Dataset Information

Training machine learning models to predict 30-day mortality in patients discharged from the emergency department: a retrospective, population-based registry study.

ABSTRACT:

Objectives

The aim of this work was to train machine learning models to identify patients at end of life with clinically meaningful diagnostic accuracy, using 30-day mortality in patients discharged from the emergency department (ED) as a proxy.

Design

Retrospective, population-based registry study.

Setting

Swedish health services.

Primary and secondary outcome measures

All cause 30-day mortality.

Methods

Electronic health records (EHRs) and administrative data were used to train six supervised machine learning models to predict all-cause mortality within 30 days in patients discharged from EDs in southern Sweden, Europe.

Participants

The models were trained using 65 776 ED visits and validated on 55 164 visits from a separate ED to which the models were not exposed during training.

Results

The outcome occurred in 136 visits (0.21%) in the development set and in 83 visits (0.15%) in the validation set. The model with highest discrimination attained ROC-AUC 0.95 (95% CI 0.93 to 0.96), with sensitivity 0.87 (95% CI 0.80 to 0.93) and specificity 0.86 (0.86 to 0.86) on the validation set.

Conclusions

Multiple models displayed excellent discrimination on the validation set and outperformed available indexes for short-term mortality prediction interms of ROC-AUC (by indirect comparison). The practical utility of the models increases as the data they were trained on did not require costly de novo collection but were real-world data generated as a by-product of routine care delivery.

SUBMITTER: Blom MC

PROVIDER: S-EPMC6701621 | biostudies-literature | 2019 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Training machine learning models to predict 30-day mortality in patients discharged from the emergency department: a retrospective, population-based registry study.

Blom Mathias Carl MC Ashfaq Awais A Sant'Anna Anita A Anderson Philip D PD Lingman Markus M

BMJ open 20190810 8

<h4>Objectives</h4>The aim of this work was to train machine learning models to identify patients at end of life with clinically meaningful diagnostic accuracy, using 30-day mortality in patients discharged from the emergency department (ED) as a proxy.<h4>Design</h4>Retrospective, population-based registry study.<h4>Setting</h4>Swedish health services.<h4>Primary and secondary outcome measures</h4>All cause 30-day mortality.<h4>Methods</h4>Electronic health records (EHRs) and administrative dat ...[more]

PMID: 31401594

Similar Datasets

Project description:BackgroundHypertension (HTN) is common in discharged emergency department (ED) patients, yet the short-term outcomes of treating HTN at ED discharge are unclear. This study aimed to investigate whether emergency physician (EP) prescription of oral antihypertensive therapy at ED discharge for hypertensive patients is associated with a decreased 30-day risk of the severe adverse events (AEs), death, and revisits to the ED.MethodsWe conducted an observational cohort study assessing the 30-day outcomes of discharged ED patients with HTN, comparing outcomes based on whether antihypertensive therapy was prescribed. All discharged adult ED patients from an eight-hospital system with a diagnosis of HTN from January 2016 to February 2020 were screened, and consisted of a mix of suburban and urban patients with broad ethnic and socioeconomic backgrounds. Patients were categorized into the treatment group if they received a prescription for antihypertensive medication at ED discharge. The primary outcome was severe composite AEs from HTN (aortic catastrophe, heart failure, myocardial infarction, hemorrhagic and ischemic stroke, or hypertensive encephalopathy) within 30 days of ED discharge. The secondary outcomes were death or ED revisit over the same period.ResultsThe study sample consisted of 93,512 ED visits; 57.5% were female, and mean age was 59.3 years. 4.7% of patients were prescribed antihypertensive treatment at ED discharge. Within 30 days, 0.7% of patients experienced an AE, 0.1% died, and 15.2% had an ED revisit. The treatment group had significantly lower odds of AE (adjusted odds ratio [aOR]: 0.224, 95%CI 0.106-0.416, p < 0.001), and ED revisits (aOR: 0.610, 95%CI 0.547-0.678, p < 0.001), adjusting for age, race, degree of HTN, ED treatment for elevated HTN, Elixhauser comorbidity index, and heart failure history. There was no difference in odds of death 30 days after discharge.Conclusion and relevancePrescription antihypertensive therapy for discharged ED patients is associated with a 30-day decrease in severe adverse events and ED revisit rate.

Project description:ObjectivesPublished literature on national emergency department (ED) revisit rates among older adults with dementia is sparse, despite anecdotal evidence of higher ED utilization. Thus we evaluated the odds ratio (OR) of 30-day ED revisits among older adults with dementia using a nationally representative sample.DesignWe assessed the frequency of claims associated with a 30-day ED revisit among Medicare beneficiaries with and without a dementia diagnosis before or at index ED visit. We used a logistic regression model controlling for dementia, age, sex, race, region, Medicaid status, transfer to a skilled nursing facility after ED, primary care physician use 12 months before index, and comorbidity.SettingA nationally representative sample of claims data for Medicare beneficiaries aged 65 and older who maintained continuous fee-for-service enrollment during 2015 and 2016. Only outpatient claims associated with an ED visit between January 2016 and November 2016 were included as a qualifying index encounter.ParticipantsWe identified 240 249 patients without dementia and 54 622 patients for whom a dementia code was recorded in the year before the index encounter in 2016.ResultsOur results indicate a significant difference in unadjusted 30-day ED revisit rates among those with an ED dementia diagnoses (22.0%) compared with those without (13.9%). Our adjusted results indicated that dementia is a significant predictor of 30-day ED revisits (P < .0001). Those with a dementia diagnosis at or before the index ED visit were more likely to have experienced an ED revisit within 30 days (OR = 1.27; 95% confidence interval = 1.24-1.31).ConclusionDementia diagnoses were a significant predictor of 30-day ED revisits. Further research should assess potential reasons why dementia is associated with markedly higher revisit rates, as well as opportunities to manage and transition dementia patients from the ED back to the community more effectively. J Am Geriatr Soc 67:2254-2259, 2019.

Project description:BackgroundUrgency triage in the emergency department (ED) is important for early identification of potentially lethal conditions and extensive resource utilization. However, in older patients, urgency triage systems could be improved by taking geriatric vulnerability into account. We investigated the association of geriatric vulnerability screening in addition to triage urgency levels with 30-day mortality in older ED patients.DesignSecondary analysis of the observational multicenter Acutely Presenting Older Patient (APOP) study.SettingEDs within four Dutch hospitals.ParticipantsConsecutive patients, aged 70 years or older, who were prospectively included.MeasurementsPatients were triaged using the Manchester Triage System (MTS). In addition, the APOP screener was used as a geriatric screening tool. The primary outcome was 30-day mortality. Comparison was made between mortality within the geriatric high- and low-risk screened patients in every urgency triage category. We calculated the difference in explained variance of mortality by adding the geriatric screener (APOP) to triage urgency (MTS) by calculating Nagelkerke R2 .ResultsWe included 2,608 patients with a median age of 79 (interquartile range = 74-84) years, of whom 521 (20.0%) patients were categorized as high risk according to geriatric screening. Patients were triaged on urgency as standard (27.2%), urgent (58.5%), and very urgent (14.3%). In total, 132 (5.1%) patients were deceased within a period of 30 days. Within every urgency triage category, 30-day mortality was threefold higher in geriatric high-risk compared to low-risk patients (overall = 11.7% vs 3.4%; P < .001). The explained variance of 30-day mortality with triage urgency was 1.0% and increased to 6.3% by adding the geriatric screener.ConclusionCombining triage urgency with geriatric screening has the potential to improve triage, which may help clinicians to deliver early appropriate care to older ED patients. J Am Geriatr Soc 68:1755-1762, 2020.

Project description:ImportanceTrauma of hospitalization refers to the depersonalizing and stressful experience of a hospital admission and is hypothesized to increase the risk of readmission after discharge.ObjectivesTo characterize the trauma of hospitalization by measuring patient-reported disturbances in sleep, mobility, nutrition, and mood among medical inpatients, and to examine the association between these disturbances and the risk of unplanned return to hospital after discharge.Design, setting, and participantsThis prospective cohort study enrolled participants between September 1, 2016, and September 1, 2017, at 2 academic hospitals in Toronto, Canada. Participants were adults admitted to the internal medicine ward for more than 48 hours. Participants were interviewed before discharge using a standardized questionnaire to assess sleep, mobility, nutrition, and mood. Responses for each domain were dichotomized as disturbance or no disturbance. Disturbance in 3 or 4 domains (the upper tertile) was considered high trauma of hospitalization, and disturbance in 0 to 2 domains (the lower 2 tertiles) was considered low trauma.Main outcome and measuresThe primary outcome was readmission or emergency department visit within 30 days of discharge. The association between trauma of hospitalization and the primary outcome was examined using logistic regression, adjusted for age; sex; length of stay; Charlson Comorbidity Index Score; Laboratory-Based Acute Physiology Score; and baseline disturbances in sleep, mobility, nutrition, and mood.ResultsA total of 207 patients participated, of whom 82 (39.6%) were women and 125 (60.4%) were men, with a mean (SD) age of 60.3 (16.8) years. Among the 207 participants, 75 (36.2%) reported sleep disturbance, 162 (78.3%) reported mobility disturbance, 114 (55.1%) reported nutrition disturbance, and 48 (23.2%) reported mood disturbance. Nearly all participants (192 [92.8%]) described a disturbance in at least 1 domain, and 61 participants (29.5%) had high trauma exposure. A statistically significant 15.8% greater absolute risk of readmission or emergency department visit was found in participants with high trauma (37.7%; 95% CI, 25.9%-51.1%) compared with those with low trauma (21.9%; 95% CI, 15.7%-29.7%), which remained statistically significant after adjusting for baseline characteristics (adjusted odds ratio, 2.52; 95% CI, 1.24-5.17; P = .01) and propensity score matching (odds ratio, 2.47; 95% CI, 1.11-5.73; P = .03).Conclusions and relevanceDisturbances in sleep, mobility, nutrition, and mood were common in medical inpatients; such trauma of hospitalization may be associated with a greater risk of 30-day readmission or emergency department visit after hospital discharge.

Project description:ObjectivesTo predict 72-h and 9-day emergency department (ED) return by using gradient boosting on an expansive set of clinical variables from the electronic health record.MethodsThis retrospective study included all adult discharges from a level 1 trauma center ED and a community hospital ED covering the period of March 2013 to July 2017. A total of 1500 variables were extracted for each visit, and samples split randomly into training, validation, and test sets (80%, 10%, and 10%). Gradient boosting models were fit on 3 selections of the data: administrative data (demographics, prior hospital usage, and comorbidity categories), data available at triage, and the full set of data available at discharge. A logistic regression (LR) model built on administrative data was used for baseline comparison. Finally, the top 20 most informative variables identified from the full gradient boosting models were used to build a reduced model for each outcome.ResultsA total of 330 631 discharges were available for analysis, with 29 058 discharges (8.8%) resulting in 72-h return and 52 748 discharges (16.0%) resulting in 9-day return to either ED. LR models using administrative data yielded test AUCs of 0.69 (95% confidence interval [CI] 0.68-0.70) and 0.71(95% CI 0.70-0.72), while gradient boosting models using administrative data yielded test AUCs of 0.73 (95% CI 0.72-0.74) and 0.74 (95% CI 0.73-0.74) for 72-h and 9-day return, respectively. Gradient boosting models using variables available at triage yielded test AUCs of 0.75 (95% CI 0.74-0.76) and 0.75 (95% CI 0.74-0.75), while those using the full set of variables yielded test AUCs of 0.76 (95% CI 0.75-0.77) and 0.75 (95% CI 0.75-0.76). Reduced models using the top 20 variables yielded test AUCs of 0.73 (95% CI 0.71-0.74) and 0.73 (95% CI 0.72-0.74).Discussion and conclusionGradient boosting models leveraging clinical data are superior to LR models built on administrative data at predicting 72-h and 9-day returns.

Dataset Information

Training machine learning models to predict 30-day mortality in patients discharged from the emergency department: a retrospective, population-based registry study.

Objectives

Design

Setting

Primary and secondary outcome measures

Methods

Participants

Results

Conclusions

Publications

Training machine learning models to predict 30-day mortality in patients discharged from the emergency department: a retrospective, population-based registry study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets