Barriers to initiating tuberculosis treatment in sub-Saharan Africa: a systematic review focused on children and youth

ABSTRACT Background: Tuberculosis (TB) is the deadliest infectious disease globally, with 10.4 million people infected and more than 1.8 million deaths in 2015. TB is a preventable, treatable, and curable disease, yet there are numerous barriers to initiating treatment. These barriers to treatment are exacerbated in low-resource settings and may be compounded by factors related to childhood. Objective: Timely initiation of tuberculosis (TB) treatment is critical to reducing disease transmission and improving patient outcomes. The aim of this paper is to describe patient- and system-level barriers to TB treatment initiation specifically for children and youth in sub-Saharan Africa through systematic review of the literature. Design: This review was conducted in October 2015 in accordance with preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. Six databases were searched to identify studies where primary or secondary objectives were related to barriers to TB treatment initiation and which included children or youth 0–24 years of age. Results: A total of 1490 manuscripts met screening criteria; 152 met criteria for full-text review and 47 for analysis. Patient-level barriers included limited knowledge, attitudes and beliefs regarding TB, and economic burdens. System-level barriers included centralization of services, health system delays, and geographical access to healthcare. Of the 47 studies included, 7 evaluated cost, 19 health-seeking behaviors, and 29 health system infrastructure. Only 4 studies primarily assessed pediatric cohorts yet all 47 studies were inclusive of children. Conclusions: Recognizing and removing barriers to treatment initiation for pediatric TB in sub-Saharan Africa are critical. Both patient- and system-level barriers must be better researched in order to improve patient outcomes.


Background
Mycobacterium tuberculosis (TB) is the leading infectious cause of death worldwide, surpassing HIV/A.I. D.S [1]. In 2015, TB killed 1.8 million people with 95% of cases and deaths in developing countries [1]. TB is an airborne infectious agent requiring at minimum an intensive six-month medication regimen for bacteriologic cure [1]. Timely initiation and correct treatment of TB are critical to reduce disease transmission and improve patient outcomes. However, barriers to treatment initiation exist at both the patient and system levels. Patient-level barriers such as perception of illness, stigma, knowledge about TB, delay in seeking care and initiating treatment, and direct and indirect costs all cause delayed treatment [2,3]. Health system barriers include resource capacity such as the availability of laboratory tests, accessibility of different levels of care, and costs. Patient costs associated with TB treatment often cause patients and families to fall into a 'medical poverty trap' [4,5].
Pulmonary TB outcomes in children are favorable when treated; however, data are limited regarding outcomes for children (0-18 years) and youth (15-24 years) [6][7][8][9][10]. This research gap is in part due to lack of standardized definitions of age cohorts (i.e. pediatric, child, adolescent, youth) as well as lack of political and community commitment to this age group [11]. Barriers to involving youth in research, coupled with developmental transitions, additional responsibilities associated with education and employment, and dependence on family commitment, may cause youth to be understudied [12].
Pediatric cases (0-18 years) account for 10% of all new and relapse cases of TB in the African region, as compared to 6.5% globally [1]. Additionally, the African region has the highest rate of TB in children and youth compared to any other region [1]. Despite the high burden of disease among younger age groups, barriers are most often studied in adult populations. As a result, barriers to treatment initiation in children and youth are less well understood [12,13]. The objective of this review is to determine patient-and system-level barriers to treatment initiation in sub-Saharan Africa (SSA) with an emphasis on children and youth diagnosed with TB, through systematic review of the literature.

Search strategy
We systematically searched six health databases for literature pertaining to pediatric and youth TB in SSA. Specified terms were agreed upon by the authors and adapted to each database. The protocol is provided in Supplement 1 and full search criteria for each database are provided in Supplement 2.

Selection criteria
Inclusion criteria were established by the authors a priori based upon preliminary literature searches. Preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines were followed for inclusion and exclusion ( Figure 1). Articles were included through 26 October 2015. All potentially eligible original research studies with abstracts in English were reviewed. Inclusion criteria were: a primary or secondary aim of the study addressing barriers to TB treatment initiation for children or youth in SSA. Definitions of barriers, age groups, treatment initiation, and other variables are provided in Table 1. All publication dates and all study designs were included.

Data extraction and analysis
All search results were entered into EndNote X7 (Thomson Reuters Scientific Inc., Carlsbad, CA, USA) and duplicates were removed. Two authors (BJS and BEE) reviewed titles and abstracts with the aim of removing publications that were unlikely to meet the inclusion criteria. After both authors reviewed 20% of articles and demonstrated greater than 95% inter-rater reliability, the remaining articles were assessed independently. Discrepancies were resolved through discussion with all three authors until agreement was reached.
Garrard's matrix strategy for abstracting was used to build a Microsoft Excel spreadsheet and abstract full texts of the remaining articles [16]. The following data were extracted from the included articles: author, publication year, title, journal, country and setting, study aim and design, outcomes and barriers measured, sample size, covariates, strengths and limitations of each study, 95% confidence intervals and effect sizes (if reported), if HIV was measured as a variable, and if pediatric and youth cases were specifically described in the findings or discussion.

Study selection
After removing duplicates, a total of 1490 articles were included. There were 1338 articles excluded and 47 articles were retained for full-text review ( Figure 1). The 47 studies were from 14 countries with 16 studies conducted in South Africa (Table 2). While all 47 studies included children or youth in their sample, only 4 studies were pediatric-only cohorts [17][18][19][20]. Barriers were identified as either patient-or system-level (or both), and whether the barrier related to (1) cost, (2) infrastructure, and/or (3) health-seeking behavior ( Table 2).
Cost as a barrier to TB treatment initiation Seven articles identified cost as a barrier to TB treatment initiation. Costs included: direct, indirect, system, or caregiver costs; as well as costs incurred prior to diagnosis and between diagnosis and treatment initiation (Supplement 3) [37,40,41,45,54,58,59]. All articles identifying costs cited patient-level costs as a barrier, while only one article considered cost as a barrier from the health system perspective [37]. All studies described direct out-of-pocket costs for patients. Five studies described indirect costs [37,40,41,54,58]. One study reported the median total of direct and indirect costs was equivalent to 45% of median annual individual incomes ($350 U.S.D) with indirect costs accounting for 85% of total costs [41].
Two articles described costs associated with caretakers in which the cost burden to guardians or caretakers was high [40,41]. In addition, one study also measured intangible costs (non-monetary costs affecting quality of life, such as pain, suffering, and social stigma) in addition to catastrophic expenditures [54]. Another study measured transaction costs, the difference between total direct costs incurred by a patient and total direct costs incurred prior to the first contact with a National TB Control Program provider [59]. In one of the only intervention studies, Abimbola et al. (2015) cited that interventions for reducing transaction costs should include effective decentralization of services to integrate TB care with primary healthcare. Studies suggested that encouraging the engagement of communities to help address health education and facilitation of referral linkages among formal and informal care providers may reduce costs. None of the cost studies specifically addressed pediatric or youth barriers beyond those of the general population.
Additionally, initial loss to follow-up was considered a structural barrier. When a patient is diagnosed but never initiates TB treatment, often it is the health system's failure to report diagnostic results or make timely follow-up, rather than a patient's unwillingness to start therapy, that is responsible. However, a mixture of lost laboratory results, long wait times between sputum collection and final culture results, and the number of providers and technicians handling the specimen can all contribute to this initial loss to follow-up [33]. Seven studies specifically measured patients with initial loss to follow-up, with rates as high as 40% [32][33][34]47,49,53,63]. No article described loss to follow-up as a function of patient age; however, Yassin et al. (2013) disaggregated symptomatic patients and smear positive pulmonary TB patients by age and sex for their community-based TB intervention, which benefitted women, children, and vulnerable groups the most [49].
The three studies of child-only cohorts addressed delays for children and the complexity and value of culture confirmation for children with TB. Waiting for culture confirmation prior to treatment can greatly increase delays in children compared to initiating treatment when a clinical diagnosis is made (median 1 day with clinical diagnosis versus median 40 days with culture diagnosis) [18]. Additionally, treatment is delayed in children when an adult source is unknown (median 58 days when adult source is known versus 123 days without a source) [19].
Follow-up in urban squatter communities was difficult, causing children in these locations to have significantly worse tracing than children in urban settled areas, rural agriculture areas, and rural settled areas [17]. All of these child-only cohort studies were conducted in South Africa.

Health-seeking behavior as a barrier to TB treatment initiation
Health-seeking behavior is complex and is influenced by knowledge, attitudes, beliefs, and accessibility of care pathways. Nineteen articles examined healthseeking behaviors specific to barriers and the subsequent delays they may cause (Supplement 4) [20,21,[23][24][25][28][29][30]35,36,38,39,42,43,51,52,55,59,61]. These behaviors include seeking care from formal and informal sectors, private and public healthcare providers, traditional healers, pharmacies or drug retailers, and private clinics. All 19 articles describing health-seeking behaviors considered patient-level behaviors (i.e. patient knowledge, attitudes, behaviors, and decisions regarding care pathways). Only three studies also considered system-level factors, including stigmatization from healthcare providers as well as provider knowledge of TB acting as barriers against patients seeking care [25,28,36].
Patients' decisions of where to first seek care were influenced by numerous factors. These decisions were made by patients themselves, by close family members, or by healthcare workers [21,24]. Distance from home and mode of transport also was a factor in where to seek care, with closer proximity to home and walking distance preferred [21]. However, patients who visited health centers, private facilities, and health posts were more likely to experience delays compared to those who visited hospitals [57]. Eastwood and Hill (2004) reported that patients who consulted with pharmacies had diagnostic delays of around one month, while patients who consulted traditional healers had delays of several months. Many studies cited multiple care provider visits prior to diagnosis and appropriate treatment of TB with one study citing upwards of six facilities being visited [28]. In cases where the mother was the source of multi-drug resistant tuberculosis (MDR-TB) infection, children were almost four times less likely to receive MDR-TB care than if the mother was not the source of infection (OR 3.78, 95% CI 1.29-11.1, p = 0.02) [20]. Thus, children with close household contact to MDR-TB received delayed care.

Discussion
This review suggests that more research in younger populations is urgently needed related to barriers to TB treatment. Previous studies have described delayed treatment for children and youth with HIV or those with TB [8,64,65]. However, there is a paucity of research specifically pertaining to TB treatment initiation in youth with or without HIV [6,66]. More specifically, the epidemiology of drugresistant TB (DR-TB) and DR-TB/HIV coinfection and treatment in youth remains unclear. Children and youth are now priority populations in TB research [67]. Thus, there should be increased research specific to children and youth. Only four of the reviewed studies were specific to children and youth. These four studies focused on barriers to timely treatment initiation with a pediatric lens. Interestingly, most of the same barriers existed for adults and children with the exception of increased social, logistic, and cultural factors contributing to pediatric non-attendance at clinics including the mother also being ill, which likely would not affect adults as much as children [20].

Cost barriers
Both direct and indirect costs pose barriers to TB treatment initiation for individuals of all ages in SSA. Additionally, intangible costs and costs associated with caretaker burden influence when and where someone seeks TB treatment. Catastrophic medical expenses and poverty can delay, and potentially prevent, individuals from initiating TB treatment [4]. Therefore, community-based TB treatment initiation, and grants and other resources for transportation, nutrition, and financial services such as through National TB Programs, may enable earlier and more adequate TB treatment. Interestingly, cost barriers were not found to be reported until 2010, suggesting that economic studies have only recently become important to researchers in SSA.
Costs unique to child and youth TB treatment initiation must be studied more specifically. For example, pediatric drug formulations are often more expensive than standard adult formulations, which may create additional financial barriers for either families if paying out of pocket or health systems if provided by National TB Programs [68]. Additionally, intangible costs such as lost days of school for a child and missed days of work for a parent double the burden. No studies assessed caretaker costs among pediatric cohorts; two studies evaluating these intangible costs only included individuals 15 years and older [40,41]. In addition, pediatric TB specialists are often in more centralized or urban areas, creating a heavier financial burden for younger children who must travel further from home, and often must be accompanied by an adult.

Infrastructure barriers
Delays were observed in children from rural or farming areas, and when errors were made by treating physicians [17]. Additionally, lack of point-of-care laboratory capacity and the lack of ability to perform certain diagnostic testing (such as cultures and drug susceptibility testing) often caused treatment delays, especially for drug-resistant TB [18]. Thus, starting same-day empiric treatment prior to culture results can greatly reduce barriers to initiating treatment.
Improving health system infrastructure through integrating TB services into existing programs was found to be critical for all ages. Deficiencies of health systems were apparent across multiple settings; application of recommended TB/HIV integration programs should be utilized [69]. The level of care at which TB services were available influenced where patients first sought treatment, with centralized services often delaying or inhibiting treatment initiation [37,43]. Geographic barriers are common and limit access to treatment. Thus, developing diagnostic and treatment options in rural areas, and training healthcare providers on the signs and symptoms of TB, could remove some geographic barriers to TB treatment initiation [39,63]. Community-based interventions have been shown to be cost-effective, such as leveraging health extension workers to educate communities on sanitation and hygiene, debunking TB and HIV myths and lessening stigma, as well as in conducting screenings [34,37]. While the public sector is by no means a panacea for TB treatment initiation, it could actively engage and educate traditional practitioners, the private sector, and community partners in order to improve services to patients closer to their homes and at more affordable costs [70,71]. Overall, practical, effective policies to strengthen health systems can create enormous benefits in TB care for all ages, and children in particular.

Health-seeking behaviors
Educating both providers and patients about TB and the importance of timely, effective treatment can greatly improve outcomes. Although Edginton et al. (2005) found that knowledge of TB was good in 63% of patients, 51% of patients were unaware of the cause of TB. Thus, 'good' knowledge should be interpreted with caution, especially when assessing knowledge across studies using different measures. Similarly, two studies found TB stigma was associated with stigma against HIV in South Africa and Zambia, two high HIV prevalence countries [28,61]. Thus, country context also affected health-seeking behaviors. Specific to children, when the source of M. DR-TB was from the mother, clinic follow-up attendance was worse [20]. Providing patient-centered TB care is one way in which some barriers can be removed from accessing treatment initiation and improving clinic follow-up after diagnosis [20,72].
Chaotic and uncoordinated services can cause delays and increase costs for patients with TB. A coordinated National TB Program can streamline services, thus improving general health education, promoting TB prevention, and regulating healthcare providers in both the private and public sectors. On the other hand, decentralized care provided through community-based organizations may facilitate easier, more cost-effective care pathways [37]. Decreasing the number of providers that patients visit prior to an accurate diagnosis and effective treatment is one way in which to facilitate more timely treatment initiation. These strategies would have even greater impact upon child and youth cases, though more research in this area is needed.

Limitations
Due to the inclusion of all study designs, there was great heterogeneity of these studies, therefore neither a pooled analysis nor a meta-analysis was conducted and no summary measures (e.g. effect size) for specific interventions could be determined. Further, inconsistencies in cost measures were found to be a limitation, as well as differences across countries. Not all studies captured data in the same manner, nor used the same definitions or timeframes for analysis. However, most articles noted in their limitations the difficulty of getting accurate income data and verifying direct and indirect costs in low-income settings. Although we undertook a systematic literature search, some studies meeting our inclusion criteria may have been missed.

Conclusion
Many patient-and system-level barriers to TB treatment initiation exist among children and youth in SSA; however, through systematic review of the literature, these barriers are more fully described for adults. To our knowledge, no study has correlated barriers to treatment initiation with patient outcomes, and more evidence in this area could benefit TB preventionand could thereby save lives [73]. The specific needs of children and youth should be prioritized in research, particularly around enhanced infrastructure such as early diagnosis and treatment initiation and community-and patient-centered approaches. We recommend more standardized language to describe barriers to TB treatment initiation within the TB research and advocacy community, in order to allow for more unified, collective, and powerful action (Table 1) [72]. Addressing both patient-and system-level barriers is vital to improving patient outcomes, especially among young populations. age groups; however, this systematic review reveals that insufficient literature assesses barriers specific to children and youth. Therefore, this paper adds valuable insights into specific barriers to tuberculosis treatment initiation in sub-Saharan Africa, emphasizing the need for additional research in vulnerable pediatric populations.