Directly Observed Therapy of TB Does Not Solve the Problem of Noncompliance


Synopsis: Noncompliance was oberved in almost one-fifth of patients receiving DOT for TB, with homelessness and alcoholism being the major risk factors.

Source: Burman WJ, et al. Noncompliance with directly observed therapy for tuberculosis. Epidemiology and effect on the outcome of treatment. Chest 1997;111:1168-1173.

Burman and colleagues reviewed their experience with directly observed therapy (DOT) of tuberculosis for the five counties in the Denver metropolitan area. Most patients in this area receive DOT, and those with fully susceptible isolates of Mycobacterium tuberculosis receive two weeks of daily therapy followed by twice weekly therapy for another 24 weeks. Patients are encouraged to receive their doses at the clinic, which has extended hours and, if needed, provides transportation. Doses of DOT are administered by outreach workers in instances when patients cannot reliably return to the clinic for twice-weekly visits. Staff members are available who speak Spanish and Vietnamese; interpreters are also available for other languages. Patients are each assigned a nurse clinician as a primary provider. Staff work closely with substance-abuse treatment programs and with social service agencies.

Of the 428 patients with tuberculosis from 1984 to 1994, 333 (78%) received DOT, 294 were suitable for analysis, and 52 (18%) of those were significantly noncompliant. Thirty-three of those were noncompliant by virtue of having missed at least two consecutive weeks of DOT (26 were absent for therapy for > 2 months), another 16 missed sufficient doses so that treatment was prolonged by 30 days or more, and three patients were incarcerated "for meeting an immediate threat to public safety."

Multivariate analysis indicated that the only independent risk factors for noncompliance were homelessness and alcohol abuse. A poor outcome of therapy (failure of sputum sterilization by four months, clinical failure, death due to TB, or relapse) occurred in 17 (32.7%) of noncompliant and in only eight (3.3%) of compliant patients.


The most common reason for failure of antituberculous therapy is patient noncompliance. (It should be noted that at least one of the authors of the paper reviewed here has argued that the use of the term compliance has a paternalistic cast, and that we should instead be speaking of adherence or non-adherence.) The most effective means of preventing noncompliance is the use of DOT and, as a consequence, the Centers for Disease Control and Prevention has recommended that all patients with active tuberculosis receive DOT (Centers for Disease Control and Prevention. MMWR Morb Mortal Wkly Rep 1990;39:369-372). This study demonstrates, however, that even one of the most comprehensive DOT programs in the nation does not assure 100% compliance.

Noncompliance in the setting of patient self-administration of a variety of drugs has been widely studied. Among the things associated with poor compliance in that setting are resolution of symptoms, prolonged duration of therapy, complexity of therapeutic regimens, and drug side effects. Compliance is ordinarily measured by patient report and/or pill count. Testing of blood or urine for the presence of drugs or their metabolites, either directly or indirectly, is also used. Some clinics take advantage of the fact that pyrazinamide increases serum uric acid concentration and that rifampin turns urine a reddish-brown color. The MEMS device has a microchip implanted into the medication dispenser which records every time the bottle, for example, was opened. Unfortunately, none of these measures is foolproof.

This study demonstrated that two factors were significantly associated with noncompliance with DOT and, by extrapolation, a poor outcome of therapy—homelessness and alcoholism. This, once again, illustrates the manner in which tuberculosis in urban society is inextricably linked to the social fabric (and its holes) in regard to its occurrence and control and also its treatment.

Examination of issues of compliance has potential value that extends beyond tuberculosis to other diseases, including HIV infection. While great concern is expressed about multidrug-resistant TB (MDR-TB), a problem with similar or even greater import is in the early stages of development—multidrug-resistant HIV infection (MDR-HIV). The major determinant of the development of resistance to HIV-1 protease inhibitors is poor patient adherence to what are very complex and difficult treatment regimens. We can be certain that the stock of resistant virus is increasing on a daily basis and that that virus is being transmitted on a daily basis. Knowledge of the means to improve patient compliance with antiretroviral therapy is as critical to the public health as is compliance with antituberculous therapy. Unfortunately, currently available antiretroviral therapy does not allow for intermittent directly observed administration.