key: cord-0702467-gm94j5he
authors: Özdemir, Serdar; Akça, Hatice Şeyma; Algın, Abdullah; Altunok, İbrahim; Eroğlu, Serkan Emre
title: Effectiveness of the rapid emergency medicine score and the rapid acute physiology score in prognosticating mortality in patients presenting to the emergency department with COVID-19 symptoms
date: 2021-06-10
journal: Am J Emerg Med
DOI: 10.1016/j.ajem.2021.06.020
sha: b7f4d8cb8d2694f3bb492ffc022db84b9d2d30e4
doc_id: 702467
cord_uid: gm94j5he

OBJECTIVE: We investigated the effectiveness of the Rapid Emergency Medicine Score and the Rapid Acute Physiology Score in identifying critical patients among those presenting to the emergency department with COVID-19 symptoms. MATERIAL AND METHODS: This prospective, observational, cohort study included patients with COVID-19 symptoms presenting to the emergency department over a two-month period. Demographics, clinical characteristics, and the data of all-cause mortality within 30 days after admission were noted, and the Rapid Emergency Medicine Score and the Rapid Acute Physiology Score were calculated by the researchers. The receiver operating characteristic curve analysis was performed to determine the discriminative ability of the scores. RESULTS: A total of 555 patients with a mean of age of 49.4 ± 16.8 years were included in the study. The rate of 30-day mortality was 3.9% for the whole study cohort, 7.2% for the patients with a positive rt-PCR test result for SARS-CoV-2, and 1.2% for those with a negative rt-PCR test result for SARS-CoV-2. In the group of patients with COVID-19 symptoms, according to the best Youden's index, the cut-off value for the Rapid Emergency Medicine Score was determined as 3.5 (sensitivity: 81.82%, specificity: 73.08%), and the area under curve (AUC) value was 0.840 (95% confidence interval 0.768–0.913). In the same group, according to the best Youden's index, the cut-off value for the Rapid Acute Physiology Score was 2.5 (sensitivity: 90.9%, specificity: 97.38%), and the AUC value was 0.519 (95% confidence interval 0.393–0.646). CONCLUSION: REMS is able to predict patients with COVID-19-like symptoms without positive rt-PCR for SARS-CoV-2 that are at a high-risk of 30-day mortality. Prospective multicenter cohort studies are needed to provide best scoring system for triage in pandemic clinics.

Providing appropriate and timely medical support is associated with reduced mortality and morbidity in patients with trauma and sepsis (1) . On the other hand, the overcrowding of the emergency department (ED) prevents healthcare workers from allocating the necessary time to patients (2) . Many scoring systems; e.g., the Emergency Department Sepsis Score, the Modified Early Warning Score (MEWS), the Rapid Emergency Medicine Score (REMS), and the Rapid Acute Physiology Score (RAPS) have been developed to prioritize patients who need emergency medical support in emergency services (2, 3) . Contrary to scores used in intensive care units, such as APACHE II, laboratory parameters are not included in the scoring system used in triage in ED to achieve prompt decisions (4) . Pulse rate, mean arterial pressure respiratory rate, and Glasgow Come Scale are using to calculate RAPS (2) . Age and peripheral oxygen saturation are being used to calculate REMS in addition to these parameters (2) .

Since SARS-CoV-2 infection was first described, more than 66 million people have been infected and more than 1.5 million people have died worldwide. The rapid spread of the disease around the world has increased the burden on health systems. Governments have had patients who had missing data or refused to have a PCR test were excluded. The information regarding the patient selection is summarized in Figure 1 .

In our study, we sought an answer to the clinical question of which is the ideal scoring system that can be used in pandemic clinics. Therefore, patients with a positive rt-PCR test for SARS-CoV-2 and those with a negative rt-PCR test for SARS-CoV-2 subgroups of the study population were created.

Nasal and pharyngeal swabs were tested by RT-PCR with the SARS-CoV-2 detection kit (Coyote Bioscience Co., Ltd) that the turn-around time was 1-4 days due to crowdedness.

To diagnose COVID 19, ORF1ab and N gene of SARS-CoV-2 were targeted and Biorad CFX 96 platform were used. Twenty-nine and above Ct values were considered positive. Tests that were positive for both ORFlab and N genes were reported as positive.

Data were collected using two sources: the pandemic triage form and the computerbased system of the hospital. The pandemic triage form was completed for each patient suspected to have COVID-19. It contains information on patient ID to define each patient in the computer-based system, COVID-19 symptoms, other nonspecific symptoms (nauseavomiting, diarrhea, headache, weakness, muscle-joint pain), Glasgow Coma Scale score, and vital parameters. Vital parameters noted on the form are blood pressure (systolic and diastolic), pulse pressure, body temperature, respiratory rate, and oxygen saturation. The data on demographics, clinical characteristics, comorbidities, laboratory findings, clinical outcome for the first 24 hours, necessity of mechanical ventilation, and 30-day mortality were extracted from the computer-based system. Comorbidities were recorded as chronic obstructive pulmonary diseases, diabetes mellitus, hypertension, coronary artery disease, congestive heart failure, active malignancy, chronic kidney disease, and immunosuppression. Laboratory findings examined included white blood cell count, platelet count, neutrophil count,   lymphocyte count, C-reactive protein, albumin, and results of rt-PCR test for SARS-CoV-2. REMS, RAPS, mean arterial pressure, C-reactive protein-to-albumin ratio, platelet-tolymphocyte ratio, and neutrophil-to-lymphocyte ratio were calculated. Parameters used in the calculation of REMS and RAPS are listed in table 1. Pulse rate, mean arterial pressure, respiratory rate, and Glasgow coma scale used in calculation of RAPS age, peripheral oxygen saturation, pulse rate, mean arterial pressure, respiratory rate, and Glasgow coma scale used in calculation of REMS. Final diagnoses of the patients with a negative rt-PCR result were recorded. Data were collected and analyzed by five independent researchers who have 5-10 years of experience in emergency medicine. Each case was analyzed by one of five independent researchers who were not blinded to the purpose of the study. Information on mortality was collected by checking the computer-based system or calling patients or their relatives if necessary. All the patients discharged from the hospital were contacted and follow up was totally completed.

All statistical analyses were performed using SPSS version 22.0 for Windows (SPSS Inc, Chicago, IL, USA). The normality analysis of continuous data was undertaken using the Kolmogorov-Smirnov test. Categorical data were presented as n (%) and compared using the chi-squared test. Quantitative variables were presented as median and interquartile range (IQR, 25 th -75 th percentile), then compared using the Mann-Whitney test or Student's t-test according to the normality of distribution for the two groups. The Bonferroni correction was used a method to counteract the problem of multiple comparisons. The univariate analyses to identify variables (RAPS and REMS) associated with 30-day mortality status were performed using the chi-square, Fisher's exact, Student's t and Mann-Whitney U tests, where appropriate. In the multivariate analysis, the effective factors identified with the univariate J o u r n a l P r e -p r o o f analyses were further examined with the logistic regression analysis to determine the independent predictors of mortality. The Hosmer-Lemeshow goodness-of-fit statistics were used to assess the model fit. Receiver operating characteristic (ROC) curves were used to assess the accuracy of RAPS and REMS to predict mortality, and the results were reported as the area under the curve (AUC) values. Youden's index was used to determine the optimal cut-off value for scores with highest sensitivity, and specificity. Likelihood ratios were calculated using sensitivity and specificity values in the evaluation of relationship between 30-day mortality and scoring systems. Statistical significance was defined at p < 0.05.

The ethical committee approval of this study was obtained from the local ethics committee with the approval number B.10.1.TKH.4.34.H.GP.0.01/127. Data collection was performed prospectively by emergency nurses. Before including the study, informed consent forms were signed by patients or their relatives. All researchers adhered to the principles of the Declaration of Helsinki throughout the study period.

Of the 558 patients included in the study, 310 (55.6%) were male. The mean of age of the 558 patients was 49.4 ± 16.8 years. A total of 22 patients died within 30 days of ED presentation. The rate of 30-day mortality was 3.9% for the whole study cohort, 7.2% for the patients with a positive rt-PCR test result for SARS-CoV-2, and 1.2% for those with a negative rt-PCR result for SARS-CoV-2. The demographic characteristics, clinical outcomes for the first 24 hours, comorbid diseases, symptoms, vital parameters at presentation, initial laboratory findings, REMS, RAPS, and mortality data are shown in Table 2 

The comparisons of the demographics, clinical characteristics and laboratory findings of the non-survivor and survivor groups are shown in Table 2 The analysis of the ROC curve was performed to determine the discriminative ability of the two scoring systems in 30-day mortality. In the group of patients with COVID-19 symptoms, according to the best Youden's index, the cut-off value for REMS was 3.5 (sensitivity: 81.82%, specificity: 73.08%), and the AUC value was 0.840 (95% confidence interval 0.768-0.913). In the same group, according to the best Youden's index, the cut-off value for RAPS was 2.5 (sensitivity: 9.09%, specificity: 97.38%), and the AUC value was 0.519 (95% confidence interval 0.393-0.646) (Table 3, Figure 2 ). Table 3 Table 3) .

The multivariate logistic regression analysis was performed to identify the independent predictors of mortality, and age and oxygen saturation were determined to be independent predictors with the p values of <0.001 and 0.001, respectively (Table 4 ).

In this study, we compared two emergency scoring systems and found REMS to be the useful tool in predicting 30-day mortality in only in the group without rt-PCR positivity.

However, both scoring systems were not useful in predicting 30-day mortality in rt-PCR positive and general patient population. On the other hand, the subgroup analysis showed that REMS could predict 30-day mortality in patients without a positive rt-PCR test for SARS-CoV-2. To the best of our knowledge, this is the first study to evaluate all patients presenting to the emergency pandemic clinics with COVID-19-like symptoms using scoring systems.

In our analysis, first, nonparametric comparison tests were used to determine the relationship between scoring systems and mortality. While REMS was significantly higher in the patients with mortality, no significant relationship was found between RAPS and mortality. A further analysis was performed based on the ROC curve to determine the two scoring systems' ability to distinguish whether a patient survived or died. AUC values less than 0.5 were evaluated as indistinguishable from random, while those close to 1 were considered close to the perfect model (7, 8) . It has been reported that the AUC value should be greater than 0.8 for a model to predict mortality well (7, 8) . In the discriminatory power analysis, we determined the AUC value of RAPS as 0.519, which was considered to be unacceptable. However, the AUC value of REMS in predicting 30-day mortality was 0.840, which indicated the predictive ability of this score for mortality. Thus, our prospective, J o u r n a l P r e -p r o o f comparative study, was demonstrated that only REMS was a predictor of 30-day mortality in patients with COVID-19 symptoms and confirmed COVID-19 according to ROC analysis. On the other hand, LRs supply the clearest data on the way in which scoring system can be used reliably (9-10). Ratios >5 or <0.2 provide of strongest evidence (9-10). In the patient group without rt-PCR positivity, LR (+) value of REMS was in this range and clinically useful.

The logistic regression analysis performed to determine the independent predictors of mortality revealed REMS as a predictor of mortality; however, RAPS was not found to be a predictive parameter. A possible explanation for this result is that among the parameters used in the calculation of scores, only age and oxygen saturation are correlated with mortality, as shown by the multivariate logistic regression analysis. Age and oxygen saturation is the only difference between RAPS and REMS.

hospital mortality in non-surgical ED patients (11, 12) . In a cohort study, the authors showed that oxygen saturation and age were the strongest prognostic parameters and added these parameters to RAPS and validated REMS as a new scoring system (11) . Subsequently, many researchers investigated this new scoring system and compared it with different scoring systems in different patient groups. In a 2019 study of 39,977 patients, REMS was shown to be a more powerful predictor of in-hospital mortality compared to RAPS and MEWS (13) . In the mentioned study, the negative predictive value and cut-off value of REMS were found to be 0.88 and 8, respectively and they did not present LRs of REMS (13).

In the current literature, scoring systems in COVID-19 patients were first evaluated in two studies by Hu et al. (14, 15) . In a retrospective study of 105 patients, they demonstrated that AUC for in-hospital mortality predictability and cut-off values were 0.841, and 6, respectively for REMS (14) . The authors suggested that REMS could be used by ED workers J o u r n a l P r e -p r o o f to prognosticate in-hospital mortality in critically ill COVID-19 patients. In a second study by Hu et al. evaluating 319 patients with COVID-19, five early warning system scoring system were determined to predict hospital discharge (15) . After Hu et al., researchers investigated different emergency alert scores or sepsis scores in confirmed COVID-19 patients in emergency department and intensive care unit (15) (16) (17) . The study indicated that REMS and NEWS scores could predict in-hospital mortality and seven-day hospitalization in the intensive care unit in patients with confirmed COVID-19 (16) . In contrast, we included all patients with COVID-19-like symptoms patients in our study and evaluated REMS and RAPS prospectively.

The data were collected from a teaching hospital declared as a pandemic hospital during the pandemic period. Despite the high number of patients presenting to the clinic and the patients being informed about the study by the researchers, only a small number of volunteers participated in the study over a two-month period, limited our cohort. Firstly, misinformation about COVID-19 spreading especially on social media negatively affects the patients' willingness to participate in studies on 19) . Secondly, low health literacy is mainly a problem in patients who admitted our department (20) . Thirdly, due to the intensity of the emergency room and pandemic conditions, there was not enough time to persuade all patients to participate in the study. Data belong to each patient were analyzed by a single researcher and there was no patient evaluated by more than one researcher. Therefore, J o u r n a l P r e -p r o o f Rapid Emergency Medicine Score (REMS) 2 (0-12) 2 (0-11) 6 (2-12) <0.001 Rapid Acute Physiology Score (RAPS) 0 (0-6) 0 (0-6) 0 (0-5) 0.676 Dear Editor,

We declare that this manuscript represents valid work and that neither this manuscript nor one with substantially similar content under the present authorship has been published or is being considered for publication elsewhere and the authorship of this article will not be contested by anyone whose name(s) is/are not listed here, and that the order of authorship as placed in the manuscript is final and accepted by the co-authors. These declarations also represent the authorship which is signed by all the authors in the order in which they are mentioned in the original manuscript.

We certify that all the data collected during the study is presented in this manuscript and no data from the study has been or will be published separately. We attest that, if requested by the editors, we will provide the data/information or will cooperate fully in obtaining and providing the data/information on which the manuscript is based, for examination by the editors or their assignees. We also certify that we have taken all necessary permissions from our institution and/or department for conducting and publishing the present work. There is no ethical problem or conflict of interest.

Sincerely yours, 

Emergency Department Triage Early Warning Score (TREWS) predicts in-hospital mortality in the emergency department

Performance Assessment of the Mortality in Emergency Department Sepsis Score, Modified Early Warning Score, Rapid Emergency Medicine Score, and Rapid Acute Physiology Score in Predicting Survival Outcomes of Adult Renal Abscess Patients in the Emergency Department

Comparing the effectiveness of three scoring systems in predicting adult patient outcomes in the emergency department

The impact of the use of the Early Warning Score (EWS) on patient outcomes: a systematic review

Promotion of scientific research on COVID-19 in Turkey

Trends in Emergency Department Visits and Hospital Admissions in Health Care Systems in 5 States in the First Months of the COVID-19 Pandemic in the US

The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology

A method of comparing the areas under receiver operating characteristic curves derived from the same cases

Systematic reviews of evaluations of diagnostic and screening tests

Use of protein: creatinine ratio measurements on random urine samples for prediction of significant proteinuria: a systematic review

Rapid Emergency Medicine score: a new prognostic tool for in-hospital mortality in nonsurgical emergency department patients

Rapid emergency medicine score can predict long-term mortality in nonsurgical emergency department patients

Comparing the effectiveness of three scoring systems in predicting adult patient outcomes in the emergency department

Comparing Rapid Scoring Systems in Mortality Prediction of Critically Ill Patients with Novel Coronavirus Disease

Predictive Value of 5 Early Warning Scores for Critical COVID-19 Patients

Prognostic Accuracy of the SIRS, qSOFA, and NEWS for Early Detection of Clinical Deterioration in SARS-CoV-2 Infected Patients

Predicting intensive care unit admission and death for COVID-19 patients in the emergency department using early warning scores

Information and Misinformation on COVID-19: a Cross-Sectional Survey Study

Health Literacy in The Emergency Department: A Cross-Sectional Descriptive Study

Blood test parameters, median (IQR) White blood cell count