| The incidence of tuberculosis (TB) disease in the United States remains five times higher than the Healthy People 2010 target, with treatment completion rates 10% below the national objective. Incomplete therapy results in drug resistance and continued transmission.; National guidelines recommend directly observed therapy (DOT) as the standard of care to ensure treatment completion. There is insufficient evidence to support its effectiveness as a universal intervention. The majority of studies implicated DOT as the major variable responsible for differences in completion rates, with limited attention given to competing explanatory variables, which ultimately may be responsible for treatment outcomes independent of DOT.; This study was a retrospective cohort analysis of 591 TB cases who started treatment in Massachusetts (excluding Boston) from 1999 through 2001, to identify factors associated with treatment completion, using data extracted from national surveillance and Massachusetts' patient management system databases.; Using a multifactorial study design, individual patient attributes, clinical and treatment factors, and care system characteristics that could potentially influence treatment completion were examined for their relationship to completion and covariance with the treatment approach. Time to treatment completion was the dependent measure, using survival analysis (Cox Proportional Hazard regression) methods, stratified by expected length of therapy.; Results demonstrated there were no overall differences in treatment completion among patients managed by self administered therapy, DOT, or a combination of both. Demographic variables were unreliable predictors of treatment completion. Clinical variables associated with longer therapy included abnormal chest radiographs, drug resistance, especially multi-drug resistant TB, and adverse reactions. Outreach, incentives, nursing consultation regarding inappropriate treatment, long-term care TB hospitalization, and treatment approach had no independent effect on completion. Findings suggest good outcomes can be achieved with different treatment approaches. Recommendations given with clinical and social policy implications included the need for patient-centered care with selective DOT based on an assessment of individual attributes, rather than universal DOT; the need to maintain the public health infrastructure; and additional variables to collect in surveillance systems. More research is needed to document the effectiveness of interventions, as well as studies that include more contextual variables, particularly those related to cultural issues. |
