Influencing factors of treatment outcomes of etiologically positive pulmonary tuberculosis patients based on Bayesian network
- Author:
ZHANG Mai
;
YANG Jingyuan
;
YU Qin
- Publication Type:Journal Article
- Keywords:
Pulmonary tuberculosis;
etiologically positive;
Bayesian network;
successful treatment rate;
influencing factor
- From:
China Tropical Medicine
2025;25(2):184-
- CountryChina
- Language:Chinese
-
Abstract:
Objective Exploring the risk factors and degree of influence on treatment regression of patients with etiologically positive pulmonary tuberculosis in Bijie city, so as to provide reference for the precise prevention and treatment of pulmonary tuberculosis epidemic. Methods The medical records of registration, treatment and management of pulmonary tuberculosis patients in Bijie City with positive etiology from 2017 to 2022 were derived from the Tuberculosis Monitoring Report Information Management System. Statistically significant variables were screened using binary logistic regression. Revealing interdependencies between factors and conditional probabilities by constructing Bayesian network models. Results A total of 21 662 cases of etiologically positive pulmonary tuberculosis were registered in Bijie City from 2017 to 2022, with etiologically positive rate of 48.60% and a successful treatment rate of 89.52%, both of which were on the rise overall (χ2trend=2 618.15, P<0.001; χ2trend=102.84, P<0.001). A multifactorial analysis showed that gender, age, ethnicity, occupation, patient origin, HIV test results, migrating population, and treatment classification were influencing factor for the successful treatment rate (P<0.05). The Bayesian network model showed that age and treatment classification were the direct factors affecting the treatment outcome, while gender, occupation and patient origin were the indirect factors. Conditional probability showed that patients<20 years of age and initial treatment had the highest rate of successful treatment (96.75%), with a 3.25% probability of having an adverse outcome; Patients≥80 years of age who were retreated had the lowest rate of successful treatment (58.70%), with a 41.30% probability of having an adverse outcome. Conducting causal effect speculation, when preventing tuberculosis in older persons aged 60 years and over, and the other patients are newly treated patients, the successful treatment rate increased from 89.50% to 93.10%, with an increase of 3.60%; Nearly 50.00% fewer farmers are without tuberculosis, and the source of the other patients is active detection (health check-ups and proactive screening), with an increase in the successful treatment rate from 89.50% to 91.70%, with a change in the conditional probability value of 2.20%. Conclusions Consultation and management of male, elderly, re-treatment, farmer, HIV-positive pulmonary tuberculosis patients in Bijie should be strengthened. Meanwhile, causal correlation effect inference using Bayesian networks is important for precise prevention, control and treatment of diseases.
- Full text:202511171019356442810.Influencing factors of treatment outcomes of etiologically positive pulmonary tuberculosis patients based on Bayesian network.pdf
