parenteral antimicrobial therapy may minimize treatment for acute bacterial skin and skin structure infections, including cellulitis. The purpose of this evaluation was to compare 30-day hospital readmission or admission due to cellulitis and economic outcomes of inpatient standard-of-care (SoC) management of acute uncomplicated cellulitis to outpatient oritavancin therapy. Methods: This retrospective, observational cohort study was conducted at a 941-bed community teaching hospital. Adult patients 18 years and older treated for acute uncomplicated cellulitis between February 2015 to December 2018 were eligible for inclusion. Information was obtained from hospital and billing department records. Patients were assigned to either inpatient SoC or outpatient oritavancin cohorts for comparison. Results: 1,549 patients were included in the study (1,348 in the inpatient SoC cohort and 201 in the outpatient oritavancin cohort). The average length of stay for patients admitted was 3.6 ± 1.5 days. The primary outcome of 30-day hospital readmission or admission due to cellulitis occurred in 49/1348 (3.6%) patients in the inpatient SoC cohort versus 1/201 (0.5%) in the outpatient oritavancin cohort (p = 0.02). The difference between costs and reimbursement was improved in the outpatient oritavancin group (p < 0.001). Conclusion: Outpatient oritavancin for acute uncomplicated cellulitis was associated with reduction in 30-day hospital readmissions or admissions compared to inpatient SoC. Beneficial economic outcomes for the outpatient oritavancin cohort were observed. Additional studies are required to confirm these findings.

Download Full-text

Risk of 30-day hospital readmission associated with medical conditions and drug regimens of polymedicated, older inpatients discharged home: a registry-based cohort study

BMJ Open ◽

10.1136/bmjopen-2021-052755 ◽

2021 ◽

Vol 11 (7) ◽

pp. e052755

Author(s):

Filipa Pereira ◽

Henk Verloo ◽

Taushanov Zhivko ◽

Saviana Di Giovanni ◽

Carla Meyer-Massetti ◽

...

Keyword(s):

Cohort Study ◽

Hospital Readmission ◽

Hospital Readmissions ◽

Drug Regimen ◽

Hospital Length ◽

Hospital Length Of Stay ◽

Day Hospital ◽

Medical Conditions ◽

Older Inpatients ◽

Drug Regimens

ObjectivesThe present study analysed 4 years of a hospital register (2015–2018) to determine the risk of 30-day hospital readmission associated with the medical conditions and drug regimens of polymedicated, older inpatients discharged home.DesignRegistry-based cohort study.SettingValais Hospital—a public general hospital centre in the French-speaking part of Switzerland.ParticipantsWe explored the electronic records of 20 422 inpatient stays by polymedicated, home-dwelling older adults held in the hospital’s patient register. We identified 13 802 hospital readmissions involving 8878 separate patients over 64 years old.Outcome measuresSociodemographic characteristics, medical conditions and drug regimen data associated with risk of readmission within 30 days of discharge.ResultsThe overall 30-day hospital readmission rate was 7.8%. Adjusted multivariate analyses revealed increased risk of hospital readmission for patients with longer hospital length of stay (OR=1.014 per additional day; 95% CI 1.006 to 1.021), impaired mobility (OR=1.218; 95% CI 1.039 to 1.427), multimorbidity (OR=1.419 per additional International Classification of Diseases, 10th Revision condition; 95% CI 1.282 to 1.572), tumorous disease (OR=2.538; 95% CI 2.089 to 3.082), polypharmacy (OR=1.043 per additional drug prescribed; 95% CI 1.028 to 1.058), and certain specific drugs, including antiemetics and antinauseants (OR=3.216 per additional drug unit taken; 95% CI 1.842 to 5.617), antihypertensives (OR=1.771; 95% CI 1.287 to 2.438), drugs for functional gastrointestinal disorders (OR=1.424; 95% CI 1.166 to 1.739), systemic hormonal preparations (OR=1.207; 95% CI 1.052 to 1.385) and vitamins (OR=1.201; 95% CI 1.049 to 1.374), as well as concurrent use of beta-blocking agents and drugs for acid-related disorders (OR=1.367; 95% CI 1.046 to 1.788).ConclusionsThirty-day hospital readmission risk was associated with longer hospital length of stay, health disorders, polypharmacy and drug regimens. The drug regimen patterns increasing the risk of hospital readmission were very heterogeneous. Further research is needed to explore hospital readmissions caused solely by specific drugs and drug–drug interactions.

Download Full-text

Perioperative Opioid Prescribing Patterns and Readmissions After Total Knee Arthroplasty in a National Cohort of Veterans Health Administration Patients

Pain Medicine ◽

10.1093/pm/pnz154 ◽

2019 ◽

Vol 21 (3) ◽

pp. 595-603 ◽

Cited By ~ 2

Author(s):

Seshadri C Mudumbai ◽

Paul Chung ◽

Nick Nguyen ◽

Brooke Harris ◽

J David Clark ◽

...

Keyword(s):

Total Knee Arthroplasty ◽

Hospital Readmission ◽

Veterans Health Administration ◽

Care Utilization ◽

Veterans Health ◽

Day Hospital ◽

Health Administration ◽

Opioid Use ◽

Short Acting ◽

Total Knee

Abstract Objective Among Veterans Health Administration (VHA) patients who undergo total knee arthroplasty (TKA) nationally, what are the underlying readmission rates and associations with perioperative opioid use, and are there associations with other factors such as preoperative health care utilization? Methods We retrospectively examined the records of 5,514 TKA patients (primary N = 4,955, 89.9%; revision N = 559, 10.1%) over one fiscal year (October 1, 2010–September 30, 2011) across VHA hospitals nationwide. Opioid use was classified into no opioids, tramadol only, short-acting only, or any long-acting. We measured readmission within 30 days and the number of days to readmission within 30 days. Extended Cox regression models were developed. Results The overall 30-day hospital readmission rate was 9.6% (N = 531; primary 9.5%, revision 11.1%). Both readmitted patients and the overall sample were similar on types of preoperative opioid use. Relative to patients without opioids, patients in the short-acting opioids only tier had the highest risk for 30-day hospital readmission (hazard ratio = 1.38, 95% confidence interval = 1.14–1.67). Preoperative opioid status was not associated with 30-day readmission. Other risk factors for 30-day readmission included older age (≥66 years), higher comorbidity and diagnosis-related group weights, greater preoperative health care utilization, an urban location, and use of preoperative anticonvulsants. Conclusions Given the current opioid epidemic, the routine prescribing of short-acting opioids after surgery should be carefully considered to avoid increasing risks of 30-day hospital readmissions and other negative outcomes, particularly in the context of other predisposing factors.

Download Full-text