Trial for the Use of Pretest Probability to Reduce Unnecessary Testing for Low-Risk Patients With Chest Pain

Overview

The purpose of this study is to evaluate if the implementation of quantitative pretest probability assessment will significantly reduce the unnecessary use of the intra-emergency department chest pain center. Specifically, the study will examine whether the PREtest Consult acute coronary syndrome (ACS) pretest probability assessment system can significantly reduce the use of chest pain unit evaluation in very low risk emergency department (ED) patients, can safely discharge patients with a pretest probability ≤ 2.0%, can reduce unnecessary procedures and lower hospital costs and will examine patient satisfaction of patients with whom pretest probability assessment was used compared to those with whom it was not used. The researchers hypothesize that patients in the control group of the study will have statistically significant reductions in mean time spent in the emergency department, mean charges billed to the patient or their insurance carrier, hospital length of stay, mean number of procedures or tests performed without a statistically significant change in patient satisfaction or adverse outcome.

Chest pain represents the second most frequent complaint among the 110 million persons who visit emergency departments in the U.S. each year. Perceived medicolegal risk compels emergency physicians to overtest for possible acute coronary syndrome (ACS), contributing to more than $20 billion in unnecessary diagnostic testing each year. The hypothesis of the present work states that quantitative pretest probability assessment can significantly and safely increase the proportion of very low risk patients with symptoms of ACS who are discharged from one emergency department. Quantitative pretest probability will be assessed with the validated, commercially available PREtest Consult ACS software device, which employs computer assisted, attribute matching. This method matches an 8-component clinical profile from any individual patient to the same profile shared by patients who were previously evaluated for ACS and whose profiles are stored in a 14,800 patient reference database. Pretest probability estimates ≤ 2% will be considered "test negative." A phase II multicenter study found that when the ACS PREtest Consult produced a pretest probability of ACS ≤ 2.0% that the actual outcome of ACS at 45 days was 0.3% (95% CI 0 to 1.8%) compared with 0.4% (0 to 0.9%) for patients discharged after negative testing in a chest pain unit (CPU) that included serial biomarkers, overnight monitoring, and cardiologist-interpreted provocative testing. Over one-quarter of all patients referred to the CPU had an estimate ≤ 2%. The present study will randomize 400 ED patients with a non-diagnostic or normal ECG and a troponin test ordered into two groups: 1. a "show me", or disclosure group, in which patients and their clinicians will receive the output of the device, and 2. a "no show" or concealed group will receive no output. The sample size will detect an 11.5% difference in rate of discharge between groups with α = 0.05 and β = 0.20. All discharged patients will undergo structured telephone and medical record follow-up at 7 and 45 days using validated methodology. Primary outcome measures will evaluate the development of acute coronary syndrome. Secondary variables will include: 1. Frequency of ACS, determined by blinded adjudicated review of follow-up data, 2. Rate of return to any ED for similar symptoms, 3. Patient satisfaction, 4. Charges. Project significance includes the potential to reduce patient exposure to unnecessary invasive procedures and to save Medicare and other insurers over $100 million in unnecessary diagnostic testing each year in the U.S.