Dataset Information

Algorithms identifying low-acuity emergency department visits: A review and validation study.

ABSTRACT:

Objective

To characterize and validate the landscape of algorithms that use International Classification of Disease (ICD) codes to identify low-acuity emergency department (ED) visits.

Data sources

Publicly available ED data from the National Hospital Ambulatory Medical Care Survey (NHAMCS).

Study design

We systematically searched for studies that specify algorithms consisting of ICD codes that identify preventable or low-acuity ED visits. We classified ED visits in NHAMCS according to these algorithms and compared agreements using the Jaccard index. We then evaluated the performance of each algorithm using positive predictive value (PPV) and sensitivity, with the reference group specified using low-acuity composite (LAC) criteria consisting of both triage and clinical components. In sensitivity analyses, we repeated our primary analysis using only triage or only clinical criteria for reference.

Data collection

We used the 2011-2017 NHAMCS data, totaling 163,576 observations before survey weighting and after dropping observations missing a primary diagnosis. We translated ICD-9 codes (years 2011-2015) to ICD-10 using a standard crosswalk.

Principal findings

We identified 15 papers with an original list of ICD codes used to identify preventable or low-acuity ED presentations. These papers were published between 1992 and 2020, cited an average of 310 (SD 360) times, and included 968 (SD 1175) codes. Pairwise Jaccard similarity indices (0 = no overlap, 1 = perfect congruence) ranged from 0.01 to 0.82, with mean 0.20 (SD 0.13). When validated against the LAC reference group, the algorithms had an average PPV of 0.308 (95% CI [0.253, 0.364]) and sensitivity of 0.183 (95% CI [0.111, 0.256]). Overall, 2.1% of visits identified as low acuity by the algorithms died prehospital or in the ED, or needed surgery, critical care, or cardiac catheterization.

Conclusions

Existing algorithms that identify low-acuity ED visits lack congruence and are imperfect predictors of visit acuity.

SUBMITTER: Chen AT

PROVIDER: S-EPMC9264468 | biostudies-literature | 2022 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Algorithms identifying low-acuity emergency department visits: A review and validation study.

Chen Angela T AT Muralidharan Madhavi M Friedman Ari B AB

Health services research 20220606 4

<h4>Objective</h4>To characterize and validate the landscape of algorithms that use International Classification of Disease (ICD) codes to identify low-acuity emergency department (ED) visits.<h4>Data sources</h4>Publicly available ED data from the National Hospital Ambulatory Medical Care Survey (NHAMCS).<h4>Study design</h4>We systematically searched for studies that specify algorithms consisting of ICD codes that identify preventable or low-acuity ED visits. We classified ED visits in NHAMCS ...[more]

PMID: 35619335

Similar Datasets

Project description:ObjectiveTo improve the performance of International Classification of Disease (ICD) code rule-based algorithms for identifying low acuity Emergency Department (ED) visits by using machine learning methods and additional covariates.Data sourcesWe used secondary data on ED visits from the National Hospital Ambulatory Medical Survey (NHAMCS), from 2016 to 2020.Study designWe established baseline performance metrics with seven published algorithms consisting of International Classification of Disease, Tenth Revision codes used to identify low acuity ED visits. We then trained logistic regression, random forest, and gradient boosting (XGBoost) models to predict low acuity ED visits. Each model was trained on five different covariate sets of demographic and clinical data. Model performance was compared using a separate validation dataset. The primary performance metric was the probability that a visit identified by an algorithm as low acuity did not experience significant testing, treatment, or disposition (positive predictive value, PPV). Subgroup analyses assessed model performance across age, sex, and race/ethnicity.Data collectionWe used 2016-2019 NHAMCS data as the training set and 2020 NHAMCS data for validation.Principal findingsThe training and validation data consisted of 53,074 and 9542 observations, respectively. Among seven rule-based algorithms, the highest-performing had a PPV of 0.35 (95% CI [0.33, 0.36]). All model-based algorithms outperformed existing algorithms, with the least effective-random forest using only age and sex-improving PPV by 26% (up to 0.44; 95% CI [0.40, 0.48]). Logistic regression and XGBoost trained on all variables improved PPV by 83% (to 0.64; 95% CI [0.62, 0.66]). Multivariable models also demonstrated higher PPV across all three demographic subgroups.ConclusionsMachine learning models substantially outperform existing algorithms based on ICD codes in predicting low acuity ED visits. Variations in model performance across demographic groups highlight the need for further research to ensure their applicability and fairness across diverse populations.

Project description:BackgroundCanadian pediatric emergency department visits are increasing, with a disproportionate increase in low-acuity visits locally (33% of volume in 2008-09, 41% in 2011-12). We sought to understand: 1) presentation patterns and resource implications; 2) parents' perceptions and motivations; and 3) alternate health care options considered prior to presenting with low-acuity problems.MethodsWe conducted a prospective cohort study at our tertiary pediatric emergency department serving two provinces to explore differences between patients with and without a primary care provider. During four, 2-week study periods over 1 year, parents of low-acuity visits received an anonymous survey. Presentation times, interventions, diagnoses and dispositions were captured on a data collection form linked to the survey by study number.ResultsParents completed 2,443 surveys (74.1% response rate), with survey-data collection form pairs available for 2,146 visits. Overall, 89.7% of respondents had a primary care provider; 68% were family physicians. Surprisingly, 40% of visits occurred during weekday office hours and 27.3% occurred within 4 hours of symptom onset; 67.5% of those early presenters were for injuries. Few parents sought care from their primary care provider (25%), health information line (20.7%), or urgent care clinic (18.5%); 36% reported that they believed their child's problem required the emergency department. Forty-five percent required only a history, physical exam and reassurance; only 11% required an intervention not available in an office setting. Patients without a primary care provider were significantly more likely to present during weekday office hours (p = 0.003), have longer symptom duration (p<0.001), and not know of other options (p = 0.001).ConclusionsMany parents seek pediatric emergency department care for low-acuity problems despite their child having a primary care provider. Ensuring timely access to these providers may help reduce pediatric emergency department overuse. Educational initiatives should inform parents about low-acuity problems and where appropriate care can/should be accessed.

Project description:PurposeLow-acuity pediatric emergency department (PED) visits are frequent in high-income countries and have a negative impact on patient care at the individual and health system levels. Knowing what drives low-acuity PED visits is crucial to inform adaptations in health care delivery. We aimed to identify factors associated with low-acuity PED visits in Switzerland, including socioeconomic status, demographic features, and medical resources of families.MethodsWe conducted a prospective, questionnaire-based study in the PEDs of two Swiss tertiary care hospitals, Bern and Lausanne. We invited all consecutive children and their caregiver attending the PED during data collection times representative of the overall PED consultation structure (e.g. day/night, weekdays/weekends) to complete a questionnaire on demographic features, socioeconomic status, and medical resources. We collected medical and administrative data about the visit and defined low-acuity visits as those meeting all of the following criteria: (1) triage category 4 or 5 on the Australasian Triage Scale, (2) no imaging or laboratory test performed, and (3) discharge home. We used a binary multiple logistic regression model to identify factors associated with low-acuity visits.ResultsWe analysed 778 PED visits (September 2019 to July 2020). Most children visiting our PEDs had a designated primary care provider (92%), with only 6% not having seen them during the last year. Fifty-five per cent of caregivers had asked for medical advice before coming to the PED. The proportion of low-acuity visits was 58%. Low-acuity visits were associated with caregiver's difficulties paying bills (aOR 2.6, 95% CI 1.6 - 4.4), having already visited a PED in the last 6 months (aOR 1.7, 95% CI 1.1 - 2.5) but not with parental education status, nor parental country of birth, parental employment status or absence of family network.ConclusionEconomic precariousness is an important driver for low-acuity PED visits in Switzerland, a high-income country with compulsory health coverage where most children have a designated primary care provider and a regular pediatric follow-up. Primary care providers and PEDs should screen families for economic precariousness and offer anticipatory guidance and connect those in financial need to social support.

Dataset Information

Algorithms identifying low-acuity emergency department visits: A review and validation study.

Objective

Data sources

Study design

Data collection

Principal findings

Conclusions

Publications

Algorithms identifying low-acuity emergency department visits: A review and validation study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets