AI辅助软件对疲劳状态下住培医生检出肋骨骨折的提升作用

The study of the enhancement effect of AI-assisted software on detection of rib fractures by resident physicians in fatigued states

目的:探讨人工智能辅助软件对疲劳状态下住培医生检出肋骨骨折的提升作用。方法:回顾性分析2021年1月—2022年5月在梅州市人民医院接受CT检查的86例肋骨骨折患者资料。住培医生按以下三种模式进行阅片:A模式为正常状态下独立阅片、B模式为疲劳状态下(即一天工作满8小时后)独立阅片、C模式为疲劳状态下应用AI辅助阅片,三种阅片模式之间间隔4周。比较住培医生不同阅片模式下对肋骨骨折的检出率及阅片时间。结果:住培医生疲劳状态下独立阅片对肋骨骨折的检出率低于正常状态独立阅片,阅片时间长于正常状态独立阅片,差异均具有统计学意义(P<0.05)。住培医生疲劳状态下应用AI辅助阅片对肋骨骨折的检出率高于疲劳状态独立阅片,阅片时间短于疲劳状态独立阅片,差异均具有统计学意义(P<0.05)。住培医生疲劳状态下应用AI辅助阅片对肋骨骨折的检出率已超过其正常状态的水平,阅片时间短于正常状态的水平,差异均具有统计学意义(P<0.05)。结论:住培医生疲劳状态对肋骨骨折的检出效能较正常状态显著降低,但在AI辅助下能显著提升疲劳状态住培医生对肋骨骨折的检出效能,甚至超过其正常状态下的水平。

Objective To explore the enhancement effect of artificial intelligence assisted software on the detection of rib fracture by resident physicians in fatigue state.Methods The data of 86 patients with rib fractures who underwent CT examination in Meizhou People's Hospital from January 2021 to May 2022 were retrospectively analyzed.The resident physicians perform CT readings in the following three methods:mode A was independent reading in normal state,mode B was independent reading in fatigue state(after working 8 hours a day),and mode C wasAI-assisted reading in fatigue state,with an interval of 4 weeks between the three modes of reading.Comparison of the detection rate and reading time of rib fractures in different reading modes by resident physicians.Results The detection rate of rib fractures by independent reading in the fatigued state of resident physicians was lower and the reading time was longer than that of independent reading in the normal state,and the differences were statistically significant(P<0.05).The detection rate of rib fractures by applying AI-assisted reading in the fatigue state of resident physicians was higher and the reading time was shorter than that of independent reading in the fatigue state,and the differences were statistically significant(P<0.05).The detection rate of rib fractures by AI-assisted reading in the fatigue state of resident physicians has exceeded the level of their normal state,and the reading time was shorter than that of the normal state,the differences were statistically significant(P<0.05).Conclusion The detection efficiency of rib fractures by resident physicians in fatigue state was significantly lower than that in normal state,but the detection efficiency of rib fractures by resident physicians in fatigue state could be significantly improved with the assistance of AI,and even exceeded the level of their normal state.