Clinical Outcomes of the Tuberculosis Control Program in Taiwan Using Propensity Score Matching

To the Editor:

T uberculosis (TB) is a single infectious agent causing more than 26% of avoidable adult deaths in the developing world, ¹ and the entire TB regimen is laborious and time-consuming. Directly observed therapy (DOT), using a strict treatment strategy, observes and records patients swallowing the correct dosage of anti-TB medicine, (eg, rifampicin), and is considered by far the most effective TB treatment strategy by the World Health Organization, which recommended this most efficient scheme for TB control. TB mortality in Taiwan has declined over time, but the prevalence/incidence rate, compared with other countries, is higher than the worldwide rate and that of the United States. DOT was universally implemented in April 2006 because of the “National TB Program Mobilization Plan” to halve the incidence of TB in 10 years. TB participants enrolled in DOT voluntarily in the initial stage; later stages focused on patients with a positive sputum smear.

The data for this retrospective study come from the Centers for Disease Control (CDC) Taiwan TB registration systems covering 4 counties in Taiwan from April 1, 2006 to March 31, 2007. There were 1328 participants in the non-DOT program and 1220 in the DOT program. Six characteristics in this study included patients' demographics (sex and age), sputum smear, sputum culture, X-ray, aboriginal status, and whether they had joined the DOT program. After using the propensity score method to match the non-DOT group (1:1), there were 1230 cases remaining from the 2 groups (615 in the DOT group and 615 in the non-DOT group). Logistic regression analysis was conducted to test the relationship for the predicators and the mortality rate.

The propensity score method is widely applied in clinical studies and the public health field to deal with selection bias issues and has proven to be an acceptable method. The factors associated with mortality rate were age, sputum smear, and the DOT program. Unsurprisingly, older TB patients had a higher likelihood of death than younger patients did (odds ratio [OR] 5.62–60.24). Patients with a positive sputum smear had 1.6 times (95% confidence interval [CI]: 1.11–2.30) the risk of death of those who did not. Most importantly, patients who joined the DOT program had a lower mortality rate than those who did not (OR, 0.31; 95% CI: 0.23–0.43).

Most findings of this study are consistent with other research. ^{2 –4} Age, race, and a positive smear are still the main factors related to TB mortality; however, joining the DOT program has an influence on reducing the risk of morality. We strongly believe that high-quality drug treatment, properly administered under DOT, can markedly reduce the burden of TB, which is demonstrated in other studies. ^{5 –7} The mortality figures for the older groups are still high and must be improved. CDC Taiwan has built a comprehensive monitoring system with DOT to reduce the lost-to-contact rate, and we look forward to adopting a seamless network of TB control efforts to reduce the burden of TB in the near future.

Although it is not easy to evaluate the effect of DOT, it generally indicates highly positive outcomes related to good health service strategy and management. It was found that standardized treatment regimens and intensive programs such as DOT provide positive outcomes in TB control, and the performance of public health responsibilities is still an essential principle of disease control.