The principal objective, NPV 96% [95% CI 89–99%], was achieved. However, this short training session was inadequate to allow ruling-out a DVT with sufficient security. Actually, 3 DVT out of 18 were not identified by the residents. A recent guideline of the American Society of Hematology recommends a NPV rate of 2% [
9]. Our study was thus not powerful enough. Furthermore, there were also 10 false-positive exams that would have possibly led to an unwarranted anticoagulant treatment. DVT prevalence, 17.6% [95% CI 11–26%], was not different when compared to other studies as reported in a meta-analysis including 2379 patients [
10] (23%). Our results showed that performances tend to be less accurate than in a majority of published studies. In the same meta-analysis [
10], pooled sensitivity and specificity were 94.8% and 96.2%, respectively, when compared to our results, 83% and 88%, respectively. Nevertheless, there was an overlap between the confidence interval for sensitivity precluding definite answer. Comparison with the Jang [
11] study exhibits similar results. However, the goal of this study was to evaluate a very short training session in residents without previous ultrasound skills. In the literature, training was either longer [
6,
12,
13] or addressed physicians with previous POCUS experience [
11,
14‐
17]. To the best of our knowledge, it is the first study which assesses performances of a very short training session in venous ultrasound for residents without previous POCUS skills. It is likely that the absence of an experiential phase on real patients was crucial to explain these results. During the training session, residents had only performed normal VLCU on one or two other participants. Confronted with real patients, they assessed a difficulty score at 3.7 ± 2 which was relatively high and were unable to conclude in 7.5% of examined sites. Furthermore, the number of exams per resident was low, 2.3 ± 2.1 which prevented acquisition of diagnostic capacities. This fact could be partially explained by the residents’ duration of stay in the ED.
Our intention is thus to modify our training pathway by including an experiential phase of 15 monitored VLCU by resident. Actually, in an article on learning curves in POCUS, conversely to other sites such as soft tissues or Focused Assessment in Trauma, the authors were unable to determine a required number of exams to reach a good accuracy [
18].
The principal limitations of this study were first, the recruitment of a convenience sample since this could not reflect the actual patients admitted to the ED for a suspicion of DVT. However, recruitment could only be performed during duty hours because of the local vascular laboratory availability and, when the ED were overcrowded, residents did not have time to recruit patients. Second, the number of exams per resident was low.