Multilevel factors affecting tuberculosis diagnosis and initial treatment

Abstract

Objectives: The study aims to assess provider adherence to national tuberculosis programme guidelines on diagnosis, initial regimens and dosages, and to examine independent effects of factors at patient, staff and hospital levels influencing adherence.

Methods: A review of 383 medical records of new tuberculosis (TB) patients and interviews with related staff were carried out. The study was conducted in 16 public hospitals of seven provinces of southern Thailand. The outcome variables were provider adherence to the guidelines on diagnostic procedure, initial regimen and dosage. Independent variables consisted of patient, staff and hospital factors. Multilevel logistic regression was used to identify factors associated with adherence.

Results: The proportions of adherence to the diagnostic procedure, initial regimen and initial dosage prescribed were 70.0%, 100.0% and 57.1%, respectively. Most of diagnosis non-adherence was anti-TB drugs being prescribed for smear-negative patients without prior antibiotic trial (12.5%). The anti-TB drug with the highest percentages of patients receiving non-adhered dosage was ethambutol (33.6%). In contrast to single-level analysis, which showed significant influence of up to five factors, multilevel analysis confirmed only strong effect of male patients receiving better adhered diagnosis and of non-doctors and TB clinics providing better dosage adherence.

Conclusions: Adherence to TB diagnostic procedures was not good, and adherence to initial dosage, especially for ethambutol, was poor. TB clinics, the key factor of adherence, should be expanded. Female patients should be reviewed more carefully because they tend to receive poorer diagnosis adherence.