Background & Setting

1. Has Directly Observed Treatment (DOT) Improved Outcomes for Patients with Tuberculosis in Southern Thailand?Pungrassami P1, Johnsen SP2, Chongsuvivatwong V3, Olsen J41Zonal Tuberculosis Centre 12, Thailand; 2Department of Clinical Epidemiology at Aarhus University Hospital and Aalborg Hospital, Aarhus, Denmark; 3Epidemiology Unit, Faculty Of Medicine, Prince Of Songkla University, Thailand; 4The Danish Epidemiology Science Centre, University of Aarhus, DenmarkAbstract: Problem Statement: The value of actual practice of Directly Observed Treatment (DOT) for tuberculosis (TB) has not been documented. Randomised controlled trials of DOT have shown conflicting results, perhaps due to information bias.Objectives: To validate the practice of DOT and evaluate its effect on treatment outcomes.Setting:The government health system in 24 districts in southern Thailand, with DOTS implemented. Methods:411 new, smear-positive, pulmonary TB patients who started treatment between February and September 1999, were followed up. Patients and/or their observers were interviewed about their actual DOT practice during the first 2 months of treatment. Treatment outcomes were evaluated at the end of the 2nd month and at the end of the treatment, based on WHO definitions.Results:Of 411 patients, 379 were assigned to DOT but only 68 practised strict DOT for every dose during the first 2 months. There was no evidence of differences in both short- and long-term outcomes between DOT and no-DOT groups (adjusted OR for “no sputum conversion” and “Unsuccessful treatment”: 1.1, 95% CI 0.6-2.1 and 1.3, 95% CI 0.6-2.8, respectively).Conclusions:Actual practice of DOT was quite different from what was intended at the assignment. The practice of strict DOT during the first 2 months was not associated with sputum conversion or treatment success in this study area.Study Funding: Thailand Research Fund. The activities of the Danish Epidemiology Science Centre are financed by a grant from the Danish National Research Foundation. The Department of Clinical Epidemiology receives financial support from the Danish Medical Research Council (grant No. 9700677).

2. Background & Setting • Directly Observed Treatment (DOT), one element of DOTS strategy, has been recommended for the improvement of patient adherence to the TB treatment. • Three maintypes of DOT observer are used in Thailand: health personnel-HP (staff members of TB clinics, hospital wards, and health centres), community members-CM (village health volunteers, community leaders, and friends), and family members-FM (close and distant relatives). • The National Guidelines recommend that the preferred choice of observer in descending order is HP, CM, and FM. However, the majority of assigned DOT observers were FM. • The study area covered 1.2 million people in 24 districts in southern Thailand with DOTS implemented since 1996.

3. Study Rationale • In spite of worldwide implementation, the efficacy of DOT remains questionable. Randomised controlled trials have given opposite results, and compliance to the DOT principle has not been reported in these studies. • It is expected to be difficult to maintain the initial allocation of DOT over several months, and analysis according to the “intention-to-treat” principle may leave a toolimited exposure contrast to produce meaningful results. Study Objectives • To quantify DOT in practice and to estimate the effect of actualDOT practice on treatment outcomes.

4. Methods 1 • Through the TB Registers at the 22 TB clinics, we identified and followed up all 455 patients with new, smear positive, pulmonary TB; who started treatment between 01/02/1999-30/09/1999. Exposure variable • Two types of DOT were defined, DOT by the initial assignment (assigned DOT) and DOT in actual practice (actual DOT). The patients initially assigned to have any types of observer were grouped as “assigned-DOT”, and those without observer assigned as “assigned-No DOT”. • The patients and/or their observers were asked whether the observers actually watched the patients swallow the medicine during the first 2 months (strict DOT). The less strict practices, such as just staying with the patient during drug intake without watching, preparing medicine for patients, and reminding patients about the drug intake were considered as “not strict DOT”. • Three cutoff points were used to define the actual practice of strict DOT: 1) every-dose vs some-dose or never, 2) more than 50% of the expected doses vs 50% or less, and 3) every-dose or some-dose vs never. We defined the former group in each cutoff point as “actual-DOT” and the latter group as “actual-No DOT”. All definitions were analysed, however we only present results according to the 1st cutoff point.

5. Methods 2 Outcome variables • We measured the effect of DOT using the following 2 endpoints; sputum negative conversion at the end of the 2nd month, and treatment success at the end of the treatment. The WHO definitions of treatment outcomes were applied to both endpoints. Potential confounders • Several potential confounders were considered and divided into 3 groups: 1) Demographic & socioeconomic status (gender, age, marital status, ethnic group, formal education, understanding Thai language, occupation, income, feasibility to be free from work or study, independence in travel, and number of living places), 2) Health services provided (type of TB clinic, drug regimen, and use of fixed dose combination), and 3) Disease condition (initial weight, initial AFB result, initial drug resistance, adverse drug effect, HIV/AIDS status, and associated diseases/conditions including diabetes mellitus, cardiovascular disease, cerebrovascular disease, liver cirrhosis, psychosis, alcoholic consumption, drug abuse, and imprisonment). • Four logistic regression models with increasing numbers of covariates were applied to determine the association between the exposure and the outcome, 1) without covariates, 2) with inclusion of the 1st group of covariates, 3) with inclusion of the 1st and 2nd groups of covariates, and 4) with inclusion of all 3 groups of covariates. For each step of adding the group of covariates, only covariates with the following criteria were retained in the model; 1) having significant association with the outcome (p <0.05) or 2) having marginal association with the outcome (p < 0.1) plus leading to a changeof more than 15% of OR for any DOT comparison in the larger model, if removed. Once a variable was included, it would be also included in the next model to ensure comparability of the log likelihood.

6. Results 1 • Of the 455 patients who started treatment, 44 were excluded because the interviewers were unable to establish contact with them or their DOT observers. Compared with the remaining patients, the excluded patients were younger (median age 31 vs 42 years), were more often HIV positive of suffered from AIDS (27% vs 11%), and had poorer treatment outcomes (sputum conversion rate 57% vs 78%; cure rate 30% vs 75%). • The remaining 411 patients were from 6 to 86 years of age (mean 44, 95% CI 42-46 years); and 75% were male. Of 323 patients who provided information on income, 76% earned less than the official “minimal daily wage” in the study area (3-3.5 US$). Among the 104 tested, 2% (95% CI 1%-3%) had multidrug resistance (resistance to at least isoniazid and rifampicin).

7. Results 2 Initial DOT assignment • Of 411 patients, 379 (92%) had been initially assigned to any type of observer and the remaining 32 patients were not assigned to any observer, because they refused to accept an observer (18), no suitable observers were available (9) or both (5). DOT was more often assigned to patients who were female, who had a living partner, who were treated at the zonal TB centre, who used fixed dose combination,orwho had initial drug resistance. Actual DOT practice • Overall, only 68 of 411 patients (17%) practised strict DOT in every-dose during the first 2 months, 97 of 393 patients with available information on the following DOT duration practised formore than 50% of expected doses (25%), whereas 154 of 411 (37%) patients took the medicine without being watched by observers.Of 379 patients assigned to DOT, 65 (17%) practised strict DOT in every-dose during the first 2 months, 93 of 362 patients with available information on the following DOT duration practised formore than 50% of expected doses (26%), whereas 133 of 379 patients (35%) never practised strict DOT. • Practice of strict DOT for every-dose during the first 2 months was found more often among the patients who hadlower income, who could be free from work/study, had any drug resistance, who were HIV positive or who had AIDS or other co-morbidity, who travelled with others, who did not use fixed-dose combination, or who were treated at the zonal TB centre.

8. Results 3 Treatment outcomes • Overall sputum conversion and treatment success rates were 78% and 85%, respectively. Sputum conversion rates were lower among male patients and those who were single, divorced or widowed. Treatment success rates were lower among patients who were male, Buddhist, never traveled alone, were treated at hospitals, did not use fixed-dose combination, with HIV/AIDS or with other co-morbidity. Multivariate analysis results • No significant differences in risk of no sputum conversion were seen between DOT and No DOT groups, regardless of DOT types or statistical models. Males had an approximately 2-fold higher odds of no sputum conversion compared with female patients. Risks of no sputum conversion were reduced to about half among the patients who had no living partner or who traveled independently. • Men had an approximately 4- to 5-fold increased risks of unsuccessful treatment than women. Risks of unsuccessful treatment were approximately 8-fold higher among those with HIV/AIDS, and approximately 2-fold among those with other co-morbidity. The risks were reduced to about 1/5 among the patients who traveled alone, and reduced to about 1/2 or 1/10 among those treated at a community hospital or the zonal TB centre, respectively, compared with a general/regional hospital. • No association was found between DOT, either assigned or actual, and chance of treatment success after adjustment for potential confounders.

9. Table 1 Treatment outcomes at the end of the 2nd month and at the end of the treatment. Results are given according to the assigned DOT at the start of treatment and the actual DOT practice during the first 2 months of the treatment.

10. Discussion 1 • The Centers for Disease Control and Prevention in the United States1 and WHO2have recommended DOT for all TB patients because it is difficult to predict whether a patient will follow the treatment. They have done so without requesting DOT to be evaluated in practice including measurement of the adherence to the DOT principle. Our findings show that actual DOT practice could be very different from the intended assignment. • The result also calls for caution when interpreting the results of the 4 randomised controlled trials analysed according to the “intention-to-treat” principle and without information on actual DOT practice The effect measures of these studies are likely biased by a difference between initial assignment and actual practice. 1U.S.Department of Health and Human Services (1994) Improving Patient Adherence to Tuberculosis Treatment Public Health Service, Centers for Disease Control and Prevention, National Center for Prevention Services, Division of Tuberculosis Elimination, Georgia. 2World Health Organization (1999) What is DOTS? A guide to Understanding the WHO-recommended TB Control Strategy Known as DOTS WHO/CDS/CPC/TB/99.270, WHO, Geneva. 3Zwarenstein M, Schoeman JH, Vundule C, Lombard CJ, & Tatley M (1998) Randomised controlled trial of self-supervised and directly observed treatment of tuberculosis. Lancet 352, 1340-1343. 4Kamolratanakul P, Sawert H, Lertmaharit S et al. (1999) Randomized controlled trial of directly observed treatment (DOT) for patients with pulmonary tuberculosis in Thailand. Transactions of the Royal Society of Tropical Medicine and Hygiene93, 552-557. 5Zwarenstein M, Schoeman JH, Vundule C, Lombard CJ, & Tatley M (2000) A randomised controlled trial of lay health workers as direct observer for treatment of tuberculosis. . The International Journal of Tuberculosis and Lung disease 4, 550-554. 6Walley JD, Khan AM, Newell JN, & Khan MH (2001) Effectiveness of the direct observation component of DOTS for tuberculosis: a randomised controlled trial in Pakistan. Lancet 357, 664-669.

11. Discussion 2 • The failure to find the significant effects of DOT supports the experience-based conclusion that “DOT is not panacea” and the clear WHO specification of DOT as a part of good case management to support TB patients to achieve cure.As demonstrated in this study, some risk factors e.g. gender and independence in travel influenced both short and long term outcomes; whereas TB services, HIV/AIDS and co-morbidity were strongly associated with the final outcome. TB services, particularly in large hospital, should optimise their services and work with existing partnerships, particularly those concerned with HIV/AIDS, in the community to provide tailored support for the patients. • Randomised trials willnot solve the problem of compliance to the DOT principle. Practical problems and the patients’ rights to choose the treatment modality they prefer will probably limit the scientific value of a randomised trial that aims at quantifying the effect of DOT. During the course of treatment, the patients may gain insight into their own ability to manage the treatment and may move out of the DOT group that would bias the comparison in disfavour of the DOT group. One should therefore not disregard the DOT principle on the basis of our findings but take the results as a reason for caution. Any DOT servicesshould try to identify the patients who actually will benefit from the DOT principleand the patients who can manage the drug intake on their own.

12. Conclusion & Recommendations • We did not find a statistically significant effect of DOT on sputum conversion or on treatment success; regardless of whether DOT was analysed according to the initial assignment or according to what was reported by the patients and/or their observers. • TB control programme should not over-invest on DOT without strengthening otherstrategies of good patient management.