Auteur: Meggie Lisette Nieuwenhuizen 6th year CRU+ medical student UMC Utrecht/UU
High urgency allocation is essential for improving ACS patients’ prognosis, but early recognition remains challenging. Presenting with many symptoms in a short time could result in a higher risk of (mis)interpretation and wrong urgency allocation.
To determine the sex-specific relationship between the number of symptoms (many versus few) and (i) urgency allocation, and (ii) ACS in patients with suggestive ACS symptoms calling the out-of-hours services in primary care.
Tape recordings of suggestive ACS patients were analysed. Follow-up data on diagnosis was collected from the patient’s general practitioner. Relationships were investigated using univariable and multivariable analyses with adjustment for sex and age.
We evaluated 3,336 patients, mean age 58.3±19.8 years, 45.1% men, and median number of three symptoms; 1,041 (31.2%) had ≤two (‘few’), and 2,295 (68.8%) had ≥three (‘many’) symptoms. 2,060 (61.8%) had follow-up data. ‘Many symptoms’ patients received a higher urgency allocation (crude OR 1.64 (95% CI 1.41-1.92), adjusted OR 2.08 (95% CI 1.76-2.45), both p<0.001). ‘Many symptoms’ had no clear relationship with ACS (crude OR 1.07 (95% CI 0.79-1.44), p=0.674, adjusted OR 1.36 (95% CI 1.00-1.86), p=0.051), except in women after adjustment for age (adjusted OR 2.35 (95% CI 1.35-4.08), p=0.002). Women had more often ‘many symptoms’ than men (72.6% vs. 64.1%, p<0.001).
Patients presenting with three or more symptoms received a high urgency allocation more often than those with less symptoms, but there was no clear relationship with ACS, except in women after adjustment for age.
Coronary heart disease (CHD) is an important global health problem, notably in developed countries such as the Netherlands [1-4]. In the Netherlands, 8,434 persons died of CHD in 2018; 5,122 because of myocardial infarction (MI) . Although CHD is still one of the major causes of death, there is a visible downward trend due to lower incidence. This is because of a reduction in (second hand) smoking and increased survival by improved treatment . A crucial step towards adequate treatment is early recognition and diagnosis [1-4]. Timing is important, especially in the acute setting; to detect those with acute coronary syndrome (ACS) among those with chest discomfort. Early treatment of ACS patients improves prognosis (‘time is muscle’) by preventing myocardial necrosis [1-3, 5].
In the Netherlands, most patients with ACS-like symptoms first contact the general practitioner (GP), or after office hours, the relevant services in primary care (OHS-PC). In a life-threatening situation, they can call the emergency number 112 for directly an ambulance . At the OHS-PC, a triage nurse assesses the urgency of the patients’ call.
For support, the triage nurse uses a semi-automatic de-
cision support system for urgency allocation, the so-
called ‘Netherlands Triage Standard’ (NTS). The NTS is an algorithm with hierarchically ordered questions popping up when the triage nurse has chosen the entry complaint out of 56 options. One of these is chest pain. Six urgency levels are used, each linked to a response time to receive medical care, considered suitable in that situation (Appendix 1). When the triage nurse does not agree with the NTS urgency level, the nurse may ‘overrule’ the NTS-generated urgency level, and either upgrade or downgrade this level [5, 6].
The NTS was created with the aim to pursue efficiency and safety of medical care . This, however, was not yet evaluated in patients with ACS-like symptoms. In OHS-PC,
a missed MI occurred in 0.006% yearly in the Netherlands .
Chest discomfort is the main ACS-like symptom, even though some suggest it to be less common in women with ACS [1-2, 7-9]. Either way, recognizing ACS by telephonic triage is challenging . Symptoms of ACS are non-specific, and multiple disorders may mimic the presentation. One of the complicating aspects could be the number of symptoms presented. Many symptoms could result in communication problems and ‘cognitive overload’ of the triage, resulting in (mis)interpretation of ACS-like symptoms and more likely submitting the wrong entry complaint for the NTS, leading to a less correct urgency allocation .
Previous studies suggested that women with ACS presented with more symptoms, received lower urgency allocation, experienced longer delay to medical care, and had higher mortality than men with ACS [1-2, 7-10]. It was suggested that recognition of ACS may be negatively influenced by multiple symptoms . However, research shows conflicting results regarding the number of symptoms presented when the patient was suspected of suffering from ACS, and the final presence or absence of ACS [3, 9-11].
Our aim was to assess the sex-specific relation between the number of symptoms (dichotomized in many and few) and (i) the urgency allocation and (ii) diagnosis of ACS in the OHS-PC setting.
This study is part of the Safety First study, an observational study aimed at improving efficiency and safety of telephone triage of patients with symptoms suggestive of ACS or transient ischemic attack (TIA)/stroke. Telephonic triage recordings, urgency allocation, and patients’ clinical outcomes were evaluated. We used data of seven OHS-PC locations in the Netherlands. Symptoms, pain score, medical his-
tory, cardiovascular risk factors, drug use, call characteristics and urgency levels were registered electronically. The final diagnosis was retrieved from the patient’s GP.
In the years 2014-2017, triage recordings of patients suspected of ACS, were collected at seven OHS-PC locations in the vicinity of Utrecht. Based on presented symptoms, the GP’s assigned an International Classification of Primary Care (ICPC) code. Several ICPC codes were considered suggestive for ACS (Appendix 2). 
Recordings included in this study consisted of a random sample per ICPC code and specific keywords used, to create a realistic variety of suspected ACS presentations. The specific keywords were: chest pain, pain on chest, complaints related to the heart, heart, myocardial infarction and heart attack. Exclusion criteria were: no telephonic triage conversation, conversation with a patient younger than eighteen years old, conversation with a passer-by, recordings with poor audio quality, missing relevant personal data, unknown GP, GP who does not provide follow-up data, no keywords and untraceable recording.
Data was collected by researches who listened to the selected recordings, assessed inclusion and exclusion criteria and processed the data per recording into a standardized case report form (CRF). This CRF included the patient’s symp-
toms and complaints and the attributed urgency level. Several answers could be labelled by the triage nurse after discussion with the patient: a question specific answer, ‘yes’, ‘no’ or ‘discussed, but the patient doesn’t know’. When an item was not discussed during the call, this item was labelled as ‘not discussed’.
Follow-up information, including the final diagnosis, was collected by the researchers by contacting the patient’s GP. When the final diagnosis concerned an ACS, more specific information about the type of ACS (unstable angina pectoris (UAP), non-ST elevation myocardial infarction (NSTEMI) or ST-elevation myocardial infarction (STEMI)) and the type of intervention (percutaneous coronary intervention or coronary artery bypass grafting) was collected. Currently, 3,096 recordings were available for analysis (Appendix 3 Flowchart).
This study focused on the number of symptoms presented during telephonic triage as a determinant of the presence of ACS. The number of symptoms was based on symptoms heard during telephone triage and registered in the CRF (Appendix 4).
Definition of outcome
Presence or absence of ACS was based on the information provided by the GP, including hospital letters. ACS was further specified in UAP, NSTEMI, STEMI or non-specified MI.
Urgency allocation was dichotomized in high urgency (urgency level 1 and urgency level 2) and low urgency (urgency level 3 through urgency level 5).
We divided patients into two groups: group 1 consisted of patients with ‘many symptoms’ (≥3) and group 2 of patients with ‘few symptoms’ (≤2) mentioned during the triage. We used the chi-square test to compare categorical variables and the Mann-Whitney U test for continuous variables. We did sub-analyses in men and women, the group with proven ACS and the group with proven non-ACS, using crosstabs.
Crosstabs, univariable and multivariable logistic regression analysis were applied to compare the number of symptoms with the urgency level allocation and the diagnosis of ACS, to compare the urgency level with the diagnoses of ACS, and to compare sex with the number of symptoms. Results were considered statistically significant if p-value <0.05. All statistical analyses were performed using SPSS software version 26 for Windows.
Approval of the Medical Ethical Committee of the University Medical Centre Utrecht was obtained. All patients and research data were processed anonymised according to the European General Data Protection Re-
258 triage recordings were screened, of which 120 could be included. These were added to the existing Safety First database, subsequently consisting of 3,336 recordings of patients with ACS-like symptoms who contacted the OHS-PC. The median total number of symptoms was 3 (IQR 2). 2,295 patients presented with ‘many symptoms’, 1,041 patients with ‘few symptoms’. For 2,060 cases, follow-up data was available (Appendix 3 Flowchart).
Table 1 shows the characteristics of patients with ACS-like symptoms at the time of OHS-PC presentation. The mean age of the population was 58.3±19.8 years, and 45.1% were male. Women more often had ‘many symptoms’ than men (72.6% vs. 64.1%, p<0.001). The group with ‘many symtoms’ was younger (mean age 56.7 vs. 62.0 years, p<0.001), had
a longer call duration (07:02 (04:19) vs. 06:32 (04:00), p<0.001), more often reported having chest pain (78.8% vs. 72.0%) and palpitations (3.5% vs. 2.2%) as entrance complaint. Patients with ‘few symptoms’ more often had a history of cardiovascular disease (Table 1).
All ACS-like symptoms were more often seen in patients with ‘many symptoms’ (Table 2). Bar chart 1 shows the distribution of the number of symptoms, with men presenting on average with fewer symptoms than women.
Predictive value of ‘many symptoms’ for urgency allocation
Patients with ‘many symptoms’ more often received a high urgency allocation (U1-U2) than those with two or fewer symptoms; OR 1.64 (95% CI 1.41-1.92, p<0.001), and adjusted OR 2.08 (95% CI 1.76-2.45, p<0.001), after adjustment for female sex and age per year. See table 3 and 4.
Relationship between ‘many symptoms’ and ACS
For 2,060 patients, follow-up data was available, out of which 234 (11.4%) had a final diagnosis of ACS. Overall, there was no relation between ‘many symptoms’ and ACS: OR 1.07 (95% CI 0.79-1.44), p=0.674 (Table 5). After adjustment for female sex and age, the adjusted OR was 1.36 (95% CI 1.00-1.86), p=0.051 (Table 6).
In men, the crude OR of ‘many symptoms’ was 0.91 (95% CI 0.63-1.33) and adjusted OR 1.01 (95% CI 0.69-1.49) after adjustment for age (Table 7 and 8). In women, the crude OR of ‘many symptoms’ was 1.66 (95% CI 0.98-2.83), p=0.060, and after adjustment for age; adjusted OR 2.35 (95% CI 1.35-4.08), p=0.002 (Table 9 and 10).
In this observational study, among 3,336 patients who contacted the OHS-PC with ACS-like symptoms, women reported having ‘many symptoms’ more often than men (72.6% vs. 64.1%, p<0.001). Patients who presented with ‘many symptoms’ received a higher urgency allocation than those with ‘few symptoms’; crude OR 1.64 (95% CI 1.41-1.92), adjusted OR 2.08 (95% CI 1.76-2.45) after adjustment for female sex and age. But, ‘many symptoms’ was not related to ACS; crude OR 1.07 (95% CI 0.79-1.44), adjusted OR 1.36 (95% CI 1.00-1.86) after adjustment for female sex and age. However, for women, ‘many symptoms’ was related to ACS after adjustment for age; adjusted OR 2.35 (95% CI 1.35-4.08).
Comparison with literature
To our knowledge, this is the first study to evaluate the relationship between the number of symptoms, the urgency level allocated, and final diagnosis in patients with ACS-like symptoms in the OHS-PC setting. Presentation of many symptoms could lead to communication problems and ‘cognitive overload’ of the triage nurse. The triage nurse’s intuition, attitude, perception and communication could be affected, and lead to less adequate urgency allocation [7, 13, 14]. Indeed, we showed that ‘many symptoms’ resulted in a high urgency allocation, while it was not related to ACS, except in women after adjustment for age. This is in line with findings that young women with ACS seem to be more difficult to diagnose among those with ACS-like symptoms .
Several studies evaluated the relationship between the number of symptoms and the presence of ACS in emergency departments, with inconsistent results [3, 9, 10, 12]. No such studies were conducted in the OHS-PC setting. Two studies of DeVon et al showed no relation between the number of symptoms and the presence of ACS, also not in women separately [10, 12], while our study showed a positive relation in women after adjustment for age per year.
Differences in symptom reporting and patient selection between these studies could have played a role in these findings. Three studies selected patients with an abnormal ECG or elevated troponin value [9, 10, 12], while another study selected patients with ‘typical’ symptom(s) . Also, the way of querying symptoms differed among studies, and could have affected symptom presentation .
Similar to our study, others showed that women presented with more symptoms than men [9, 10]. Khan et al showed a positive relation for the number of prodromal symptoms, such as unusual fatigue, sleep disturbances, anxiety and arm weakness/discomfort, presented by female ACS patients .
In our study, older people with ACS-like symptoms more often had ‘few symptoms’ than younger people. This is in line with previous studies performed in patients suspected of ACS who were seen in the emergency department . Two studies demonstrated a reduction in ‘atypical’ symptoms (defined as non-chest pain symptoms) in relation to an increase in age [10, 18]. We did not try to separate ‘typical’ from ‘atypical’ symptoms because there is an ongoing debate whether these exist in the acute setting. This differs from patients with stable angina, in which typical and atypical are well defined.
There are several strengths to this study. Firstly, the study consisted of a large number of participants, with real-life conversations, and assessment of symptoms blinded to the final medical outcome which prevented information bias. Secondly, the large majority of patients with ACS-like symptoms in the Netherlands first contact the GP or OHS-PC, and therefore results are generalizable to the Dutch primary care setting and even to the whole domain of people with ACS-like symptoms. Thirdly, the blinding of GP’s, who provided follow-up data for the analyses of the recordings, prevented selection bias.
This study also has several limitations. Firstly, there was missing data on symptom variables, as these were either ‘not discussed’ or not known by the patient or caller. We considered these missing variables as an absence of these symptoms. This may have caused an underestimation of the number of symptoms in our study. Secondly, not all symptoms were registered in the CRF used by the student researchers, e.g. presence or absence of fatigue. Some patients had zero symptoms, which cannot be true. Thirdly, multiple student researchers listened to the recordings and scored the CRFs, this may have led to differences in interpretation of symptoms. Furthermore, not all GPs provided follow-up information about the diagnoses, which has led to less data for analyses, and thus a low risk of selection bias. Lastly, confounders other than sex and age were not considered.
Implications for research and practice
This study shows that triage nurses more often allocate high urgency to patients with ‘many symptoms’, while they do not have a higher risk of ACS than patients with ‘few symptoms’.
To make triage safer and more efficient, there is no need to consider the number of symptoms for urgency allocation in general. Although, based on this study, including the number of symptoms for urgency allocation in women could be considered. However, more research is recommended on the women-specific relationship between the number of symptoms and the presence of ACS in the OHS-PC setting given the inconsistent results in several studies, and the lack of studies on this topic in the OHS-PC setting.
Patients presenting with three or more symptoms received more often a high urgency allocation than those with fewer symptoms, but there was no clear relationship with ACS, except in women after adjustment for age.
LIST OF ABBREVIATIONS
ACS Acute coronary syndrome
CHD Coronary heart disease
CI Confidence interval
CRF Case report form
GP General practitioner
ICPC International Classification of Primary Care
MI Myocardial infarction
NSTEMI Non-ST elevation myocardial infarction
NTS Netherlands Triage Standard
OHS-PC Out-of-hours services in primary care
OR Odds ratio
STEMI ST-elevation myocardial infarction
TIA Transient ischemic attack
U(1-5) Urgency level (1-5)
UAP Unstable angina pectoris