[Reliability and validity of four-level and three-district triage standards in emergency department: a randomized sampling cross-sectional study of 1 106 adult patients].

Objective: To evaluate the reliability and validity of three-district and four-level triage standards in adult emergency department.
Methods: A randomized sampling cross-sectional study was conducted. A total of 1 106 emergency adult patients admitted to the Second Xiangya Hospital of Central South University in Hunan Province from December 2015 to April 2016 were enrolled. The triage was independently performed by 12 nurses according to the emergency triage criteria. Based on the shift style, 2 nurses were assigned to each shift as the triage guider and assistant respectively, who did the triage for every patient independently. The clinical data were recorded as follows: the demographic data, emergency information (triage time, emergency way, complaints, vital signs, and conscious state), triage information (triage level, admitted department), waiting time, treatment time, destination and outcomes. The reliability of three-district and four-level triage standards was analyzed by Spearman correlation, and the receiver operating characteristic curve (ROC) was plotted to evaluate its validity.
Results: (1) A total of 254 patients were enrolled for reliability evaluation in the first 2 weeks of the study. The overall internal consistency rate of the triage instructors and the triage assistants was 72%, the total Kappa value was 0.686 [95% confidence interval (95%CI) = 0.608-0.757, P < 0.001]. (2) Validity analysis showed that in the 1 125 emergency patients collected during the study, a total of 1 106 patients were finally enrolled in the analysis excluding the patients who refused to accept the treatment, whose data was incomplete and who was diagnosed as prehospital death. With the increase of three-district and four-level triage level, a significant increase was showed in the waiting time of patients, the treatment time, and the retention rate; on the contrary, the salvage rate, the hospitalization rate, hospitalization time, emergency mortality, in-hospital mortality and total mortality rate were decreased [the waiting time of patients from triage level 1 to 4 (minutes) was 1.00 (1.00, 1.75), 1.00 (1.00, 5.00), 8.00 (2.00, 23.00), 10.00 (4.50, 28.00), the treatment received time (minutes) was 1.00 (1.00, 10.00), 6.00 (1.00, 23.00), 48.00 (25.00, 105.00), 87.00 (41.00, 140.00), the retention rate was 4.76%, 10.94%, 55.91%, 42.86%, the salvage rate was 95.24%, 87.94%, 20.81%, 0%, the hospitalization rate was 57.14%, 70.98%, 53.62%, 20.41%, the hospitalization time (days) was 19.50 (9.75, 28.00), 11.00 (8.00, 17.00), 12.00 (8.25, 17.00), 10.50 (8.75, 15.25), the emergency mortality was 19.05%, 6.92%, 1.41%, 0%, the in-hospital mortality was 16.67%, 15.09%, 6.25%, 0%, and the total mortality rate was 28.57%, 17.63%, 4.76%, 0%, all P < 0.05]. ROC curve analysis showed that the area under ROC curve (AUC) of three-district and four-level triage standards for identifying patients needed an immediate intervention (triage level 1 to 2) was 0.854 (95%CI = 0.831-0.878), and the sensitivity and specificity were 78.62% and 89.89%, respectively, the misdiagnosis rate was 10.11%, and the missed diagnosis rate was 21.38%.
Conclusions: The three-district and four-level triage standards were proved to be a reliable and valid instrument, which can distinguish the severity of the disease and help nurses to triage patients correctly.