Study on safety performance and condition-suggestion accuracy of the symptom assessment mobile applications
10.3760/cma.j.cn114798-20220627-00703
- VernacularTitle:数字化症状评估程序应用于疾病诊断及安全分诊准确性的探讨
- Author:
Shanzhu ZHU
1
;
Sunfang JIANG
;
Juan SHOU
;
Zhigang PAN
;
Yu ZHANG
;
Minghui PENG
;
Hua YANG
;
Stephen GILBERT
Author Information
1. 复旦大学附属中山医院全科医学科，上海 200032
- Keywords:
Artificial intelligence;
Diagnosis, differential;
Decision support systems;
Clinical
- From:
Chinese Journal of General Practitioners
2023;22(3):288-294
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To compare the breadth of condition coverage, accuracy of suggested conditions and appropriateness of urgency advice of the 8 symptom assessment mobile applications (APPs) available on the Chinese market.Methods:The APPs were assessed using 200 primary care vignettes and were measured against the vignettes′ standard. The primary outcome measures were proportion of conditions covered by an APP, proportion of vignettes with the correct primary diagnosis,and proportion of safe urgency advice.Results:For APPs assessed,condition-coverage was from 29.0%(58/200)to 99.5%(199/200), top-3 suggestion accuracy was from 8.5%(17/200) to 61.5%(123/200), the proportion of safe urgency advice was from 84.8%(167/197) to 99.5% (198/199).Conclusions:The APPs showed a wide range of coverage, safety performance and condition-suggestion accuracy. Symptom assessment APPs with good performance could be used by general practitioners as supporting tools. However, even symptom assessment APPs with excellent performance need to be further assessed in a real clinical environment.
