Canadian Paediatric Triage and Acuity Scale: assessment in a European pediatric emergency department

Size: px

Start display at page:

Download "Canadian Paediatric Triage and Acuity Scale: assessment in a European pediatric emergency department"

Sybil Porter
5 years ago
Views:

1 ORIGINAL ARTICLE Canadian Paediatric Triage and Acuity Scale: assessment in a European pediatric emergency department ANA FERNÁNDEZ 1, JOSÉ IGNACIO PIJOAN 2, MARÍA ISABEL ARES 1, SANTIAGO MINTEGI 1, FRANCISCO JAVIER BENITO 1 1 Servicio Urgencias de Pediatría. 2 Unidad de Epidemiología Clínica. Hospital de Cruces. Barakaldo. Bizkaia, Spain. CORRESPONDENCE: Dra. Ana Fernández Landaluce Urgencias de Pediatría Hospital de Cruces Plaza de Cruces, s/n Cruces-Barakaldo Bizkaia, Spain ana.fernandez@euskalnet.net RECEIVED: ACCEPTED: CONFLICT OF INTEREST: None Objectives: To assess the functionality and reliability of a computerized version of the Canadian Paediatric Triage and Acuity Scale (PaedCTAS) in a European pediatric emergency department. Methods: Retrospective review of emergency cases in We collected data on the seriousness of each emergency, mean patient stay in the department, rate of admission, and tests ordered. The reliability of the scale was estimated by calculating the agreement between raters theoretical classification of cases and the reference assessment. Results: A total of emergencies were analyzed. The distribution by levels of severity were as follows: level 1, 0.07%; level 2, 1.73%; level 3, 43.1%; level 4, 50%; and level 5, 5.1%. The frequency of orders for tests ranged from 63.6% at level 1 to 8% at level 5 (P<.01). The mean duration of stay increased progressively from 72 minutes for level 5 to 373 minutes for level 1. Admission rates also differed by severity level (from 84% at level 1 to 1.6% at level 5, P<.001). The reliability study showed that triage based on the Paediatric Assessment Triangle was exact in 93.6% of the cases; the κ index was 0.77 (95% confidence interval [CI] ). The scale correctly assigned the level of severity or included only a slight discrepancy in 80.3% of the cases. The κ index was 0.47 (95% CI, ). Conclusions: The PaedCTAS seems to be applicable to European pediatric emergency department use although certain features of the scale could be improved for application by our emergency department. [Emergencias 2010;22: ] Key words: Triage. Canadian Paediatric Triage and Acuity Scale (PaedCTAS). Reliability. Introduction The progressive increase in the rate of consultation in pediatric emergency departments (PED) has been reported worldwide. In 2003, the rate of PED consultation in USA of children under 1 year was 97.5%, and in children aged 1-12 and years this was 40% (higher than in any other age group consulting the emergency department (ED), excluding patients older than 65 years) 1. In our case, the number of PED emergencies increased from 38,659 in 1995 to over 60,000 in 2004 (an increase of approximately 55%) 2. ED saturation has led to the development of strategies to effectively manage the increase of patient flow 3. These include triage or classification of patients on arrival at the ED. Triage has been performed in a formal or informal manner ever since the first EDs were established and is inherent in emergency work, but how to do it and the scales used vary widely around the world and even within countries. In 1993, the Australian Medical Committee, in an attempt to unify the different scales used in triage, proposed a single national 5-level triage scale 4. Since then another four systems of 5-level triage have been developed, supported by published concordance studies 5-8 : the Canadian Triage and Acuity Scale (CTAS), The Manchester Triage Scale, The Emergency Severity Index, and the Escala Andorrana de Triaje-Sistema español de triaje (EAT-SET). Over the last decade, pediatric emergency medicine has been widely developed and defined. The recognition that children requiring urgent attention have special needs involves both specialized personnel and specific tools for this work, Emergencias 2010; 22:

2 A. Fernández et al. one of which is pediatric triage protocols 9. In Europe, in 2008, according to a 2008 survey by the REPEM group (Research in European Pediatric Emergency Medicine) of the European Society of Emergency Medicine, conducted at 53 tertiary hospitals in 14 countries, 88% of PED had adopted pediatric triage protocols and most of them (69.8%) had a specific room for triage 10. That same year, the Working Group on Quality of the Spanish Society of Emergency Pediatrics published the results of a nationwide survey showing that 79% of respondents had adopted systematic classification of child patients 11. Derived from the adult Canadian Triage and Acuity Scale, the pediatric version (PaedCTAS) was developed and published in Since then several studies have evaluated it and shown a moderate degree of concordance and good internal validity demonstrated by the correlation between triage levels and severity markers Most of these studies have been conducted in Canadian PED. And the adult version (CTAS) has been validated at hospitals outside Canada with good results, but the PaedCTAS has yet to be validated in our pediatric population 18. In 2007, a new system of electronic triage was introduced in our PED, based on the Canadian PaedCTAS. In this context, the objective of this study was to evaluate the performance and reliability of the computerized version of PaedCTAS in a European PED. Method The study was carried out in a PED of tertiary university hospital in an urban area, attending approximately 60,000 patients per year. The PED is a staffed by a medical team dedicated exclusively to emergency attention of children age 0-14 years, along with other specialty pediatricians and medical internal residents (MIR) specializing in Pediatrics. On arrival at the PED, patients are assessed and classified by a senior nurse on triage duty. Throughout the year 2006, as part of a specific management plan for the triage area of our PED, we designed a triage system based on PaedCTAS. For this, we developed a scale adapted to our environment, and developed a triage software tool which appears as a mandatory screen showing the entire triage process, with automatic assignment of triage level and area of care location of the patient according to the information typed into the boxes. Prior to implementation, staff nurses, pediatricians and MIR physicians received theoretical and practical training on triage in general and this system in particular (5h course for nurses and 1h for pediatricians). The new computerized system of triage was launched in January According to this scale, patient assessment involves 3 steps. First, and most important, is pediatric triage assessment (PTA). Second, the Main Complaint, and third, physiological assessment (vital signs). The PTA is a quick, simple and practical tool to evaluate pediatric patients. Its three components (appearance, skin perfusion and respiratory effort) reliably reflect the overall physiological status of the patient. Regarding the third step, for logistical reasons (we have not yet established a specific triage bay) our PED has no systematic register of the vital constants in triage assessment. All patients classified as Level I (resuscitation required) or Level II (emergency) are taken directly to the relevant treatment area where a more thorough assessment is made. We carried out a retrospective observational study of the first year of operation of this system. We reviewed the records of all patients attended in our PED between January 22 and December 31, The following variables were analyzed: time to triage (time from the administrative record to triage), triage process time, time to be assessed by a physician, percentage of patients seen within recommended time for each level, level and care area assigned, mean stay time (MST) in the ED (time from the administrative record until discharge), additional tests requested, and final destination (hospitalization rate: percentage of patients requiring admission to the observation unit or a hospital ward). Reliability and inter-observer agreement were studied by comparing the classification of cases and examining 2 aspects: the degree of agreement between the ED nurse classifications and the accuracy of their responses with respect to a reference assessment made by experts. The participating members of staff (38 people) were nurses trained and experienced in pediatric emergencies who had completed the new triage training course and worked with the computerized triage system during at least 3 months. The study was conducted using written simulated cases (29), which the participants classified in real time, using the new triage system. The simulated cases were habitual pediatric situations in an ED, with basic information on age, sex and reason for consultation and a description of the patient on arrival at the ED. For each scenario, the participants recorded their initial assessment, the problem group and the severity level assigned. 356 Emergencias 2010; 22:

3 CANADIAN PAEDIATRIC TRIAGE AND ACUITY SCALE: ASSESSMENT IN A EUROPEAN PEDIATRIC EMERGENCY DEPARTMENT A committee of experts (PED physicians and senior nurses responsible for triage) classified all the cases, and this classification was used as the gold standard of reference. The statistical program used for analysis was Windows SPSS version 15.0 (SPSS Inc.). We used chi-square test or Fisher exact test for comparison of qualitative variables, and Student t test or Mann Whitney test for quantitative variables. Differences with a p value less than 0.05 were considered significant. Accuracy and agreement were calculated for each item separately (initial assessment, problem group and level assigned). To analyze the accuracy of responses with respect to the reference assessment, the following terms were defined: a) Total Agreement: a response showing exact match for all sections; b) Reasonable/Sufficient Agreement: a response with irrelevant disagreement; c) Minimal/Acceptable Disagreement: difference in group problem, one-level over-triage in level V patients, one-level under-triage in level IV patients, one-level over-triage in level I-III patients, d) Relevant/Unacceptable Disagreement: two-level under-triage or over-triage of any patient, or onelevel under-triage in level I-III patients. The degree of agreement for all items was calculated using linear weighted kappa index (Table 1). Given the extensive range of responses (16 groups of problems) and the inability to weight the difference between them, the kappa index was not calculated for the section entitled reason for consultation. Results During the study period there were 57,617 visits to our PED, 56% males and 50.7% under 36 months (mean age 51.9 months). All (99.9%) were classified and assigned an area of care within the first 10 minutes after arrival and in 98% of cases the process of triage lasted less than 5 minutes. Table 2 shows the distribution of severity levels, and the percentage of patients attended within the recommended time for each level. Thirteen percent of patients required further laboratory or radiologic tests. This percentage differed for each severity level, as follows: 63.6% for level I, 41.4% for level II; 16.8% for level III, 9.4% and 8.0% for levels IV and V (with significant differences P < 0.05 for all pairs of severity levels). About 6% of patients needed to be in the observation unit or a hospital ward. Hospitalization rate ranged from 84% for level I to Table 1. Kappa index values and level of agreement Kappa value 1.6% in levels IV and V. Figure 1 shows the rates of hospitalization and mean stay time for the different levels of severity. During the study period 247 patients left the PED before being assessed by the pediatrician (0.4%). Concerning reliability, in 93.6% of cases the response was accurate in relation to the reference assessment. The kappa index between the different nursing assessments was 0.77 (95% 0.75 to 0.79). For the chief complaint, 75.8% of the responses were accurate. And for severity level, complete or acceptable agreement (including irrelevant disagreement) with the reference rating was found in 80.3% of cases. The weighted kappa was 0.47 (95% ). Discussion Level of agreement Negligible Poor Moderate Good > 0.8 Very good This study appears to demonstrate good correlation between the PaedCTAS level of severity and MST, as well as hospitalization rates and the number of additional tests, in line with the results of recent Canadian studies. Gravel et al., using a computerized version, found a strong association between severity level, MST, hospital admissions and intensive care unit admissions. Ma et al. additionally found good correlation between severity, costs and the use of diagnostic resources Some of the results of our study deserve special mention. First, the so-called digital fingerprint (number of patients per level) in our PED differs from that reported in the literature regarding level II patients (1.73% vs. 5-10% in the literature) and level IV patients (50% in our PED versus 37-46%) 15,17,19. Although these differences may be due to incorrect application of the system leading to misclassification, the acceptable correlation between levels and severity suggest simply a difference in the profile of patients treated. Analysis of our results indicates a possible problem in differentiating between levels IV and V, because although the percentage of tests and the MST were different, the rate of hospitalization was the same for both levels. This problem is not mentioned by other authors, which suggests an Emergencias 2010; 22:

4 A. Fernández et al. Table 2. Digital footprint (number of patients per level) and percentage of patients seen within time recommended for each level Triage Level Nº of patients Percentage attended (%) within time recommended I 44 (0.1) 100 II 1,001 (1.7) 83 III 24,845 (43.1) 87 IV 28,804 (50) 89 V 2,923 (5.1) 97 anomaly of the scale in our ED that may be corrected by improved training of the staff performing triage (there may be over-triage of some patients classified as level V to avoid prolonging their waiting time). The user s guide on using the scale offers some reference rates of admission according to severity level. In our study, these rates were met for levels I and II, but were lower for levels III, IV and V. This fact has also been picked up by other authors. Gouin et al. found that hospitalization rates were lower than the reference rates for all levels and suggested that PaedCTAS seemed less accurate (compared with a previously used scale) in predicting the need for admission. Gravel and Ma also obtained similar results in more recent studies 15,17,19. The reliability, accuracy and global consistency of the computerized version of PaedCTAS were only moderate although the pediatric triage assessment (a fundamental component of identifying critically ill patients) showed a high level of agreement. Other studies (some using the computerized version of the scale) show similar results. In the study with theoretical scenarios, Bergeron et al. found that the kappa index was higher among nurses (0.51, 95% CI: 0.50 to 0.52) than ED pediatricians (0.39, 95% CI: ) 12. In 2007, Gravel et al., published a 2-stage experimental study that compared inter-observer agreement on nurse triage of theoretical scenarios using a traditional version of the PaedCTAs in the first phase and a computerized version of the same scale (Staturg) in the second phase. The computerized tool facilitated better concordance (kappa index 0.55 versus 0.51 for the traditional format) 13. A year later, the same team measured the level of agreement of this triage tool for classifying real patients and obtained a linear weighted kappa of 0.55, and a biquadratic kappa of 0.61) 14. Our reliability study had some limitations that may have influenced the results. No real patients were used, only simulated scenarios. This implies uniformity of data for the participant assessors, but also makes it difficult to extrapolate the results to actual practice. In other studies the level of inter-observer agreement was calculated for pairs of observers, whereas in our case this was measured in multiple observers, making it difficult to compare results. In general the overall correlation between triage level and severity markers means that the validity of the PaedCTAS was good, but future studies are needed to validate the different reasons for consultation, as recommended in the user guide updated and published in May 2008 by the CTAS National Working Group. In an attempt to further improve the concordance results and adapt the scale to new technologies (computerization of the scale), this update proposes a simplified version of the scale to unify the process of triage of adults and children, based on reasons for consultation and 1st and 2nd degree physiological modifiers 20. According to the results of the present study, the PaedCTAS seems applicable in a European PED but certain aspects could be improved, especially reliability (which may be corrected with amendments recently added) and its application in our PED (which should improve with training Figure 1. Hospitalization rate (%) and length of stay (mean and standard error of the mean). 358 Emergencias 2010; 22:

5 CANADIAN PAEDIATRIC TRIAGE AND ACUITY SCALE: ASSESSMENT IN A EUROPEAN PEDIATRIC EMERGENCY DEPARTMENT and greater experience in managing the system). Despite this, our experience in the use of PaedC- TAS is clearly positive; it has markedly improved the dynamics of work, care of patients and resource management in our PED. References 1 Hostetler MA, Mace S, Brown K, Finkler J, Hernandez D, Krug SE, et al. Subcommittee on Emergency Department Overcrowding and Children, Section of Paediatric Emergency Medicine, American College of Emergency Physicians. Emergency department overcrowding and children. Pediatr Emerg Care. 2007;23: Mintegi Raso S, Benito Fernández J, García González S. Demanda y asistencia en un servicio de urgencias hospitalario. An Esp Pediatr. 2004;61: Yen K, Gorelick MH. Strategies to improve flow in the paediatric emergency department. Pediatr Emerg Care. 2007;23: FitzGerald G. Emergency Department Triage [thesis]. Queensland, Australia: University of Queensland; Wuerz RC, Milne LW, Eitel DR, Travers D, Gilboy N. Reliability and validity of a new five-level triage instrument. Acad Emerg Med. 2000;7: Cooke MW, Jinks S. Does the Manchester triage system detect the critically ill? J Accid Emerg Med. 1999;16: Gómez J, Becerra O, Boneu F, Albert E, Ferrando JB, Medina M, et al. Validación clínica de la nueva versión del Programa de Ayuda al Triaje (web_e-pat v3) del Modelo Andorrano de Triaje (MAT) y Sistema Español de Triaje (SET). Fiabilidad, utilidad y validez en la población pediátrica y adulta. Emergencias. 2006;18: Beveridge R, Ducharme J, Janes L, Beanlieu S, Walter S. Reliability of the Canadian Emergency Department Triage and Acuity Scale: Interrater Agreement. Ann Emerg Med. 1999;34: Care of Children in the Emergency Department: Guidelines for Preparedness American Academy of Paediatrics, Committee on Paediatric Emergency Medicine and American College of Emergency Physicians, Paediatric Committee. Pediatrics. 2001;107: Mintegi S, Shavit I, Benito J. The REPEM group (Research in European Paediatric Emergency Medicine). Paediatric Emergency Care in Europe: A Descriptive Survey of 53 Tertiary Medical Centres. Pediatr Emerg Care. 2008;24: Luaces C, Ortiz J, Trenchs V, Pou J. Encuesta nacional sobre las urgencias pediátricas. Aspectos organizativos y funcionales. Emergencias. 2008;20: Bergeron S, Gouin S, Bailey B, Amre DK, Patel H. Agreement among paediatric health care professionals with the paediatric Canadian triage and acuity scale guidelines. Pediatr Emerg Care. 2004;20: Gravel J, Gouin S, Bailey B, Roy M, Bergeron S, Amre DK. Reliability of a computerized version of the Paediatric Canadian Triage and Acuity Scale. Acad Emerg Med. 2007;14: Gravel J, Gouin S, Manzano S, Arsenault M, Amre DX. Interrater Agreement between Nurses for the Paediatric Canadian Triage and Acuity Scale in a Tertiary Care Center. Acad Emerg Med. 2008;15: Ma W, Gafni A, Goldman RD. Correlation of the Canadian Paediatric Emergency Triage and Acuity Scale to ED resource utilization. Am J Emerg Med. 2008;26: Ma W, Jarvis AD, Goldman RD. Paediatric Canadian Triage and Acuity Scale as a predictor for outcome and resource utilization. CJEM. 2007;9: Gravel J, Manzano S, Arsenault M. Validity of the Canadian Paediatric Triage and Acuity Scale in a tertiary care hospital. CJEM. 2009;11: Gómez Jiménez J, Murray MJ, Beveridge R, Pons J, Albert E, Ferrando JB, et al. Implementation of the Canadian Triage and Acuity Scale (CTAS) in the Principality of Andorra: Can triage parameters be used as Emergency Department Quality Indicators? CJEM. 2003;5: Gouin S, Gravel J, Amre D, Bergeron S. Evaluation of the Paediatric Canadian Triage and Acuity Scale in a paediatric ED. Am J Emerg Med. 2005;23: Warren D, Jarvis A, LeBlanc L, Gravel J. CTAS National Working Group (NWG). Revisions to the Canadian triage and acuity scale paediatric guidelines (PAEDCTAS). CJEM 2008;10: Emergencias 2010; 22:

6 A. Fernández et al. Evaluación de la escala canadiense de triaje pediátrico en un servicio de urgencias de pediatría europeo Fernández A, Pijoan JI, Ares MI, Mintegi S, Benito FJ Objetivo: Evaluar el funcionamiento y fiabilidad de una versión informatizada de la Canadian Paediatric Triage and Acuity Scale (PaedCTAS) en un servicio de urgencias pediátrico (SUP) europeo. Método: Revisión retrospectiva de los episodios de pacientes valorados en el 2007, recogiendo datos sobre el nivel de gravedad, la estancia media, la tasa de hospitalización y pruebas complementarias. Evaluamos la fiabilidad que comparaba la clasificación de casos teóricos (concordancia entre diferentes evaluadores y exactitud con respecto a una valoración de referencia). Resultados: Se analizaron episodios cuya distribución por niveles fue la siguiente: nivel I 0,1%; II 1,7%; III 43,1%; IV 50% y V 5,1%. El porcentaje de pruebas complementarias realizadas varió desde un 63,6% en el nivel I a un 8% en el nivel V (p < 0,01). El tiempo de estancia media (TEM) aumentó progresivamente desde 72 minutos para el nivel V a 373 min. para el I. Las tasas de hospitalización fueron también diferentes para cada nivel de gravedad (desde 84% en el nivel I a 1,6% en el nivel V; p < 0,001). En el estudio de fiabilidad, la valoración del triángulo de evaluación pediátrica fue exacta en el 93,6% de los casos y el índice Kappa fue de 0,77 (IC 95% 0,75-0,79). Se asignó correctamente el nivel de gravedad (o incluyó un desacuerdo irrelevante en el 80,3% de los casos), y el índice Kappa fue 0,47 (IC 95% 0,46-0,48). Conclusiones: La PaedCTAS parece aplicable en un SUP europeo aunque son mejorables determinados aspectos de la escala y de su aplicación en nuestro servicio. [Emergencias 2010;22: ] Palabras clave: Triaje. PaedCTAS. Fiabilidad. 360 Emergencias 2010; 22:

Emergency care workload units: A novel tool to compare emergency department activity

Bond University epublications@bond Faculty of Health Sciences & Medicine Publications Faculty of Health Sciences & Medicine 10-1-2010 Emergency care workload units: A novel tool to compare emergency department