A pilot study to assess feasibility of evaluating the safety and efficiency of a simplified diagnostic approach for pulmonary embolism

Abstract

Background: Decision rules designed to reduce the need for pulmonary embolism (PE) imaging are seldom used by emergency physicians because of their complexity and poor credibility. We designed a simple, modified age-adjusted decision rule for PE testing in the emergency department (ED) – the ‘Adjust-Unlikely’ rule. Given increasing pressure on EDs, alternative methods to in-person recruitment are needed to validate the diagnostic performance of the Adjust-Unlikely rule in a full-scale study. Objectives: The primary objective of this research was to assess the feasibility of a protocol consenting patients tested for suspected PE with the Adjust-Unlikely rule by telephone after they left the ED. The secondary objectives were to estimate safety and efficiency of the Adjust-Unlikely rule for diagnosing PE in the ED and to assess whether our methods introduced spectrum bias. Methods: A prospective management pilot study was conducted in one ED and one urgent care centre. Adult patients tested for PE using the Adjust-Unlikely rule were recruited by telephone and followed for 90 days to identify subsequent testing for venous thromboembolism. PE and deep vein thrombosis (DVT) testing during follow-up was independently adjudicated. The feasibility outcomes were recruitment rate, missed eligible rate and overall follow-up rate. To assess for spectrum bias, multivariable logistic regression was used to compare included patients to missed eligible patients and patients who declined participation. Results: Two hundred patients were included. On average, 9 patients were recruited per site, per week and 6 patients were missed per site, per week. The overall follow-up rate was 96.5%. Of the 143 patients with a negative Adjust-Unlikely rule on index visit, 1 patient was diagnosed with PE during follow-up. We found evidence of spectrum bias in our study population. Female patients and older patients were over-represented in our study population compared to patients who were eligible but not included. Conclusions: Telephone recruitment did not meet our predefined thresholds for feasibility. This research has informed the design and planning of a full-scale study that will prospectively validate the safety and efficiency of the Adjust-Unlikely rule for diagnosing PE in the ED across Ontario.