Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Effects of Psycho-Emotional and Socio-Economic Support for Tuberculosis Patients on Treatment Adherence and Treatment Outcomes – A Systematic Review and Meta-Analysis

Abstract

Background

There is uncertainty about the contribution that social support interventions (SSI) can have in mitigating the personal, social and economic costs of tuberculosis (TB) treatment on patients, and improving treatment outcomes.

Objective

To identify psycho-emotional (PE) and socio-economic (SE) interventions provided to TB patients and to assess the effects of these interventions on treatment adherence and treatment outcomes.

Search strategy

We searched PubMed and Embase from 1 January 1990–15 March 2015 and abstracts of the Union World Conference on Lung Health from 2010–2014 for studies reporting TB treatment adherence and treatment outcomes following SSI.

Selection criteria

Studies measuring the effects of PE or SE interventions on TB treatment adherence, treatment outcomes, and/or financial burden.

Data collection and analysis

Two reviewers independently assessed titles and abstracts for inclusion of articles. One reviewer reviewed full text articles and the reference list of selected studies. A second reviewer double checked all extracted information against the articles.

Main results

Twenty-five studies were included in the qualitative analysis; of which eighteen were included in the meta-analysis. Effects were pooled from 11 Randomized Controlled Trials (RCTs), including 9,655 participants with active TB. Meta-analysis showed that PE support (RR 1.37; CI 1.08–1.73), SE support (RR 1.08; CI 1.03–1.13) and combined PE and SE support (RR 1.17; CI 1.12–1.22) were associated with a significant improvement of successful treatment outcomes. Also PE support, SE support and a combination of these types of support were associated with reductions in unsuccessful treatment outcomes (PE: RR 0.46; CI 0.22–0.96, SE: RR 0.78; CI 0.69–0.88 and Combined PE and SE: RR 0.42; CI 0.23–0.75). Evidence on the effect of PE and SE interventions on treatment adherence were not meta-analysed because the interventions were too heterogeneous to pool. No evidence was found to show whether SE reduced the financial burden for TB patients.

Discussion and Conclusions

Our review and meta-analysis concluded that PE and SE interventions are associated with beneficial effects on TB treatment outcomes. However, the quality of evidence is very low and future well-designed evaluation studies are needed.

Background

In 2013, 9 million people developed TB and 1.5 million died from this disease [1,2]. TB is the most common cause of death in people with HIV [1]. The treatment duration for TB is long, at least 6 months for drug-susceptible TB and 18–24 months for multidrug-resistant tuberculosis (MDR-TB) that does not respond to the two most effective anti-TB drugs isoniazid and rifampicin. The long treatment, adverse drug reactions during treatment, stigma and financial burden of TB contribute to non-adherence to treatment and unsuccessful treatment outcomes [38]. In addition, ensuring patient adherence to treatment through facility-based directly observed therapy (DOT) competes with work related priorities of patients, adding to the financial burden coming from out-of-pocket and indirect costs related to treatment [7,9], even though anti-TB drugs are provided free of charge in most countries [1,10]. The quick improvement of TB symptoms early in treatment also contributes to patients’ stopping treatment prematurely (i.e. loss to follow-up) as competing interests take priority [9,11]. Poor treatment adherence and loss to follow-up increase morbidity, mortality, and the risk of drug resistance development, and can lead to prolonged transmission of TB [1217].

Adherence to tuberculosis treatment improves the chance of cure and reduces acquisition of drug resistance and ongoing transmission of TB. The use of DOT through a patient-centered approach, which often requires enablers, is recommended to encourage adherence to TB treatment [18,19]. In some settings and circumstances, incentives alone or in addition to enablers are used to motivate patients to adhere to and complete their full course of treatment [9,16,2022]. Social support through various educational, emotional, and/or material (in-kind or services) interventions are being provided by numerous TB programmes to remove or alleviate barriers to treatment adherence [9,20,2325], including the financial burden associated with TB illness and its treatment. Despite the fact that different types of social support interventions (SSI) are implemented, countries still struggle to develop systems that are able to provide SSI in an efficient, effective and sustainable way [26]. WHO guidelines for the programmatic management of drug resistant TB and the new End TB Strategy recommend the use of SSI in TB patients, though WHO has not yet systematically assessed the evidence to support such a recommendation [2,19,27]. Hence, a systematic review of relevant literature on the effects of SSI on TB treatment adherence, treatment outcomes, and financial burden will be informative for national and global policy making.

The primary aim of this systematic review was to identify SSI provided to TB and MDR-TB patients and assess the evidence of their effects on treatment adherence, treatment outcomes and financial burden related to TB illness. The secondary aim was to describe the funding sources for and ownership of local organizations in the identified interventions.

Methods

This review followed standard methods as defined by the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [28,29]. The PRISMA checklist is enclosed in the supporting information (S1 PRISMA Checklist).

Literature search

In this review we searched for two main categories of SSI, namely PE support and SE support. PE support includes both emotional support through psychological interventions (e.g. counseling by health care workers) and companionship support through provision of help for patients to participate in a social network (e.g. peer counseling for patients and their support network)[19]. We did not consider interventions aimed only at providing improved information or education to TB patients, given the recent systematic review showing a lack of evidence related to TB treatment [17]. In addition, reminder systems were not considered social support interventions [30]. SE support entails delivering services, material goods and/or financial assistance [19,31,32]. Financial assistance was categorized according to Richter et al. [7] as”direct transfers of money, such as cash paid as part of a social security system or a program incentive, transport reimbursements, treatment allowances, and the like that are paid directly to affected individuals”. Indirect assistance was defined as: “indirect transfers through, for example, food packages or vouchers, travel vouchers, and payment of health insurance for individuals, households or families”. Some forms of indirect assistance may also be converted into cash. We included tax exemption under indirect assistance. Enterprise assistance was defined as”training programs or microcredit that aim to assist individuals or families to generate income” [7]. We searched for studies assessing the effects of socio-economic and/or psycho-emotional interventions on treatment adherence and/or treatment outcomes and/or financial burden. The study population consisted of patients initiated on anti-TB treatment, including treatment for MDR-TB.

Outcome measures

Treatment adherence, treatment outcomes and financial burden were considered as the primary outcome measures. Adherence was calculated as the percentage of prescribed doses actually taken. Treatment outcomes were defined according to WHO definitions, where cure and completed treatment are defined as successful treatment outcomes [1]. Unsuccessful treatment outcomes for active TB treatment included death, treatment failure and loss to follow-up (previously named default). Patients with transfer-out or missing treatment outcomes were excluded from the analysis. As timing of loss to follow-up per individual was not available for studies reporting on treatment outcomes but not treatment adherence, for these studies loss to follow-up was not included in calculation of treatment adherence. Financial burden was reported according to the definitions used in the individual studies. We also extracted information about how the SSI were financed and organized.

Search strategy

We systematically searched PubMed and Embase for primary articles and reviews reporting on SSI and tuberculosis treatment for human subjects, published from 01 January 1990–15 March 2015, on the grounds that relevant old information would emerge from previous reviews and references lists. We reviewed the reference lists of identified articles, editorials and reviews. Additionally, we hand searched the 2010–2014 abstract books of the Union World Conference on Lung Health to identify recent studies that were not published in the literature yet. Databases were searched using the full text search strategy as described in S1 Web annex. We contacted authors when we were not able to extract required information from the identified publication on the SSI provided and its effects.

Eligibility criteria

Eligibility of studies was based on predetermined inclusion criteria. Original studies including a description of SSI had to be in place, as well as an evaluation of the association of SSI on treatment adherence, treatment outcome and/or financial burden. This was evaluated either by means of a comparison between outcomes of an intervention group and a group receiving standard support (which could be none or a more limited package), or by means of a comparison of the occurrence of interventions in those with positive and negative outcomes (case-control studies). The search strategy was restricted to certain languages including publications in Dutch, English, French, German, Portuguese, Russian and Spanish. No age restriction was applied. We chose not to exclude studies that did not provide DOT to their patients as there is no hard evidence that DOT in a strict sense (i.e. direct observation of medication ingestion) without the DOT provider supporting the patient through education and counseling improves treatment outcome under programmatic conditions [22,33].

Data collection and analysis

Selection of studies and data extraction.

One reviewer conducted the literature search (RH) based on the search strategy developed by all authors. Subsequently, two reviewers (SH, RH) independently examined titles and abstracts retrieved by the search. One reviewer (RH) reviewed full texts and the reference lists of selected articles, and extracted study data, which were then verified by a second reviewer (SH). For data extraction and management, a pre-piloted form was developed to list study characteristics including: study design and study aim, type(s) of patients, type(s) of TB treatment, descriptions of intervention and control group, descriptions of intervention and routine support, coverage of patients that received the intervention, results of the intervention and control group and differences between these groups. Duplicate publications of included studies were taken into account if they provided additional information. When disagreements occurred, a third independent reviewer was consulted and discrepancies were resolved by consensus among the three.

Risk of bias and quality of evidence.

Risk of bias was assessed separately for Randomized Controlled Trial (RCTs) and Non Randomized Studies (NRS). We used the Newcastle Ottawa Scale for NRS [34] and The Cochrane Collaboration’s Tool for RCTs [35]. Furthermore, an additional assessment was made for Cluster Randomized Trials on recruitment bias, baseline imbalance and loss of clusters [36]. For NRS, we considered <10% of subjects lost as indicative of low risk of bias. The quality of evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) tool [3740].

Data analysis.

All SSI were described, irrespective of inclusion in the meta-analysis. We analyzed the dichotomous outcomes using Risk Ratios (RR) for RCTs and cohort studies, and Odds Ratios (OR) for case-control studies, together with corresponding 95% confidence intervals. Ratios were (re)calculated from the data provided in the publications. Subsequently, the (calculated) intervention effects were combined in the meta-analysis. Studies were assessed on clinical diversity (e.g. differences in patient spectrum, type and dose of treatment) and methodological diversity (e.g. differences in methods: blinding of patients, concealment and randomization). Additionally, (statistical) heterogeneity was examined with the I2 test along with the visual assessment of the forest plots [28,41]. An I2 of 0–40% was considered as low heterogeneity, 30–60% was defined as moderate heterogeneity, 50–90% substantial heterogeneity and 75–100% as high heterogeneity [42]. Furthermore, the I2 was interpreted along with the directions and magnitudes of the different studies observed in the forest plots. A p-value for the Chi2 test of ≤0.10 was considered as a cut-off point for statistically significant heterogeneity. In case of statistically significant heterogeneity, sensitivity analysis were performed based on patient type (e.g. MDR-TB or not) and risk of bias (e.g. low vs. high risk of bias)[42]. Funnel plots were created to assess for publication bias. To execute the meta-analysis, a random effects model was used, considering the diversity in participants (e.g., susceptible TB-patients and MDR-patients) and interventions (e.g. self-help groups and counseling). The DerSimonian Laird method is based on the inverse-variance approach [42]. Due to the potential heterogeneity of the interventions (PE support, SE support and combined PE and SE support) also stratified analyses were performed [43]. Stata (STATA/SE 13.1) was used to perform the meta-analysis. To visualize the risk of bias assessment, Review Manager (Review Manager (RevMan) 5.3, The Nordic Cochrane Centre, Copenhagen) was used.

Results

In total, we identified 2443 articles. After removal of 694 duplicates, two reviewers screened titles and abstracts of the 1752 citations. Twenty-five articles were eligible for inclusion in the description of included studies (Fig 1).

Description of included studies

Fourteen NRS and eleven RCTs were included in the description of interventions from 15 different countries. Study populations ranged from 46 to 4,091 participants. Eight studies included both children and adults [4451]. Three studies explicitly included adults [5254]. For the other studies the age range was not reported, however mean age was provided frequently [20,5564]. Most studies were conducted in middle income countries, 9 in upper middle income countries and 7 in lower middle income countries [65]. Six studies were performed in high income countries and the remaining three studies in low income countries. Eleven studies provided SE support only, seven studies provided only PE support, while the remaining seven studies provided a combination of PE and SE support [44,52,56,57,61,66,67] (Table 1). Table 2 includes a comprehensive summary of studies including the frequency of the intervention provided and sustainability of the below described interventions.

thumbnail
Table 1. Overview on types of support and inclusion in the quantitative analysis.

https://doi.org/10.1371/journal.pone.0154095.t001

thumbnail
Table 2. Summary table for all studies included in the qualitative analysis.

https://doi.org/10.1371/journal.pone.0154095.t002

Psycho-emotional support.

Seven studies provided counseling, exclusively [46,53] or in combination with other PE and or SE interventions [44,51,52,61,67]. The scope of the additional interventions varied from food supplementation [44] combined with home visits [67], direct economic support constituted after an exploratory quality study [52], cash coupons at every monthly visit and at the end of treatment [61], arrangement of a self-chosen treatment supporter [51]. See Table 2 for details.

Furthermore, 2 studies organized self-help groups [50,59], one of these studies along with stigma reduction and home visits [59]. TB clubs were raised in the form of self-help groups in combination with support to reduce stigma and home visits to get insight in the social network of the patients and to plan activities to support the patient [59]. In the second study, the patients could choose the number of meetings and the topics discussed [50]. Another 6 studies arranged home visits together with other interventions [51,57,59,6668].

Socio-economic support.

Eight studies provided food supplementation consisting of fresh food supplies [58,60], hot meals [44] and/or food packages [44,45,49,54,60,67,68]. Four of them exclusively provided food supplementation [45,49,58,60]. Other studies also provided food supplementation, in combination with direct economic support and/or other material support through provision of e.g. clothing and legal support [44], assistance in providing documentation for health care access and social security [54], or establishing a supportive social network of organizations that could provide support to the local community, such as public day care centers and employment agencies [68]. One study additionally provided PE support [67].

Four studies provided indirect economic support including food and transport vouchers [20,47,56,61]. Coupons varying from 5 to 15 US$ were given when attending each appointment or at drug collection each month. Some studies provided additional coupons varying from 40 to 60 US$ after completion of 3 months of treatment or at the end of treatment [56,61]. Seven studies granted direct economic support, mainly financial support varying from 19 to 240 US$ per month [44,48,52,57,6264]. Four studies provided direct economic support exclusively [48,6264]. Other studies only provided economic support for the first three months and 5 US$ per month for travel expenses [57] or arranged reimbursement of travel for an unknown amount of money, combined with food supplementation, other material support and psycho-emotional support [44]. The remaining three studies also combined socio-economic support with psycho-emotional support. No studies on ‘enterprise assistance’ were found. Details on economic support provided per study are retrievable in Table 2.

Funding sources and organization.

Information on funding sources and involvement of local bodies in the organization of the interventions can be found in Table 2. Seven SSIs were financed through governmental funding or local authorities. Another nine interventions were funded by foreign donor assistance (e.g. WHO, Unicef). Three interventions received combined funding (local and foreign donor assistance). For the remaining five interventions the funding source was unknown.

In total nine studies provided information on the organization of interventions, including six RCTs [46,5052,55,67] and three NRS [44,59,66]. A study from Russia organized and implemented support by regional TB services and a local international organization[23] and a study from Nicaragua raised TB clubs organized by TB patients, with the help of local non-governmental organizations [59]. Community involvement was integrated into regular patient management in Burkina Faso [44,59,67]. The remaining studies reported very limited information on organizational sustainability.

Incentives and enablers.

All the RCTs defined their support as incentives. Incentives are rewards for adherence while enablers assist patients to overcome barriers to treatment adherence. Most studies provided support to all TB patients. In studies where only poor patients were supported [64]; it may be that the support in fact was in the form of enablers.

Risk of bias and quality of evidence

Risk of bias was assessed for all included RCTs, including six Cluster Randomized Trials [47,5052,60,67]. Only five out of eleven RCTs described an adequate randomization approach [5052,58,60]. For the majority of the studies it was not described whether investigators were blinded to the outcome, and assessment of reporting bias was not possible due to a lack of information. None of the Cluster Randomized Trials assessed baseline imbalances between clusters or took random effects into account in the analysis. Ten NRS were assessed on risk of bias, including eight cohort studies and two case-control studies. Four studies [20,56,63,66] were not included in the meta-analysis and risk of bias assessment; reasons for exclusion are described in Table 3. Only three NRS adjusted for one or more confounders in the analysis [44,48,53]. Five additional studies were not included because of inadequacy of follow-up and/or assessment of outcome measures [44,48,53,62,68]. More information on the risk of bias assesment of the RCTs and NRS can be found in the supportive information S1S3 Tables. Quality of evidence was assessed for the included RCTs per outcome measure. The quality of evidence for the RCTs was downgraded with one level for risk of bias, two levels on indirectness of studies and one level for limitations in consistency of the results. Hence, the overall quality of evidence of this systematic review is considered to be very low [40,6974]. The quality of evidence per outcome measure is similar to the overall quality of evidence and retrievable in the summary of findings table (Table 4). No rating up for the overall quality of evidence was possible. Based on the funnel plot for the results of the ten RCTs included in the meta-analysis, it was not possible to determine whether publication bias was present (Fig 2)[28]

thumbnail
Fig 2. Funnel plot to evaluate publication bias in Randomized Controlled Trials on the effects of social support interventions on treatment outcomes.

https://doi.org/10.1371/journal.pone.0154095.g002

Meta-analysis

Eleven RCTs, eight cohort studies, and two case-control studies were included in the meta-analysis, including 17 743 patients (9655 patients participating in RCTs and 8088 patients in NRS). Most data originated from Brazil, China, Russia, Senegal and South Africa. No evidence was found concerning the effect of SSI on financial burden. Only one NRS measured the cost-effectiveness ratio of the provided economic support [64]. Studies assessing the effect of SSI on treatment adherence were too heterogeneous to pool. Meta-analysis of different outcome measures are presented separately (Figs 3 and 4).

thumbnail
Fig 3. The effects of social support on treatment success by type of intervention in Randomized Controlled Trials.

https://doi.org/10.1371/journal.pone.0154095.g003

thumbnail
Fig 4. The effects of social support on unsuccessful treatment outcomes by type of intervention in Randomized Controlled Trials.

https://doi.org/10.1371/journal.pone.0154095.g004

Treatment outcomes.

In total, nine RCTs had treatment success as an outcome measure (Fig 3). The overall effect of these studies showed a significant positive effect (RR 1.17; CI 1.09–1.25), however significant heterogeneity was observed (I2 of 72.8%, P = <0.001). Stratified analyses were performed for the different types of interventions. Three studies provided PE support [50,52,55] including counseling, psychotherapy and the organization of self-help groups. A significant pooled effect was found for this intervention (RR 1.37; CI 1.08–1.73). The association between SE support and treatment success was examined by four studies [47,49,58,60] providing food supplementation and economic support. A significant pooled effect was found for this intervention (RR 1.08; CI 1.03–1.13). Combined support was provided by three studies [51,52,67]. Also, a significant pooled effect was found for these interventions on successful treatment outcomes (RR 1.17; CI 1.12–1.22). No significant heterogeneity was observed in two of three stratified analyses (SE: I2 of 14%, P = 0.32; combined: I2 of 0%, P = 0.42). Studies that provided PE support were substantially heterogenic and the p-value for the Chi2 test was significant (I2 of 78%, P = 0.01) (Fig 3). A sensitivity analysis was performed on the effect of PE support on treatment success, comparing high vs. low risk of bias studies. Omitting one high risk of bias study removed heterogeneity (I2 of 0%, P = 0.53) (data not shown), and did not change effect size (RR 1.20; CI 1.07–1.35) [55]. Sensitivity analysis on MDR-TB patients vs. non-MDR-TB patients did not change the effect size and statistical significance (data not shown).

Nine studies had unsuccessful treatment outcomes as an outcome measure including seven also having treatment success as an outcome measure (Fig 4). An overall significant protective effect was found (RR 0.53; CI 0.41–0.70), however, substantial heterogeneity was observed (I2 of 80.2% and P = <0.001). Stratified analyses were performed on the different interventions provided. Four studies investigated the effect of PE support on unsuccessful treatment outcomes, including counseling, psychotherapy and the organization of self-help groups [46,50,52,55]. Two studies examined the effect of SE support, including food supplementation and economic support [47,58] and four studies assessed the effect of combined support [51,52,61,67]. A significant reduction in unsuccessful treatment outcomes was found for all three stratified analyses: PE support (RR 0.46; CI 0.22–0.96), SE support (RR 0.78; CI 0.69–0.88) and a combination of PE and SE support (RR 0.42; CI 0.23–0.75). Heterogeneity was considered to be very low for the studies that provided SE support interventions (I2 of 0% and P = 0.37). The studies that provided PE support and combined support were substantially heterogenic (PE: I2 of 85%, P = <0.001 and combined: I2 of 64% (P = 0.03) (Fig 4). A sensitivity analysis was performed in the PE stratum on the basis of higher risk of bias compared to the other studies [46,55]. Removal of one high-risk of bias study [46] decreased the I2 to 0% (P = 0.54) and the effect size changed but remained statistically significant (RR 0.33; CI 0.22–0.50). Omitting both biased studies did not change heterogeneity or the effect size. Sensitivity analysis on risk of bias was not possible in the studies providing a combination of PE and SE support, due to the fact that 3 out of 4 studies were classified as biased studies. Sensitivity analyses on MDR-TB patients vs. non-MDR TB patients did not change the effect size or heterogeneity significantly (data not shown).

Treatment adherence.

Three RCTs assessed the effect of PE and/or SE on treatment adherence. A PE-intervention study conducted in Mexico showed a significant improvement in treatment adherence (RR 1.20; CI 1.03–1.39). A study from the USA did not show significantly higher levels of adherence in the intervention group compared to the group that received usual care (RR 1.11; CI 0.92–1.33). A third study from Timor-Leste showed no effect for patients that received SE support compared to patients that did not receive this support (RR 1.01; CI .0.85–1.21). Above-described interventions were not pooled as they were too heterogeneous.

Financial burden.

None of the RCTs examined the effect of PE or SE support on financial burden for TB patients.

Non-randomized studies.

Due to the fact that the studies’ characteristics were heterogeneous on several levels and at higher risk of bias than the RCTs, we chose not to pool the effects for these studies (S1 and S3 Figs) [28,75]. Seven NRSs reported an effect of social support on successful treatment outcomes. Effects of interventions on successful treatment outcomes (RR) ranged from 1.03 to 2.51 (CI 0.96–2.99). Five of seven NRSs reported significant effect sizes [48,54,57,64,68]. Two studies found no significant effects [45,59]. Furthermore, six NRSs examined the effect of social support on unsuccessful treatment outcomes. Effect sizes varied from RR 0.32–0.96 (CI 0.18–3.49). Five out of six NRSs showed significant beneficial effects [45,54,62,64,68]. Only one study reported a non-significant effect [59]. In addition, two case-control studies investigated the effect of social support on unsuccessfull treatment outcomes. Both studies showed significant beneficial effects (RR 0.51 (CI 037–0.70) and RR 0.10 (CI 0.05–0.20)).

Discussion

This review found that PE and SE support did improve treatment outcomes across a variety of settings and patient populations, with a tendency towards better outcomes with PE interventions or a combined approach. However, the quality of evidence was classified as “very low” under the GRADE approach. Food supplementation and counselling were commonly included in the package of support. PE, SE and combined interventions improved treatment outcomes; only for interventions including SE support exclusively there was no significant improvement in treatment success. Overall, support interventions were associated with significantly higher treatment success (overall RR 1.08; CI 1.03–1.13) and reductions in unsuccessful treatment outcomes (overall RR 0.53; CI 0.41–0.70). Hardly any studies assessed the effect of interventions on treatment adherence. However, improved treatment adherence is an intermediate goal with the final aim to improve treatment outcomes, which was shown to improve.

A recent systematic review concluded that the economic burden for patients is considered to be high, loss of income is an important indirect cost factor for TB patients, and transport and nutritional supplementation were important direct cost components [8]. A study in Peru evaluated the expenses for MDR-TB patients that received free treatment and found that having MDR-TB was associated with high costs, which was associated with adverse outcomes (population attributable fraction 18–20%) [76]. In line with our review, these two studies suggest that economic support is of great importance for improving treatment outcomes. Some of the findings of this review however differ from those from other SSI-related reviews. A recent review [77] on RCTs assessing the effect of material incentives on TB treatment adherence and completion of TB treatment identified two trials, both included in our review as well [47,60], and neither demonstrated a clear benefit. However, in one trial the incentive was not well received by the patients and in the other trial fidelity to the intervention was low. A review of Sinclair et al. did not find any evidence that food supplementation had a beneficial impact on treatment outcomes [78]. This may be explained by their focus on micronutrient supplementation alone as reflected in their search strategy. In a systematic review about strategies to reduce loss to follow-up in drug-resistant patients, a comprehensive package of interventions (e.g. financial support and food supplementation) was associated with reduced loss to follow-up [79]. Our review included studies focusing on all TB patients, not only those with MDR-TB [79]. As mentioned in the methods section, we did not consider interventions aimed only at providing improved information or education to TB patients, given the recent systematic review showing a lack of its evidence related to TB treatment [17]. Some of the intervention packages included in our review included an information or education component, but it was not possible to delineate the effects of this specific component in our review. We also did not include interventions focusing only on reminder systems, as these are not considered PE or SE support. However, reminder systems can be integrated into SSI programs to enhance its effects since pre-appointment reminder phone calls and letters or home visits did have a small but potentially relevant effect on treatment completion [30].

There were some limitations to our review. Only a limited number of studies were available on the effect of PE/SE support interventions on TB treatment outcomes and very limited evidence on treatment adherence and financial burden. Within the identified studies, we were not able to stratify results by the type of organization and quality of health service delivery due to insufficient information, although it is known that organization and quality of health service delivery influence treatment adherence [9]. Some NRSs only provided support to subgroups of patients including poor patients [64], patients that already received support before referral to the intervention studied [66] and non-adherent patients [20]. This precludes conclusions on the effects of these interventions when provided to all patients. Such patient selection may have led to overestimations in the observed effect of the PE/SE interventions. On the other hand, selecting patients most in need seems prudent and is in practice applied in resource-limited settings. Although the number of studies included in the meta-analysis was small, the optimal size criterion was sufficient both for the overall meta-analysis and stratified analyses as examined by calculation of the sample size for the overall effect and subgroup analyses [72]. We could not examine for a dose response rate across all included studies, as most studies did not include a comprehensive description of interventions. However, one study did show a positive dose-response within their study regarding provision of indirect economic support: among patients in the intervention group who received the voucher at least once, treatment success rates significantly improved [47]. Furthermore, the more frequent the vouchers were received by patients, the higher their probability of treatment success [47]. Plausible heterogeneity was observed and seven out of eleven RCTs had a high risk of bias on one or two domains. However, we did not exclude studies on the basis of heterogeneity only, as this may introduce bias [42].

Conclusions

This review provides evidence to endorse implementation of SSI in order to improve treatment outcomes. Firstly, PE and combined PE/SE support have a beneficial impact on treatment success. Secondly, SE support and a combination of PE/SE support are associated with reductions in unsuccessful treatment outcomes. No conclusions can be drawn considering the overall effect of PE and/or SE support on treatment adherence and financial burden due to a lack of evidence. Our findings need to be interpreted with caution, as the quality of the evidence included in the meta-analysis is “very low” based on the GRADE approach. In addition, most support included multifaceted types of interventions, so no conclusions can be drawn on the effect of individual interventions. Simultaneously, this might signify that multifaceted types of interventions are needed to improve treatment outcomes. High quality evidence, from well-designed randomized studies in larger sized populations, would provide more certainty on the effects of different PE and SE interventions. Cluster-randomized studies would provide an opportunity to compare differential packages and delineate the importance of specific components. In addition, more systematic data collection on PE and SE as already used by TB programs to monitor implementation and evaluate its effects and qualitative data collection in both studies and program settings to assess which interventions are most appreciated and most feasible to implement on a wide scale, would be useful. Reports should include information on costs and sustainability to provide information on efficiency and scalability.

Supporting Information

S1 Fig. The effects of social support on treatment success in non-randomized cohort studies.

https://doi.org/10.1371/journal.pone.0154095.s002

(PNG)

S2 Fig. The effects of social support on unsuccessful treatment outcomes in non-randomized cohort studies.

https://doi.org/10.1371/journal.pone.0154095.s003

(PNG)

S3 Fig. The effects of social support on unsuccessful treatment outcomes in Case-control studies.

https://doi.org/10.1371/journal.pone.0154095.s004

(PNG)

S1 Table. Risk of bias assessment–Cochrane collaborations tool for randomized controlled trials.

https://doi.org/10.1371/journal.pone.0154095.s005

(DOCX)

S2 Table. Risk of bias assessment–New-castle Ottawa scale for non-randomized studies.

https://doi.org/10.1371/journal.pone.0154095.s006

(DOCX)

S3 Table. Risk of bias assessment–New-castle Ottawa scale for case-control studies.

https://doi.org/10.1371/journal.pone.0154095.s007

(DOCX)

S1 Web Annex. Full text search strategy per database.

https://doi.org/10.1371/journal.pone.0154095.s008

(DOCX)

Acknowledgments

We thank Nathan Ford and D’Arcy Richardson for review of previous versions of the manuscript.

Author Contributions

Conceived and designed the experiments: SH RH. Performed the experiments: SH RH. Analyzed the data: SH RH. Wrote the paper: SH RH DC EJ AG.

References

  1. 1. WHO (2014) Global tuberculosis report 2014. Geneva: World Health Organization.
  2. 2. WHO (reviewed March 2015) Fact sheet N°104.
  3. 3. Baral S, Karki D, Newell J (2007) Causes of stigma and discrimination associated with tuberculosis in Nepal: a qualitative study. BMC Public Health 7: 211. pmid:17705841
  4. 4. Vijay S, Kumar P, Chauhan LS, Vollepore BH, Kizhakkethil UP, Rao SG (2010) Risk factors associated with default among new smear positive TB patients treated under DOTS in India. PLoS One 5: e10043. pmid:20386611
  5. 5. Eastwood S, Hill P (2004) A gender-focused qualitative study of barriers to accessing tuberculosis treatment in the Gambia, West Africa. Int J Tuberc Lung Dis 8: 70–75. pmid:14974748
  6. 6. Torun T, Gungor G, Ozmen I, Bolukbasi Y, Maden E, Bicakci B, et al. (2005) Side effects associated with the treatment of multidrug-resistant tuberculosis. Int J Tuberc Lung Dis 9: 1373–1377. pmid:16468160
  7. 7. Richter LM, Lonnroth K, Desmond C, Jackson R, Jaramillo E, Weil D (2014) Economic Support to Patients in HIV and TB Grants in Rounds 7 and 10 from the Global Fund to Fight AIDS, Tuberculosis and Malaria. PLoS One 9: e86225. pmid:24489702
  8. 8. Tanimura T, Jaramillo E, Weil D, Raviglione M, Lonnroth K (2014) Financial burden for tuberculosis patients in low- and middle-income countries: a systematic review. Eur Respir J.
  9. 9. Munro SA, Lewin SA, Smith HJ, Engel ME, Fretheim A, Volmink J (2007) Patient adherence to tuberculosis treatment: a systematic review of qualitative research. PLoS Med 4: e238. pmid:17676945
  10. 10. Ngamvithayapong-Yanai J, Luangjina S, Nedsuwan S, Kantipong P, Wongyai J, Ishikawa N (2013) Engaging women volunteers of high socioeconomic status in supporting socioeconomically disadvantaged tuberculosis patients in Chiang Rai, Thailand. Western Pac Surveill Response J 4: 34–38. pmid:23908953
  11. 11. Kaona FA, Tuba M, Siziya S, Sikaona L (2004) An assessment of factors contributing to treatment adherence and knowledge of TB transmission among patients on TB treatment. BMC Public Health 4: 68. pmid:15625004
  12. 12. WHO (2009) Guidelines for surveillance of drug resistance in tuberculosis– 4th ed. WHO/HTM/TB/2009.422. Geneva: World Health Organisation.
  13. 13. Zignol M, Hosseini MS, Wright A, Weezenbeek CLv, Nunn P, Watt CJ, et al. (2006) Global Incidence of Multidrug-Resistant Tuberculosis. Journal of Infectious Diseases 194: 479–485. pmid:16845631
  14. 14. MMWR (2003) Treatment of Tuberculosis, American Thoracic Society, CDC, and Infectious Diseases Society of America. American Journal of Respiratory and Critical Care Medicine 167: 603–662. pmid:12588714
  15. 15. Pablos-Mendez A, Knirsch CA, Barr RG, Lerner BH, Frieden TR (1997) Nonadherence in tuberculosis treatment: predictors and consequences in New York City. Am J Med 102: 164–170. pmid:9217566
  16. 16. Frieden TR, Munsiff SS (2005) The DOTS strategy for controlling the global tuberculosis epidemic. Clin Chest Med 26: 197–205, v. pmid:15837105
  17. 17. M'Imunya J M, Kredo T, Volmink J (2012) Patient education and counselling for promoting adherence to treatment for tuberculosis. Cochrane Database Syst Rev 5: Cd006591. pmid:22592714
  18. 18. WHO (2015) Pursue high-quality DOTS expansion and enhancement. World Health Organization.
  19. 19. WHO (2014) Companion handbook to the WHO guidelines for the programmatic management of drug-resistant tuberculosis. In: Rich M, Jaramillo E, editors. Geneva, Switzerland: WHO Document Production Services.
  20. 20. Bock NN, Sales RM, Rogers T, DeVoe B (2001) A spoonful of sugar…: improving adherence to tuberculosis treatment using financial incentives. Int J Tuberc Lung Dis 5: 96–98. pmid:11263524
  21. 21. Frieden TR (2000) Directly observed treatment, short-course (DOTS): ensuring cure of tuberculosis. Indian J Pediatr 67: S21–27. pmid:11129903
  22. 22. Volmink J, Matchaba P, Garner P (2000) Directly observed therapy and treatment adherence. Lancet 355: 1345–1350. pmid:10776760
  23. 23. Jakubowiak WM, Bogorodskaya EM, Borisov SE, Danilova ID, Lomakina OB, Kourbatova EV (2007) Social support and incentives programme for patients with tuberculosis: experience from the Russian Federation. Int J Tuberc Lung Dis 11: 1210–1215. pmid:17958983
  24. 24. Belo MT, Selig L, Luiz RR, Hanson C, Luna AL, Teixeira EG, et al. (2006) Choosing incentives to stimulate tuberculosis treatment compliance in a poor county in Rio de Janeiro state, Brazil. Med Sci Monit 12: Ph1–5. pmid:16641886
  25. 25. Garner P, Smith H, Munro S, Volmink J (2007) Promoting adherence to tuberculosis treatment. Bull World Health Organ 85: 404–406. pmid:17639229
  26. 26. Wise J (1998) WHO identifies 16 countries struggling to control tuberculosis. BMJ 316: 955.
  27. 27. WHO (Emergency update, 2008) Guidelines for the programmatic management of drug-resistant tuberculosis Geneva: World Health Organization.
  28. 28. Higgins JPT, Green S (Updated March 2011) Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0: The Cochrane Collaboration.
  29. 29. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 151: 264–269, w264. pmid:19622511
  30. 30. Liu Q, Abba K, Alejandria MM, Sinclair D, Balanag VM, Lansang MA (2014) Reminder systems to improve patient adherence to tuberculosis clinic appointments for diagnosis and treatment. Cochrane Database Syst Rev 11: Cd006594. pmid:25403701
  31. 31. Mattson M, Hall J (2011) Health as Communication Nexus: A Service-Learning Approach.: Kendall Hunt Publishing Company.
  32. 32. van den Hof S, Collins D, Leimane I, Jaramillo E, Gebhard A (2014) Lessons Learned from Best Practices in Psycho-Socio-Economic Support for Tuberculosis Patients.
  33. 33. Karumbi J, Garner P (2015) Directly observed therapy for treating tuberculosis. Cochrane Database Syst Rev 5: Cd003343. pmid:26022367
  34. 34. Wells GA, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses.
  35. 35. Higgins J, Altman D, Sterne J (Updated September 2008) Chapter 8: Assessing risk of bias in included studies. In: Higgins JPT, Green S (editors), Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1: The Cochrane Collaboration.
  36. 36. Higgins J, Deeks J, Altman D (Updated September 2008) Chapter 16: Special topics in statistics. In: Higgins JPT, Green S (editors), Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1: The Cochrane Collaboration.
  37. 37. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. (2008) GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. Bmj 336: 924–926. pmid:18436948
  38. 38. Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, et al. (2011) GRADE guidelines: 3. Rating the quality of evidence. Journal of clinical epidemiology 64: 401–406. pmid:21208779
  39. 39. Guyatt GH, Oxman AD, Vist G, Kunz R, Brozek J, Alonso-Coello P, et al. (2011) GRADE guidelines: 4. Rating the quality of evidence—study limitations (risk of bias). Journal of clinical epidemiology 64: 407–415. pmid:21247734
  40. 40. Schünemann HJ, Oxman AD, Vist GE, Higgins JPT, Deeks JJ, Glasziou P, et al. (Updated September 2008) Chapter 12: Interpreting results and drawing conclusions. In: Higgins JPT, Green S (editors), Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1: The Cochrane Collaboration.
  41. 41. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. Bmj 327: 557–560. pmid:12958120
  42. 42. Deeks J, Higgins J, Altman D (Updated September 2008) Chapter 9: Analysing data and undertaking meta-anlysis. In: Higgins JPT, Green S (editors), Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1: The Cochrane Collaboration.
  43. 43. JBIEBNM (2000) Evidence Based Practice Information Sheets for Health Professionals: Appraising Systematic Reviews, Changing Practice Sup. 1. The Joanna Briggs Insititute for evidence based nursing and midwifery.
  44. 44. Jakubowiak WM, Bogorodskaya EM, Borisov SE, Danilova ID, Kourbatova EV (2007) Risk factors associated with default among new pulmonary TB patients and social support in six Russian regions. Int J Tuberc Lung Dis 11: 46–53. pmid:17217129
  45. 45. Cantalice Filho JP (2009) Food baskets given to tuberculosis patients at a primary health care clinic in the city of Duque de Caxias, Brazil: effect on treatment outcomes. J Bras Pneumol 35: 992–997. pmid:19918632
  46. 46. Liefooghe R, Suetens C, Meulemans H, Moran MB, De Muynck A (1999) A randomised trial of the impact of counselling on treatment adherence of tuberculosis patients in Sialkot, Pakistan. Int J Tuberc Lung Dis 3: 1073–1080. pmid:10599010
  47. 47. Lutge E, Lewin S, Volmink J, Friedman I, Lombard C (2013) Economic support to improve tuberculosis treatment outcomes in South Africa: a pragmatic cluster-randomized controlled trial. Trials 14: 154. pmid:23714270
  48. 48. Lu H, Yan F, Wang W, Wu L, Ma W, Chen J, et al. (2013) Do transportation subsidies and living allowances improve tuberculosis control outcomes among internal migrants in urban Shanghai, China? Western Pac Surveill Response J 4: 19–24.
  49. 49. Sudarsanam TD, John J, Kang G, Mahendri V, Gerrior J, Franciosa M, et al. (2011) Pilot randomized trial of nutritional supplementation in patients with tuberculosis and HIV-tuberculosis coinfection receiving directly observed short-course chemotherapy for tuberculosis. Trop Med Int Health 16: 699–706. pmid:21418447
  50. 50. Alvarez Gordillo Gdel C, Alvarez Gordillo JF, Dorantes Jimenez JE (2003) [Educational strategy for improving patient compliance with the tuberculosis treatment regimen in Chiapas, Mexico]. Rev Panam Salud Publica 14: 402–408. pmid:14769157
  51. 51. Thiam S, LeFevre AM, Hane F, Ndiaye A, Ba F, Fielding KL, et al. (2007) Effectiveness of a strategy to improve adherence to tuberculosis treatment in a resource-poor setting: a cluster randomized controlled trial. Jama 297: 380–386. pmid:17244834
  52. 52. Baral SC, Aryal Y, Bhattrai R, King R, Newell JN (2014) The importance of providing counselling and financial support to patients receiving treatment for multi-drug resistant TB: mixed method qualitative and pilot intervention studies. BMC Public Health 14: 46. pmid:24438351
  53. 53. Finlay A, Lancaster J, Holtz TH, Weyer K, Miranda A, van der Walt M (2012) Patient- and provider-level risk factors associated with default from tuberculosis treatment, South Africa, 2002: a case-control study. BMC Public Health 12: 56. pmid:22264339
  54. 54. Garden B, Samarina A, Stavchanskaya I, Alsterlund R, Ovregaard A, Taganova O, et al. (2013) Food incentives improve adherence to tuberculosis drug treatment among homeless patients in Russia. Scand J Caring Sci 27: 117–122. pmid:22671304
  55. 55. Janmeja AK, Das SK, Bhargava R, Chavan BS (2005) Psychotherapy improves compliance with tuberculosis treatment. Respiration 72: 375–380. pmid:16088280
  56. 56. Davidson H, Schluger NW, Feldman PH, Valentine DP, Telzak EE, Laufer FN (2000) The effects of increasing incentives on adherence to tuberculosis directly observed therapy. Int J Tuberc Lung Dis 4: 860–865. pmid:10985655
  57. 57. Farmer P, Robin S, Ramilus SL, Kim JY (1991) Tuberculosis, poverty, and "compliance": lessons from rural Haiti. Semin Respir Infect 6: 254–260. pmid:1810004
  58. 58. Jahnavi G, Sudha CH (2010) Randomised controlled trial of food supplements in patients with newly diagnosed tuberculosis and wasting. Singapore Med J 51: 957–962. pmid:21221502
  59. 59. Macq J, Solis A, Martinez G, Martiny P (2008) Tackling tuberculosis patients' internalized social stigma through patient centred care: an intervention study in rural Nicaragua. BMC Public Health 8: 154. pmid:18466604
  60. 60. Martins N, Morris P, Kelly PM (2009) Food incentives to improve completion of tuberculosis treatment: randomised controlled trial in Dili, Timor-Leste. Bmj 339: b4248.
  61. 61. Morisky DE, Malotte CK, Choi P, Davidson P, Rigler S, Sugland B, et al. (1990) A patient education program to improve adherence rates with antituberculosis drug regimens. Health Educ Q 17: 253–267. pmid:2228629
  62. 62. Sripad A, Castedo J, Danford N, Zaha R, Freile C (2014) Effects of Ecuador's national monetary incentive program on adherence to treatment for drug-resistant tuberculosis. Int J Tuberc Lung Dis 18: 44–48. pmid:24365551
  63. 63. Wei X, Zou G, Yin J, Walley J, Yang H, Kliner M, et al. (2012) Providing financial incentives to rural-to-urban tuberculosis migrants in Shanghai: an intervention study. Infect Dis Poverty 1: 9. pmid:23849348
  64. 64. Zou G, Wei X, Witter S, Yin J, Walley J, Liu S, et al. (2013) Incremental cost-effectiveness of improving treatment results among migrant tuberculosis patients in Shanghai. Int J Tuberc Lung Dis 17: 1056–1064. pmid:23827030
  65. 65. (2015) The World Bank Group. Available: http://data.worldbank.org/country.
  66. 66. Gelmanova IY, Taran DV, Mishustin SP, Golubkov AA, Solovyova AV, Keshavjee S (2011) 'Sputnik': a programmatic approach to improve tuberculosis treatment adherence and outcome among defaulters. Int J Tuberc Lung Dis 15: 1373–1379. pmid:22283898
  67. 67. Drabo M, Zerbo R, Berthe A, Ouedrago L, Konfe S, Mugishe E, et al. (2009) [Community involvement in tuberculosis care in three rural health districts of Burkina Faso]. Sante Publique 21: 485–497. pmid:20229641
  68. 68. Soares EC, Vollmer WM, Cavalcante SC, Pacheco AG, Saraceni V, Silva JS, et al. (2013) Tuberculosis control in a socially vulnerable area: a community intervention beyond DOT in a Brazilian favela. Int J Tuberc Lung Dis 17: 1581–1586. pmid:24200272
  69. 69. Balshem H, Helfand M, Schunemann HJ, Oxman AD, Kunz R, Brozek J, et al. (2011) GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol 64: 401–406. pmid:21208779
  70. 70. Guyatt GH, Oxman AD, Vist G, Kunz R, Brozek J, Alonso-Coello P, et al. (2011) GRADE guidelines: 4. Rating the quality of evidence—study limitations (risk of bias). J Clin Epidemiol 64: 407–415. pmid:21247734
  71. 71. Guyatt GH, Oxman AD, Montori V, Vist G, Kunz R, Brozek J, et al. (2011) GRADE guidelines: 5. Rating the quality of evidence—publication bias. J Clin Epidemiol 64: 1277–1282. pmid:21802904
  72. 72. Guyatt GH, Oxman AD, Kunz R, Brozek J, Alonso-Coello P, Rind D, et al. (2011) GRADE guidelines 6. Rating the quality of evidence—imprecision. J Clin Epidemiol 64: 1283–1293. pmid:21839614
  73. 73. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand M, et al. (2011) GRADE guidelines: 7. Rating the quality of evidence—inconsistency. J Clin Epidemiol 64: 1294–1302. pmid:21803546
  74. 74. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand M, et al. (2011) GRADE guidelines: 8. Rating the quality of evidence—indirectness. J Clin Epidemiol 64: 1303–1310. pmid:21802903
  75. 75. Reeves B, Deeks J, Higgins J, Wells G (Updated September 2008) Chapter 13: Including non-randomized studies. In: Higgins JPT, Green S (editors), Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1: The Cochrane Collaboration.
  76. 76. Wingfield T, Boccia D, Tovar M, Gavino A, Zevallos K, Montoya R, et al. (2014) Defining Catastrophic Costs and Comparing Their Importance for Adverse Tuberculosis Outcome with Multi-Drug Resistance: A Prospective Cohort Study, Peru. PLoS Med 11: e1001675. pmid:25025331
  77. 77. Lutge EE, Wiysonge CS, Knight SE, Sinclair D, Volmink J (2015) Incentives and enablers to improve adherence in tuberculosis. Cochrane Database Syst Rev 9: Cd007952. pmid:26333525
  78. 78. Sinclair D, Abba K, Grobler L, Sudarsanam TD (2011) Nutritional supplements for people being treated for active tuberculosis. Cochrane Database Syst Rev: Cd006086. pmid:22071828
  79. 79. Toczek A, Cox H, du Cros P, Cooke G, Ford N (2013) Strategies for reducing treatment default in drug-resistant tuberculosis: systematic review and meta-analysis. Int J Tuberc Lung Dis 17: 299–307. pmid:23211716