The data of 37854 PTB and suspected PTB patients, who should be referred, reported by NTI of Hefei in 2011-2016 the TB information management system were collected from TB Information Management System. All analysis (e.g. collation, summary and statistical analysis) was performed using Excel and SPSS (version 19.0). The constituent ratio was the main evaluation index. Logistic regression was used to analyze the related factors and P values less than 0.05 was considered statistically significant.

Among the PTB and suspected PTB patients who should be referred reported by all kinds of NTI in 2011-2016, 36286 (95.86%) referral cases arrived. A total of 16484 cases should be tracked, 14916 (90.49%) cases arrived and 1561 (9.47%) cases didn’t arrived, and 7 (0.04%) cases had no track information. The reasons for 1561 cases did not arrive included factors of patients (45.04%, 703/1561), factors of NTI (29.34%, 458/1561) and factors of TB control institution (25.62%, 400/1561). Single factor analysis showed that the age of the patient, the type of reported hospital, location of the present address, occupation, and the classification of patients had different level of effects on whether the patients could be tracked (χ ² values were 33.19, 938.83, 206.13, 26.71, 24.64, Ps<0.001). Multivariate logistic regression analysis showed that the age of the patient, the type of reported hospital, location of the present address, and occupation played important roles in tracking failure (Wald χ ² values were 39.65, 621.62, 288.81 and 25.65, Ps<0.001).

The age of the patient, the type of reported hospital, the location of the present address and occupation are important factors of tracking failure. We should increase the communication of TB control institution with NTI and strengthen the training and monitor of NTI on TB Centralized Management in order to improve the overall arrival rate of TB patients.