Objective   To evaluate the performance of the detection and management mode of tuberculosis (TB) patients by active screening, hospitalization in infectious phase and “centralized medication + nutritional breakfast” in Xinjiang, and analyze the difference in the performance in areas with different TB burden and provide reference for the improvement of TB control programs and the formulation of health intervention policies in Xinjiang.

  Methods  The registration data of TB patients in Aksu and Bayingoleng of Xinjiang from January 2012 to December 2021 were collected. Interrupted time series (ITS) was used to analyze the change of TB registration rate before and after the implementation of the TB detection and management mode. Differences-in-Differences (DID) model was used to analyze the difference in the implementation effect of the mode between high-burden areas and low-burden areas.

  Results  From 2012 to 2021, the registration rate of TB in Aksu was higher than that in Bayingoleng. The registration rates of TB in both Aksu and Bayinguoleng showed upward trends before 2018, reached the peak in 2018, and continued to decline rapidly after 2018. In the year when the mode was taken, the TB registration rate in Aksu increased significantly by 248.800/100 000 (P<0.001); The analysis on long-term trend after the implementation of the mode indicated that the registration rates in Aksu and Bayingoleng all showed downward trends (β₃<0, P<0.05), the registration rate in Aksu decreased more rapidly compared with Bayingoleng (β₁+β₃: 139.747>48.712). There were significant differences in the number of hospital beds per 10 000 people, number of medical institutions, number of doctors per 10 000 people, GDP per capita, disposable income of rural residents and basic medical insurance for employees between Aksu and the control area (P<0.05). The results of DID model constructed by taking them as covariates showed that after the implementation of the mode the average effect difference between Aksu and Bayinguoleng was 355.04, which had no significance (P=0.205), but the average effect difference was significant in 2018 and 2019, indicating that the intervention effects in Aksu in 2018 and 2019 were greater than those in Bayinguoleng.

  Conclusion  In the areas with high and low TB burden in Xinjiang, the detection and management mode of TB patients by active screening, hospitalization in infectious phase and “centralized medication + nutritional breakfast” had long-term effects on reducing the registration rate of TB, and it is expected that continuous implementation of the mode can effectively curb the spread of TB.