Emergency. 2017; 5 (1): e31 OR I G I N A L RE S E A RC H Worthing Physiological Score vs Revised Trauma Score in Outcome Prediction of Trauma patients; a Comparative Study Babak Nakhjavan-Shahraki1, Mahmoud Yousefifard2, Mohammad Javad Hajighanbari3, Parviz Karimi4, Masoud Baikpour5, Jalaledin Mirzay Razaz6, Mehdi Yaseri7, Kavous Shahsavari8, Fatemeh Mahdizadeh9, Mostafa Hosseini7∗ 1. Sina Trauma and Surgery Research Center, Tehran University of Medical Sciences, Tehran, Iran. 2. Physiology Research Center and Department of Physiology, Faculty of Medicine, Iran University of Medical Sciences, Tehran, Iran. 3. Department of Emergency Medicine, Hafte Tir Hospital, Iran University of Medical Sciences, Tehran, Iran. 4. Department of Emergency Medicine, Robatkarim Hospital, Iran University of Medical Sciences, Tehran, Iran. 5. Department of Medicine, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran. 6. Department of Community Nutrition, Faculty of Nutrition and Food Technology, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 7. Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran. 8. Road Traffic Injury Research Center, Tabriz University of Medical Sciences, Tabriz, Iran. 9. Department of Emergency Medicine, Ilan University of Medical Sciences, Ilam, Iran. Received: October 2016; Accepted: November 2016; Published online: 11 January 2017 Abstract: Introduction: Awareness about the outcome of trauma patients in the emergency department (ED) has become a topic of interest. Accordingly, the present study aimed to compare the rapid trauma score (RTS) and worthing physiological scoring system (WPSS) in predicting in-hospital mortality and poor outcome of trauma patients. Methods: In this comparative study trauma patients brought to five EDs in different cities of Iran during the year 2016 were included. After data collection, discriminatory power and calibration of the models were assessed and compared using STATA 11. Results: 2148 patients with the mean age of 39.50±17.27 years were included (75.56% males). The AUC of RTS and WPSS models for prediction of mortality were 0.86 (95% CI: 0.82-0.90) and 0.91 (95% CI: 0.87-0.94), respectively (p=0.006). RTS had a sensitivity of 71.54 (95% CI: 62.59-79.13) and a specificity of 97.38 (95% CI: 96.56-98.01) in prediction of mortality. These measures for the WPSS were 87.80 (95% CI: 80.38-92.78) and 83.45 (95% CI: 81.75-85.04), respectively. The AUC of RTS and WPSS in predicting poor outcome were 0.81 (95% CI: 0.77-0.85) and 0.89 (95% CI: 0.85-0.92), respectively (p<0.0001). Conclusion: The findings showed a higher prognostic value for the WPSS model in predicting mortality and severe disabilities in trauma patients compared to the RTS model. Both models had good overall performance in prediction of mortality and poor outcome. Keywords: Trauma Severity Indices; Prognosis; Trauma; emergency department; decision support techniques © Copyright (2017) Shahid Beheshti University of Medical Sciences Cite this article as: Nakhjavan-Shahraki B, Yousefifard M, Hajighanbari M, Karimi P, Baikpour M, Mirzay Razaz J, Yaseri M, S, Shahsavari K, Mahdizadeh F, Hosseini M. Worthing Physiological Score vs Revised Trauma Score in Outcome Prediction of Trauma patients; a Comparative Study. Emergency. 2017; 5(1): e31. ∗Corresponding Author: Mostafa Hosseini, Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sci- ences, Poursina Ave, Tehran, Iran; Email: mhossein110@yahoo.com; Tel: This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com B. Nakhjavan-Shahraki et al. 2 1. Introduction Q uick assessment of trauma patients and knowledge about the severity of their injuries can significantly affect the outcome of these patients, decrease their mortality rates and their associated disabilities (1-7). Aware- ness about the final outcome of trauma patients in the emer- gency setting has become a topic of discussion in recent years and various methods have been proposed to address this is- sue. In this regard, different scoring systems have been de- veloped (8-10). Over the years these scoring systems be- came so popular among physicians that encouraged further development of these models. Application of these scor- ing systems help in identifying high-risk patients (9), which leads to a better controlled management and treatment of patients. Nevertheless, each of these scoring systems have their own shortcomings, some of which include numerous variables involved in the model, complicated calculations needed to reach a conclusion (e.g. injury severity score) and their validity and reliability not having been assessed in different clinical settings. These limitation encouraged re- searchers to design better systems, the examples of which are the revised trauma score (RTS), rapid acute physiology score (RAPS), rapid emergency medicine score (REMS) and Wor- thing Physiological Scoring System (WPSS) (11-14). RTS is a scoring system based on physiologic variables of Glasgow coma scale (GCS), systolic blood pressure (SBP) and respi- ratory rate (RR), in which the GCS has higher weight com- pared to the other two variables. However, its low prognostic value for outcome of trauma patients pushed the researchers to search for other scoring systems (12, 14). WPSS was an- other scoring system presented in the year 2007. The model was designed based on a study conducted on 3184 patients that found the 6 factors of RR, pulse rate, SBP, body temper- ature, the oxygen saturation and the level of consciousness assessed on arrival of the patients to be able to predict their mortality (11). However, little information is available on the overall validity of this model. Accordingly, the present study was designed to assess and compared the value of WPSS and RTS models in prediction of in-hospital mortality and poor outcome in trauma patients presenting to the emergency de- partments. 2. Methods 2.1. Study design and setting In this prospective cross-sectional study, trauma patients brought to five emergency departments in different cities +982188989125; Fax: +982188989127 Table 1: Baseline characteristics of studied patients Variable Value Age (year) 39.50 ± 17.27 Gender(n, %) Male 1623 (75.56) Female 525 (24.44) Trauma mechanism Motorcycle accident 591 (27.51) Car rider accident 518 (24.12) Pedestrian 378 (17.60) Falls more than 3 meters 152 (7.08) Falls less than 3 meters 201 (9.36) Other 308 (14.34) GCS 14.4 ± 2.19 HR (beat/minute) 87.60 ± 15.63 SBP (mmHg) 115.38 ± 15.36 DBP (mmHg) 73.49 ± 10.07 O2 saturation 94.78 ± 5.80 Temperature (Celsius) 36.81 ± 0.90 RR (number/minute) 16.46 ± 6.15 Outcome Good recovery 1630 (75.88) Moderate disability 342 (15.92) Severe disability 53 (2.47) Death 123 (5.73) Data were presented as mean ± standard deviation or fre- quency and percentage; GCS: Glasgow coma scale; HR: heart rate; SBP: systolic blood pressure; DBP: diastolic blood pres- sure; O2 saturation: arterial oxygen saturation; RR: respira- tory rate. of Iran (Tehran, Ilam, Jahrom, Tabriz and Urmia) from May to October 2016 were included. Completed checklists were posted to Tehran and reviewed by the senior researcher. Af- ter verifying their validity, gathered data were analyzed using the statistical software. The Ethics Committee of Tehran Uni- versity of Medical Sciences reviewed and approved the study protocol. The guidelines laid down by Declaration of Helsinki were adhered to by all the authors throughout the survey and all the included patients or their family members signed an informed written consent for participating in the study. 2.2. Participants Trauma patients older than 18 years of age brought to the designated emergency departments were included as the study population through a convenience sampling method. Pregnancy and death before admission to the emergency de- partment were considered as the exclusion criteria. 2.3. Data gathering Gathered information included age, gender, trauma mecha- nism, vital signs, arterial oxygen saturation level, and level of consciousness on admission. The patients were followed throughout their hospital stay and their final outcome (ex- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 3 Emergency. 2017; 5 (1): e31 Figure 1: Area under the curve (AUC) of revised trauma score (RTS) and worthing physiological scoring system (WPSS) in prediction of in- hospital mortality and poor outcome. pired vs. alive) along with the condition in which the patient was discharged from the hospital (full recovery, moderate disability, severe disability or vegetative state) were recorded. 2.4. Assessed outcomes Glasgow outcome scale (GOS) was used to assess the final outcome of the patient when being discharged from the hos- pital (20). In-hospital mortality was considered as the pri- mary outcome and discharge with a severe disability (based on GOS) was considered as the secondary outcome. 2.5. Statistical Analysis In order to calculate the minimum sample size needed for this survey, the rate of in-hospital mortality in trauma patients was considered as 5.2% based on previous reports (21). Accordingly, the minimum sample size was estimated at 1894 patients based on a 95% confidence interval (CI) (α=0.05), a 90% power (β=0.1) and a maximum error of 1.5% (d=0.015). Data analysis was performed by STATA 11.0 software. Severity of trauma were calculated for each patient based on RTS and WPSS models and the prognostic value of the systems was compared according to the discrimination power, calibration and overall performance. Discrimination was evaluated by measuring the area under the curve (AUC) of the receiver operating characteristic (ROC) curve and calculating the sensitivity, specificity, positive and nega- tive likelihood ratios with 95% CI. The method proposed by Cleves and Rick was used for comparing AUC for the two models (22). Calibration plot was constructed for as- sessment of general calibration in which the frequency of observed versus predicted mortality or poor outcome were compared. Overall performance was assessed by evaluating the predictive reliability and predictive accuracy based on the calculated Brier score. Finally, in order to assess the concordance between RTS-predicted and WPSS-predicted percent of mortality and poor outcome, Spearman’s rank coefficient was computed. A p value less than 0.05 was considered as statistically significance level in all analyses. 3. Results A total of 2148 patients with the mean age of 39.50±17.27 year were included in this survey (75.56% male). Motor- cycle accident was the most common trauma mechanism (75.65%). GCS ranged from 3-8 in 63 patients (2.98%), 9-12 in 36 patients (1.7%) and it was higher than 13 in 2014 cases (95.3%). Table 1 presents the basic characteristics of the studied subjects. Follow up of the subjects revealed that only 2.47% of the patients were discharged with severe disabilities and 5.73% of the cases expired. 3.1. Performance of RTS and WPSS in prediction of mortality 3.1.1. Discrimination The AUC of the two RTS and WPSS models for prediction of patients’ mortality was calculated to be 0.86 (95% CI: 0.82- 0.90) and 0.91 (95% CI: 0.87-0.94), respectively (p=0.006). The optimum cut-off level was found to be 1 for the RTS and 4 for This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com B. Nakhjavan-Shahraki et al. 4 Figure 2: Calibration plots of revised trauma score (RTS) and worthing physiological scoring system (WPSS) in prediction of in-hospital mor- tality and poor outcome. the WPSS. The sensitivity and specificity of the RTS model for predicting patients’ mortality was calculated to be 71.54 (95% CI: 62.59-79.13) and 97.38 (95% CI: 96.56-98.01), respectively. These measures for the WPSS were found to be 87.80 (95% CI: 80.38-92.78) and 83.45 (95% CI: 81.75-85.04), respectively (Figure 1 and Table 2). 3.1.2. Calibration Both scoring systems had good calibration (agreement be- tween observed and predicted rate of mortality) in prediction of mortality. Calibration plot of the RTS model had a slope of 1.04 and an intercept of 0.02. The mentioned measured were calculated to be 1.02 and 0.01 for the WPSS model, respec- tively (Figure 2). 3.1.3. Overall performance Brier score and scaled reliability of the RTS model in predic- tion of mortality were 0.024 and zero, respectively. These measures were found to be 0.031 and 0.0003 for the WPSS model, respectively. The findings exhibit the high predictive accuracy and reliability of both models (Table 3). 3.2. Performance of RTS and WPSS in prediction of poor outcome 3.2.1. Discrimination The RTS model had an AUC of 0.81 (95% CI: 0.77-0.85) in predicting poor outcome, which was significantly lower than that of WPSS model with an AUC of 0.89 (95% CI: 0.85-0.92) (p<0.0001). The sensitivity and specificity of the RTS model for predicting poor outcome was found to be 61.93 (95% CI: 54.29-69.05) and 98.38 (95% CI: 97.69-98.87) considering the cut-off value of 1, respectively. These figures for the WPSS model with a cut-off level of 4 were calculated to be 82.95 (95% CI: 76.40-88.03) and 84.95 (95% CI: 83.27-86.47), respec- tively (Figure 1 and Table 2). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 5 Emergency. 2017; 5 (1): e31 Figure 3: Concordance between revised trauma score (RTS) predicted and worthing physiological scoring system (WPSS) predicted percent of mortality and poor outcome. Table 2: Screening performance characteristics of revised trauma score (RTS) and worthing physiological scoring system (WPSS) in prediction of mortality and poor outcome Characteristics Mortality Poor outcome RTS WPSS RTS WPSS True positive 88 108 109 146 True negative 1972 1690 1940 1675 False positive 53 335 32 297 False negative 35 15 67 30 Sensitivity 71.54 (62.59-79.13) 87.80 (80.38-92.78) 61.93 (54.29-69.05) 82.95 (76.40-88.03) Specificity 97.38 (96.56-98.01) 83.45 (81.75-85.04) 98.38 (97.69-98.87) 84.94 (83.27-86.47) PositiveLR 27.34 (20.49-36.46) 5.31 (4.72-5.97) 38.16 (28.56-54.85) 5.51 (4.86-6.24) Negative LR 0.29 (0.22-0.39) 0.15 (0.09-0.24) 0.39 (0.32-0.47) 0.20 (0.14-0.28) ∗ Data are presented as estimated value and 95% confidence interval. LR: Likelihood ratio. Table 3: Overall performance of revised trauma score (RTS) and worthing physiological scoring system (WPSS) in prediction of in-hospital mortality and poor outcome Characteristics Mortality Poor outcome RTS WPSS RTS WPSS Brier score 0.026 0.031 0.038 0.045 Scaled reliability <0.0001 0.0003 <0.0001 0.001 3.2.2. Calibration Both scoring systems had good calibration in predicting poor outcome of patients as well. The slope and intercept of the RTS model’s calibration plot were 1.05 and 0.04, respectively. The mentioned measures were 0.87 and 0.01 for the WPSS model’s calibration plot (Figure 2). 3.2.3. Overall performance Brier score and scaled reliability calculated for RTS model in predicting patients’ poor outcome were 0.034 and zero, while these measures were found to be 0.045 and 0.001 for the WPSS model, respectively (Table 3). Both RTS and WPSS models have good overall performance in prediction of poor outcome. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com B. Nakhjavan-Shahraki et al. 6 3.3. Concordance between RTS and WPSS There was good concordance between RTS and WPSS mod- els in prediction of mortality (r=0.63; p <0.001) and poor out- come (r=0.68; p <0.001) (Figure 3). 4. Discussion: Classifying the severity of trauma in emergency settings is a challenging issue for the physicians. Scoring system can help to diagnosis of high risk patient. However, each scor- ing systems have specific advantages and limitations. The present study compared the two physiologic scoring systems of RTS and WPSS and found that the value of WPSS model in predicting mortality and occurrence of severe disabilities in trauma patients in higher than that of the RTS model. Al- though the RTS model involves simple criteria for estimating the severity of injuries, its prognostic value is at a moderate level. An acceptable scoring system for prediction of an out- come should have a high screening value along with a high sensitivity. The sensitivity of RTS model in prediction of mor- tality and poor outcome were 71.54% and 61.93%, respec- tively, while similar figures for the WPSS model were found to be 82.95% and 87.8%. Despite the greater number of vari- ables included in the WPSS model compared to RTS model, its application is easy (11). WPSS is a physiologic scoring sys- tem which incorporates the respiratory rate, pulse rate, body temperature, arterial oxygen saturation and the level of con- sciousness. These factors can be easily assessed and are rou- tinely evaluated in the emergency departments. The only factor that is not precisely measured in the emergency set- tings is the body temperature. In the busy hours of an emer- gency department, physicians or nurses might not pay ade- quate attention to accurate measuring of the patients’ body temperature, while assessment of this factor plays an impor- tant role in predicting the outcome of patients. Therefore, it is suggested that more attention be paid to the body tem- perature as a physiologic factor in patients referring to emer- gency departments. Few studies have assessed the prognos- tic value of WPSS for patients’ mortality. The findings of the present survey is congruent with the results of the study con- ducted by Duckitt et al. which has shown that the WPSS model is a better index for predicting patients’ mortality compared to the early-warning scoring system (11). Ha et al. also reported that both rapid emergency medicine score and WPSS have good prognostic values for mortality of patients in the emergency department, with the latter slightly supe- rior to the former scoring system (23). Similarly, Brabrand et al. referred to the WPSS model as a scoring system with ac- ceptable discriminatory power and calibration in predicting patients’ mortality (24). In this regard, it seems that the WPSS model can be used as a screening tool for classifying trauma patients in the emergency departments. The large sample size of the present study and its multi-center setting could be considered as the strengths of this survey warranting its power. Moreover, the results of this study can be generalized to the whole Iranian population since patients were included from emergency departments located in five different cities of Tehran, Ilam, Jahrom, Tabriz and Urmia. 5. Limitations The findings might be subject to selection bias due to the convenience sampling method used for inclusion of patients. Another factor that might have confounded the results of this survey was the probably inaccurate measurement of the pa- tients’ axillary body temperature in the overcrowded emer- gency departments. 6. Conclusion The findings showed a higher prognostic value for the WPSS model in predicting mortality and severe disabilities in trauma patients compared to the RTS model. Both models had good overall performance in prediction of mortality and poor outcome. 7. Appendix 7.1. Acknowledgements The authors wish to acknowledge the cooperation of Tehran, Tabriz, Urmia, Jahrom and Ilam emergency departments in providing patient data. 7.2. Author contribution All authors passed four criteria for authorship contribution based on recommendations of the International Committee of Medical Journal Editors. 7.3. Funding This research has been supported by a Tehran University of Medical Sciences and Health Services grant and Sina Trauma and Surgery research center, Tehran University of Medical Sciences grant (grant number: 94-04- 38-30757). 7.4. Conflict of interest The authors report no declarations of interest. References 1. Hosseini M, Ghelichkhani P, Baikpour M, Tafakhori A, Asady H, Ghanbari MJH, et al. Diagnostic Accuracy of Ultrasonography and Radiography in Detection of Pulmonary Contusion; a Systematic Review and Meta- Analysis. Emergency. 2015;3(4):127-36. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 7 Emergency. 2017; 5 (1): e31 2. Rahimi-Movaghar V, Yousefifard M, Ghelichkhani P, Baikpour M, Tafakhori A, Asady H, et al. Application of Ultrasonography and Radiography in Detection of Hemothorax; a Systematic Review and Meta-Analysis. Emergency. 2016;4(3):116-26. 3. Safari S, Yousefifard M, Baikpour M, Rahimi-Movaghar V, Abiri S, Falaki M, et al. Validation of thoracic injury rule out criteria as a decision instrument for screening of chest radiography in blunt thoracic trauma. Journal of clinical orthopaedics and trauma. 2016;7(2):95-100. 4. Safari S, Yousefifard M, Hashemi B, Baratloo A, Forouzan- far MM, Rahmati F, et al. The value of serum crea- tine kinase in predicting the risk of rhabdomyolysis- induced acute kidney injury: a systematic review and meta-analysis. Clinical and experimental nephrology. 2016;20(2):153-61. 5. Yousefifard M, Baikpour M, Ghelichkhani P, Asady H, Darafarin A, Amini Esfahani MR, et al. Comparison of Ul- trasonography and Radiography in Detection of Thoracic Bone Fractures; a Systematic Review and Meta-Analysis. Emergency. 2016;4(2):55-64. 6. Yousefifard M, Baikpour M, Ghelichkhani P, Asady H, Darafarin A, Esfahani MRA, et al. Comparison of Ultra- sonography and Radiography in Detection of Thoracic Bone Fractures; a Systematic Review and Meta-Analysis. Emergency. 2016;4(2):55-64. 7. Yousefifard M, Baikpour M, Ghelichkhani P, Asady H, Nia KS, Jafari AM, et al. Screening Performance Char- acteristic of Ultrasonography and Radiography in De- tection of Pleural Effusion; a Meta-Analysis. Emergency. 2015;3(4):127–36. 8. Della Corte F, Vignazia G, Cavaglia M. Trauma scoring systems. Anaesthesia, Pain, Intensive Care and Emer- gency Medicine–APICE: Springer; 2002. p. 351-61. 9. Lefering R. Trauma scoring systems. Current opinion in critical care. 2012;18(6):637-40. 10. McEvoy S, Walker SM. Trauma scoring systems. The sci- entific basis of injury prevention and control. 2004:62-74. 11. Duckitt R, Buxton-Thomas R, Walker J, Cheek E, Be- wick V, Venn R, et al. Worthing physiological scoring system: derivation and validation of a physiological early-warning system for medical admissions. An obser- vational, population-based single-centre study. British journal of anaesthesia. 2007;98(6):769-74. 12. Gabbe BJ, Cameron PA, Finch CF. Is the revised trauma score still useful? ANZ journal of surgery. 2003;73(11):944-8. 13. Olsson T, Terent A, Lind L. Rapid Emergency Medicine Score: a new prognostic tool for inâĂŘhospital mortality in nonsurgical emergency department patients. Journal of internal medicine. 2004;255(5):579-87. 14. Roorda J, Van Beeck E, Stapert J, Ten Wolde W. Eval- uating performance of the revised trauma score as a triage instrument in the prehospital setting. Injury. 1996;27(3):163-7. 15. de Pennington J, Laurenson J, Lebus C, Sihota S, Smith P. Evaluation of early warning systems on a medical admissions unit. Journal of the Intensive Care Society. 2005;6(2):19-21. 16. Goldhill D. Preventing surgical deaths: critical care and intensive care outreach services in the postoperative pe- riod. British journal of anaesthesia. 2005;95(1):88-94. 17. Olsson T, Lind L. Comparison of the Rapid Emer- gency Medicine Score and APACHE II in Nonsurgical Emergency Department Patients. Academic Emergency Medicine. 2003;10(10):1040-8. 18. Olsson T, Terent A, Lind L. Rapid Emergency Medicine Score Can Predict Long-term Mortality in Nonsurgical Emergency Department Patients. Academic Emergency Medicine. 2004;11(10):1008-13. 19. Williams E, Subbe C, Gemmell L, Morgan R, Park G, McEl- ligot M, et al. Outreach critical care–cash for no ques- tions? British journal of anaesthesia. 2003;90(5):699-702. 20. Anderson SI, Housley AM, Jones PA, Slattery J, Miller JD. Glasgow Outcome Scale: an inter-rater reliability study. Brain Injury. 1993;7(4):309-17. 21. Imhoff BF, Thompson NJ, Hastings MA, Nazir N, Mon- cure M, Cannon CM. Rapid Emergency Medicine Score (REMS) in the trauma population: a retrospective study. BMJ Open. 2014;4(5). 22. Cleves MA, Rock L. From the help desk: Comparing ar- eas under receiver operating characteristic curves from two or more probit or logit models. The Stata Journal. 2002;2(3):301-13. 23. Ha DT, Dang TQ, Tran NV, Vo NY, Nguyen ND, Nguyen TV. Prognostic performance of the Rapid Emergency Medicine Score (REMS) and Worthing Physiological Scoring system (WPS) in emergency department. Inter- national journal of emergency medicine. 2015;8(1):18. 24. Brabrand M, Hallas J, Knudsen T. Nurses and Physicians in a Medical Admission Unit Can Accurately Predict Mor- tality of Acutely Admitted Patients: A Prospective Cohort Study. PLoS ONE. 2014;9(7):e101739. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com Introduction Methods Results Discussion: Limitations Conclusion Appendix References