Triage of acute chest pain evaluated in urgent primary care (TRACE)

BACKGROUND Chest pain is a common symptom, and identifying patients with chest pain who require urgent care can be quite challenging. Making the correct assessment during telephone triage is even harder. Temporal trends show that the referral threshold of telephone triage has lowered over time, resulting in overcrowding of first-responders and emergency services. While various stakeholders advocate for a more efficient triage system, careful evaluation of the performance of the current triage in primary care is lacking. PRIMARY AIMS 1) to evaluate the performance of the current triage system for chest pain (“Nederlandse Triage Standaard” = NTS) ; and 2) to study the performance of clinical prediction rules for chest pain (Marburg Heart Score and INTERCHEST score) alone or when embedded/integrated in the current triage system. STUDY DESIGN The TRACE study involves a large cohort study designed to collect information on the triage of chest pain in urgent primary care, as well as baseline, diagnostic work-up and clinical outcomes data. First, we evaluate the performance of the NTS using a retrospective cohort with subsequent follow-up. Secondly, we aim to evaluate prospectively the performance of clinical prediction rules that are based on anamnestic questions and that have been validated in primary care (INTERCHEST, Marburg Heart Score). Data on triage will be collected from a large, regional urgent primary care facility. Follow-up information will be obtained from affiliated day-time primary care practices that keep up-to-date records of their patients, including data on diagnostic work-up, hospital admission and discharge letters, medications, and clinical outcomes. STUDY POPULATION The study consists of consecutive patients of at least 18 years of age with a primary or secondary symptom of chest pain who contacted urgent primary care in Alkmaar, a facility that provides out-of-hours primary care service for a region covering over 240,000 individuals. The region is representative of the Dutch population in terms of ethnicity, as the majority of the inhabitants are of Dutch or European ancestry and 12.8% being of non-Western/non-Caucasian descent (mainly Turkish, Moroccan, Surinamese and Antillean). NTS QUESTIONS AND CLINICAL PREDICTION RULES The NTS questions for chest pain involve chest pain characteristics (type, duration, severity, course, location), presence of radiation, sweating, nausea, or vomiting. Additionally, triage nurses can explore the medical history, medication use or the presence of accompanying symptoms, such as dyspnea or dizziness. The Marburg Heart Score and INTERCHEST score assign points to a priori risk factors (age >=65 in women and >=55 in men, history of cardiovascular disease) and chest pain characteristics (relationship with exercise, palpation, ”pressure-type” pain), and whether a cardiac cause is suspected (by the triage-nurse/physician or the patient). CLINICAL OUTCOMES The primary outcome is the occurrence of a ‘major event’, which is defined as the occurrence of acute coronary syndrome or other high-risk diagnoses in which delay is inadmissible and hospitalization is necessary. PRIMARY ANALYSIS We will evaluate the performance of the NTS triage system, clinical prediction rules alone, and clinical prediction rules when integrated in NTS triage. We will evaluate performance by assessing the ability to correctly classify patients regarding urgency (accuracy), the proportion of safe actions following triage (safety) as well as correct use of ambulance deployments (effectiveness). SAMPLE SIZE / DATA ANALYSIS We aim to include 2,200 patients in total, of which 700 patients will be enrolled prospectively. We anticipate that this will render a total of 150-200 cases with an urgent diagnosis. This number is deemed sufficient to evaluate the variables of NTS (retrospective cohort of 1,500 patients) , Marburg Heart Score and INTERCHEST. PROPOSED SECONDARY ANALYSES First, we will evaluate whether sex-based differences exist in the performance of triage for chest pain. Second, we aim to evaluate the impact of ‘nurse/physician overruling of the NTS-suggested urgency’ on safety, effectiveness and accuracy. Third, we will explore whether we will be able to build a reliable triage model using novel machine learning techniques.