Justification and evidence
The recommendations in this section address three PICO questions (see
Annex 1).
PICO question 10 (DS-TB, 2017). In patients with TB, are any interventions to promote adherence to TB treatment more or less likely to lead to the outcomes listed below?38
PICO question 7 (DR-TB, 2011). Among patients with MDR-TB, is ambulatory therapy, compared with inpatient treatment, more or less likely to lead to the outcomes listed below?39
PICO question 11 (DS-TB, 2017). Is decentralized treatment and care for MDR-TB patients more or less likely to lead to the outcomes listed below?40
Treatment supervision. Currently, WHO defines DOT as any person observing the patient taking medications in real time. The treatment observer does not need to be a health-care worker, but could be a friend, a relative or a lay person who works as a treatment supervisor or supporter. Observed treatment may also be achieved with real-time video observation and video recording. However, in this document, DOT refers to treatment administered under direct observation by another person. Adherence definitions varied across the studies. However, in general, adherence was defined as taking >90% of medications under conditions of direct observation by another person.
The systematic review conducted in support of this guideline was based on synthesis of data from RCTs (89–96) and from observational studies (97–110), with preference given to the results of RCTs. Outcomes of DOT and SAT given under standard TB practice and without any additional support were compared. DOT could be administered by a health-care worker, a family member or a community member and either at home, in the patient’s community or at a clinic. DOT was generally administered daily. The GDG focused preferentially on RCT data from the systematic review. When the data from RCTs were limited or not available, observational study data were examined and their results presented. Interpretation of the associations, however, needs caution due to limitations of the observational data when the associations are confounded by different factors. In uncontrolled observational studies, for instance, patients with more severe disease or higher risk of non-adherence are likely to be assigned DOT and patients who are less sick or less likely to be non-compliant are assigned SAT. The same may apply to the selection of DOT location, DOT provider or other interventions in cohort studies.
When DOT alone was compared with SAT, patients who were on DOT had better rates of treatment success, adherence and 2-month sputum conversion; and also had slightly lower rates of loss to follow up and acquired drug resistance. However, patients on DOT had a slightly higher relapse rate. The GDG considered that, overall, the evidence was inconsistent in showing a clear advantage of DOT alone over SAT or vice versa. However, the evidence showed that some subgroups of patients (e.g. TB patients living with HIV) with factors affecting treatment adherence are likely to benefit more from DOT than other patients; or specific types of DOT delivery (e.g. locations of DOT or DOT providers) are likely to work better than others. The evidence also showed that when patients received treatment adherence interventions (e.g. different combinations of patient education, staff education, material support, psychological support, tracer and use of medication monitor) in conjunction with DOT or SAT, treatment outcomes were significantly improved compared to DOT or SAT alone (see below). Only cohort studies were available to examine DOT and SAT in HIV-positive TB patients (111–127), and many of these studies were conducted in the pre-ART era or shortly after the introduction of early ART for HIV-positive TB patients (123–126). As above, DOT could have been administered by a variety of people in a variety of settings, including homes and clinics, and occasionally, during the initial intensive phase of treatment, it was hospital-based. A few studies provided incentives and enablers or provided DOT only for persons considered to be at higher risk of loss to follow up. HIV-positive TB patients on SAT had lower rates of treatment success, treatment completion and cure. They had higher rates of mortality, treatment failure and loss to follow up. The evidence showed that HIV-positive TB patients, as a subgroup, benefit more from DOT than general TB patients do, and that SAT alone is not advisable in HIV-positive TB patients. Reasons such as increased rates of drug–drug interactions and more severe disease in this cohort may cause DOT to offer a significant advantage over SAT. DOT and SAT in MDR-TB patients were also examined in the systematic review. However, very limited data were available from a cohort study (114). There were higher rates of mortality and non-adherence and lower rates of treatment completion in MDR-TB patients on SAT compared with those on DOT, although the differences were not significant.
DOT provider. RCTs (91,93–95) and observational studies (98,99,102,104,109,112,117,119,120,122,123,127) were available for examination of the effect of DOT providers versus SAT. Providers were grouped as health-care workers, lay providers and family members. The health-care worker group was varied and included personnel working at different levels of health-care systems and who had received health training. Health-care workers could be nurses, physicians or trained community health workers. Lay providers were also varied and could include teachers, community volunteers or traditional healers. DOT by lay providers had higher rates of treatment success and cure, and a slightly lower rate of loss to follow up compared with SAT. Patients receiving DOT from a family member had higher rates of treatment success and lower rates of loss to follow up compared with patients using SAT. When DOT provided by a health-care worker was compared to SAT, there were higher rates of cure and adherence, and lower rates of relapse and acquisition of drug resistance with health-care worker DOT. The effect that different types of DOT provider had on outcomes was also examined. DOT provided by health-care workers and DOT provided by lay persons were compared. Only observational studies were available in the literature (99,102,119,128–132). Slightly higher rates of success, and lower rates of mortality, failure and loss to follow up were observed among patients who had DOT administered by a lay provider versus a health-care worker, although the difference was not statistically significant.
When provision of DOT by a family member was compared to provision of DOT by a health-care worker, there were higher rates of mortality, loss to follow up and failure, and lower rates of successful treatment, cure and treatment adherence among patients who had DOT administered by family members. Therefore, although DOT by a health-care worker, trained lay provider and family member showed advantages compared to SAT, provision by trained lay providers and health-care workers are the preferred options for DOT and a family member is the least preferred DOT provider.
DOT location. RCTs and observational studies examined how DOT location affected treatment outcome. Locations were grouped by community- or home-based DOT, and health facility-based DOT (91,93,95,97,104,109,117,119,122,123,133–170). Community- or home-based DOT was defined as DOT delivered in the community close to the patient’s home or workplace. In general, community- or home-based DOT was provided close to the patients. Health facility-based DOT was defined as DOT delivered at a health centre, clinic or hospital. There were some instances of community- or home-based DOT being provided by health-care workers. When comparing DOT locations, community- or home-based DOT had higher rates of treatment success, cure, treatment completion and 2-month sputum conversion. Community- or home-based DOT also had lower rates of mortality and lower rates of unfavourable outcomes compared with health facility-based DOT. When comparing community-/home-based DOT or health facility-based DOT with SAT, there were no significant differences across the outcomes in RCTs. However, cohort studies showed higher rates of treatment success and adherence, and a lower rate of loss to follow up with community-/home-based DOT compared with SAT. Observational data from cohort studies also showed lower rates of treatment completion, and slightly higher rates of failure and loss to follow up in health-facility DOT compared to SAT. Therefore, community- or home-based DOT is the preferred option rather than health facility-based DOT and SAT. Combining the evidence on DOT provider and DOT location, DOT should preferably be delivered at home or in the community and by a health-care worker or trained lay provider. DOT delivered at a health facility, DOT provided by a family member and unsupervised treatment are not preferred options.
Video-observed treatment (VOT). For VOT there were only two cohort studies from high-income countries and no data from low- and middle-income countries (171,172). These studies compared in-person DOT with VOT done in real time. Patients who were provided with VOT had no statistically significant difference in treatment completion and mortality compared to patients who had in-person DOT. Although there is some concern as to the indirectness of evidence for VOT, as the studies were conducted in high-income countries and the uncertainty of evidence surrounding the use of VOT, the results from the two cohort studies showed that in-person DOT was not better than VOT. DOT has been the standard of care that many programmes aim for, even if in practice they have to resort to SAT in many patients because of lack of resources. The advantages of using VOT are its potential to observe adherence to treatment from a distance – and even when people travel and cannot visit or be visited by a DOT provider. VOT is also more flexible to people’s schedules by offering virtual observation at different times of the day. VOT could help achieve better levels of patient interaction at a much lower cost and less inconvenience when compared with in-person DOT. VOT can be used as an addition to, or interchangeable with, in-person DOT or other treatment administration options. For instance, it is not expected that a patient receives VOT as the sole option of supervision during the whole duration of treatment. Furthermore, the technology required for VOT (broadband Internet and smartphone availability) is becoming increasingly available in resource-constrained settings. Moreover, VOT delivery options are evolving (e.g. enhanced possibility for real-time communication in addition to recorded video), and therefore evidence and best practices are likely to develop further in the coming years, especially from ongoing RCTs. The benefits of VOT may become more apparent as programmes are able to choose forms of VOT that best meet their needs. In fact, VOT may be particularly useful for easing the burden on the health-care system in low- and middle-income countries.
Package of combined treatment adherence interventions. Both RCTs and observational studies examining the effects of combined treatment adherence interventions were reviewed (137–143,172–178). When patients receiving combined treatment adherence interventions along with DOT or SAT were compared to those receiving DOT or SAT alone, patients who received combined treatment adherence interventions had higher rates of treatment success, treatment completion, cure and adherence, and lower rates of mortality and loss to follow up. The mixture of types of adherence interventions was varied (). These included different combinations of patient education, staff education, material support (e.g. food, financial incentives, transport fees, bonuses for reaching treatment goals), psychological support and counselling. The treatment adherence interventions also included tracers such as home visits, use of digital health communication (e.g. SMS, telephone calls) or a medication monitor. The interventions should be selected on the basis of assessment of the individual patient’s needs, providers’ resources and conditions for implementation.
Treatment adherence interventions.
Tracers and digital health interventions rather than VOT. Varied tracers were included in RCTs and observational studies (171,172,179–191). These interventions could include SMS, telephone calls or automated telephone reminders. Patients who missed appointments or failed to collect their medication received reminder letters or home visits by health-care workers. Medication monitors or computer systems in the clinic were also used to aid health-care workers in tracing patients. Medication monitors can measure the time between openings of the pill box, give audio reminders, record when the pill box is opened or send SMS reminders to take medications. There were higher rates of treatment success, treatment adherence and 2-month sputum conversion, and lower rates of mortality, loss to follow up and drug resistance acquisition with tracers, either through home visits or mobile telephone communication (SMS or telephone call). When mobile telephone interventions (SMS or telephone call) were examined separately, there were higher rates of treatment success, cure and 2-month sputum conversion, and lower rates of treatment failure, loss to follow up, poor adherence and unfavourable outcomes with mobile telephone reminders as opposed to no intervention. Medication monitors had better rates of adherence and favourable outcomes, and combined interventions of SMS and medication monitors also showed better adherence compared to no intervention. It should be noted, however, that only a small number of studies were available for all digital health interventions. With all the digital interventions and tracers, including VOT, patient support and the ability of the patient to interact with health-care workers should be preserved. In fact, these interventions should be considered as tools to enable better communication with the health-care provider rather than as replacements for other adherence interventions. In practice, it is expected that SMS, telephone calls and VOT may replace in-person DOT for periods of time rather than for the entire duration and that they promote patient-centred approaches to care. Mobile telephone interventions, tracers and VOT may also increase health equity if the need to travel to a health clinic or to a patient’s home is reduced. However, the ability of patients to participate in these programmes depends on the patients living in an area with a good telecommunication infrastructure.
Material support for patients. The effects of material support were examined both with RCTs (150–153) and observational studies (159,192–199). The interventions included giving meals with DOT, monthly food vouchers, food baskets, food supplements and vitamins. Food support for patients and family members is an important incentive for TB patients and it also helps protect patients from the catastrophic costs associated with TB. Food may be an incentive, but it may also improve outcome biologically due to reduction in malnutrition and consequent improvement in immune function. Other material support could be financial support in the form of financial incentives, transport subsidies, living allowance, housing incentives or financial bonuses after reaching treatment targets. There were higher rates of treatment success, completion and sputum conversion in patients who received material support, and lower rates of treatment failure and loss to follow up in patients who did not receive material support. It is of note that all of these studies were in low- and middle-income countries, so presumably these incentives were of significant value to the patients in these settings. However, material support would be of significant value to TB patients even in higher-income countries, especially in countries that do not have a good social welfare system, as TB is a disease of poverty. The studies in this review found that material support was usually given to the most vulnerable groups, and therefore health equity was presumably improved by this intervention. However, if these incentives are not applied equitably, health disparities may be increased. The distribution of material support is likely to depend on the country context and may have different effects within and between countries.
Patient education or educational counselling. Analysis of the benefit of patient education included RCTs (145–148) and an observational study (156). Patients who received education or educational counselling had better rates of treatment success, treatment completion, cure and treatment adherence, and had lower rates of loss to follow up. It should be noted in this case that “counselling” refers to educational counselling and not psychological counselling. Patient education could include oral or written education via health-care workers or pharmacists. The education could be one-time at discharge from the intensive phase of therapy or at each presentation for follow-up care. The educational session might include only the health-care worker or it might involve the patients’ social network and family members. It is important to make sure that education and counselling are done in a culturally appropriate manner. Additionally, specific marginalized populations may require special educational efforts.
Staff education. Staff education may include peer training, visual aids to help initiate conversations with patients, other tools to aid in decision-making and as reminders, and the education of laboratory staff. This intervention was examined in both RCTs and observational studies (149,150,199,200). There were higher rates of treatment success and slightly lower rates of mortality and loss to follow up with staff education. With better staff education, treatment for patients is likely to improve and any stigma that health-care workers may hold towards patients would decrease, as health-care workers better understand TB disease and TB treatment.
Psychological support. Psychological support was varied and could include self-help groups, alcohol cessation counselling and TB clubs (137,155,201). Patients who had access to psychological support had higher rates of treatment completion and cure, as well as lower rates of treatment failure and loss to follow up. However, the GDG had concerns about confounding in these studies due to the severity of illness in the groups receiving support. Additionally, allocation of patients to the support groups was not always randomized. When considering these data, it should also be noted that psychological support types are very broad and may not be adequately represented in this review. To maximize health equity, psychological support should be targeted at the most marginalized populations.
Ambulatory care. Outcomes from models of MDR-TB care based mainly on clinic-based ambulatory treatment were compared with those using mainly hospital-based inpatient treatment. The data used came from cost–effectiveness studies in four countries (Estonia and the Russian Federation [Tomsk oblast] (202), Peru (203) and the Philippines (204)). The design of these observational studies did not allow direct comparison of effects between models of care. Given that none of the studies were RCTs, the evidence was considered of very low quality. Cost–effectiveness was modelled for all possible WHO Member States in a probabilistic analysis of the data from the four countries (205).
A high value was placed on conserving resources and on patient outcomes such as preventing death and transmission of MDR-TB as a result of delayed diagnosis and inpatient treatment. There should always be provision for a back-up facility to manage patients who need inpatient treatment. This may be necessary in certain patient groups at particular risk, such as children during the intensive phase, among whom close monitoring may be required for a certain period of time.
Decentralized care. As the use of Xpert MTB/RIF expands, more patients will be diagnosed and enrolled on MDR-TB treatment. Having treatment and care provided in decentralized health-care facilities is a practical approach to scaling up treatment and care for patients who are eligible for MDR-TB treatment. Therefore, a systematic review of the treatment and care of bacteriologically confirmed or clinically diagnosed MDR-TB patients in decentralized versus centralized systems was conducted to gather evidence on whether the quality of treatment and care is likely to be compromised with a decentralized approach. Data from both RCTs and observational studies were analysed, the majority being from low- and middle-income countries (201,206–213). The review provided additional value to the recommendation in the previous guidelines (2) on ambulatory over hospitalized models of care for MDR-TB patients, where the evidence was examined only for treatment and care of patients outside or inside hospitals. In the review, decentralized care was defined as care provided in the local community where the patient lives, by non-specialized or peripheral health centres, by community health workers or nurses, non-specialized doctors, community volunteers or treatment supporters.
Care could occur at local venues or at the patient’s home or workplace. Treatment and care included DOT and patient support, in addition to injections during the intensive phase. In this group, a brief phase of hospitalization of less than 1 month was accepted for patients who were in need in the initial phase of treatment or when they had any treatment complications. Centralized care was defined as inpatient treatment and care provided solely by specialized DR-TB centres or teams for the duration of the intensive phase of therapy or until culture or smear conversion. Afterwards, patients could receive decentralized care. Centralized care was usually delivered by specialist doctors or nurses and could include centralized outpatient clinics (outpatient facilities located at or near the site of the centralized hospital). Analysis of the data showed that treatment success and loss to follow up improved with decentralized care compared to centralized care. The risks of death and treatment failure showed minimal differences between patients undergoing decentralized care and centralized care.
There were limited data on adverse reactions, adherence, acquired drug resistance and cost. Both HIV-negative and HIV-positive persons were included in the reviewed studies; however, the studies did not stratify patients according to HIV status. There was some discussion regarding the quality of the data. The GDG expressed concerns that health-care workers may have selected for the centralized care groups those patients who they thought might have a worse prognosis. None of the studies controlled for this risk of bias.
Implementation considerations
Treatment adherence interventions. As treatment supervision alone is not likely to be sufficient to ensure good TB treatment outcomes, additional treatment adherence interventions need to be provided. Patient education should be provided to all patients on TB treatment. A package of the other treatment adherence interventions also needs to be offered to patients on TB treatment. The interventions should be selected on the basis of an assessment of the individual patient’s needs, provider’s resources and conditions for implementation. With regard to telephone or video-assisted interventions, there may be reluctance to use new technology, making implementation more difficult. There may be privacy concerns surrounding security of telephone data, so encryption and other measures to safeguard privacy will need to be considered. The feasibility of implementing these types of interventions depends on telecommunication infrastructure, telephone availability and connection costs. Multiple organizations have initiated programmes such as these, so TB programmes may find it helpful to collaborate and communicate with other medical service delivery programmes that have already set up infrastructure. There may be reluctance on the part of implementers (e.g. national or local governments, health partners) to pay for incentives. Implementers may be more willing to pay for material support for smaller subgroups with particularly high risk (e.g. patients with MDR-TB). However, one of the components of the End TB Strategy (206) is to provide “social protection and poverty alleviation” for patients with TB. This publication specifically calls for measures to “alleviate the burden of income loss and non-medical costs of seeking and staying in care”. Included in these suggested protections are social welfare payments, vouchers and food packages. The benefit of material support found in this review supports these components of the End TB Strategy (206). In order to distribute the material support, government and/or nongovernment organization (NGO) infrastructure would need to be in place, including anti-fraud mechanisms (e.g. reliable unique personal identifiers) and appropriate accounting to ensure that incentives are distributed equitably and to the people who need them most. Countries should choose incentives that are the most appropriate for their situation.
Treatment administration. Community-based or home-based DOT has more advantages than health facility-based DOT, though family members should not be the first or only option for administering DOT. DOT is better provided at home or in the community and by trained lay providers or health-care workers. There may be challenges in providing community- or home-based DOT by health-care workers because of the increased number of health-care workers required and the increased costs of staff time and daily travel to the community or patient’s home. DOT provision in the community or at home by trained local lay persons is more feasible. A combination of lay provider and health-care worker for provision of community- or home-based DOT is also an option. Community-based or home-based DOT is more likely to be acceptable and accessible to patients than other forms of DOT. However, stigma may continue to be an issue with community- or home-based DOT. Having a health-care worker coming regularly to a patient’s house may be stigmatizing and the feeling of being “watched over” may be disempowering for patients. Other forms of DOT (e.g. administered by an emotionally supportive relative or close friend) may be more acceptable but may still be stigmatizing. Given complex family social dynamics, family members may not always be the best people to supervise treatment, and the suitability of such treatment adherence supervisors needs to be carefully analysed in each national or local context. If family members are providing DOT, careful identification and training of those persons is required. Additional supervision of local supporters or health-care workers is still needed, as family members cannot be depended on as the only option for care. Patients will continue to need social support, even if family members are providing DOT. Assessment of potential risk factors for poor adherence must be taken into account by health-care workers at the start of treatment in order to decide which treatment administration option should be selected for the patient. Some groups of patients who are less likely to adhere to treatment may benefit more from DOT than others. Another factor to consider when selecting treatment administration options is that some patients with inflexible work or family responsibilities may not be able to do DOT. Any option of treatment administration offered to a patient must be provided in conjunction with proper medical care, including regular pick-up of TB drugs, consultations with a physician or other health-care workers when necessary, TB treatment that is free of charge, and provision to the patient of essential information on TB treatment.
Ambulatory care. Cost varied widely across the modelled settings. The cost per disability-adjusted life year (DALY) averted by an ambulatory model in one setting was sometimes higher than the cost per DALY averted by a hospitalization model in another setting. However, cost per DALY averted was lower under outpatient-based care than under inpatient-based care in the vast majority (at least 90%) of settings for which cost–effectiveness was modelled. The variation in cost–effectiveness among settings correlated most strongly with the variation in the cost of general health-care services and other non-drug costs. Despite the limitations in the data available, there was no evidence that was in conflict with the recommendation and which indicated that treatment in a hospital-based model of care leads to a more favourable treatment outcome.
The overall cost–effectiveness of care for a patient receiving treatment for MDR-TB can be improved with an ambulatory model. The benefits include reduced resource use, and at least as many deaths avoided among primary and secondary cases compared with hospitalization models. This result is based on clinic-based ambulatory treatment (patients attend a health-care facility); in some settings, home-based ambulatory treatment (provided by a worker in the community) might improve cost–effectiveness even further. The benefit of reduced transmission can be expected only if proper infection control measures are in place in both the home and the clinic. Potential exposure to people who are infectious can be minimized by reducing or avoiding hospitalization where possible, reducing the number of outpatient visits, avoiding overcrowding in wards and waiting areas, and prioritizing community-care approaches for TB management (214). The regimen used in one of the studies on ambulatory care was from a time when the combinations of medicines were not yet optimized, so outcomes achieved were probably inferior to those that can be obtained with the regimens in use today. Admission to hospitals for patients who do not warrant it may also have important social and psychological consequences that need to be taken into account.
There may be some important barriers to accessing clinic-based ambulatory care, including distance to travel and other costs to individual patients. Shifting costs from the service provider to the patient has to be avoided, and implementation may need to be accompanied by appropriate enablers. While placing patients on adequate therapy would be expected to decrease the bacterial load and transmission of DR-TB, infection control measures for home-based and clinic-based measures will need to be part of an ambulatory model of care to decrease the risk of transmission in households, the community and clinics. TB control programmes will have to consider whether they are capable of reallocating resources from hospital to ambulatory care support in order to undertake the necessary changes in patient management. The choice between these options will affect the feasibility of implementing the recommendation in a particular programme.
Decentralized care. National TB programmes should have standardized guidelines regarding which patients are eligible for decentralized care. Patient preference should be given a high value when choosing centralized or decentralized care.
Decentralized care for MDR-TB patients requires appropriate treatment supervision, patient education and social support, staff training, infection control practices and quality assurance. The optimal treatment supervision options and treatment adherence interventions recommended in this section should be considered for MDR-TB patients on decentralized care.
Several of the studies in the review addressed treatment costs. However, the cost estimates were found to vary widely and no concrete recommendations could be made on the basis of cost. Resource requirements are likely to vary because TB treatment programmes are highly variable, so costs for these programmes vary across different countries. The GDG raised several issues for TB programmes to consider. Although hospitalization is generally thought to be more expensive than outpatient care, the costs of good outpatient programmes can also be significant. Additionally, outpatient costs may vary significantly according to the services provided. A cost-saving measure to consider in decentralized care is that patients may be able to receive treatment faster. The financial benefits of decentralized care would include finding patients before they are very ill and require more medical care, while treating people before TB can be transmitted to contacts would be a public health benefit.
If a patient is living with a person from a high-risk group, such as a PLHIV or a young child, there may be complications in sending the patient home for treatment. However, the risk posed to these high-risk groups varies significantly, depending on whether the TB programme gives preventive treatment to high-risk persons. Studies involving preventive therapy for MDR-TB therapy are ongoing.
An additional implementation issue to consider is that it may be illegal in some settings to treat MDR-TB patients in a decentralized setting, especially when the treatment involves injections. Such legal concerns need to be addressed.
