Background
Adherence in tuberculosis management
Effective drugs for tuberculosis have been available since the 1940s, but two million people continue to die each year, mostly in low-income countries (Dye 1999; Netto 1999). People with tuberculosis require treatment for at least six to eight months. Many find it difficult to complete their course of treatment and this serves as a major constraint to eradicating the disease (Fox 1958; Addington 1979; Cuneo 1989). Poor adherence to treatment can lead to prolonged infectiousness, drug resistance, relapse of tuberculosis, or even death. Incomplete treatment thus poses a serious risk for the individual as well as the community.
There are three groups of people for whom adherence is important: people under evaluation for suspected tuberculosis (to ensure they complete the diagnostic regimen or start treatment); people receiving prophylaxis (preventive therapy), where antituberculous drugs are given to people exposed to tuberculosis or thought to be at particular risk; and people with diagnosed tuberculosis in whom completion of treatment helps ensure cure.
Adherence to a tuberculosis treatment programme requires accessible and appropriate health care. People need to be diagnosed correctly, provided with information about their disease and the need for completion of treatment, and supplied with appropriate outpatient drugs. But even where these services are available, people may not adhere to the intended regimen. Healthcare providers have responded by developing a variety of specific measures to improve adherence (Cuneo 1989; CDC 1993; Sbarbaro 1994). These interventions are aimed at influencing the behaviour of healthcare personnel, the organization of the service, or the behaviour of the person with suspected or confirmed tuberculosis. Originally we included all these interventions in one Cochrane Review, but this approach did not allow us to consider the particulars of each of the following interventions, and why in one set of circumstances it may be effective and another it may not. This Cochrane Review is one of several planned or in progress to evaluate each type of intervention:
- Directly observed therapy (DOT): an appointed agent (health worker, community volunteer, family member) directly monitors people swallowing their antituberculous drugs (this review).
- Staff motivation and supervision: training and management processes that aim to improve how providers care for people with tuberculosis.
- Reminder systems and late patient tracers in the diagnosis and management of tuberculosis: routinely reminding patients to keep an appointment and actions taken when patients fail to keep an appointment (Liu 2007).
- Education and counselling for promoting adherence to the treatment of active tuberculosis: provision of information or one-to-one or group counseling about tuberculosis and the need to attend for treatment (M'Imunya 2007).
- Incentives and reimbursements: money or cash in kind to reimburse expenses of attending services, or to improve the attractiveness of visiting the service.
- Contracts: written or verbal agreements to return for an appointment or course of treatment (Bosch-Capblanch 2007).
- Peer assistance: people from the same social group helping someone with tuberculosis return to the health service by prompting or accompanying them.
DOT
DOT seeks to improve the adherence of people to tuberculosis treatment through health workers, family members, or community members directly observing them taking their antituberculous drugs. This approach was first adopted in studies in Madras, India, and Hong Kong as early as the 1960s (Bayer 1995), and a number of specialists now widely recommend DOT for the control of tuberculosis (Bass 1994; Maher 1997; Chaulk 1998; Enarson 2000). Indeed, Frieden and Sbarbaro state that it is essential and that it prevents relapse occurring and drug resistance developing (Frieden 2007).
The advantages of DOT are that people can be closely monitored and that there is a social process with peer pressure that may improve adherence. On the other hand, the disadvantages associated with DOT are that it moves away from adherence models of communication with cooperation between patient and provider back to a traditional medical approach with the patient as the passive recipient of advice and treatment (Donovan 1992; Sumartojo 1993); resource implications for such a policy are substantial, particularly in low-income and middle-income countries where the case load is high; and it may make adherence worse if it is rigidly applied in an authoritarian setting or where people are expected to travel considerable distances to have their treatment supervised
Programmes that include DOT
Over the years DOT has come to mean much more than the supervised swallowing of drugs, causing considerable confusion. In the USA, DOT programmes are complex and consist of several components including social support, housing, food tokens, and jail for recalcitrant people (ATS 1992; Volmink 2000b).
The World Health Organization (WHO) promotes another version of DOT called 'directly observed therapy, short course' (DOTS). This is a comprehensive tuberculosis management programme that focuses on low-income countries. DOTS is a five-element strategy for the control of tuberculosis that consists of political commitment, improved laboratory analysis, direct patient observation while swallowing each dose of medication, a drug supply that provides for the correct complete short course antituberculous drug combination for free, and a reporting system that documents the progress in curing the patient (WHO 1997).
The WHO believes that direct observation is a key element for the success of DOTS and has retained it in more recent definitions of the DOTS strategy, although their documentation has clearly taken into account criticisms of their blanket policy. For example, the WHO states that the third component of their expanded strategic framework is "standardized short-course chemotherapy to all cases of TB [tuberculosis] under proper case-management conditions including direct observation of treatment" (WHO 2002). The 2004 progress report mentions direct observation for "at least" the first two months of treatment (WHO 2004). Even so, how this is interpreted in countries varies, and direct observation by a health worker remains national policy in China, for example.
In contrast, we have previously suggested that the benefits associated with DOT programmes found in observational studies may be attributable to simultaneous inputs rather than direct observation specifically (Volmink 1997a). An informed debate is important because direct observation has considerable resource implications; it therefore important to determine how important this intervention is for improving adherence to tuberculosis treatment and ensuring cure. Since our initial Cochrane Review (Volmink 1997b), which documented the absence of randomized controlled trials of DOT, several trials have been conducted aimed at disaggregating the effects of this specific intervention from those of accompanying inputs, and they are included in this review update.
Not only is there a debate about whether DOT is effective, there is also a debate about who should provide this. This may be contingent practically on an individual's circumstances, but the specialists are divided in their recommendations: some consider family members can help, whereas others regard family observation as a "seductive but risky concept" (Frieden 2007). We therefore also summarize trials comparing DOT through different providers and settings.
Objectives
To compare DOT with self administration of treatment or different DOT options for people requiring treatment for clinically active tuberculosis or prevention of active disease.
Methods
Criteria for considering studies for this review
Types of studies
Randomized and quasi-randomized controlled trials.
Types of participants
People requiring treatment for clinically active tuberculosis or medication for preventing active disease (prophylaxis or preventive therapy).
Types of interventions
Intervention
Health worker, family member, or community volunteer routinely observes participants taking their antituberculous drugs.
Control
Routine self administration of treatment at home, with intermittent clinic visits for drugs with or without treatment adherence checks.
Where researchers explored different methods of implementing direct observation, the experimental method was allocated to the intervention group and the standard method to the control.
Types of outcome measures
Primary
- Cure.
- Completion of treatment.
- Development of clinical tuberculosis (in trials of drug prophylaxis).
Secondary
Keeping outpatient appointments.
Search methods for identification of studies
We attempted to identify all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).
Databases
We searched the following databases using the search terms and strategy described in Appendix 1: Cochrane Infectious Diseases Group Specialized Register (May 2007); Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library (2007, Issue 2); MEDLINE (1966 to May 2007); EMBASE (1974 to May 2007); and LILACS (1982 to May 2007). We also searched the metaRegister of Controlled Trials (mRCT) using 'tuberculosis AND DOT*' (May 2007).
Researchers and organizations
For unpublished and ongoing trials, we contacted individual researchers working in the field and the following organizations: World Health Organization (1997 and 2004), the International Union Against Tuberculosis and Lung Disease (IUATLD) (1997), and the Centers for Disease Control and Prevention (1997).
Reference lists
We also checked the reference lists of all studies identified by the above methods.
Data collection and analysis
Selection of studies
Both authors independently applied the inclusion criteria to all identified trials. We used the titles and abstracts of the identified citations to exclude trials that clearly did not meet the inclusion criteria. If either author judged that the trial might be eligible for inclusion, we obtained the full paper. We independently screened the full articles of selected trials to confirm eligibility and resolved any disagreements by discussion.
Data extraction and management
We independently extracted the data and checked whether authors had conducted an intention-to-treat analysis. Trialists were contacted to supply missing information and to clarify issues. We resolved discrepancies through discussion.
Assessment of risk of bias in included studies
We independently evaluated the methodological quality of each trial, classifying the generation of allocation sequence and concealment of allocation as adequate, inadequate, or unclear according to Jüni 2001. We classified blinding as adequate if steps were taken to ensure the people recording the main outcome of the study were blind to the assigned interventions and inadequate if this was not the case or if there was no mention of attempts to blind the observers. We assessed completeness of follow up as adequate if 90% or more, inadequate if less than 90%, or unclear if not mentioned.
Data synthesis
We used Review Manager 5 to analyse the data, using relative risk (RR) with 95% confidence intervals (CI) to assess estimates of effect. We used the fixed-effect model when there was no statistically significant heterogeneity (chi square P > 0.1).
Results
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies.
Eleven trials with 5609 participants met the inclusion criteria (see 'Characteristics of included studies'). Twelve studies that initially seemed to fit the inclusion criteria were eventually excluded from our review; reasons for these decisions are provided in the 'Characteristics of excluded studies'.
Eight of the 11 trials were conducted in low-income and middle-income countries and evaluated DOT for treating people with active tuberculosis: Pakistan (Walley 2001); South Africa (Zwarenstein 1998; Zwarenstein 2000); Tanzania (Lwilla 2003; Wandwalo 2004); Nepal (Newell 2006); Thailand (Kamolratanakul 1999); and Swaziland (Wright 2004). Three trials were in high-income countries: one in Australia (MacIntyre 2003); and two in the USA that examined prophylaxis in intravenous drug users (Chaisson 2001; Malotte 2001). Two trials were cluster randomized (Lwilla 2003; Newell 2006). One trial used a quasi-random method of allocation (MacIntyre 2003).
Services for general populations
DOT versus self administration of treatment
Five trials compared DOT with self administration of treatment. There is some overlap of data in two of these trial reports, but we have taken this into account in the analyses: Zwarenstein 1998 combined data from two communities in which trials compared direct observation by nurses at clinics with self administration of treatment, while Zwarenstein 2000 described data from one of these trials that had three arms of direct observation by nurses, lay health workers, and self administration. Kamolratanakul 1999 allowed participants to choose between DOT by a health worker, community leader, or family member; most chose the latter. Walley 2001 compared DOT by a health worker or community health worker with DOT by a family member and with self administration of treatment. MacIntyre 2003 evaluated DOT by a family member.
Alternative DOT delivery options
A cluster-randomized trial from Tanzania (Lwilla 2003) compared a community health worker observing people at home with DOT at a health facility. Three trials compared DOT by a family member with either DOT by a health worker at a health facility (Wandwalo 2004) or DOT by a community health worker (Wright 2004; Newell 2006).
Services for intravenous drug users
Two trials from the USA evaluated DOT in drug users on prophylaxis for latent tuberculosis (Chaisson 2001; Malotte 2001). Malotte 2001 studied intravenous and crack cocaine users, and compared DOT by an outreach worker at a site chosen by the participant (with or without a US$5 at each visit) with DOT at a community clinic with a US$5 at each visit. Chaisson 2001 involved intravenous drug users, and studied DOT by an outreach nurse with self administration either with monthly peer support or monthly clinic visits.
Outcomes
The numbers of people cured, cured and completed treatment, or completed treatment were the main outcomes assessed in the included trials.
Adjustment for clustering
Both cluster trials adjusted for clustering appropriately: standard error of the coefficients for clustering on units corrected using the Huber-White-Sandwich method (Lwilla 2003); and, in Newell 2006, using the coefficient of variation between clusters.
Risk of bias in included studies
See Table 1 for a summary of the assessment.
Generation of allocation sequence
Seven trials used adequate methods − computer-generated random sequences (Zwarenstein 1998; Zwarenstein 2000; Walley 2001; Chaisson 2001), a random-number table (Kamolratanakul 1999), coin tossing (Wandwalo 2004), and random draws of paper from a basket (Newell 2006). One trial used alternate allocation, an inadequate method (MacIntyre 2003). The remaining trial reports did not provide information (Malotte 2001; Lwilla 2003; Wright 2004).
Allocation concealment
Four trials employed adequate methods for concealing allocation (Zwarenstein 1998; Zwarenstein 2000; Walley 2001; Malotte 2001). Allocation concealment was unclear in three trials (Chaisson 2001; Lwilla 2003; Wright 2004) and not used in the remaining trials (MacIntyre 2003; Kamolratanakul 1999; Wandwalo 2004; Newell 2006).
Blinding
Outcome assessment was blind in only four trials (Walley 2001; MacIntyre 2003; Wright 2004; Newell 2006).
Completeness of follow up
In two trials, more than 10% of participants were excluded from the analysis (Chaisson 2001; Lwilla 2003). A further four trials did not provide sufficient information to assess this aspect of study quality (Zwarenstein 2000; Malotte 2001; Walley 2001; MacIntyre 2003). In the rest of the trials follow up was adequate.
Effects of interventions
1. Services for general public
DOT versus self administration of treatment
Treatment outcomes were similar among participants in the DOT and self administration of treatment arms. There was no statistically significant difference between the interventions for the number of people cured (1603 participants, 4 trials, Analysis 1.1), cured or completed treatment (1603 participants, 4 trials, Analysis 1.2), or completed treatment (173 participants, 1 trial, Analysis 1.3). There was significant heterogeneity between the trials for cure (chi squared 8.41; df = 3), but the point estimate of the RR was close to one, and no difference was demonstrated with either fixed-effect or random-effects models ( Analysis 1.1).
We explored whether different effect sizes were associated with home-based and clinic-based DOT. Effect size was similar in the two groups ( Analysis 2.1 and Analysis 2.2 ). Home-based DOT was strongly influenced by the trial from Thailand (Kamolratanakul 1999), and there was a statistically significant improvement in cure, but the difference was small (RR 1.10, 95% CI 1.02 to 1.18; 1365 participants, 3 trials, Analysis 2.1).
DOT: family member or community health worker versus at clinic
Wandwalo 2004 found cure or treatment completion to be similar for those observed by a family member compared with a health worker (587 participants, 1 trial, Analysis 3.1). A cluster-randomized trial, Lwilla 2003, which evaluated DOT by a community health worker, found no difference for sputum conversion at two months (OR 0.62, 95% CI 0.23 to 1.71) or cure at the end of treatment (OR 1.58, 95% CI 0.32 to 7.88; trial authors adjusted for design effects).
DOT: family member versus by a community health worker
Wright 2004 found no statistically significant difference in the number of people cured or who completed treatment (1326 participants, Analysis 4.1. A cluster-randomized trial, Newell 2006, which compared community-based DOT by a community health worker or village health worker with family-based DOT, found no statistically significant difference in success rates (cure and treatment completion) (85% vs 89%; OR 0.67, 95% CI 0.41 to 1.10; trial authors adjusted for design effects).
2. Services for intravenous drug users
Chaisson 2001 found no statistically difference in the number of people who completed twice-weekly clinic-based DOT and those on daily self administration of treatment (199 participants, Analysis 5.1). For participants receiving prophylaxis, Malotte 2001 studied people receiving prophylaxis and found no statistically significant difference in the number who completed treatment between those allowed to choose their DOT location and those receiving DOT at a community clinic (108 participants, Analysis 6.1).
Discussion
Direct observation of people taking their antituberculous drugs is widely advocated and forms part of the World Health Organization's 'directly observed therapy, short course' (DOTS) strategy. While this strategy includes a number of useful components, the available evidence does not provide strong support for the routine adoption of direct observation in favour of self administration of treatment either for people with active tuberculosis or those with latent tuberculosis requiring prophylaxis. There is also no evidence that one form of direct observation is better than another: direct, randomized comparisons between clinic-based DOT and community-based DOT did not demonstrate a difference; and, within community-based DOT, comparisons between DOT provided by a family member versus a community health worker had similar outcomes.
Given the prevailing support for DOT-based programmes, these findings are important. We have previously suggested that the benefits associated with DOT programmes in observational studies may be attributable to simultaneous interventions rather than direct observation being the key adherence-promoting strategy (Volmink 2000b). A qualitative study notes that the implementation of DOT is in the process of shifting from being a rigid model involving observation of drug swallowing to one that includes an array of incentives and enablers for supporting the patient (Macq 2003). Within such a package of patient-centred interventions it remains to be established whether direct observation is necessary at all. Of interest in this regard are the findings of a cluster-randomized trial in a rural South Africa in which motivation and support from a lay health worker (with or without DOT) was shown to be more effective in ensuring treatment than a conventional DOT-based service (Clarke 2005). People with tuberculosis are often poor and encounter numerous barriers to treatment adherence. Strategies aimed at reducing social and health system barriers may therefore be preferable to coercive approaches that impact negatively on patient autonomy. The encouraging results of a recent trial in Senegal using a multifaceted approach to address these challenges, which also included a flexible approach to DOT, lend support to this notion (Thiam 2007). Further rigorous trials, in particular those testing interventions social and family barriers to adherence, are needed (Garner 2007).
One of the trials included in our review found a modest improvement with DOT for cure and treatment completion (Kamolratanakul 1999), and one may speculate about the reasons for these findings. The Thai health system is relatively well resourced and has a tuberculosis control programme that is well organized. This is in contrast to the trials in South Africa (Zwarenstein 1998; Zwarenstein 2000) and Pakistan (Walley 2001) where the trials were conducted in areas with high disease burden, overcrowded tuberculosis clinics, and poorly motivated staff. On the other hand, DOT did not improve treatment completion rates in either Australia (MacIntyre 2003) or the USA (Chaisson 2001; Malotte 2001), both of which are low burden, highly resourced settings. A second reason one might posit is that cultural responses to supervision may vary with Thai people being more responsive to this approach. Finally, Kamolratanakul 1999 is so far the only trial to have allowed participants to choose their supervisor, and this patient-centred approach may also have influenced the effects.
Can adopting a DOT policy make adherence worse? The authors of the South African trial suggest an increasingly negative and demoralizing effect of direct observation on participants with tuberculosis (Zwarenstein 1998). This trial found that in participants with a first episode of tuberculosis, the outcomes were equivalent in DOT and self administration of treatment arms, while 'retreatment' participants who were assigned to DOT fared worse than those who self administered treatment. Given the small numbers of participants in the retreatment group, further research is warranted to confirm the findings.
Bias could have influenced the results in some of the trials. As outlined in the methodological quality assessment, the number of trials with adequate quality criteria was limited: four with adequate method for concealing treatment allocation; four with blinding of the outcome assessors; and two trials that excluded more than 10% of participants from the analysis. Nevertheless, these trials are not easy to implement and are a credit to the researchers who have worked hard to develop the evidence base for rational decision making.
Authors' conclusions
Implications for practice Randomized controlled trials provide no assurance that the routine use of DOT in low- and middle-income countries improves cure or treatment completion in people with tuberculosis. There is also no rigorous evidence to support the use of DOT for prophylaxis in people with latent tuberculosis. There appears to be no sound reason to advocate the allocation of resources to the routine use of DOT until we better understand the situations in which it may be beneficial. In the meantime, it could reasonably be argued that resources should be invested in interventions that have been shown to be effective for improving adherence, such as providing patient motivation and support, incentives, and defaulter action. |
Implications for research The relation between DOT and treatment outcome is complex. Factors that determine its usefulness in various settings require further study. The extent to which the quality of interaction between patients and their observers influences outcome would be a particularly fruitful topic for future research. It will also be worth testing whether DOT efficacy differs in people receiving tuberculosis treatment for the first time compared with those requiring retreatment, and in men compared with women. Comparisons of DOT in relation to other strategies aimed at improving adherence should be determined. |
Acknowledgements
To the Global Health Council and the Department of Arts, Culture, Science and Technology of South Africa for supporting Jimmy Volmink in the preparation of the review in the period before the 2006 update. This document is an output from a project funded by the UK Department for International Development (DFID) for the benefit of developing countries. The views expressed are not necessarily those of DFID.
Data and analyses
- Top of page
- Background
- Objectives
- Methods
- Results
- Discussion
- Authors' conclusions
- Acknowledgements
- Data and analyses
- Appendices
- What's new
- History
- Contributions of authors
- Declarations of interest
- Sources of support
- Notes
- Index terms
| |||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||
| |||||||||||||||
| |||||||||||||||
| |||||||||||||||
Appendices
Appendix 1. Search methods: detailed search strategies
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
What's new
Last assessed as up-to-date: 13 August 2007.
| |||||||||||||||
History
Protocol first published: Issue 4, 2001
Review first published: Issue 4, 2001
| |||||||||||||||||||||||||||||||||
Contributions of authors
Jimmy Volmink initiated the review, wrote the protocol, applied the inclusion criteria, conducted the data extraction, and drafted the review. Paul Garner helped write the protocol, applied the inclusion criteria, constructed comparisons, and helped draft the review. Both authors remain actively involved in updating.
Declarations of interest
None known.
Sources of support
Internal sources
- South African Medical Research Council, South Africa.
- Liverpool School of Tropical Medicine, UK.
External sources
- Department for International Development, UK.
Notes
2011, Issue 9:
The Cochrane Infectious Diseases Group are piloting a system to indicate whether the question is currently relevant, and the status of the review with regards to being up to date.
For relevance, we classify reviews into:
- historical question, where an intervention or policy has been superseded by new medical developments (such as a new drug),
- current question, which are still relevant to current policy or practice.
For status, we have three categories, with an explanation after each: “up to date”; “update pending”; “no update intended”.
For this review, we have categorised the review as: Current question - no update intended (Results conclusive).
Index terms
Medical Subject Headings (MeSH)
*Directly Observed Therapy; Antitubercular Agents [*therapeutic use]; Randomized Controlled Trials as Topic; Tuberculosis, Pulmonary [*drug therapy]
MeSH check words
Humans
* Indicates the major publication for the study