Predicting Inpatient Medication Orders From Electronic Health Record Data.
Ontology highlight
ABSTRACT:
SUBMITTER: Rough K
PROVIDER: S-EPMC7325318 | biostudies-literature | 2020 Jul
REPOSITORIES: biostudies-literature
Publications
Predicting Inpatient Medication Orders From Electronic Health Record Data.
Rough Kathryn K Dai Andrew M AM Zhang Kun K Xue Yuan Y Vardoulakis Laura M LM Cui Claire C Butte Atul J AJ Howell Michael D MD Rajkomar Alvin A
Clinical pharmacology and therapeutics 20200411 1
In a general inpatient population, we predicted patient-specific medication orders based on structured information in the electronic health record (EHR). Data on over three million medication orders from an academic medical center were used to train two machine-learning models: A deep learning sequence model and a logistic regression model. Both were compared with a baseline that ranked the most frequently ordered medications based on a patient's discharge hospital service and amount of time sin ...[more]
Similar Datasets
Project description:BACKGROUND:Readmissions after hospitalization for pneumonia are common, but the few risk-prediction models have poor to modest predictive ability. Data routinely collected in the electronic health record (EHR) may improve prediction. OBJECTIVE:To develop pneumonia-specific readmission risk-prediction models using EHR data from the first day and from the entire hospital stay ("full stay"). DESIGN:Observational cohort study using stepwise-backward selection and cross-validation. SUBJECTS:Consecutive pneumonia hospitalizations from 6 diverse hospitals in north Texas from 2009-2010. MEASURES:All-cause nonelective 30-day readmissions, ascertained from 75 regional hospitals. RESULTS:Of 1463 patients, 13.6% were readmitted. The first-day pneumonia-specific model included sociodemographic factors, prior hospitalizations, thrombocytosis, and a modified pneumonia severity index; the full-stay model included disposition status, vital sign instabilities on discharge, and an updated pneumonia severity index calculated using values from the day of discharge as additional predictors. The full-stay pneumonia-specific model outperformed the first-day model (C statistic 0.731 vs 0.695; P = 0.02; net reclassification index = 0.08). Compared to a validated multi-condition readmission model, the Centers for Medicare and Medicaid Services pneumonia model, and 2 commonly used pneumonia severity of illness scores, the full-stay pneumonia-specific model had better discrimination (C statistic range 0.604-0.681; P < 0.01 for all comparisons), predicted a broader range of risk, and better reclassified individuals by their true risk (net reclassification index range, 0.09-0.18). CONCLUSIONS:EHR data collected from the entire hospitalization can accurately predict readmission risk among patients hospitalized for pneumonia. This approach outperforms a first-day pneumonia-specific model, the Centers for Medicare and Medicaid Services pneumonia model, and 2 commonly used pneumonia severity of illness scores. Journal of Hospital Medicine 2017;12:209-216.
| S-EPMC6296251 | biostudies-literature
Project description:ImportanceSecondary prevention interventions to reduce post-stroke cognitive impairment (PSCI) can be aided by the early identification of high-risk individuals who would benefit from risk factor modification.ObjectiveTo develop and evaluate a predictive model to identify patients at increased risk of PSCI over 5 years using data easily accessible from electronic health records.DesignCohort study with patients enrolled between 2003-2016 with follow-up through 2022.SettingPrimary care practices affiliated with two academic medical centers.ParticipantsIndividuals 45 years or older, without prior stroke or prevalent cognitive impairment, with primary care visits and an incident ischemic stroke between 2003-2016 (development/internal validation cohort) or 2010-2022 (external validation cohort).ExposuresPredictors of PSCI were ascertained from the electronic health record.Main outcomeThe outcome was incident dementia/cognitive impairment within 5 years and beginning 3 months following stroke, ascertained using ICD-9/10 codes. For model variable selection, we considered potential predictors of PSCI and constructed 400 bootstrap samples with two-thirds of the model derivation sample. We ran 10-fold cross-validated Cox proportional hazards models using a least absolute shrinkage and selection operator (LASSO) penalty. Variables selected in >25% of samples were included.ResultsThe analysis included 332 incident diagnoses of PSCI in the development cohort (n=3,741), and 161 and 128 incident diagnoses in the internal (n=1,925) and external (n=2,237) validation cohorts. The c-statistic for predicting PSCI was 0.731 (95% CI: 0.694-0.768) in the internal validation cohort, and 0.724 (95% CI: 0.681-0.766) in the external validation cohort. A risk score based on the beta coefficients of predictors from the development cohort stratified patients into low (0-7 points), intermediate (8-11 points), and high (12-35 points) risk groups. The hazard ratios for incident PSCI were significantly different by risk categories in internal (High, HR: 6.2, 95% CI 4.1-9.3; Intermediate, HR 2.7, 95% CI: 1.8-4.1) and external (High, HR: 6.1, 95% CI: 3.9-9.6; Intermediate, HR 2.8, 95% CI: 1.9-4.3) validation cohorts.Conclusions and relevanceFive-year risk of PSCI can be accurately predicted using routinely collected data. Model output can be used to risk stratify and identify individuals at increased risk for PSCI for preventive efforts.
| S-EPMC10863024 | biostudies-literature
Project description:BackgroundElectronic health record (EHR) system transitions are challenging for healthcare organizations. High-volume, safety-critical tasks like barcode medication administration (BCMA) should be evaluated, yet standards for ensuring safety during transition have not been established.ObjectiveIdentify risks in common and problem-prone medication tasks to inform safe transition between BCMA systems and establish benchmarks for future system changes.DesignStaff nurses completed simulation-based usability testing in the legacy system (R1) and new system pre- (R2) and post-go-live (R3). Tasks included (1) Hold/Administer, (2) IV Fluids, (3) PRN Pain, (4) Insulin, (5) Downtime/PRN, and (6) Messaging. Audiovisual recordings of task performance were systematically analyzed for time, navigation, and errors. The System Usability Scale measured perceived usability and satisfaction. Post-simulation interviews captured nurses' qualitative comments and perceptions of the systems.ParticipantsFifteen staff nurses completed 2-3-h simulation sessions. Eleven completed both R1 and R2, and seven completed all three rounds. Clinical experience ranged from novice (< 1 year) to experienced (> 10 years). Practice settings included adult and pediatric patient populations in ICU, stepdown, and acute care departments.Main measuresTask completion rates/times, safety and non-safety-related use errors (interaction difficulties), and user satisfaction.Key resultsOverall success rates remained relatively stable in all tasks except two: IV Fluids task success increased substantially (R1: 17%, R2: 54%, R3: 100%) and Downtime/PRN task success decreased (R1: 92%, R2: 64%, R3: 22%). Among the seven nurses who completed all rounds, overall safety-related errors decreased 53% from R1 to R3 and 50% from R2 to R3, and average task times for successfully completed tasks decreased 22% from R1 to R3 and 38% from R2 to R3.ConclusionsUsability testing is a reasonable approach to compare different BCMA tasks to anticipate transition problems and establish benchmarks with which to monitor and evaluate system changes going forward.
| S-EPMC10593701 | biostudies-literature
Project description:Predictive analytics with electronic health record (EHR) data holds promise for improving outcomes of psychiatric care. This study evaluated models for predicting outcomes of psychotherapy for depression in a clinical practice setting. EHR data from two large integrated health systems (Kaiser Permanente Colorado and Washington) included 5,554 new psychotherapy episodes with a baseline Patient Health Questionnaire (PHQ-9) score ≥ 10 and a follow-up PHQ-9 14-180 days after treatment initiation. Baseline predictors included demographics and diagnostic, medication, and encounter history. Prediction models for two outcomes-follow-up PHQ-9 score and treatment response (≥ 50% PHQ-9 reduction)-were trained in a random sample of 70% of episodes and validated in the remaining 30%. Two methods were used for modeling: generalized linear regression models with variable selection and random forests. Sensitivity analyses considered alternate predictor, outcome, and model specifications. Predictions of follow-up PHQ-9 scores poorly estimated observed outcomes (mean squared error = 31 for linear regression, 40 for random forest). Predictions of treatment response had low discrimination (AUC = 0.57 for logistic regression, 0.61 for random forest), low classification accuracy, and poor calibration. Sensitivity analyses showed similar results. We note that prediction model performance may vary for settings with different care or EHR documentation practices. In conclusion, prediction models did not accurately predict depression treatment outcomes despite using rich EHR data and advanced analytic techniques. Health systems should proceed cautiously when considering prediction models for psychiatric outcomes using baseline intake information. Transparent research should be conducted to evaluate performance of any model intended for clinical use.
| S-EPMC8448296 | biostudies-literature
Project description:ObjectivesSince the 1970s, a plethora of tools have been introduced to support the medication use process. However, automation initiatives to assist pharmacists in prospectively reviewing medication orders are lacking. The review of many medications may be protocolized and implemented in an algorithmic fashion utilizing discrete data from the electronic health record (EHR). This research serves as a proof of concept to evaluate the capability and effectiveness of an electronic prospective medication order review (EPMOR) system compared to pharmacists' review.Materials and methodsA subset of the most frequently verified medication orders were identified for inclusion. A team of clinical pharmacist experts developed best-practice EPMOR criteria. The established criteria were incorporated into conditional logic built within the EHR. Verification outcomes from the pharmacist (human) and EPMOR (automation) were compared.ResultsOverall, 13 404 medication orders were included. Of those orders, 13 133 passed pharmacist review, 7388 of which passed EPMOR. A total of 271 medication orders failed pharmacist review due to order modification or discontinuation, 105 of which passed EPMOR. Of the 105 orders, 19 were duplicate orders correctly caught by both EPMOR and pharmacists, but the opposite duplicate order was rejected, 51 orders failed due to scheduling changes.DiscussionThis simulation was capable of effectively discriminating and triaging orders. Protocolization and automation of the prospective medication order review process in the EHR appear possible using clinically driven algorithms.ConclusionFurther research is necessary to refine such algorithms to maximize value, improve efficiency, and minimize safety risks in preparation for the implementation of fully automated systems.
| S-EPMC10822119 | biostudies-literature
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.
| S-EPMC6207111 | biostudies-literature
Project description:BackgroundGuidelines recommend breast and colorectal cancer screening for older adults with a life expectancy >10 years. Most mortality indexes require clinician data entry, presenting a barrier for routine use in care. Electronic health records (EHR) are a rich clinical data source that could be used to create individualized life expectancy predictions to identify patients for cancer screening without data entry.ObjectiveTo develop and internally validate a life expectancy calculator from structured EHR data.DesignRetrospective cohort study using national Veteran's Affairs (VA) EHR databases.PatientsVeterans aged 50+ with a primary care visit during 2005.Main measuresWe assessed demographics, diseases, medications, laboratory results, healthcare utilization, and vital signs 1 year prior to the index visit. Mortality follow-up was complete through 2017. Using the development cohort (80% sample), we used LASSO Cox regression to select ~100 predictors from 913 EHR data elements. In the validation cohort (remaining 20% sample), we calculated the integrated area under the curve (iAUC) and evaluated calibration.Key resultsIn 3,705,122 patients, the mean age was 68 years and the majority were male (97%) and white (85%); nearly half (49%) died. The life expectancy calculator included 93 predictors; age and gender most strongly contributed to discrimination; diseases also contributed significantly while vital signs were negligible. The iAUC was 0.816 (95% confidence interval, 0.815, 0.817) with good calibration.ConclusionsWe developed a life expectancy calculator using VA EHR data with excellent discrimination and calibration. Automated life expectancy prediction using EHR data may improve guideline-concordant breast and colorectal cancer screening by identifying patients with a life expectancy >10 years.
| S-EPMC8858374 | biostudies-literature
Project description:BackgroundMedication-wide association studies (MWAS) have been applied to assess the risk of individual prescription use and a wide range of health outcomes, including cancer, acute myocardial infarction, acute liver failure, acute renal failure, and upper gastrointestinal ulcers. Current literature on the use of preconception and periconception medication and its association with the risk of multiple gestation pregnancies (eg, monozygotic and dizygotic) is largely based on assisted reproductive technology (ART) cohorts. However, among non-ART pregnancies, it is unknown whether other medications increase the risk of multifetal pregnancies.ObjectiveThis study aimed to investigate the risk of multiple gestational births (eg, twins and triplets) following preconception and periconception exposure to prescription medications in patients who delivered at Penn Medicine.MethodsWe used electronic health record data between 2010 and 2017 on patients who delivered babies at Penn Medicine, a health care system in the Greater Philadelphia area. We explored 3 logistic regression models: model 1 (no adjustment); model 2 (adjustment for maternal age); and model 3-our final logistic regression model (adjustment for maternal age, ART use, and infertility diagnosis). In all models, multiple births (MBs) were our outcome of interest (binary outcome), and each medication was assessed separately as a binary variable. To assess our MWAS model performance, we defined ART medications as our gold standard, given that these medications are known to increase the risk of MB.ResultsOf the 63,334 distinct deliveries in our cohort, only 1877 pregnancies (2.96%) were prescribed any medication during the preconception and first trimester period. Of the 123 medications prescribed, we found 26 (21.1%) medications associated with MB (using nominal P values) and 10 (8.1%) medications associated with MB (using Bonferroni adjustment) in fully adjusted model 3. We found that our model 3 algorithm had an accuracy of 85% (using nominal P values) and 89% (using Bonferroni-adjusted P values).ConclusionsOur work demonstrates the opportunities in applying the MWAS approach with electronic health record data to explore associations between preconception and periconception medication exposure and the risk of MB while identifying novel candidate medications for further study. Overall, we found 3 novel medications linked with MB that could be explored in further work; this demonstrates the potential of our method to be used for hypothesis generation.
| S-EPMC9214620 | biostudies-literature
Project description: Not available
| S-EPMC10751588 | biostudies-literature
Project description:There is controversy over the impact of electronic health record (EHR) systems on cost of care and safety. The authors studied the effects of an inpatient EHR system with computerized provider order entry on selected measures of cost of care and safety. Laboratory tests per week per hospitalization decreased from 13.9 to 11.4 (18%; p < 0.001). Radiology examinations per hospitalization decreased from 2.06 to 1.93 (6.3%; p < 0.009). Monthly transcription costs declined from $74,596 to $18,938 (74.6%; p < 0.001). Reams of copy paper ordered per month decreased from 1668 to 1224 (26.6%; p < 0.001). Medication errors per 1000 hospital days decreased from 17.9 to 15.4 (14.0%; p < 0.030), while near misses per 1000 hospital days increased from 9.0 to 12.5 (38.9%; p < 0.037), and the percentage of medication events that were medication errors decreased from 66.5% to 55.2% (p < 0.007). In this manuscript, we demonstrate that the implementation of an inpatient EHR with computerized provider order entry can result in rapid improvement in measures of cost of care and safety.
| S-EPMC3116263 | biostudies-literature