The South African triage scale (adult version) provides valid acuity ratings when used by doctors and enrolled nursing assistants

摘要

Highlights ? The South African Triage Scale is a valid tool that may be further implemented in similar resource limited settings. ? Mis-prediction characteristics are determined by the setting in which triage takes place (primary, secondary or tertiary health facilities) and show up both potential poor care and/or resource wastage in these different settings. Abstract Objective To estimate the validity of triage ratings by South African nurses and doctors with training and practical experience using the South African Triage Scale. Methods Five emergency physicians and 10 enrolled nursing assistants, who had been trained in the use of the South African Triage Scale, were selected via convenience sampling to retrospectively triage adult emergency centre vignettes. Participants independently assigned triage ratings to 100 written vignettes unaware of the ratings given by others. Triage ratings were compared with ratings of two experts from the South African Triage Group. Standard validity indicators including sensitivity, specificity, positive predictive value and negative predictive value were used to estimate the validity for the combined group of emergency physicians and enrolled nursing assistants. Associated percentages for over-/under-triage were used to further assess validity within the South African context and over-/under-prediction to further assess practical application of the South African Triage Scale. Results On average over all acuity levels, sensitivity was 75%, specificity 91%, under-triage occurred 10% and over-triage 15% of the time. The positive predictive value was 74% and negative predictive value 91%. Conclusion The results of this study fall within the accepted range of over-/under-triage and indicate that the South African Triage Scale is valid when used by emergency physicians and nurses to triage emergency centre vignettes under South African conditions. Further research into appropriate reference ranges for extent of over-/under-triage and over-/under-prediction within each acuity level is recommended.