The effectiveness of the implementation of the Cape Triage Score at the emergency department of the National District Hospital, Bloemfontein

Abstract

Background: The need for an appropriate South African triage system led to the development of the Cape Triage Score (CTS), a system to prioritise emergency departments’ workloads. This study evaluated the effectiveness of the CTS at the National District Hospital emergency department, Bloemfontein.

Methods: In this retrospective, descriptive, observational study, files of adult patients triaged at the emergency department during February 2013 were randomly selected. Inclusion was subject to the availability of the files. Physiological parameter results were used to calculate the Triage Early Warning Score (TEWS). The side-room investigations and South African Triage Score (SATS) were recorded. Other information gathered included diagnosis, outcome, and times the patient was triaged and seen by the doctor.

Results: A total of 396 files were included in the study. Patients, of whom 57.8% were women, were between the ages of 16–89 years. More than half (52%) of side-room investigations were omitted or inappropriate. The adjustment of the TEWS to the SATS was done incorrectly in 52% of cases. The majority of patients (69.9%) were discharged home after treatment, although 88% were SATS orange coded. Over-triage occurred in 8.1% of TEWS and 67.8% of SATS cases. The mean waiting time from triage until patient was seen by the doctor was 2 hours.

Conclusions: The TEWS proved to predict outcome better than the SATS. Due to inaccurate triage, only 8% of patients were seen within the recommended waiting times. The CTS has not been effectively implemented at the National District Hospital emergency department.

Keywords:

• Cape Triage Score
• effectiveness
• implementation
• National District Hospital
• triage

Introduction

The need for a triage system, appropriate to the South African setting, led to the development of the Cape Triage Score (CTS), synonymous to the South African Triage Score (SATS). Before the implementation of the CTS, the triage system used varied from area to area and there was no national triage system that was practised consistently, displaying the obvious need thereof.¹ The CTS, thus, culminated from the dire need for an accurate triage system that could be used with speed to prioritise the large emergency department workload of both trauma and medical patients. International triage tools include the Manchester Triage, Canadian Triage Assessment Scale (CTAS), and the Australian Triage Score (ATS). However, the implementation and adoption of these triage tools in South Africa are difficult as considerable training is needed.²

According to Rosedale et al.,³ South African emergency departments’ workload is one of the highest worldwide due to poorly resourced, overcrowded, under-staffed, and under-funded hospitals. Emergency departments have to cope with the so-called ‘quadruple burden of disease’: communicable diseases, non-communicable chronic diseases, injuries, and HIV/AIDS. Patients with HIV and tuberculosis (TB) usually score higher in the SATS, hence these chronically ill patients are prioritised over acute or injured patients.³

The Emergency Medicine Society of South Africa adopted the CTS as the standard triage tool in South Afica.⁴ The CTS is a valid and reliable triage tool used by emergency department doctors and nurses since 2006.⁵ The CTS comprises a physiologically based scoring system, the Triage Early Warning Score (TEWS), and a list of discriminators such as mechanism of injury, hyperglycaemia or abdominal pain, bleeding history, and so forth.⁶

The TEWS consists of physiological parameters where a patient is scored on seven parameters and a score of between 0–3 is allocated to each parameter. According to the TEWS a patient is then allocated a colour code: red, orange, yellow, green, or blue (see Table 1). Under each colour code there are different discriminators that will allow the person doing the triage to adjust the colour code (then referred to as the SATS colour code) according to additional information. This additional information includes the presenting complaint as well as triage aids (side-room investigations). Indications for side-room investigations are, for example, pulse oximetry measures for respiratory rate scores of one or more, finger prick glucotests for patients with a history of diabetes mellitus, altered consciousness, or active seizures, etc.⁶

Table 1: Description of the colour codes used in the Cape Triage Score²

Colour code	Description
Red	Resuscitation/physiologically unstable patient
Orange	Serious cases with potentially unstable physiology or potential life/limb threatening pathology
Yellow	‘Physiologically stable’ cases with reasonably serious medical or trauma problems
Green	Minor injuries/illness
Blue	Clearly dead patients

The emergency department of the National District Hospital (NDH) was established and formally opened on 30 August 1999, with services rendered at Primary Level 1 care to Bloemfontein and surrounding areas. The emergency department was faced with a myriad of challenges that included infrastructure, referral systems, and development of triage protocols. In an attempt to overcome these challenges, many changes were implemented including the introduction of the CTS. However, capacity has remained a growing burden, placing much strain on the triage system.

Aim

The aim of this study was to evaluate the effectiveness of the Cape Triage Score at the emergency department of the National District Hospital in Bloemfontein.

Objectives

• To determine the profile of adult patients seen at the emergency department of the NDH;

• To determine how accurately the CTS system has been implemented at the NDH. The following criteria were used to assess this objective:

  • To evaluate the accuracy of the initial assessment,

  • To evaluate the relationship between assessment category and outcomes of patients (be it discharge, admission, referral, or death), and

  • To evaluate whether waiting times are in accordance with the CTS guidelines; and

• To make recommendations on how to improve or adapt the CTS system so that it can be more relevant to the situation at the NDH.

Methods

This was a retrospective, descriptive, observational study done by reviewing the files of adult patients triaged at the emergency department of the NDH, Bloemfontein, during February 2013.

Information was retrieved from emergency department registers and patient files were randomly selected over the 1-month study period. According to the NDH emergency department’s local statistics of 2010–2012, the approximate average number of patients seen per month is 2000.

The name of every fifth patient recorded on the emergency department register was selected through systematic random sampling, as a representative sample. A list was compiled with the patient names and registration numbers. These files were then sourced from the hospital records department. The first file was randomly selected from the first five patients seen for the month and, thereafter, every fifth file that met the inclusion criteria was chosen. Inclusion was subject to the availability of the file and, if the file was untraceable, no additional file was included.

The only exclusion criteria for this study were:

• Children younger than 16 years, as there is a different triage tool for children.

• Unavailability of the selected patient file.

• Insufficient clinical information in the patient file.

Measurement

Information from the patient files was recorded on data forms. Physiological parameter results were used to calculate the TEWS, as well as confirm the accuracy of the TEWS. The side-room investigations and the appropriateness thereof in the light of the presenting complaint were recorded. The SATS colour code, as well as the accuracy thereof, was also recorded.

Other information gathered included basic patient information, diagnosis and outcome of the patient, number of medical personnel on duty, day of the week, and specific shift that the patient entered the emergency department, as well as the time the patient was triaged, seen by the doctor, and left the emergency department.

Pilot study

A pilot study was done on five files from patients that visited the emergency department in the month preceding the selected study period. The order of some of the items on the data form was changed to follow the flow of the emergency department notes and to make data collection easier. Some items, like the date, were removed and more information regarding different waiting times was added. Pilot study data were not included in the analysis.

Statistical analysis

Statistical analysis was done by the Department of Biostatistics, Faculty of Health Science at the University of the Free State using theSAS/STAT software, Version 9.3 of the SAS System for Windows (©2010 SAS Institute, Inc.). Data are summarised by frequencies, tables, graphs, percentages, and means.

Ethical approval

The Research and Ethics Committee of the Faculty of Health Sciences at the University of the Free State and the Member of the Executive Council of the Free State Health Department granted approval for the study.

No patient names were recorded on the data forms, and all patient and medical personnel information was treated confidentially.

Results and discussion

During February 2013, a total of 2 153 patients were seen at the emergency department of the NDH. A total of 431 files were systematically selected for the study. Due to logistical reasons, 33 files were untraceable and two files did not contain enough clinical data. A total of 396 files were, therefore, included in the study. This is a response rate of 93% and a sampling of 18.4% of all patients.

Profile of adult patients seen at the emergency department of the NDH

The first objective of the study was to determine the profile of adult patients seen at the emergency department of the NDH (see Table 2). This objective was necessary to assess whether the triage system in use is appropriate for the profile of adult patients that is seen at the emergency department, as well as to compare the results with that of other studies.

Table 2: Baseline characteristics of patients included in the study

Characteristics	Frequency
	n	%
Gender
Men	167	42.3
Women	225	57.7
Age
< 20 years	15	3.8
20–40 years	200	50.6
41–65 years	129	32.7
> 65 years	51	12.9
Residential area
Bloemfontein	322	81.5
Outside Bloemfontein	73	18.5

The ages of the patients included in the study varied between 16–89 years, with a median age of 37 years, and a mean age of 41.9 years. This finding can be explained by the number of patients with trauma and HIV in this age group and is consistent with the literature⁷ and previous hospital statistics.

Women contributed to 57.7% of the patients. This finding is substantiated by Dr Lindi le Roux⁸ in her dissertation done at the NDH emergency department and comparable to a study at St Rita’s Hospital emergency department in Limpopo.⁹ However, in other studies done at district hospitals in South Africa,^3,10 as well as in the UK,⁷ more men were seen at the emergency departments. This difference can be attributed to a dedicated trauma facility at a nearby secondary hospital, which manages most of the trauma cases of the area.

Emergency Medical Services were used by 39.2% (n = 155) of patients for transport, rather than emergencies, to the NDH emergency department, of which only 13.4% (n = 53) were referred to the hospital as emergencies. Most areas do not have 24-hour primary care facilities; therefore, patients flock to the emergency department at the NDH as the first point of entrance to medical services.

The diagnoses of patients were grouped together in the following categories: trauma, HIV-related, respiratory conditions, abdominal conditions, neurological conditions, chronic conditions, psychiatric conditions, other emergencies, and other. Only patients known to be HIV positive and with HIV-related conditions were grouped under HIV-related. Patients in the ‘other’ category had very diverse diagnoses, and the number of patients per diagnosis was too small to be seen as a specific category. The distribution of patients per diagnosis and age group is displayed in Table 3.

Table 3: Distribution of patients per diagnosis and age group

Diagnosis	Age				Percentage of patients (%)
	< 20 years (n = 15)	20–40 years (n = 200)	41–65 years (n = 129)	> 65 years (n = 51)
Trauma	4	38	14	5	15.4
HIV-related	0	49	19	4	18.2
Respiratory conditions	2	17	20	3	10.6
Abdominal conditions	1	18	19	11	12.4
Neurological conditions	0	6	4	2	3.0
Chronic conditions	0	12	19	15	11.6
Psychiatric conditions	1	20	7	0	7.1
Other emergencies	3	12	6	2	5.8
Other	4	28	21	9	15.7

The category ‘HIV-related’ complications accounted for almost 20% of the emergency department visits, with category ‘trauma’ second, at 15.4%. HIV-positive patients presented mainly with acute or chronic conditions and could be managed at a lower level of care, with little immediate intervention.

The profile of adult patients seen at the NDH emergency department is unique and, therefore, the triage protocol needs to be adjusted to this profile.

Results of the initial assessment

On arrival at the emergency department, patients are questioned about their main complaints. Basic physiological parameters (side-room observations) are assessed and the TEWS calculated. In this study basic parameters were measured in 98% of cases.

The TEWS is then adjusted according to a list of discriminators, including specific complaints and relevant side-room investigations. This is then called the SATS colour codes. Figure 1 represents the change from TEWS to SATS once the list of discriminators was taken into account. After the adjustment of the TEWS, 88% of patients were assigned to the SATS orange code.

Figure 1: Percentage of patients assigned to a colour code category before and after adjustment. [SATS: South African Triage Score, TEWS: Triage Early Warning Score]

In 13% of cases the TEWS were not adjusted after other discriminators were added to the assessment. The adjustment according to the CTS guidelines of the TEWS to the SATS colour code was done incorrectly in 52% of cases.

Side-room investigations are part of the initial steps and are indicated for specific signs and symptoms after the TEWS assessment. In Table 4 the results of the use of side-room investigations are displayed. Results are displayed as appropriate and inappropriate according to CTS guidelines, where inappropriate use of side-room investigations is defined as tests either indicated and not done, or not indicated but done.

Table 4: Appropriateness of the side-room investigations as indicated per Cape Triage Score guidelines

Test	Appropriateness		Proportion inappropriate (%)
	Appropriate (n)	Inappropriate (n)	Lower 95% CI	Proportion	Upper 95% CI
Random glucose test	154	240	56.0	60.9	65.6
Haemoglobin test	154	240	56.0	60.9	65.6
Oxygen saturation	192	202	46.3	51.3	56.2
Urine dipstick	208	186	47.9	52.8	57.7

Table 4 shows the proportion of side-room investigations inappropriately performed. The 95% confidence intervals show that, for both the Random glucose test and the Haemoglobin test, more than half of the tests were inappropriately performed, even given the possibility of sampling error. Thus, many of the side-room investigations were inappropriate, meaning the incorrect tests were done or the indicated tests were omitted. Despite this omission, the TEWS was still adjusted.

The 88% of patients assigned to the SATS orange code is much higher than in other published studies done in South African hospitals, namely 14.5% in the Paarl,¹⁰ 6% in Limpopo,⁹ and 21% in GF Jooste Hospital in the Western Cape.¹¹ From the results two possible factors could contribute to these out-of-range results. Firstly, side-room investigations were indicated per CTS guidelines in many of the cases and necessary to adjust the TEWS. However, in Table 4 it can be seen that appropriate side-room investigations were done in less than 50% of cases and, therefore, it was impossible to correctly adjust the TEWS. Secondly, the TEWS should be adjusted to SATS according to specific guidelines. This was done incorrectly in 52% of cases.

Relationship between outcomes and assessment category

The majority of patients (69.9%) were discharged home after treatment, while 20.7% were admitted to the NDH and 9.3% were referred to specialists. Of the 321 orange coded patients, according to SATS classification, the majority (71%) were discharged home after treatment.

In this and other studies it was expected that a green coded patient would be discharged from the emergency department, a yellow coded patient might either be discharged or admitted to a district hospital. An orange or red coded patient might either die or be admitted to a specialist hospital.^1,3,5,11,12 From Figure 2 it can be seen that TEWS alone was a better predictor of outcome than SATS when the same outcome criteria were expected for a specific colour code. Sun et al.¹² investigated the use of TEWS alone to triage patients in resource-limited countries and found it to be acceptable.

Figure 2: A comparison of outcomes between the TEWS and SATS colour coded patients. [SATS: South African Triage Score, TEWS: Triage Early Warning Score]

The term ‘under-triage’ is used when the outcome of a patient is under-estimated and ‘over-triage’ is used when the outcome is over-estimated for that colour code category.^3,5 In Table 5, purple indicates patient outcome was as expected, brown indicates under-triage and blue indicates over-triage. Under-triage occurred in 13.2% (52/394) of cases for TEWS and 0.3% (1/366) for SATS. Over-triage occurred in 8.1% (32/395) of TEWS and 67.9% (248/365) of SATS cases.

Table 5: Number of patients per colour code category compared to the actual outcome

As the SATS colour codes were incorrectly adjusted in more than half of the patients, the researcher expected the outcomes compared to the SATS colour codes to also be inaccurate. This can clearly be seen in Figure 2, where 30% of orange coded patients and only 37% of red coded patients were admitted either to the NDH or referred to secondary or tertiary hospitals. As only two patients stayed in the green code after adjustment, no conclusion can be made from this finding.

The use of TEWS alone may under-triage patients when compared to the expected outcome.^1−3,11 Twomey et al.⁵ concluded that a range of 10% under-triage and 15% over-triage was acceptable. From the results and as displayed in Table 5, over-triage was a considerable problem when using SATS colour codes. The first reason for this finding is the incorrect SATS colour codes allocated to patients as discussed above. Another reason could be the profile of adult patients seen at the emergency department of the NDH. A study by Rosedale et al.³ conducted in KwaZulu-Natal found that patients with TB contributed to 24% of SATS over-triaged patients. In this study, chronic diseases like HIV and TB contributed to 18.3% of patients seen at the emergency department at the NDH.

Waiting times

The mean time from triage until a patient file was opened was 29 minutes, with a maximum of 180 minutes. From triage until the patient was seen by the doctor, the mean waiting time was 121 minutes with a maximum of 635 minutes.

According to the CTS guidelines the target time to see a red coded patient should be immediately, an orange coded patient within 10 minutes, a yellow coded patient within 60 minutes, and a green coded patient within 240 minutes.⁶ As only 8% of patients were seen within the timeframe indicated in the guidelines, possible reasons were sought. Firstly, the inaccurate SATS colour codes made it impossible to stay within the timeframes, as 98% of patients would need to have been seen within 10 minutes from entering the emergency department. Secondly, the overcrowding of the emergency department with patients already seen, but that had to wait up to 12 hours to be referred to the secondary and tertiary hospitals, and 48 hours to be admitted to the NDH. Thirdly, the study results showed that the time from triage until a file was opened ranged from 29 minutes up to a maximum of 180 minutes, and this alone could significantly impact on waiting time. Finally, ambulances are currently used to transport non-acute patients from home to the emergency department due to the lack of 24-hour primary care facilities and are, therefore, not available to transport acute patients to referral hospitals.

No statistical significant difference could be found between the waiting time during day or night shifts (p = 0.43), or number of doctors and nurses on duty (p = 0.66). There was also no difference between the waiting times during different days of the week.

The number of patients inside the emergency department seems to result in overcrowding, but this was an impression and could not be measured. However, the time it took from when the decision was made to admit a patient to the NDH until the patient left the emergency department ranged from 1–48 hours, with a mean of 8 hours and a median of 6 hours. The time it took from when the referral was arranged with the specialist until the patient left the emergency department ranged from less than 1 hour to 12 hours, with a mean of 6 hours. In the meantime patients stayed in the emergency department, occupying examination space and beds.

Study limitations

• Data needed for the study were dependent on accurate record keeping. Note keeping was not up to standard in some of the files. Therefore, interpretation of some clinical notes was difficult and was left to the discretion of the researcher, which could have led to measurement errors.

• Only clinical data written in the files could be used and in several cases some information was missing, which led to incomplete data collection.

• The availability of patient files and records posed a challenge, but eventually most were found after searching the wards, pharmacy, records, and accounts departments. All files selected through the systematic sampling process not available or untraceable were excluded from the study.

Conclusion

The NDH emergency department presented with a unique adult patient profile with a majority of female patients, less trauma cases and an overload of HIV-related conditions compared to other emergency departments.

Initial assessment: Of the 395 patient files reviewed it was found that the SATS colour coding was incorrectly adjusted in 52% of the patients and 88% of patients were assigned to the orange code. In addition, side-room investigations were not done when required in more the 50% of cases and, therefore, the appropriate initial steps were not followed. It is clear that the CTS system has not been effectively implemented at the NDH emergency department.

Outcomes: Over-triage occurred in 67.9% of SATS cases. Therefore, the SATS colour coding was not accurate to predict the outcomes of patients seen at the NDH emergency department. The TEWS proved to be a better indication of outcome, with an over-triage rate of 8.1%.

Timeframe: Due to the inaccurate triage, only 8% of patients were seen within the recommended waiting times according to the CTS guidelines. On average patients waited 2 hours to be seen by the doctors, even though 88% of the patients were orange coded, which has a recommended 10-minute waiting time.

There were no statistically significant differences between waiting times of different shifts or number of medical personnel on duty.

Recommendations

To improve and adapt the CTS system so that it can be more relevant to the situation at the NDH emergency department the following recommendations are made:

• Chronically ill patients (especially those with HIV-related complaints) can receive minor interventions at initial triage assessment (intravenous fluids, O₂ mask) and can be down-triaged to yellow code. This will dramatically decrease the over-triage to orange code.

• Patients who refuse to or are unable to visit primary healthcare facilities (usually due to no doctors, no treatment, rude staff, line cuts, working clients with after-hour needs, and only daytime health facilities), should be re-directed back to appropriate level of care to alleviate over-crowding in the emergency department. More 24-hour primary care facilities should be available in referring areas.

• Nursing staff need adequate continuous training in triage and the implementation of the CTS guidelines. This should be a continuous process that is frequently reinforced.

• There should be a dedicated emergency department admissions office in order to open files immediately and prevent long queuing with the main stream of patients. Waiting for a file should not hamper emergency patients from receiving immediate care.

• Doctors need to be involved in the initial triage assessment as well as the down-triaging of chronic patients and re-direction of green coded patients to clinics.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgement

Mrs T. Mulder, medical writer, School of Medicine, University of the Free State, for technical and editorial preparation of the manuscript.