In emergency department patients with septic shock, afebrile patients received lower rates of emergency department antibiotic administration, lower mean IV fluids volume, and suffered higher in-hospital mortality.
Objective
Linking emergency medical services (EMS) electronic patient care reports (ePCRs) to emergency department (ED) records can provide clinicians access to vital information that can alter management. It can also create rich databases for research and quality improvement. Unfortunately, previous attempts at ePCR and ED record linkage have had limited success. In this study, we use supervised machine learning to derive and validate an automated record linkage algorithm between EMS ePCRs and ED records.
Materials and Methods
All consecutive ePCRs from a single EMS provider between June 2013 and June 2015 were included. A primary reviewer matched ePCRs to a list of ED patients to create a gold standard. Age, gender, last name, first name, social security number, and date of birth were extracted. Data were randomly split into 80% training and 20% test datasets. We derived missing indicators, identical indicators, edit distances, and percent differences. A multivariate logistic regression model was trained using 5-fold cross-validation, using label k-fold, L2 regularization, and class reweighting.
Results
A total of 14 032 ePCRs were included in the study. Interrater reliability between the primary and secondary reviewer had a kappa of 0.9. The algorithm had a sensitivity of 99.4%, a positive predictive value of 99.9%, and an area under the receiver-operating characteristic curve of 0.99 in both the training and test datasets. Date-of-birth match had the highest odds ratio of 16.9, followed by last name match (10.6). Social security number match had an odds ratio of 3.8.
Conclusions
We were able to successfully derive and validate a record linkage algorithm from a single EMS ePCR provider to our hospital EMR.
In this study, more than half of the patients had elevated triage BP (≥ 90th percentile), which was rarely recognized by emergency department practitioners regardless of specialty or experience. Early recognition of elevated triage BP offers opportunities for diagnosis of hypertension and related disorders but is challenging to accomplish.
IntroductionStrategies to identify high-risk emergency department (ED) patients often use markedly abnormal vital signs and serum lactate levels. Risk stratifying such patients without using the presence of shock is challenging. The objective of the study is to identify independent predictors of in-hospital adverse outcomes in ED patients with abnormal vital signs or lactate levels, but who are not in shock.MethodsWe performed a prospective observational study of patients with abnormal vital signs or lactate level defined as heart rate ≥130 beats/min, respiratory rate ≥24 breaths/min, shock index ≥1, systolic blood pressure <90mm/Hg, or lactate ≥4mmole/L. We excluded patients with isolated atrial tachycardia, seizure, intoxication, psychiatric agitation, or tachycardia due to pain (ie: extremity fracture). The primary outcome was deterioration, defined as development of acute renal failure (creatinine 2× baseline), non-elective intubation, vasopressor requirement, or mortality. Independent predictors of deterioration after hospitalization were determined using logistic regression.ResultsOf 1,152 consecutive patients identified with abnormal vital signs or lactate level, 620 were excluded, leaving 532 for analysis. Of these, 53/532 (9.9±2.5%) deteriorated after hospital admission. Independent predictors of in-hospital deterioration were: lactate >4.0mmol/L (OR 5.1, 95% CI [2.1–12.2]), age ≥80 yrs (OR 1.9, CI [1.0–3.7]), bicarbonate <21mEq/L (OR 2.5, CI [1.3–4.9]), and initial HR≥130 (OR 3.1, CI [1.5–6.1]).ConclusionPatients exhibiting abnormal vital signs or elevated lactate levels without shock had significant rates of deterioration after hospitalization. ED clinical data predicted patients who suffered adverse outcomes with reasonable reliability.
Clinical data can predict the presence of sepsis causing shock in the ED in most patients. The remaining diagnostic uncertainty provides an opportunity for adding novel diagnostic testing.
Conflicts of interest: SH receives grant funding by Philips Healthcare in the areas of heart failure risk stratification, imaging analysis, and big data. The submitted manuscript has no relationship to the grants. LAN has stock in Forerun Systems, an emergency department information system. The submitted manuscript did not use this system. SH and LAN conceived and designed the study. CR and AT collected the data. AT, YH, DAS, and SH performed the analysis. CR, AT, and SH drafted the manuscript, and all authors contributed substantially to its revision. SH takes responsibility for the paper as a whole.
AbstractBackground: Linking EMS electronic patient care reports (ePCRs) to ED records can provide clinicians access to vital information that can alter management. It can also create rich databases for research and quality improvement. Unfortunately, previous attempts at ePCR -ED record linkage have had limited success.Objective: To derive and validate an automated record linkage algorithm between EMS ePCR's and ED records using supervised machine learning.Methods: All consecutive ePCR's from a single EMS provider between June 2013 and June 2015 were included. A primary reviewer matched ePCR's to a list of ED patients to create a gold standard. Age, gender, last name, first name, social security number (SSN), and date of birth (DOB) were extracted. Data was randomly split into 80%/20% training and test data sets. We derived missing indicators, identical indicators, edit distances, and percent differences. A multivariate logistic regression model was trained using 5k fold cross-validation, using label kfold, L2 regularization, and class re-weighting.Results: A total of 14,032 ePCRs were included in the study. Inter-rater reliability between the primary and secondary reviewer had a Kappa of 0.9. The algorithm had a sensitivity of 99.4%, a PPV of 99.9% and AUC of 0.99 in both the training and test sets. DOB match had the highest odd ratio of 16.9, followed by last name match (10.6). SSN match had an odds ratio of 3.8.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.