Dataset Information

Development of a Machine Learning Model Using Electronic Health Record Data to Identify Antibiotic Use Among Hospitalized Patients.

ABSTRACT:

Importance

Comparisons of antimicrobial use among hospitals are difficult to interpret owing to variations in patient case mix. Risk-adjustment strategies incorporating larger numbers of variables haves been proposed as a method to improve comparisons for antimicrobial stewardship assessments.

Objective

To evaluate whether variables of varying complexity and feasibility of measurement, derived retrospectively from the electronic health records, accurately identify inpatient antimicrobial use.

Design, setting, and participants

Retrospective cohort study, using a 2-stage random forests machine learning modeling analysis of electronic health record data. Data were split into training and testing sets to measure model performance using area under the curve and absolute error. All adult and pediatric inpatient encounters from October 1, 2015, to September 30, 2017, at 2 community hospitals and 1 academic medical center in the Duke University Health System were analyzed. A total of 204 candidate variables were categorized into 4 tiers based on feasibility of measurement from the electronic health records.

Main outcomes and measures

Antimicrobial exposure was measured at the encounter level in 2 ways: binary (ever or never) and number of days of therapy. Analyses were stratified by age (pediatric or adult), unit type, and antibiotic group.

Results

The data set included 170 294 encounters and 204 candidate variables from 3 hospitals during the 3-year study period. Antimicrobial exposure occurred in 80 190 encounters (47%); 64 998 (38%) received 1 to 6 days of therapy, and 15 192 (9%) received 7 or more days of therapy. Two-stage models identified antimicrobial use with high fidelity (mean area under the curve, 0.85; mean absolute error, 1.0 days of therapy). Addition of more complex variables increased accuracy, with largest improvements occurring with inclusion of diagnosis information. Accuracy varied based on location and antibiotic group. Models underestimated the number of days of therapy of encounters with long lengths of stay.

Conclusions and relevance

Models using variables derived from electronic health records identified antimicrobial exposure accurately. Future risk-adjustment strategies incorporating encounter-level information may make comparisons of antimicrobial use more meaningful for hospital antimicrobial stewardship assessments.

SUBMITTER: Moehring RW

PROVIDER: S-EPMC8008288 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BACKGROUND:The limitations of existing HIV risk prediction tools are a barrier to implementation of pre-exposure prophylaxis (PrEP). We developed and validated an HIV prediction model to identify potential PrEP candidates in a large health-care system. METHODS:Our study population was HIV-uninfected adult members of Kaiser Permanente Northern California, a large integrated health-care system, who were not yet using PrEP and had at least 2 years of previous health plan enrolment with at least one outpatient visit from Jan 1, 2007, to Dec 31, 2017. Using 81 electronic health record (EHR) variables, we applied least absolute shrinkage and selection operator (LASSO) regression to predict incident HIV diagnosis within 3 years on a subset of patients who entered the cohort in 2007-14 (development dataset), assessing ten-fold cross-validated area under the receiver operating characteristic curve (AUC) and 95% CIs. We compared the full model to simpler models including only men who have sex with men (MSM) status and sexually transmitted infection (STI) positivity, testing, and treatment. Models were validated prospectively with data from an independent set of patients who entered the cohort in 2015-17. We computed predicted probabilities of incident HIV diagnosis within 3 years (risk scores), categorised as low risk (<0·05%), moderate risk (0·05% to <0·20%), high risk (0·20% to <1·0%), and very high risk (?1·0%), for all patients in the validation dataset. FINDINGS:Of 3?750?664 patients in 2007-17 (3?143?963 in the development dataset and 606?701 in the validation dataset), there were 784 incident HIV cases within 3 years of baseline. The LASSO procedure retained 44 predictors in the full model, with an AUC of 0·86 (95% CI 0·85-0·87) for incident HIV cases in 2007-14. Model performance remained high in the validation dataset (AUC 0·84, 0·80-0·89). The full model outperformed simpler models including only MSM status and STI positivity. For the full model, flagging 13?463 (2·2%) patients with high or very high HIV risk scores in the validation dataset identified 32 (38·6%) of the 83 incident HIV cases, including 32 (46·4%) of 69 male cases and none of the 14 female cases. The full model had equivalent sensitivity by race whereas simpler models identified fewer black than white HIV cases. INTERPRETATION:Prediction models using EHR data can identify patients at high risk of HIV acquisition who could benefit from PrEP. Future studies should optimise EHR-based HIV risk prediction tools and evaluate their effect on prescription of PrEP. FUNDING:Kaiser Permanente Community Benefit Research Program and the US National Institutes of Health.

Project description:ObjectivesBacteremia and fungemia can cause life-threatening illness with high mortality rates, which increase with delays in antimicrobial therapy. The objective of this study is to develop machine learning models to predict blood culture results at the time of the blood culture order using routine data in the electronic health record.DesignRetrospective analysis of a large, multicenter inpatient data.SettingTwo academic tertiary medical centers between the years 2007 and 2018.SubjectsAll hospitalized patients who received a blood culture during hospitalization.InterventionsThe dataset was partitioned temporally into development and validation cohorts: the logistic regression and gradient boosting machine models were trained on the earliest 80% of hospital admissions and validated on the most recent 20%.Measurements and main resultsThere were 252,569 blood culture days-defined as nonoverlapping 24-hour periods in which one or more blood cultures were ordered. In the validation cohort, there were 50,514 blood culture days, with 3,762 cases of bacteremia (7.5%) and 370 cases of fungemia (0.7%). The gradient boosting machine model for bacteremia had significantly higher area under the receiver operating characteristic curve (0.78 [95% CI 0.77-0.78]) than the logistic regression model (0.73 [0.72-0.74]) (p < 0.001). The model identified a high-risk group with over 30 times the occurrence rate of bacteremia in the low-risk group (27.4% vs 0.9%; p < 0.001). Using the low-risk cut-off, the model identifies bacteremia with 98.7% sensitivity. The gradient boosting machine model for fungemia had high discrimination (area under the receiver operating characteristic curve 0.88 [95% CI 0.86-0.90]). The high-risk fungemia group had 252 fungemic cultures compared with one fungemic culture in the low-risk group (5.0% vs 0.02%; p < 0.001). Further, the high-risk group had a mortality rate 60 times higher than the low-risk group (28.2% vs 0.4%; p < 0.001).ConclusionsOur novel models identified patients at low and high-risk for bacteremia and fungemia using routinely collected electronic health record data. Further research is needed to evaluate the cost-effectiveness and impact of model implementation in clinical practice.

Project description:BACKGROUND:Stroke severity is an important predictor of patient outcomes and is commonly measured with the National Institutes of Health Stroke Scale (NIHSS) scores. Because these scores are often recorded as free text in physician reports, structured real-world evidence databases seldom include the severity. The aim of this study was to use machine learning models to impute NIHSS scores for all patients with newly diagnosed stroke from multi-institution electronic health record (EHR) data. METHODS:NIHSS scores available in the Optum© de-identified Integrated Claims-Clinical dataset were extracted from physician notes by applying natural language processing (NLP) methods. The cohort analyzed in the study consists of the 7149 patients with an inpatient or emergency room diagnosis of ischemic stroke, hemorrhagic stroke, or transient ischemic attack and a corresponding NLP-extracted NIHSS score. A subset of these patients (n?=?1033, 14%) were held out for independent validation of model performance and the remaining patients (n?=?6116, 86%) were used for training the model. Several machine learning models were evaluated, and parameters optimized using cross-validation on the training set. The model with optimal performance, a random forest model, was ultimately evaluated on the holdout set. RESULTS:Leveraging machine learning we identified the main factors in electronic health record data for assessing stroke severity, including death within the same month as stroke occurrence, length of hospital stay following stroke occurrence, aphagia/dysphagia diagnosis, hemiplegia diagnosis, and whether a patient was discharged to home or self-care. Comparing the imputed NIHSS scores to the NLP-extracted NIHSS scores on the holdout data set yielded an R2 (coefficient of determination) of 0.57, an R (Pearson correlation coefficient) of 0.76, and a root-mean-squared error of 4.5. CONCLUSIONS:Machine learning models built on EHR data can be used to determine proxies for stroke severity. This enables severity to be incorporated in studies of stroke patient outcomes using administrative and EHR databases.

Project description:BackgroundAccurate, pragmatic risk stratification for postoperative delirium (POD) is necessary to target preventative resources toward high-risk patients. Machine learning (ML) offers a novel approach to leveraging electronic health record (EHR) data for POD prediction. We sought to develop and internally validate a ML-derived POD risk prediction model using preoperative risk features, and to compare its performance to models developed with traditional logistic regression.MethodsThis was a retrospective analysis of preoperative EHR data from 24,885 adults undergoing a procedure requiring anesthesia care, recovering in the main post-anesthesia care unit, and staying in the hospital at least overnight between December 2016 and December 2019 at either of two hospitals in a tertiary care health system. One hundred fifteen preoperative risk features including demographics, comorbidities, nursing assessments, surgery type, and other preoperative EHR data were used to predict postoperative delirium (POD), defined as any instance of Nursing Delirium Screening Scale ≥2 or positive Confusion Assessment Method for the Intensive Care Unit within the first 7 postoperative days. Two ML models (Neural Network and XGBoost), two traditional logistic regression models ("clinician-guided" and "ML hybrid"), and a previously described delirium risk stratification tool (AWOL-S) were evaluated using the area under the receiver operating characteristic curve (AUC-ROC), sensitivity, specificity, positive likelihood ratio, and positive predictive value. Model calibration was assessed with a calibration curve. Patients with no POD assessments charted or at least 20% of input variables missing were excluded.ResultsPOD incidence was 5.3%. The AUC-ROC for Neural Net was 0.841 [95% CI 0. 816-0.863] and for XGBoost was 0.851 [95% CI 0.827-0.874], which was significantly better than the clinician-guided (AUC-ROC 0.763 [0.734-0.793], p < 0.001) and ML hybrid (AUC-ROC 0.824 [0.800-0.849], p < 0.001) regression models and AWOL-S (AUC-ROC 0.762 [95% CI 0.713-0.812], p < 0.001). Neural Net, XGBoost, and ML hybrid models demonstrated excellent calibration, while calibration of the clinician-guided and AWOL-S models was moderate; they tended to overestimate delirium risk in those already at highest risk.ConclusionUsing pragmatically collected EHR data, two ML models predicted POD in a broad perioperative population with high discrimination. Optimal application of the models would provide automated, real-time delirium risk stratification to improve perioperative management of surgical patients at risk for POD.

Project description:RationalePatients transferred from the intensive care unit to the wards who are later readmitted to the intensive care unit have increased length of stay, healthcare expenditure, and mortality compared with those who are never readmitted. Improving risk stratification for patients transferred to the wards could have important benefits for critically ill hospitalized patients.ObjectivesWe aimed to use a machine-learning technique to derive and validate an intensive care unit readmission prediction model with variables available in the electronic health record in real time and compare it to previously published algorithms.MethodsThis observational cohort study was conducted at an academic hospital in the United States with approximately 600 inpatient beds. A total of 24,885 intensive care unit transfers to the wards were included, with 14,962 transfers (60%) in the training cohort and 9,923 transfers (40%) in the internal validation cohort. Patient characteristics, nursing assessments, International Classification of Diseases, Ninth Revision codes from prior admissions, medications, intensive care unit interventions, diagnostic tests, vital signs, and laboratory results were extracted from the electronic health record and used as predictor variables in a gradient-boosted machine model. Accuracy for predicting intensive care unit readmission was compared with the Stability and Workload Index for Transfer score and Modified Early Warning Score in the internal validation cohort and also externally using the Medical Information Mart for Intensive Care database (n = 42,303 intensive care unit transfers).ResultsEleven percent (2,834) of discharges to the wards were later readmitted to the intensive care unit. The machine-learning-derived model had significantly better performance (area under the receiver operating curve, 0.76) than either the Stability and Workload Index for Transfer score (area under the receiver operating curve, 0.65), or Modified Early Warning Score (area under the receiver operating curve, 0.58; P value < 0.0001 for all comparisons). At a specificity of 95%, the derived model had a sensitivity of 28% compared with 15% for Stability and Workload Index for Transfer score and 7% for the Modified Early Warning Score. Accuracy improvements with the derived model over Modified Early Warning Score and Stability and Workload Index for Transfer were similar in the Medical Information Mart for Intensive Care-III cohort.ConclusionsA machine learning approach to predicting intensive care unit readmission was significantly more accurate than previously published algorithms in both our internal validation and the Medical Information Mart for Intensive Care-III cohort. Implementation of this approach could target patients who may benefit from additional time in the intensive care unit or more frequent monitoring after transfer to the hospital ward.

Project description:BackgroundPythia is an automated, clinically curated surgical data pipeline and repository housing all surgical patient electronic health record (EHR) data from a large, quaternary, multisite health institute for data science initiatives. In an effort to better identify high-risk surgical patients from complex data, a machine learning project trained on Pythia was built to predict postoperative complication risk.Methods and findingsA curated data repository of surgical outcomes was created using automated SQL and R code that extracted and processed patient clinical and surgical data across 37 million clinical encounters from the EHRs. A total of 194 clinical features including patient demographics (e.g., age, sex, race), smoking status, medications, comorbidities, procedure information, and proxies for surgical complexity were constructed and aggregated. A cohort of 66,370 patients that had undergone 99,755 invasive procedural encounters between January 1, 2014, and January 31, 2017, was studied further for the purpose of predicting postoperative complications. The average complication and 30-day postoperative mortality rates of this cohort were 16.0% and 0.51%, respectively. Least absolute shrinkage and selection operator (lasso) penalized logistic regression, random forest models, and extreme gradient boosted decision trees were trained on this surgical cohort with cross-validation on 14 specific postoperative outcome groupings. Resulting models had area under the receiver operator characteristic curve (AUC) values ranging between 0.747 and 0.924, calculated on an out-of-sample test set from the last 5 months of data. Lasso penalized regression was identified as a high-performing model, providing clinically interpretable actionable insights. Highest and lowest performing lasso models predicted postoperative shock and genitourinary outcomes with AUCs of 0.924 (95% CI: 0.901, 0.946) and 0.780 (95% CI: 0.752, 0.810), respectively. A calculator requiring input of 9 data fields was created to produce a risk assessment for the 14 groupings of postoperative outcomes. A high-risk threshold (15% risk of any complication) was determined to identify high-risk surgical patients. The model sensitivity was 76%, with a specificity of 76%. Compared to heuristics that identify high-risk patients developed by clinical experts and the ACS NSQIP calculator, this tool performed superiorly, providing an improved approach for clinicians to estimate postoperative risk for patients. Limitations of this study include the missingness of data that were removed for analysis.ConclusionsExtracting and curating a large, local institution's EHR data for machine learning purposes resulted in models with strong predictive performance. These models can be used in clinical settings as decision support tools for identification of high-risk patients as well as patient evaluation and care management. Further work is necessary to evaluate the impact of the Pythia risk calculator within the clinical workflow on postoperative outcomes and to optimize this data flow for future machine learning efforts.

Project description:BACKGROUND:Accurate anesthesiology procedure code data are essential to quality improvement, research, and reimbursement tasks within anesthesiology practices. Advanced data science techniques, including machine learning and natural language processing, offer opportunities to develop classification tools for Current Procedural Terminology codes across anesthesia procedures. METHODS:Models were created using a Train/Test dataset including 1,164,343 procedures from 16 academic and private hospitals. Five supervised machine learning models were created to classify anesthesiology Current Procedural Terminology codes, with accuracy defined as first choice classification matching the institutional-assigned code existing in the perioperative database. The two best performing models were further refined and tested on a Holdout dataset from a single institution distinct from Train/Test. A tunable confidence parameter was created to identify cases for which models were highly accurate, with the goal of at least 95% accuracy, above the reported 2018 Centers for Medicare and Medicaid Services (Baltimore, Maryland) fee-for-service accuracy. Actual submitted claim data from billing specialists were used as a reference standard. RESULTS:Support vector machine and neural network label-embedding attentive models were the best performing models, respectively, demonstrating overall accuracies of 87.9% and 84.2% (single best code), and 96.8% and 94.0% (within top three). Classification accuracy was 96.4% in 47.0% of cases using support vector machine and 94.4% in 62.2% of cases using label-embedding attentive model within the Train/Test dataset. In the Holdout dataset, respective classification accuracies were 93.1% in 58.0% of cases and 95.0% among 62.0%. The most important feature in model training was procedure text. CONCLUSIONS:Through application of machine learning and natural language processing techniques, highly accurate real-time models were created for anesthesiology Current Procedural Terminology code classification. The increased processing speed and a priori targeted accuracy of this classification approach may provide performance optimization and cost reduction for quality improvement, research, and reimbursement tasks reliant on anesthesiology procedure codes.