Abstract:

Objective: To investigate the current status of delayed diagnosis of pulmonary tuberculosis (PTB) at Wuhan Fourth Hospital, analyze its influencing factors of delayed diagnosis and develop a risk prediction model, thereby providing evidence for optimizing the prevention and control strategies of PTB. Methods: A total of 125 PTB patients reported by Wuhan City between January 1, 2020 to December 31, 2022, and who had medical visits to Wuhan Fourth Hospital within six months before reporting were selected from the Chinese Disease Prevention and Control Information System based on inclusion criteria. Demographic characteristics and clinical data of the patients were collected from the hospital’s medical record information system. Descriptive statistics, univariate analysis and multivariate logistic regression were used to analyze the influencing factors associated with diagnostic delay. A risk prediction model was constructed and evaluated using the area under the receiver operating characteristic curve (AUC), decision curve analysis (DCA), and calibration curve (CC). Results: The median (interquartile range, IQR) of reporting interval was 49 (13, 110) days. The overall diagnostic delay rate was 74.40% (93/125), and 66 patients (52.80%) received imaging or TB tests. Among those reported by Wuhan Fourth Hospital itself (n=34, 27.20%), the median reporting interval was 10 (5, 81) days, with a diagnostic delay rate of 44.12% (15/34) and 100.00% (34/34) received imaging or TB tests. Among patients reported by other hospitals (n=91, 72.80%), the median reporting interval was 58 (23, 114) days, the diagnostic delay rate was 85.71% (78/91), and only 35.16% (32/91) received imaging or TB tests. Statistically significant differences were observed between our hospital and other hospitals in terms of reporting interval, proportion of imaging or TB tests, and diagnostic delay rate (Z=-3.199, P=0.001; χ²=41.750, P=0.001; χ²=22.486, P=0.001). The delay rate in the respiratory department (33.33% (7/21)) was significantly lower than that in other departments (82.69% (86/104)) (χ²=22.349, P<0.001). Multivariate logistic regression analysis identified first consultation in the respiratory department (OR=0.182, 95%CI: 0.055-0.597) and ordering of imaging or TB-related tests at the initial hospital visit (OR=0.196, 95%CI: 0.065-0.588) as independent protective factors against diagnostic delay. The AUC of the risk prediction model was 0.820, and 0.833 in the internal validation set. DCA demonstrated that the model was clinically applicable across a threshold probability range of 6% to 80%. The calibration curve indicated good agreement between predicted and actual outcomes. Conclusion: The delayed diagnosis rate of PTB in our hospital is high. The independent risk factors for diagnostic delay are the first visit in non-respiratory medicine department and the absence of imaging or tuberculosis test. The constructed risk prediction model demonstrated favorable predictive performance. It is recommended to strengthen the first diagnosis of tuberculosis, improve the ability of doctors in general hospitals, especially non-respiratory specialists, to identify PTB through systematic and long-term training, and improve the timeliness of hospital report.

Key words: Tuberculosis, pulmonary, Early diagnosis, Hospitals, general, Epidemiologic studies, Factor analysis, statistical