Archives of Academic Emergency Medicine. 2023; 11(1): e34 OR I G I N A L RE S E A RC H Modified Shock Index as a Predictor of Admission and In- hospital Mortality in Emergency Departments; an Analysis of a US National Database Bachar Hamade1, Jamil D. Bayram2, Yu-Hsiang Hsieh2, Basem Khishfe3, Nour Al Jalbout4∗ 1. Center for Emergency Medicine, Main Campus and Department of Intensive Care and Resuscitation, Cleveland Clinic Foundation, Cleveland, Ohio. 2. Department of Emergency Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland. 3. Department of Emergency Medicine, St. Elizabeth’s Hospital, O’Fallon, Illinois. 4. Department of Emergency Medicine, Massachusetts General Hospital, Boston, Massachusetts. Received: February 2023; Accepted: March 2023; Published online: 29 April 2023 Abstract: Introduction: The modified shock index (MSI) is the ratio of heart rate to mean arterial pressure. It is used as a pre- dictive and prognostic marker in a variety of disease states. This study aimed to derive the optimal MSI cut-off that is associated with increased likelihood (likelihood ratio, LR) of admission and in-hospital mortality in patients present- ing to emergency department (ED). Methods: We retrospectively reviewed data from the National Hospital Ambulatory Medical Care Survey between 2005 and 2010. Adults>18 years of age were included regardless of chief complaint. Basic patient demographics, initial vital signs, and outcomes were recorded for each patient. Then the optimal MSI cut-off for prediction of admission and in-hospital mortality in ED was calculated. LR ≥ 5 was considered clinically significant. Results: 567,994,402 distinct weighted adult ED patient visits were included in the analysis. 15.7% and 2.4% resulted in admissions and in-hospital mortality, respectively. MSI > 1.7 was associated with a moderate increase in the likelihood of both admission (Positive LR (+LR) = 6.29) and in-hospital mortality (+LR = 5.12). +LR for hospital admission at MSI >1.7 was higher for men (7.13; 95% CI 7.11-7.15) compared to women (5.49; 95% CI 5.47-5.50) and for non-white (7.92; 95% CI 7.88-7.95) compared to white patients (5.85; 95% CI 5.84-5.86). For MSI <0.7, the +LRs were not clinically signif- icant for admission (+LR = 1.07) or in-hospital mortality (LR = 0.75). Conclusion: In this largest retrospective study, to date, on MSI in the undifferentiated ED population, we demonstrated that an MSI >1.7 on presentation is predictive of admission and in-hospital mortality. The use of MSI could help guide accurate acuity designation, resource allocation, and disposition. Keywords: Modified shock index; Hospitalization; Inpatients; Hospital Mortality; Emergency Service, Hospital; Probability Cite this article as: Hamade B, Bayram JD, Hsieh Y, Khishfe B, Al Jalbout N. Modified Shock Index as a Predictor of Admission and In-hospital Mortality in Emergency Departments; an Analysis of a US National Database. Arch Acad Emerg Med. 2023; 11(1): e34. https://doi.org/10.22037/aaem.v11i1.1901. 1. Introduction The shock index (SI) is a clinical metric obtained by divid- ing the heart rate (HR) by the systolic blood pressure (SBP). It was first described by Allgower and Burri (1) in 1967 and has been proposed to serve as an early indicator of clinical dete- rioration (2, 3). Originally investigated in shock states, the SI has been studied as a prognostic metric in a variety of disease ∗Corresponding Author: Nour Al Jalbout; Department of Emergency Medicine, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114. Email: naljalbout@mgh.harvard.edu, Tel: 001 (617) 726-7622, ORCID: https://orcid.org/0000-0002-9369-0260. states including trauma, pneumonia, sepsis, gastrointestinal bleeding, and ST segment elevation myocardial infarction (4-11). In the undifferentiated emergency department (ED) population, one recent large retrospective multicenter study found that SI>1.3 was associated with a higher likelihood of hospital admission and in-hospital mortality (12). Currently, most EDs measure blood pressure non-invasively via automated blood pressure machines that rely on os- cillometry. These machines measure mean arterial pres- sure (MAP) and extrapolate SBP and diastolic blood pres- sure (DBP), overestimating SBP compared to invasive arte- rial monitoring (13), which can lead to false hemodynamic reassurance. MAP is an important clinical metric driven This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index B. Hamade et al. 2 Table 1: Baseline characteristics of studied emergency department patient visits Variable Number (%) Age (years) 18-44 279,660,193 (53.1) 45-64 144,066,977 (27.4) ≥65 102,054,146 (19.5) Sex Male 299,216,051 (56.9) Female 226,565,265 (43.1) Race Non-white 137,426,759 (26.1) White 388,354,557 (73.9) Disposition Admission 82,501,250 (15.7) In-hospital mortality 1,917,737 (2.4) by early changes in DBP, which defines perfusion pressure, and guides therapy in septic patients and early hemorrhagic states (14-16). In 2012, Liu et al. introduced the modified shock index (MSI), which is the HR divided by the MAP. They demonstrated that MSI <0.7 or >1.3 were significantly asso- ciated with higher in-hospital mortality in a single center ED population (17). In another large prospective study of 9860 trauma patients, the same MSI cut-offs outperformed SI and traditional vital signs as predictors of mortality (18). Since then, the predictive and prognostic utilities of MSI have been studied in a variety of disease states such as sepsis and my- ocardial infarction (19-25). However, literature regarding its optimal predictive cut-offs and utility in the general undiffer- entiated ED population is limited (26, 27). The main objective of this study is to derive an MSI threshold that predicts both admission and in-hospital mortality in the undifferentiated ED population based on the initial present- ing HR and MAP, irrespective of clinicians’ subjective clinical judgements. The secondary objective is to analyze the per- formance of different MSI thresholds when performing sub- group analysis based on age, sex, and race. 2. Methods 2.1. Study design and setting We retrospectively analyzed ED visits from 2005 to 2010 from the National Hospital Ambulatory Medical Care Survey (NHAMCS), the largest nationally representative database of utilization and provision of ED services in the United States, using a weighted sample of all U.S. ED visits reported in the survey. The study was granted exempt status by our institu- tion’s review board (IRB00151493). The NHAMCS is a nationally representative survey con- ducted by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (28). It includes an estimated 120 million encounters annually ob- tained from probability sampling of ED visits in the 50 states and the District of Columbia, excluding visits to federal, mil- itary and Veteran Administration hospitals. 2.2. Participants Adult patient visits above 18 years of age were included. En- counters with insufficient data to calculate the MSI in ad- dition to encounters resulting in ED deaths were excluded (since our main outcomes of interest were admissions and in-hospital mortality). 2.3. Data gathering Data is collected from ED visit medical records during a ran- domly assigned four-week period by NHAMCS personnel un- der the supervision of field representatives. Staff members independently checked 10% of the data for accuracy with an error rate ranging from 0.3% to 0.9% for various items on the survey (28). Between 352 and 389 hospitals agreed to partic- ipate in the survey from the year 2005 to 2010, which consti- tuted our study period. This data was publicly available with free access. The NHAMCS does not include information on specific ED clinical management and condition fluctuation details. For each patient visit, basic patient demographics including age, sex and race, initial HR, SBP and DBP, and visit outcomes were recorded by trained coders following CDC’s National Center for Health Statistics standardized NHAMCS data col- lection protocol (28). From the initial SBP and DBP, MAP and subsequently the MSI were calculated. The SBP and DBP were measured non- invasively at triage, as this is the standard across Emergency Rooms. For this exploratory study, effect modifiers were not considered in the data analysis plan. MSI, the HR divided by the MAP, which was calculated as (2*DBP+SBP)/3, was the predictor of the primary outcomes. 2.4. Outcome measures The primary outcome measures of our study were hospital admission and in-hospital mortality, which were based on the data coding on ED disposition and hospital discharge sta- tus, respectively. 2.5. Data analysis Descriptive data analysis was first performed to summarize key descriptive statistics of the study population. Likelihood ratios (LRs) for a broad range of thresholds of MSI for both outcomes were then calculated. Positive LRs (+LRs) were used since they represent statistically robust measures of di- agnostic accuracy independent of pretest probability, unlike positive and negative predictive values (29). Since prior data suggests different likelihood of admission based on age, sex This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 3 Archives of Academic Emergency Medicine. 2023; 11(1): e34 and race, these variables were considered as potential con- founders a priori. For this exploratory study, we performed sub-group analysis stratified by race (white vs. non-white), sex (women vs. men) and age group (18-44, 45-64 and ≥65 years) for the outcome of hospital admission to minimize the bias resulting from the potential confounding effect by age, sex, and race (30-32). With regards to in-hospital mortality, the original weighted sample size was less than 30 in each subgroup, and thus prevented stratifying the data by demo- graphics due to the concern of generating invalid estimates. Values of + LR are considered to be clinically significant when they are ≥ 5, with the range between 5 and 10 considered moderate increase in the likelihood of the outcome and ≥10 considered large and often conclusive for the likelihood of the outcome (29). 3. Results 3.1. Baseline characteristics of ED visits We identified 567,994,402 distinct weighted adult ED pa- tient visits between the years 2005 and 2010. We ex- cluded 660,207 (0.12%) visits of patients who died in the ED in addition to 42,213,086 (7.43%) visits due to incomplete records. 525,781,316 weighted patient visits were included in the final analysis, of which 299,216,051 (57%) were female, 388,354,557 (74%) were white and 279,660,193 (53%) were between the ages of 18 and 44 years. Table1 summarizes the baseline characteristics of the studied ED visits. Out of the 525,781,316 weighted patient visits included in this study, 82,501,250 (15.7%) unique ED visits resulted in in-hospital admissions, and 1,917,737 (2.4%) resulted in in-hospital mor- tality. 3.2. Predictive value of MSI for admission and in-hospital mortality An initial MSI value of >1.7 was associated with a moderate increase in the likelihood of both admission and in-hospital mortality with a +LR of 6.29 (95% CI: 6.28-6.31) and 5.12 (95% CI: 5.10-5.14), respectively. For patients with an initial MSI value of <0.7, the +LRs for admission and in-hospital mor- tality were not clinically significant with values of 1.07 (95% CI: 1.07-1.07) and 0.75 (95% CI: 0.75-0.75), respectively. Like- wise, for patients with MSI value of >1.3, the +LRs for admis- sion and in-hospital mortality were 2.55 (95% CI: 2.55-2.55) and 2.69 (95% CI: 2.68-2.70), respectively (Tables 2 and 3). Out of the 82,501,250 admitted patients, 2,011,388 (2.4%) had an initial MSI value >1.7, and out of 1,917,737 patients who died, 215,337 (11.2%) had an MSI>1.7 (Tables 2 and 3). A total of 23,713,531 (4.51%) admitted patient encounters had MSI <0.7 or MSI >1.3 (Table 2). Among them, 800,831 (0.15%) pa- tients died during hospitalization (Table 3). The frequency of both outcomes at every MSI cut-off is presented in Tables 2 and 3. The areas under the receiver operating characteristic (ROC) curve of MSI in predicting the admission and in-hospital mortality were 0.454 (95%CI: 0.450-0.459) and 0.366 (95%CI: 0.339-0.393), respectively (Figure 1). 3.3. +LR of MSI stratified by age, sex, and race When stratified by age, sex, and race, +LR for hospital admis- sion at MSI >1.7 was higher for men (7.13; 95% CI: 7.11-7.15) compared to women (5.49; 95% CI: 5.47-5.50) and for non- white (7.92; 95% CI: 7.88-7.95) compared to white patients (5.85; 95% CI: 5.84-5.86) (Table 4). Admission +LRs for all age groups increased with increasing MSI values, with the high- est +LR being for patients aged 45-64 years (7.48; 95% CI 7.45- 7.51) (Table 4). Sub-group analysis for in-hospital mortality outcome was not performed due to the small sample size, rendering +LR estimates unreliable. 4. Discussion In the undifferentiated ED population, our study shows an association between an initial MSI value >1.7 and a clini- cally significant increase in the likelihood of both outcomes, admission and in-hospital mortality, with +LRs of 6.29 and 5.12, respectively. To the best of our knowledge, this is the largest multicenter retrospective study of the general ED population, with 525,781,316 weighted ED patient encoun- ters. These results are in line with a recent study describing an MSI of >1.7 as a strong predictor of mortality. In this study, Smischney et al. showed that within the first 24 hours of ICU admission, patients with an elevated MSI have a significant mortality risk (23). Our study results indicate that the MSI cut-off values of <0.7 and >1.3 introduced by Liu et al. are not reliable predictors of mortality in the general ED population, with +LRs being significantly less than 5. It is important to note that there are important differences in the methodology between our study and that of Liu et al. that may explain the differences in MSI values. While both studies are similar in calculating the MSI from the triage vital signs and defining one of the primary outcomes as in-hospital mortality, our study was a large mul- ticenter study encompassing a more diverse patient popula- tion with a variety of disease states and complaints. Liu et al. included only those patients that received intravenous (IV ) fluids because they were considered “real emergency” pa- tients. The study of Liu et al. analyzed a smaller total number of patients (22,161) in a different setting (China), with a wider age range, including those aged 10 years up to 100 years. Our study analyzed more than 525 million ED visits, and was lim- ited to those 18 years old and above. In our study, only 4.51% of the admitted patients and 0.15% of those who died as in- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index B. Hamade et al. 4 Table 2: Screening performance characteristics of modified shock index (MSI) in different cut-offs for predicting the need for hospital admis- sion MSI Number* Admitted PLR Sensitivity Specificity <0.3 781,336 126,089 1.03 (1.03-1.04) 0.15 (0.15-0.15) 99.85 (99.85-99.85) <0.4 1,462,308 374,241 1.85 (1.84-1.85) 0.45 (0.45-0.46) 99.75 (99.75-99.76) <0.5 6,658,809 1,623,417 1.73 (1.73-1.73) 1.97 (1.96-1.97) 98.86 (98.86-98.87) <0.6 30,810,315 6,097,638 1.33 (1.32-1.33) 7.39 (7.39-7.40) 94.43 (94.42-94.43) <0.7 89,440,729 14,817,168 1.07 (1.07-1.07) 17.96 (17.95-17.97) 83.17 (83.16-83.17) 0.5≤ & ≤0.7 83,149,111 69,917,576 0.98 (0.98-0.98) 15.77 (15.77-15.78) 83.96 (83.16-83.17) >0.7 435,973,396 67,646,298 0.99 (0.99-0.99) 81.99 (81.99 – 82.0) 16.91 (16.91-16.91) >0.8 344,098,058 55,261,912 1.03 (1.03-1.03) 66.98 (66.97-66.99) 34.84 (34.84-34.85) >0.9 239,969,027 41,744,323 1.13 (1.13-1.13) 50.60 (50.59-50.61) 55.28 (55.28-55.29) >1.0 148,962,698 29,618,270 1.33 (1.33-1.33) 35.90 (35.89-35.91) 73.08 (73.07-73.08) >1.1 87,173,982 20,246,957 1.63 (1.62-1.63) 24.54 (24.53-24.55) 84.90 (84.90-84.91) >1.2 49,130,138 13,537,356 2.04 (2.04-2.05) 16.41 (16.40-16.42) 91.97 (91.97-91.97) >1.3 27,631,575 8,896,363 2.55 (2.55-2.55) 10.78 (10.77-10.79) 95.77 (95.77-95.78) >1.4 15,649,057 6,045,761 3.38 (3.38-3.39) 7.33 (7.32-7.33) 97.83 (97.83-97.83) >1.5 9,131,844 3,972,421 4.14 (4.13-4.14) 4.81 (4.81-4.82) 98.84 (98.84-98.84) >1.6 5,678,149 2,779,613 5.15 (5.14-5.16) 3.37 (3.37-3.37) 99.35 (99.35-99.35) >1.7 3,728,673 2,011,388 6.29 (6.28-6.31) 2.44 (2.43 -2.44) 99.61 (99.61-99.61) >1.8 2,530,265 1,375,089 6.40 (6.38-6.41) 1.66 (1.66-1.67) 99.74 (99.74-99.74) Data are presented with 95% confidence interval. *Denotes the number of patient encounters included in the study in each specific MSI category. PLR: positive likelihood ratio. Figure 1: Area under the receiver operating characteristic (ROC) curve of modified shock index in predicting the need for admission (left) and in-hospital mortality (right). patients had an MSI<0.7 or >1.3. Finally, our study showed that MSI <0.7 was not a strong predictor of admission and in- patient mortality, which could mean that in the general ED population, a hypodynamic circulatory state is much more indicative of serious underlying pathology than a hyperdy- namic state as Liu et al. concluded. When stratified by groups, +LR for hospital admission strat- ified by race and sex were highest for non-white (7.92; 95% CI 7.88-7.95) and male patients (7.13; 95% CI 7.11-7.15). We could not identify specific causes for these differences with- out substratifying for confounders like age and comorbidi- ties; however prior data demonstrates higher admission rates for non-white patients and males in certain disease states (30-32). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 5 Archives of Academic Emergency Medicine. 2023; 11(1): e34 Table 3: Screening performance characteristics of modified shock index (MSI) in different cut-offs for predicting the in-hospital mortality MSI Number* Admitted PLR Sensitivity Specificity <0.3 113,160 1,737 0.58 (0.55-0.61) 0.09 (0.09-0.10) 99.84 (99.84-99.84) <0.4 319,606 9,786 1.17 (1.15-1.20) 0.51(0.50-0.52) 99.56 (99.56-99.57) <0.5 1,437,215 42,745 1.14 (1.13-1.15) 2.22 (2.21-2.25) 98.04 (98.04-98.04) <0.6 5,464,331 110,687 0.77 (0.76-0.77) 5.77 (5.74-5.80) 92.47 (92.47-92.49) <0.7 13,181,503 260,138 0.75 (0.74-0.75) 13.56 (13.52-13.61) 81.84 (81.84-81.86) 0.5≤ & ≤0.7 11,782,072 11,564,679 1.43 (1.43-1.44) 16.24 (16.24-16.25) 88.66 (88.62-88.71) >0.7 59,889,401 1,657,599 1.06 (1.06-1.06) 86.43 (86.39-86.48) 18.20 (18.19-18.21) >0.8 48,768,358 1,466,088 1.15 (1.15-1.15) 76.44 (76.39-76.51) 33.55 (33.54-33.57) >0.9 36,826,926 1,299,048 1.36 (1.36-1.36) 67.73 (67.67-67.80) 50.09 (50.08-50.11) >1.0 26,225,851 1,106,980 1.64 (1.63-1.63) 57.72 (57.65-57.79) 64.71 (64.71-64.73) >1.1 17,910,296 894,677 1.95 (1.95-1.96) 46.65 (46.58-46.72) 76.09 (76.09-76.11) >1.2 12,014,175 717,380 2.36 (2.36-2.36) 37.40 (37.34-37.48) 84.13 (84.12-84.14) >1.3 8,004,628 540,693 2.69 (2.68-2.70) 28.19 (28.13-28.26) 89.51 (89.51-89.52) >1.4 5,458,320 453,297 3.36 (3.35-3.37) 23.63 (23.58-23.70) 92.96 (92.96-92.98) >1.5 3,596,224 319,161 3.62 (3.60-3.62) 16.64 (16.59-16.70) 95.39 (95.39-95.40) >1.6 2,519,176 274,000 4.53 (4.51-4.54) 14.28 (14.24-14.34) 96.84 (96.84-96.85) >1.7 1,777,416 215,337 5.12 (5.10-5.14) 11.22 (11.18-11.27) 97.80 (97.80-97.81) >1.8 1,223,325 156,472 5.44 (5.42-5.47) 8.15 (8.12-8.20) 98.50 (98.50-98.50) Data are presented with 95% confidence interval. *Denotes the number of patient encounters included in the study in each specific MSI category. PLR: positive likelihood ratio. Table 4: Positive likelihood of modified shock index (MSI) in different cut-offs for predicting the need for admission stratified by sex, race, and age category MSI Sex Race Age (year) Female Male Non-white White 18-44 45-64 ≥65 <0.3 0.74 (0.73-0.75) 1.45 (1.44-1.46) 0.45 (0.45-0.46) 1.14 (1.13-1.15) 0.71 (0.70-0.72) 0.76 (0.75-0.77) 1.11 (1.10-1.11) <0.4 1.65 (1.64-1.66) 2.03 (2.02-2.04) 1.67 (1.66-1.69) 1.87 (1.87-1.88) 0.82 (0.81-0.83) 1.65 (1.64-1.66) 1.35 (1.34-1.35) <0.5 1.67 (1.66-1.67) 1.74 (1.74-1.74) 1.85 (1.84-1.85) 1.70 (1.70-1.71) 1.74 (1.73-1.75) 1.35 (1.34-1.35) 0.87 (0.87-0.87) <0.6 1.55 (1.55-1.55) 1.12 (1.12-1.12) 1.32 (1.32-1.33) 1.33 (1.33-1.34) 1.34 (1.34-1.34) 0.93 (0.93-0.93) 0.76 (0.76-0.76) <0.7 1.23 (1.23-1.23) 0.91 (0.91-0.91) 1.03 (1.03-1.04) 1.08 (1.08-1.08) 0.98 (0.97-0.98) 0.81 (0.81-0.81) 0.74 (0.74-0.74) 0.5≤ &≤0.7 0.83 (0.83-0.83) 1.18 (1.1-1.18) 1.02 (1.02-1.03) 0.97 (0.96-0.97) 1.07 (1.07-1.07) 1.30 (1.30-1.30) 1.37 (1.37-1.37) >0.7 0.96 (0.96-0.96) 1.02 (1.03-1.03) 0.99 (0.99-0.99) 0.98 (0.98-0.98) 1.00 (1.00-1.00) 1.05 (1.05-1.05) 1.11 (1.10-1.11) >0.8 0.98 (0.97-0.98) 1.11 (1.11-1.11) 1.05 (1.05-1.05) 1.02 (1.02-1.02) 1.05 (1.05-1.05) 1.16 (1.16-1.16) 1.26 (1.26-1.26) >0.9 1.05 (1.05-1.05) 1.29 (1.29-1.29) 1.15 (1.15-1.15) 1.12 (1.12-1.12) 1.13 (1.13-1.13) 1.38 (1.38-1.38) 1.48 (1.48-1.48) >1.0 1.19 (1.19-1.19) 1.61 (1.61-1.61) 1.39 (1.38-1.39) 1.31 (1.31-1.31) 1.32 (1.32-1.32) 1.71 (1.71-1.72) 1.76 (1.76-1.76) >1.1 1.40 (1.40-1.40) 2.10 (2.10-2.10) 1.68 (1.68-1.68) 1.60 (1.60-1.60) 1.52 (1.52-1.52) 2.13 (2.13-2.13) 2.22 (2.22-2.22) >1.2 1.73 (1.73-1.73) 2.72 ( 2.71-2.72) 2.20 (2.20-2.20) 1.99 (1.99-1.99) 1.88 (1.88-1.88) 2.74 (2.74-2.74) 2.56 (2.56-2.56) >1.3 2.17 (2.16-2.17) 3.34 (3.34-3.35) 2.77 (2.76-2.77) 2.48 (2.48-2.48) 2.30 (2.30-2.30) 3.27 (3.27-3.28) 2.83 (2.83-2.84) >1.4 2.87 (2.87-2.87) 4.35 (4.34-4.35) 4.01 (4.00-4.02) 3.20 (3.20-3.20) 3.18 (3.17-3.18) 4.18 (4.17-4.18) 3.19 (3.18-3.19) >1.5 3.46 (3.46-3.47) 5.28 (5.27-5.29) 5.53 (5.52-5.55) 3.78 (3.77-3.78) 4.16 (4.15-4.16) 4.70 (4.69-4.71) 3.47 (3.46-3.47) >1.6 4.40 (4.39-4.40) 6.17 (6.16-6.19) 7.05 (7.03-7.08) 4.67 (4.66-4.68) 4.84 (4.82-4.85) 5.53 (5.51-5.55) 4.18 (4.16-4.19) >1.7 5.49 (5.47-5.50) 7.13 (7.11-7.15) 7.92 (7.88-7.95) 5.85 (5.84-5.86) 5.85 (5.83-5.87) 7.48 (7.45-7.51) 4.82 (4.80-4.83) >1.8 5.73 (5.70-5.75) 6.98 (6.96-7.00) 8.79 (8.74-8.83) 5.80 (5.79-5.82) 6.03 (6.00-6.05) 7.13 (7.10-7.17) 4.78 (4.76-4.80) Data are presented with 95% confidence interval. +LR for admission increased with increasing MSI values in each age group; however, elderly patients ≥65 years demon- strated the lowest +LR of 4.81 (95% CI 4.79-4.83) for MSI>1.7 compared to +LRs of 5.85 (95% CI 5.83-5.87) and 7.48 (95% CI 7.45-7.51) in those 18-44 years old and 45-64 years old, re- spectively. An explanation to this finding could be the use of antihypertensives in elderly, including beta blockers that may blunt HR response, which results in lower MSI values. Another explanation could be the lower threshold of admit- ting elderly patients regardless of MSI due to polypharmacy and multiple comorbidities. Our study showed promising results and the findings can be used in the general ED population, to help identify sick pa- tients and allow the ED team to allocate resources effectively and efficiently. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index B. Hamade et al. 6 5. Limitations Our study has certain limitations. This was a retrospec- tive study that used a representative national data, of which 41,552,879 (7.23 %) encounters were excluded due to miss- ing data. The cohort was not uniform in terms of diagnosis, and thus patients presenting with disease states that do not affect stroke volume might have potentially skewed our re- sults and generated a different MSI threshold than described in previous studies. As data was not available to us, we also could not account for pre-ED interventions, such as crystal- loid and blood product administration, nor could we account for medication profiles, both of which could have potentially affected our results; specifically, the use of beta blockers al- tering the heart rate, or oral peripheral vasoconstrictors and other antihypertensive medications that would affect phys- iological responses to illness, thus affecting MSI values. We could not account for the way MAP and BP were measured non-invasively by using automated machines such as oscil- lometry or manually via sphygmomanometer, which some- times yields different values (13). Data on serial vital signs was not available to study serial MSI measurements, which can have significant prognostic reliability. We also did not have data regarding level of care upon admission. Our study derived MSI thresholds for all ED comers and more research is needed in specific patient populations, such as trauma and sepsis, as different MSI threshold maybe derived. In addition, our research focused on one compound variable -the MSI- and future studies are needed conducting multivariate anal- ysis including the MSI in order to construct a more compre- hensive best fit model. Finally, mortality could have been a result of different complications during the hospital course, and not a direct effect of the presenting ED complaint, but the limitation is due to our date set. 6. Conclusion Our large nationally representative retrospective study of all adult ED patients suggests that an initial MSI >1.7 in the gen- eral adult ED population is an important predictor of admis- sion and in-hospital mortality. Further studies on MSI are needed in various specific patient populations to further as- sess its relationship to the current triage systems, its impact on resource utilization and allocation in the ED, and on the decision making regarding the level of care of the admitting unit. 7. Declarations 7.1. Acknowledgments None. 7.2. Conflict of interest None. 7.3. Fundings and supports None. 7.4. Authors’ contribution BH, JDB, and NAJ conceived the study. JDB and NAJ obtained IRB approval. BH, JDB, and NAJ designed the study. YH ob- tained the data and provided statistical analysis of the data. BH, JDB and NAJ interpreted the data analysis. BH, JDB, BK, and NAJ drafted various sections of the manuscript, and all authors contributed substantially to its revision. BH takes re- sponsibility for the paper as a whole. The corresponding au- thor (NAJ) attests that all listed authors meet authorship cri- teria and that no others meeting the criteria have been omit- ted. References 1. Allgower M, Burri C. ["Shock index"]. Dtsch Med Wochenschr. 1967;92(43):1947-50. German. 2. Rady MY, Nightingale P, Little RA, Edwards JD. Shock in- dex: a re-evaluation in acute circulatory failure. Resusci- tation. 1992;23(3):227-34. 3. Rady MY, Smithline HA, Blake H, Nowak R, Rivers E. A comparison of the shock index and conventional vital signs to identify acute, critical illness in the emergency department. Ann Emerg Med. 1994;24(4):685-90. 4. Cannon CM, Braxton CC, Kling-Smith M, Mahnken JD, Carlton E, Moncure M. Utility of the shock index in predicting mortality in traumatically injured patients. J Trauma. 2009;67(6):1426-30. 5. McNab A, Burns B, Bhullar I, Chesire D, Kerwin A. An analysis of shock index as a correlate for outcomes in trauma by age group. Surgery. 2013;154(2):384-7. 6. Montoya KF CJ, Calle-Toro JS, Nunez LR, Poveda G. Shock inex as a mortality predictor in patients with acute poly- trauma. J Acute Dis. 2015;4(3):202-4. 7. Sankaran P, Kamath AV, Tariq SM, Ruffell H, Smith AC, Prentice P, et al. Are shock index and adjusted shock index useful in predicting mortality and length of stay in community-acquired pneumonia? Eur J Intern Med. 2011;22(3):282-5. 8. Berger T, Green J, Horeczko T, Hagar Y, Garg N, Suarez A, et al. Shock index and early recognition of sepsis in the emergency department: pilot study. West J Emerg Med. 2013;14(2):168-74. 9. Nakasone Y, Ikeda O, Yamashita Y, Kudoh K, Shigematsu Y, Harada K. Shock index correlates with extravasation on angiographs of gastrointestinal hemorrhage: a lo- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 7 Archives of Academic Emergency Medicine. 2023; 11(1): e34 gistics regression analysis. Cardiovasc Intervent Radiol. 2007;30(5):861-5. 10. Huang B, Yang Y, Zhu J, Liang Y, Tan H, Yu L, et al. Useful- ness of the admission shock index for predicting short- term outcomes in patients with ST-segment elevation myocardial infarction. Am J Cardiol. 2014;114(9):1315-21. 11. Reinstadler SJ, Fuernau G, Eitel C, de Waha S, Desch S, Metzler B, et al. Shock Index as a Predictor of Myocardial Damage and Clinical Outcome in ST-Elevation Myocar- dial Infarction. Circ J. 2016;80(4):924-30. 12. Al Jalbout N, Balhara KS, Hamade B, Hsieh YH, Kelen GD, Bayram JD. Shock index as a predictor of hospital admis- sion and inpatient mortality in a US national database of emergency departments. Emerg Med J. 2019, 36(5):293- 297. 13. Hamade B, DT H. Hemodynamic Monitoring. Tintinalli’s Emergency Medicine: A Comprehensive Study Guide. 9th ed. New York: McGraw-Hill Education; 2019. 14. Kirkman E, Watts S. Haemodynamic changes in trauma. Br J Anaesth. 2014;113(2):266-75. 15. Thooft A, Favory R, Salgado DR, Taccone FS, Donadello K, De Backer D, et al. Effects of changes in arterial pres- sure on organ perfusion during septic shock. Crit Care. 2011;15(5):R222. 16. Shapiro DS, Loiacono LA. Mean arterial pressure: thera- peutic goals and pharmacologic support. Crit Care Clin. 2010;26(2):285-93. 17. Liu YC, Liu JH, Fang ZA, Shan GL, Xu J, Qi ZW, et al. Mod- ified shock index and mortality rate of emergency pa- tients. World J Emerg Med. 2012;3(2):114-7. 18. Singh A, Ali S, Agarwal A, Srivastava RN. Correlation of shock index and modified shock index with the outcome of adult trauma patients: a prospective study of 9860 pa- tients. N Am J Med Sci. 2014;6(9):450-2. 19. Wang IJ, Bae BK, Park SW, Cho YM, Lee DS, Min MK, et al. Pre-hospital modified shock index for prediction of mas- sive transfusion and mortality in trauma patients. Am J Emerg Med. 2019, 38(2):187-190. 20. Abreu G, Azevedo P, Galvao Braga C, Vieira C, Alvares Pereira M, Martins J, et al. Modified shock index: A bed- side clinical index for risk assessment of ST-segment el- evation myocardial infarction at presentation. Rev Port Cardiol. 2018;37(6):481-8. 21. Jayaprakash N, Gajic O, Frank RD, Smischney N. Ele- vated modified shock index in early sepsis is associated with myocardial dysfunction and mortality. J Crit Care. 2018;43:30-5. 22. Shangguan Q, Xu JS, Su H, Li JX, Wang WY, Hong K, et al. Modified shock index is a predictor for 7-day outcomes in patients with STEMI. Am J Emerg Med. 2015;33(8):1072- 5. 23. Smischney NJ, Seisa MO, Heise KJ, Schroeder DR, Weis- ter TJ, Diedrich DA. Elevated Modified Shock Index Within 24 Hours of ICU Admission Is an Early Indica- tor of Mortality in the Critically Ill. J Intensive Care Med. 2018;33(10):582-8. 24. Terceros-Almanza LJ, Garcia-Fuentes C, Bermejo- Aznarez S, Prieto-Del Portillo IJ, Mudarra-Reche C, Saez-de la Fuente I, et al. Prediction of massive bleeding. Shock index and modified shock index. Med Intensiva. 2017;41(9):532-8. 25. Trivedi S, Demirci O, Arteaga G, Kashyap R, Smischney NJ. Evaluation of preintubation shock index and mod- ified shock index as predictors of postintubation hy- potension and other short-term outcomes. J Crit Care. 2015;30(4):861 e1-7. 26. Torabi M, Mirafzal A, Rastegari A, Sadeghkhani N. Asso- ciation of triage time Shock Index, Modified Shock Index, and Age Shock Index with mortality in Emergency Sever- ity Index level 2 patients. Am J Emerg Med. 2016;34(1):63- 8. 27. Kim SY, Hong KJ, Shin SD, Ro YS, Ahn KO, Kim YJ, et al. Validation of the Shock Index, Modified Shock Index, and Age Shock Index for Predicting Mortality of Geriatric Trauma Patients in Emergency Departments. J Korean Med Sci. 2016;31(12):2026-32. 28. National Center For Health Statistics. The National Ambulatory Medical Care Survey (NAMCS) description [Available from: http://www.cdc.gov/nchs/ahcd/ahcd_questionannaires. htm. 29. Grimes DA, Schulz KF. Refining clinical diagnosis with likelihood ratios. Lancet. 2005;365(9469):1500-5. 30. Husaini BA, Levine RS, Norris KC, Cain V, Bazargan M, Moonis M. Heart Failure Hospitalization by Race/Ethnicity, Gender and Age in California: Im- plications for Prevention. Ethn Dis. 2016;26(3):345-54. 31. Talbott EO, Rager JR, Brink LL, Benson SM, Bilonick RA, Wu WC, et al. Trends in acute myocardial infarction hos- pitalization rates for US States in the CDC tracking net- work. PLoS One. 2013;8(5):e64457. 32. Ramirez L, Kim-Tenser MA, Sanossian N, Cen S, Wen G, He S, et al. Trends in Acute Ischemic Stroke Hospitaliza- tions in the United States. J Am Heart Assoc. 2016,11;5(5). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index Introduction Methods Results Discussion Limitations Conclusion Declarations References