Triage for out-of-hours primary care calls: a reliability study of a new French-language algorithm, the SALOMON rule.

Triage systems for out-of-hours primary care physician (PCP) calls have been implemented empirically but no triage algorithm has been validated to date. A triage algorithm named SALOMON (Système Algorithmique Liégeois d'Orientation pour la Médecine Omnipraticienne Nocturne) was developed to guide triage nurses. This study assessed the performance of the algorithm using simulated PCP calls. Ten nurses were involved in 130 simulated PCP call scenarios, allowing the determination of SALOMON's inter-rater agreement by comparing the actual choices of a specific triage flowchart and the level of care selected as compared with reference assignments. Intra-rater agreement was estimated by comparing triage after training (T1) and 3 to 6 months after SALOMON use in clinical practice (T2). Overall selection of flowcharts was accurate for 94 .1% of scenarios at T1 and 98.7% at T2. Level of triage was adequate for 93.4% of scenarios at T1 and 98.5% at T2. Both flowchart and triage level accuracy improved significantly from T1 to T2 ( < 0.0001). SALOMON algorithm use is associated with a 0.97/0.99 sensitivity and 0.97/0.99 specificity, at T1/T2 respectively. Results revealed that using the SALOMON algorithm is valid for out-of-hours PCP calls triage by nurses. The criterion validity of this algorithm should be further evaluated through its implementation in a real life setting.