Search

Aims: Worsening heart failure (WHF) events occurring in non‐inpatient settings are becoming increasingly recognized, with implications for prognostication. We evaluate the performance of a natural language processing (NLP)‐based approach compared with traditional diagnostic coding for non‐inpatient clinical encounters and left ventricular ejection fraction (LVEF). Methods and results: We compared characteristics for encounters that did vs. did not meet WHF criteria, stratified by care setting [i.e. emergency department (ED) and observation stay]. Overall, 8407 (22%) encounters met NLP‐based criteria for WHF (3909 ED visits and 4498 observation stays). The use of an NLP‐derived definition adjudicated 3983 (12%) of non‐primary HF diagnoses as meeting consensus definitions for WHF. The most common diagnosis indicated in these encounters was dyspnoea. Results were primarily driven by observation stays, in which 2205 (23%) encounters with a secondary HF diagnosis met the WHF definition by NLP. Conclusions: The use of standard claims‐based adjudication for primary diagnosis in the non‐inpatient setting may lead to misclassification of WHF events in the ED and overestimate observation stays. Primary diagnoses alone may underestimate the burden of WHF in non‐hospitalized settings. [ABSTRACT FROM AUTHOR]

Hospital admissions for patients with acute heart failure (AHF) remain high. There is an opportunity to improve alignment between patient risk and admission decision. We recently developed a machine learning (ML)-based model that stratifies emergency department (ED) patients with AHF based on predicted risk of a 30-day severe adverse event. Prior to deploying the algorithm and paired clinical decision support, we sought to understand barriers and opportunities regarding successful implementation. We conducted semi-structured interviews with eight front-line ED providers and surveyed 67 ED providers. Audio-recorded interviews were transcribed and analyzed using thematic analysis, and we had a 65% response rate to the survey. Providers wanted decision support to be streamlined into workflows with minimal disruptions. Most providers wanted assistance primarily with ED disposition decisions, and secondarily with medical management and post-discharge follow-up care. Receiving feedback on patient outcomes after risk tool use was seen as an opportunity to increase acceptance, and few providers (<10%) had significant hesitations with using an ML-based tool after education on its use. Engagement with key front-line users on optimal design of the algorithm and decision support may contribute to broader uptake, acceptance, and adoption of recommendations for clinical decisions. [ABSTRACT FROM AUTHOR]

INTRODUCTION: We sought to investigate the association between receipt of an opioid pain reliever (OPR) in the emergency department (ED) and downstream acute health care utilization. METHODS: Within Kaiser Permanente Northern California, we identified opioid-naïve patients, ages 18-64, who were treated and discharged from the ED for a painful, low-severity condition between January 1, 2017, and December 31, 2017. We also identified patients who received an OPR, either administered in the ED or obtained at a Kaiser Permanente Northern California pharmacy within 7 days of ED arrival, and investigated subsequent acute care utilization in cases with at least 1 ED, urgent care, or inpatient visit within 1 month or 3 months of the index encounter or 2 visits within 12 months. RESULTS: Of the 39,468 adults included in our study, 50.7% were female, 55.0% were non-White, and 25.2% received an OPR in association with their index ED encounter. After adjustment, we found that patients who received an OPR had greater odds of downstream acute care utilization than those who did not, with odds ratios of 1.68, 1.53, and 1.50 at 1, 3, and 12 months, respectively (all p < 0.05). CONCLUSION: Patients who received an OPR at their index encounter had substantially increased odds of a subsequent ED, urgent care, or inpatient visit. This effect was most pronounced early in followup and persisted for the duration of the study period. Receipt of an OPR among opioid-naïve adults for a painful, low-severity condition is associated with increased downstream acute care utilization. [ABSTRACT FROM AUTHOR]

Objectives: Coronary risk scores are commonly applied to emergency department patients with undifferentiated chest pain. Two prominent risk score–based protocols are the Emergency Department Assessment of Chest pain Score Accelerated Diagnostic Protocol (EDACS‐ADP) and the History, ECG, Age, Risk factors, and Troponin (HEART) pathway. Since prospective documentation of these risk determinations can be challenging to obtain, quality improvement projects could benefit from automated retrospective risk score classification methodologies. Methods: EDACS‐ADP and HEART pathway data elements were prospectively collected using a Web‐based electronic clinical decision support (eCDS) tool over a 24‐month period (2018–2019) among patients presenting with chest pain to 13 EDs within an integrated health system. Data elements were also extracted and processed electronically (retrospectively) from the electronic health record (EHR) for the same patients. The primary outcome was agreement between the prospective/eCDS and retrospective/EHR data sets on dichotomous risk protocol classification, as assessed by kappa statistics (ĸ). Results: There were 12,110 eligible eCDS uses during the study period, of which 66 and 47% were low‐risk encounters by EDACS‐ADP and HEART pathway, respectively. Agreement on low‐risk status was acceptable for EDACS‐ADP (ĸ = 0.73, 95% confidence interval [CI] = 0.72 to 0.75) and HEART pathway (ĸ = 0.69, 95% CI = 0.68 to 0.70) and for the continuous scores (interclass correlation coefficients = 0.87 and 0.84 for EDACS and HEART, respectively). Conclusions: Automated retrospective determination of low risk status by either the EDACS‐ADP or the HEART pathway provides acceptable agreement compared to prospective score calculations, providing a feasible risk adjustment option for use in large data set analyses. [ABSTRACT FROM AUTHOR]

Introduction The Pulmonary Embolism Severity Index (PESI) is a validated prognostic score to estimate the 30-day mortality of emergency department (ED) patients with acute pulmonary embolism (PE). A simplified version (sPESI) was derived but has not been as well studied in the U.S. We sought to validate both indices in a community hospital setting in the U.S. and compare their performance in predicting 30-day all-cause mortality and classification of cases into low-risk and higher-risk categories. Materials and methods This retrospective cohort study included adults with acute objectively confirmed PE from 1/2013 to 4/2015 across 21 community EDs. We evaluated the misclassification rate of the sPESI compared with the PESI. We assessed accuracy of both indices with regard to 30-day mortality. Results Among 3006 cases of acute PE, the 30-day all-cause mortality rate was 4.4%. The sPESI performed as well as the PESI in identifying low-risk patients: both had similar sensitivities, negative predictive values, and negative likelihood ratios. The sPESI, however, classified a smaller proportion of patients as low risk than the PESI (27.5% vs. 41.0%), but with similar low-risk mortality rates (< 1%). Compared with the PESI, the sPESI overclassified 443 low-risk patients (14.7%) as higher risk, yet their 30-day mortality was 0.7%. The sPESI underclassified 100 higher-risk patients (3.3%) as low risk who also had a low mortality rate (1.0%). Conclusions Both indices identified patients with PE who were at low risk for 30-day mortality. The sPESI, however, misclassified a significant number of low-mortality patients as higher risk, which could lead to unnecessary hospitalizations. [ABSTRACT FROM AUTHOR]