- Email: [email protected]
A retrospective chart review of adult mortality characteristics of patients presenting to a principal tertiary emergency department
Accident and Emergency Nursing (2005) 13, 122–125
Accident and Emergency Nursing www.elsevierhealth.com/journals/aaen
A retrospective chart review of adult mortality characteristics of patients presenting to a principal tertiary emergency department Margaret Fry RN, CITN, BASc, M.Ed (Clinical Nurse Consultant) *, Ann Rhodes-Sutton RN, ENB998, DPNS, BSc (Hons) Emergency Department, St George Hospital, Kogarah, Sydney, NSW 2217, Australia Received 3 March 2004; accepted 20 January 2005
KEYWORDS
Summary Aim: To determine the incidence of death after arrival to one metropolitan emergency department (ED) for each triage code, day of arrival, diagnosis and length of stay. Method: A 12-month retrospective chart review was conducted in a 550 bed principal referral hospital of all patients that presented to the ED and subsequently died. The ED annual attendance rate for the study year was 46,017 patients. The data collected included patient demographics, diagnosis, triage code, length of stay and ward disposition. The data are expressed as proportions, percentages, frequencies or the statistic chi-square (v2). The proportion of ED deaths also examined included those patients admitted through the ED to a ward area. Results: The ED mortality rate for the study period was 1.3% (n = 595). One hundred and twenty-three patients (21%) died in the unit, while 472 (79%) were admitted through the ED to a ward area. The total annual hospital mortality rate was 1.7% (n = 790). The mean age of patients that died was 76 years. Of ED deaths, 329 (55%) were male and 266 (45%) female. Triage code 1 (60%) was allocated more commonly to patients that died in the ED compared with triage code 3 (46%) for ED-Ward deaths. Conclusions: Examination of ED mortality rate provides a deeper understanding of service utilisation and nursing staff work performance. Triage codes are beneficial in evaluating ED services, casemix, policies and quality assurance activities, but the application of this indicator to in-patient services appears limited. c 2005 Elsevier Ltd. All rights reserved.
Emergency deaths; Triage code; Mortality rate
*
Corresponding author. Tel: +611 02 9350 1650x817. E-mail address: [email protected] (M. Fry).
0965-2302/$ - see front matter c 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.aaen.2005.01.002
Adult mortality characteristics of patients
Introduction In Australia, little is known of the characteristics of emergency department (ED) patient deaths or of those patients (ED-Ward) that died following admission through the ED. Auditing of ED deaths, while recommended as a clinical indicator, largely remains focused on those patients that died within the ED. As a result the ward patient that dies following admission through the ED remains concealed within the hospital mortality rate and often not reviewed by ED services. While the ED mortality rate proportional to total patient presentations is generally very low, it is important to review all ED deaths not just the group that dies within the unit. To date ED mortality profiles are under reported in the literature. In 1994, to assist the triage nurse in the early recognition of patient urgency, the Australasian triage scale (ATS) was introduced (ACEM, 2002). Early recognition of patient acuity has been shown to improve patient outcome and reduce mortality rate. Triage codes are the main method to identify patient clinical urgency. However, association of triage codes with patient co-morbidity, complexity and illness severity is weak. While triage codes provide a reliable indication of patient case-mix, service utilisation and efficiency, the relevance of triage codes with ED-Ward mortality rate remains unclear.
Objective The objectives of this study were: (i) to identify the incidence and characteristics of ED patient deaths; and (ii) to identify and compare day of arrival, diagnosis and length of stay for ED and ED-Ward death groups.
Methods A 12-month retrospective chart review of all patients that died following presentation to one Sydney metropolitan ED was conducted from July 2001 to June 2002. The study was conducted in a 550 bed principal referral hospital with an ED attendance rate of 46,017 for the same period. The data collected included patient demographics, diagnosis, triage code, length of stay and ward disposition. Cause of death was identified using the patient discharge summary. Causes of death were grouped into similar categories used by other researchers
123 (Dent and Rofe, 1999; Doherty et al., 2003). Inpatient hospital data was extracted using (TrendstarTM). ED data were obtained from the emergency department information system (EDISTM). Data were analysed using the statistical package for social sciences (SPSS 10.0 1999). The statistical data are expressed as proportions, percentages or frequencies where appropriate. The statistic chi-square (v2) was used to examine for association between triage code and day of presentation.
Results During the study period 46,017 patients presented to the ED. Of these 17,139 (37%) were admitted. Of ED presentations, 123 died in the unit while 472 died following admission. Of the 595 ED patients that died, 266 (45%) were female and 329 (55%) male. The mean age was 76 years. Total ED mortality rate for the study year was 1.3% (n = 595). The hospital mortality rate, for the same period, was 1.7% (n = 790). Of the hospital deaths, 60% (n = 472) had been admitted through the ED. A profile of ED presentations and mortality rate per triage code category is shown in Table 1. Sixty percent of patients that died in the ED were more commonly allocated triage code 1 compared with triage code 3 (46%) for ED-Ward deaths (see Table 2). Triage code 5 (the lowest urgency code) was not allocated to any ED patient who died. Of the 472 patients transferred from the ED to ward area, 250 (53%) went to general medical units, 73 (15.5%) went to critical care areas, 73 (15.5%) went to the cardiac unit, 69 (14.6%) went to general surgical wards and 7 (1.4%) went to a short stay unit. The most frequent cause of death was acute cardiac event (n = 149; 25%). When comparing ED patient and ED-Ward mortality profiles atraumatic cerebral events, chronic respiratory/cardiac conditions and sepsis contributed most frequently to mortality rate. Table 3 shows the causes of death for each group in the study. The average length of stay for ED deaths was 7 h: 42 min. Average length of stay of triage code 1 was 5 h: 48 min, triage code 2 was 10 h: 20 min, triage code 3 was 11 h: 37 min, and triage code 4 was 11 h: 36 min. For ED-Ward deaths the average length of stay was less than 7 days (n = 264; 56%) (see Table 4). Chi squared (v2) was used to examine proportional differences between total ED presentations and death rate per day of the week and weekends. There was no significant difference between the
124
M. Fry, A. Rhodes-Sutton
Table 1
Comparison of ED presentations with mortality rates by triage code category
Australasian triage scalea
ED presentations
ED deaths
1 2 3 4 5
488 3808 17,848 22,562 1311
133 125 233 104 0
46,017
595
(seen (seen (seen (seen (seen
immediately) in 10 min) in 30 min) in 60 min) in 120 min)
Total a
(27.2%) (3.2%) (1.3%) (0.4%)
The Australasian triage scale is a nationally recognised and authorised triage category tool used in emergency departments.
days of the week (p = 0.9) or when comparing weekdays with weekends (p = 0.8) for ED deaths.
Discussion The 12 month study identified a 1.3% ED mortality rate. Two Australian studies (Dent and Rofe, 1999; Doherty et al., 2003) examined ED-Ward deaths and identified mortality rates of 1.2% and 0.54%, respectively. The higher mortality rate found in our study may be due to the inclusion of both ED and ED-Ward patient groups. The different mortality rate found by our study compared with the previously cited studies may also be the result of a different casemix, trauma load or being a designated metropolitan principal referral hospital. The findings of this study showed that mortality rate statistically was not affected by the day of week. This is in contrast to an American study that found patients with life-threatening conditions were more likely to die in hospital if admitted on a weekend (Bell and Redelmeier, 2001). As an explanation for our findings, emergency physicians work weekends and evening shifts supervising and coordinating the care of ED patients. Also, nursing unit managers are on each shift and oversee the work of all staff. These staff features may go some way to explaining the lack of association between mortality rate and presentation day. Table 2 Comparison of ED with ED-Ward mortality rates by triage code category Australasian triage scale Triage Triage Triage Triage Triage Total
code code code code code
1 2 3 4 5
ED deaths
ED-Ward deaths
74 33 15 1 0
59 92 218 103 0
123
472
Of those that died in the ED, the lowest urgency code (5) was not allocated to any patient. Ninetynine percent of ED deaths were allocated a triage code 1 2 or 3. Within our ED this finding supports that triage nurses are recognising patient urgency and appropriately allocating triage codes. This study identified that 46% of ED-Ward deaths were allocated triage code 3 compared with Dent and Rofe’s (1999) study that found triage code 4 was more frequent. Researchers have identified that the mortality rate for triage code 3 or 4 admitted patients is higher when compared to patients allocated triage code 1 or 2 (Dent and Rofe, 1999; Doherty et al., 2003). While these studies infer that clinical urgency is a poor predictor of mortality rate for admitted patients it suggests that triage code 3 and 4 patients are using EDs appropriately. However, examination of triage code and ED-Ward deaths may only be relevant for specific sub-groups. Therefore, we suggest that the review process for ED deaths include patients that die within 24–48 h of ED presentation. We believe the ongoing investigation of this subset of patients would benefit service evaluation processes for both emergency and in-patient areas. The Australasian triage scale has shown to be sensitive for patient urgency, but the scale does not accommodate for patient complexity or comorbidity, in that, patients allocated a lower triage code can still have a serious or significant medical condition. Given that the average length of stay for ED-Ward deaths was 10 days, interventions such as a medical emergency team may make a significant contribution towards reducing mortality rate and improving patient outcome (Bellomo et al., 2003; Parr et al., 2001). In Australia, triage codes remain an important measure of ED service performance and while triage code benchmarking is relevant during the early stages of a patient’s presentation, the significance of measuring triage codes for ED-Ward deaths beyond 24–48 h may be limited.
Adult mortality characteristics of patients Table 3
125
Causes of ED deaths and ED-Ward deaths
Diagnosis
ED deaths
ED-Ward deaths
Total
Acute cardiac events Chronic cardio/respiratory Atraumatic cerebral event Sepsis Other Multi-trauma Gastrointestinal Cancer/haematology Metabolic condition Acute abdomen
72 (58.5%) 12 (9.7%) 11 (8.9%) 7 (6%) 5 (4%) 5 (4%) 2 (2%) 0 (0.0%) 4 (3%) 5 (4%)
77 (12%) 136 (7.4%) 55 (0.0%) 43 (10%) 41 (0.8%) 36 (7.4%) 37 (7.6%) 38 (12.5%) 7 (4.4%) 2 (3.4%)
149 (25.0%) 147 (24.9%) 66 (11.0%) 50 (8.4%) 46 (7.9%) 40 (6.7%) 39 (6.5%) 38 (6.3%) 11 (1.8%) 9 (1.5%)
Table 4 of stay
A comparison of ED-Ward deaths by length
Length of stay
ED-Ward deaths
<1 day, <7 days, >7 < 14 days, >15 < 30 days >31 days
91 173 111 69 28
Total
472
(19%) (37%) (24%) (14%) (6%)
Limitation The study was conducted at one site; hence the findings may not be applicable to other similar designated principal referral hospitals. We did not explore the potential for avoidable deaths due to resources, but it is an interesting area for further research. Another limitation included the use of discharge summaries given the unavailability of death certificates or post-mortem reports.
Conclusions Our study has made explicit the ED death rate per presentation, day of the week and by triage code. Triage codes remain a valuable performance indicator for emergency departments, although the
application of this indicator to patients who have subsequently been transferred to an inpatient area offers limited value. Examination of triage code and mortality rate provides for a deeper understanding of service utilisation and nursing staff triage work performance.
References ACEM, 2002. The Australasian triage scale. Emergency Medicine 14(3) 335–336. Bell, C., Redelmeier, D., 2001. Mortality among patients admitted to hospitals on weekends as compared with weekdays. The New England Journal of Medicine 345 (9), 663–668. Bellomo, R., Goldsmith, D., Uchino, S., Buckmaster, J., Hart, H., Opdam, H., Silvester, W., Doolan, L., Gutteridge, G., 2003. A prospective before-and-after trial of a medical emergency team. Medical Journal of Australia. 179 (6), 283– 287. Dent, A., Rofe, G., 1999. Which triage category patients die in hospital after being admitted through emergency departments? A study in one teaching hospital. Emergency Medicine 11 (2), 68–71. Doherty, S., Hore, C.T., Curran, S.W., 2003. Inpatient mortality as related to triage category in three New South Wales regional base hospitals. Emergency Medicine 15 (4), 334– 340. Parr, M.J., Hadfield, J.H., Flabouris, A., Bishop, G., Hillman, K., 2001. The Medical Emergency Team: 12 month analysis of reasons for activation, immediate outcome and not-forresuscitation orders. Resuscitation 50 (1), 39–44.
Accident and Emergency Nursing www.elsevierhealth.com/journals/aaen
A retrospective chart review of adult mortality characteristics of patients presenting to a principal tertiary emergency department Margaret Fry RN, CITN, BASc, M.Ed (Clinical Nurse Consultant) *, Ann Rhodes-Sutton RN, ENB998, DPNS, BSc (Hons) Emergency Department, St George Hospital, Kogarah, Sydney, NSW 2217, Australia Received 3 March 2004; accepted 20 January 2005
KEYWORDS
Summary Aim: To determine the incidence of death after arrival to one metropolitan emergency department (ED) for each triage code, day of arrival, diagnosis and length of stay. Method: A 12-month retrospective chart review was conducted in a 550 bed principal referral hospital of all patients that presented to the ED and subsequently died. The ED annual attendance rate for the study year was 46,017 patients. The data collected included patient demographics, diagnosis, triage code, length of stay and ward disposition. The data are expressed as proportions, percentages, frequencies or the statistic chi-square (v2). The proportion of ED deaths also examined included those patients admitted through the ED to a ward area. Results: The ED mortality rate for the study period was 1.3% (n = 595). One hundred and twenty-three patients (21%) died in the unit, while 472 (79%) were admitted through the ED to a ward area. The total annual hospital mortality rate was 1.7% (n = 790). The mean age of patients that died was 76 years. Of ED deaths, 329 (55%) were male and 266 (45%) female. Triage code 1 (60%) was allocated more commonly to patients that died in the ED compared with triage code 3 (46%) for ED-Ward deaths. Conclusions: Examination of ED mortality rate provides a deeper understanding of service utilisation and nursing staff work performance. Triage codes are beneficial in evaluating ED services, casemix, policies and quality assurance activities, but the application of this indicator to in-patient services appears limited. c 2005 Elsevier Ltd. All rights reserved.
Emergency deaths; Triage code; Mortality rate
*
Corresponding author. Tel: +611 02 9350 1650x817. E-mail address: [email protected] (M. Fry).
0965-2302/$ - see front matter c 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.aaen.2005.01.002
Adult mortality characteristics of patients
Introduction In Australia, little is known of the characteristics of emergency department (ED) patient deaths or of those patients (ED-Ward) that died following admission through the ED. Auditing of ED deaths, while recommended as a clinical indicator, largely remains focused on those patients that died within the ED. As a result the ward patient that dies following admission through the ED remains concealed within the hospital mortality rate and often not reviewed by ED services. While the ED mortality rate proportional to total patient presentations is generally very low, it is important to review all ED deaths not just the group that dies within the unit. To date ED mortality profiles are under reported in the literature. In 1994, to assist the triage nurse in the early recognition of patient urgency, the Australasian triage scale (ATS) was introduced (ACEM, 2002). Early recognition of patient acuity has been shown to improve patient outcome and reduce mortality rate. Triage codes are the main method to identify patient clinical urgency. However, association of triage codes with patient co-morbidity, complexity and illness severity is weak. While triage codes provide a reliable indication of patient case-mix, service utilisation and efficiency, the relevance of triage codes with ED-Ward mortality rate remains unclear.
Objective The objectives of this study were: (i) to identify the incidence and characteristics of ED patient deaths; and (ii) to identify and compare day of arrival, diagnosis and length of stay for ED and ED-Ward death groups.
Methods A 12-month retrospective chart review of all patients that died following presentation to one Sydney metropolitan ED was conducted from July 2001 to June 2002. The study was conducted in a 550 bed principal referral hospital with an ED attendance rate of 46,017 for the same period. The data collected included patient demographics, diagnosis, triage code, length of stay and ward disposition. Cause of death was identified using the patient discharge summary. Causes of death were grouped into similar categories used by other researchers
123 (Dent and Rofe, 1999; Doherty et al., 2003). Inpatient hospital data was extracted using (TrendstarTM). ED data were obtained from the emergency department information system (EDISTM). Data were analysed using the statistical package for social sciences (SPSS 10.0 1999). The statistical data are expressed as proportions, percentages or frequencies where appropriate. The statistic chi-square (v2) was used to examine for association between triage code and day of presentation.
Results During the study period 46,017 patients presented to the ED. Of these 17,139 (37%) were admitted. Of ED presentations, 123 died in the unit while 472 died following admission. Of the 595 ED patients that died, 266 (45%) were female and 329 (55%) male. The mean age was 76 years. Total ED mortality rate for the study year was 1.3% (n = 595). The hospital mortality rate, for the same period, was 1.7% (n = 790). Of the hospital deaths, 60% (n = 472) had been admitted through the ED. A profile of ED presentations and mortality rate per triage code category is shown in Table 1. Sixty percent of patients that died in the ED were more commonly allocated triage code 1 compared with triage code 3 (46%) for ED-Ward deaths (see Table 2). Triage code 5 (the lowest urgency code) was not allocated to any ED patient who died. Of the 472 patients transferred from the ED to ward area, 250 (53%) went to general medical units, 73 (15.5%) went to critical care areas, 73 (15.5%) went to the cardiac unit, 69 (14.6%) went to general surgical wards and 7 (1.4%) went to a short stay unit. The most frequent cause of death was acute cardiac event (n = 149; 25%). When comparing ED patient and ED-Ward mortality profiles atraumatic cerebral events, chronic respiratory/cardiac conditions and sepsis contributed most frequently to mortality rate. Table 3 shows the causes of death for each group in the study. The average length of stay for ED deaths was 7 h: 42 min. Average length of stay of triage code 1 was 5 h: 48 min, triage code 2 was 10 h: 20 min, triage code 3 was 11 h: 37 min, and triage code 4 was 11 h: 36 min. For ED-Ward deaths the average length of stay was less than 7 days (n = 264; 56%) (see Table 4). Chi squared (v2) was used to examine proportional differences between total ED presentations and death rate per day of the week and weekends. There was no significant difference between the
124
M. Fry, A. Rhodes-Sutton
Table 1
Comparison of ED presentations with mortality rates by triage code category
Australasian triage scalea
ED presentations
ED deaths
1 2 3 4 5
488 3808 17,848 22,562 1311
133 125 233 104 0
46,017
595
(seen (seen (seen (seen (seen
immediately) in 10 min) in 30 min) in 60 min) in 120 min)
Total a
(27.2%) (3.2%) (1.3%) (0.4%)
The Australasian triage scale is a nationally recognised and authorised triage category tool used in emergency departments.
days of the week (p = 0.9) or when comparing weekdays with weekends (p = 0.8) for ED deaths.
Discussion The 12 month study identified a 1.3% ED mortality rate. Two Australian studies (Dent and Rofe, 1999; Doherty et al., 2003) examined ED-Ward deaths and identified mortality rates of 1.2% and 0.54%, respectively. The higher mortality rate found in our study may be due to the inclusion of both ED and ED-Ward patient groups. The different mortality rate found by our study compared with the previously cited studies may also be the result of a different casemix, trauma load or being a designated metropolitan principal referral hospital. The findings of this study showed that mortality rate statistically was not affected by the day of week. This is in contrast to an American study that found patients with life-threatening conditions were more likely to die in hospital if admitted on a weekend (Bell and Redelmeier, 2001). As an explanation for our findings, emergency physicians work weekends and evening shifts supervising and coordinating the care of ED patients. Also, nursing unit managers are on each shift and oversee the work of all staff. These staff features may go some way to explaining the lack of association between mortality rate and presentation day. Table 2 Comparison of ED with ED-Ward mortality rates by triage code category Australasian triage scale Triage Triage Triage Triage Triage Total
code code code code code
1 2 3 4 5
ED deaths
ED-Ward deaths
74 33 15 1 0
59 92 218 103 0
123
472
Of those that died in the ED, the lowest urgency code (5) was not allocated to any patient. Ninetynine percent of ED deaths were allocated a triage code 1 2 or 3. Within our ED this finding supports that triage nurses are recognising patient urgency and appropriately allocating triage codes. This study identified that 46% of ED-Ward deaths were allocated triage code 3 compared with Dent and Rofe’s (1999) study that found triage code 4 was more frequent. Researchers have identified that the mortality rate for triage code 3 or 4 admitted patients is higher when compared to patients allocated triage code 1 or 2 (Dent and Rofe, 1999; Doherty et al., 2003). While these studies infer that clinical urgency is a poor predictor of mortality rate for admitted patients it suggests that triage code 3 and 4 patients are using EDs appropriately. However, examination of triage code and ED-Ward deaths may only be relevant for specific sub-groups. Therefore, we suggest that the review process for ED deaths include patients that die within 24–48 h of ED presentation. We believe the ongoing investigation of this subset of patients would benefit service evaluation processes for both emergency and in-patient areas. The Australasian triage scale has shown to be sensitive for patient urgency, but the scale does not accommodate for patient complexity or comorbidity, in that, patients allocated a lower triage code can still have a serious or significant medical condition. Given that the average length of stay for ED-Ward deaths was 10 days, interventions such as a medical emergency team may make a significant contribution towards reducing mortality rate and improving patient outcome (Bellomo et al., 2003; Parr et al., 2001). In Australia, triage codes remain an important measure of ED service performance and while triage code benchmarking is relevant during the early stages of a patient’s presentation, the significance of measuring triage codes for ED-Ward deaths beyond 24–48 h may be limited.
Adult mortality characteristics of patients Table 3
125
Causes of ED deaths and ED-Ward deaths
Diagnosis
ED deaths
ED-Ward deaths
Total
Acute cardiac events Chronic cardio/respiratory Atraumatic cerebral event Sepsis Other Multi-trauma Gastrointestinal Cancer/haematology Metabolic condition Acute abdomen
72 (58.5%) 12 (9.7%) 11 (8.9%) 7 (6%) 5 (4%) 5 (4%) 2 (2%) 0 (0.0%) 4 (3%) 5 (4%)
77 (12%) 136 (7.4%) 55 (0.0%) 43 (10%) 41 (0.8%) 36 (7.4%) 37 (7.6%) 38 (12.5%) 7 (4.4%) 2 (3.4%)
149 (25.0%) 147 (24.9%) 66 (11.0%) 50 (8.4%) 46 (7.9%) 40 (6.7%) 39 (6.5%) 38 (6.3%) 11 (1.8%) 9 (1.5%)
Table 4 of stay
A comparison of ED-Ward deaths by length
Length of stay
ED-Ward deaths
<1 day, <7 days, >7 < 14 days, >15 < 30 days >31 days
91 173 111 69 28
Total
472
(19%) (37%) (24%) (14%) (6%)
Limitation The study was conducted at one site; hence the findings may not be applicable to other similar designated principal referral hospitals. We did not explore the potential for avoidable deaths due to resources, but it is an interesting area for further research. Another limitation included the use of discharge summaries given the unavailability of death certificates or post-mortem reports.
Conclusions Our study has made explicit the ED death rate per presentation, day of the week and by triage code. Triage codes remain a valuable performance indicator for emergency departments, although the
application of this indicator to patients who have subsequently been transferred to an inpatient area offers limited value. Examination of triage code and mortality rate provides for a deeper understanding of service utilisation and nursing staff triage work performance.
References ACEM, 2002. The Australasian triage scale. Emergency Medicine 14(3) 335–336. Bell, C., Redelmeier, D., 2001. Mortality among patients admitted to hospitals on weekends as compared with weekdays. The New England Journal of Medicine 345 (9), 663–668. Bellomo, R., Goldsmith, D., Uchino, S., Buckmaster, J., Hart, H., Opdam, H., Silvester, W., Doolan, L., Gutteridge, G., 2003. A prospective before-and-after trial of a medical emergency team. Medical Journal of Australia. 179 (6), 283– 287. Dent, A., Rofe, G., 1999. Which triage category patients die in hospital after being admitted through emergency departments? A study in one teaching hospital. Emergency Medicine 11 (2), 68–71. Doherty, S., Hore, C.T., Curran, S.W., 2003. Inpatient mortality as related to triage category in three New South Wales regional base hospitals. Emergency Medicine 15 (4), 334– 340. Parr, M.J., Hadfield, J.H., Flabouris, A., Bishop, G., Hillman, K., 2001. The Medical Emergency Team: 12 month analysis of reasons for activation, immediate outcome and not-forresuscitation orders. Resuscitation 50 (1), 39–44.