Patient support for tuberculosis patients in low-incidence countries: A systematic review

Background Patient support during tuberculosis treatment is expected to be more often available and more customized in low tuberculosis incidence, high-resource settings than in lower-resource settings. The aim of this systematic review is to provide an overview of tuberculosis patient support interventions implemented in low-incidence countries and an evaluation of their effects on treatment-related outcomes as well as their acceptability by patients and providers. Methods PubMed, Social Science Citation Index and Cumulative Index to Nursing and Allied Health and Literature were searched for the period 01.2006–05.2016 on publications describing tuberculosis patient support interventions in low-incidence countries (<20 patients per 100,000 population). Results Through our search strategy, 1875 unique publications were identified. Forty publications were included: 17 evaluated patient support quantitatively, 9 qualitatively and 14 only described the patient support. Nineteen publications assessed treatment supervision options only, 21 assessed (combinations of) treatment supervision, socio-economic, psycho-emotional, health-educational and other support. Of eight studies quantitatively evaluating the effects of support with a control group, four showed positive effects: two out of three that used combinations of patient support and two out of five that compared treatment supervision options. Heterogeneity of interventions precluded pooling of results. Qualitative and descriptive studies showed that patients appreciated individualized support including treatment supervision, psycho-emotional and socio-economic support; and digital health interventions. Conclusion Our review shows that a variety of patient support interventions is implemented in low-incidence countries. Although only a few interventions were evaluated quantitatively, we identified potential best practices. The scarcity of evidence on effectiveness, however, indicates the need for further research to evaluate potential best practices.


Introduction
Ensuring adherence to anti-tuberculosis (TB) treatment is a challenge, not only in high TB burden countries but also in countries with a low TB incidence and well-financed health care systems. Addressing barriers to TB diagnosis and treatment adherence for vulnerable and hard-to-reach groups are priority action areas for low-incidence countries progressing towards TB elimination. Detection of TB is high in most low-incidence countries through good access to care and enhanced case finding among risk groups [1]. Adherence to TB treatment, however, remains a challenge given the long treatment duration with multiple antibiotics; at least six months for drug-susceptible TB and 18-24 months for multidrug-resistant TB (MDR-TB), i.e. TB resistant to at least the two key first-line anti-TB drugs isoniazid and rifampicin [2,3]. Adverse drug reactions, early improvement of symptoms during treatment and socio-economic difficulties of the patient contribute to non-adherence to treatment, even in countries with relatively well-financed health care systems [4][5][6][7][8][9][10][11]. Interventions to ensure TB treatment adherence are not only essential in increasing the chance of cure but also in reducing the risk of further TB transmission and of drug resistance development [12][13][14][15].
There are various forms of patient support interventions to alleviate or remove barriers to treatment adherence including use of treatment supervision, also known as directly observed therapy (DOT), and other treatment administration support forms such as provision of pill boxes; health education; socio-economic support; and psycho-emotional support. Patient support interventions have been systematically reviewed and evaluated, but not yet specifically focussing on low-incidence countries [16][17][18][19][20][21]. Most of the studies reviewed, focused on high TB burden countries with limited financial resources or only on special risk groups. Patient support in low-incidence countries, however, may differ from support in high-incidence countries as in the former setting there usually are less patients and more resources. Patient support in high-resource countries may be more tailored to the individual patient's needs and make more use of digital innovations, such as video observed treatment (VOT) [1,7]. In VOT medication intake is monitored through videos recorded by the patients or through video calls. In high-incidence countries interventions need to be provided to a larger number of patients representing a broader spectrum of the society, while resources are more limited.
In the Netherlands, all TB patients are entitled to support by a specialized TB nurse. This support may comprise health education, counselling, incentives and enablers. TB nurses also act as a case manager for the patient and coordinate the organization of treatment supervision, further socio-economic and psycho-emotional support. DOT can be provided by the TB nurse and/or a selected third party such as home nursing services. TB nurses determine the nature and intensity of support individually per patient based on an assessment at the start of treatment. For this assessment, the TB nurses interview patients in a structured way on diseaserelated factors, such as symptoms and co-morbidities, as well as on socio-economic and psycho-social factors. In the Netherlands, these factors are regarded essential for successful treatment. The patient should be well-informed (empowered), motivated to complete TB treatment, should be in financial and social stable living conditions, and side-effects and comorbidities should be managed [22,23].
Treatment results in the Netherlands are satisfactory with 88% successful treatment outcome for rifampicin-sensitive TB for the years 2009-2013 [24]. As current patients support practices are not evidence-based, however, the question has been raised as to what are the most effective and efficient means of providing TB patient support. To develop an evidencebased Dutch guideline on patient support which may also be informative for other low-incidence countries, KNCV Tuberculosis Foundation initiated the project 'Improving Patient Support Interventions'. This project includes a systematic review on evidence for patient support interventions in low-incidence countries, a European survey on countries' patient support policies, and a qualitative study capturing the current Dutch practices.

Objective
The aim of this review is to provide an overview of published TB patient support interventions in low-incidence countries, including their effectiveness in improving treatment adherence as well as their acceptability by patients and providers.

Eligibility criteria
As the aim of our systematic review was to identify evidence on patient support most relevant to low TB incidence settings such as the Netherlands, we only included publications describing TB patient support interventions in low-incidence countries [25]. Low incidence was defined as less than 20 per 100,000 population in line with the definitions of the European Centres of Disease Control [26]. Publications from all countries that had reached low TB incidence by 2014 were included [27]. Patient support interventions were defined as any intervention aiming to increase treatment adherence. Studies were excluded if they only compared adherence under different regimen options such as daily vs. thrice weekly doses. All publications measuring any qualitative or quantitative outcomes of patient support were included. Additionally, we included publications describing patient support as an ancillary intervention to the primary intervention under study, and the patient support component not being evaluated. Only articles published since 2006 were included as we aimed to capture the impact of recent evidence and practices, e.g. utilizing online interventions. We included publications in Dutch, English, German, Portuguese, Spanish, Italian and Russian.
Search strategy. The bibliographic database PubMed was searched on April 15 th, 2016 and the databases Cumulative Index to Nursing and Allied Health and Literature (CINAHL) and Social Science Citation Index (SSCI) were searched on May 2 nd, 2016 to identify relevant publications on patient support interventions. Search terms were developed by the authors and included combinations of three domains: (i) "tuberculosis" and related terms, (ii) "patient support interventions" and related terms and (iii) "treatment adherence" and related terms. A complete list of search terms is provided in S1 Appendix. Reference lists of relevant systematic reviews and of the included publications were scanned to identify additional publications.
Publications eligible for inclusion were selected in three steps: (i) screening of titles, (ii) assessment of abstracts and (iii) assessment of full texts. Titles and abstracts were assessed on eligibility independently by two researchers (SB and SH). Full text selection and review of reference lists was performed by one researcher (SB) and verified by a second reviewer (SH). Discrepancies between reviewers were resolved by discussion.
Data collection and analysis. Characteristics of patient support interventions were entered into a pre-piloted form. Extracted characteristics included the country where the intervention was provided, a description of the study design including, on availability, a description of the intervention and the control group. We divided patient support interventions into five categories: Treatment supervision, treatment administration support other than DOT, health educational (HE) support, socio-economic (SE) support and psycho-emotional (PE) support. These categories were defined as follows: • Treatment supervision: Direct observation of medication intake by any person, at any location in any frequency, also known as directly observed therapy (DOT) • Treatment administration support other than DOT: Tracer, pill boxes and other measures other than DOT to promote adherence • HE support: Presentation of information to the patients aiming to educate them about their disease and its' treatment as well as related health issues • SE support: Provision of (social) services, material goods and/ or financial support • PE support: Interventions focusing on the patients' feelings, emotions or social relationships and social support The following information was compiled on the interventions described in the publications: intervention categories included, reasons for patient support, patients eligible for the support, support provider(s), level of implementation (e.g. national/regional), duration of patient support provision, and experiences with provision. If available, also quantitative treatment adherence-related outcomes of the intervention were extracted. Treatment adherence-related outcomes as defined and calculated in the respective studies were adopted without changes.
Risk of bias in individual studies and quality of evidence. Risk of bias for studies quantitatively evaluating effects of patient support interventions against a comparison group was assessed using The Cochrane Collaboration's Tool for Randomized Controlled Trials (RCT) for RCT and Newcastle Ottawa Scale (NOS) for Non-Randomized Studies (NRS). For NRS <10% of subjects lost was considered as indicative of low risk of attrition bias.
Summary measures and synthesis of results. Patient support interventions were described per category. Quantitative outcomes among patients receiving the interventions were described and, if applicable, compared to those among a comparison group of patients. Dichotomous outcomes were described using risk ratios (RR) for cohort studies and controlled trials and odds ratios (OR) for case-control studies, both with corresponding 95% confidence intervals. If not provided in the publication, ratios were calculated from the data provided in the publications, if possible. For non-dichotomous outcomes, absolute and relative differences (d r ) were calculated. Ratios and relative differences were calculated using Microsoft Excel (2013). ethnography [60], a focused ethnography [61], and an interpretive phenomenology [62]); 3 experimental studies (RCTs [63,64]), and a non-randomized controlled trial [65]) and 1 quasiexperimental study (historical before-after comparison) [66].

Quantitative evidence on patient support described
Patient support is quantitatively evaluated in 17 studies; in 8 with comparison to a control group and in 9 without comparison to a control group.
Studies quantitatively evaluating patient support compared to a control group. Of the eight studies that compared a patient support intervention with a control group [30,35,39,48,[63][64][65][66], five evaluated different forms of treatment supervision options [35,39,48,[63][64][65][66], one evaluated treatment supervision combined with reminders [65]; and two evaluated provision of HE support combined with reminders [63,66]. In four studies, interventions were associated with a significant improvement in treatment adherence. Patients whose treatment was supervised by peers had a lower risk of treatment failure compared to patients whose treatment was supervised by health workers (RR = 1.40 [CI: 1.08-1.82]) [64]. Patients whose treatment was supervised at a location convenient to them were more likely to successfully complete treatment compared to patients on self-administered treatment (SAT) (RR = 1.14 [CI: 1.07-1.22]) [35]. Patients who received HE support combined with reminders were more likely to attend 100% of the follow-up visits and to complete treatment, respectively, compared to patients receiving routine care (RR = 1.83 [CI: 1.1-2.9]; RR = 1.16 [CI: 1.11-1.20]) [63,66]. In the other four studies [30,39,48,65], patient support interventions had no statistically significant effect on treatment adherence. In one of these studies, treatment outcomes were not influenced by the type of DOT provider, i.e. health workers versus other providers [48]. In two of these studies, VOT was not superior compared to in-person DOT in terms of treatment completion (RR = 1. 49 [30,48]. An overview of the outcomes of the interventions is provided in Table 2. Studies quantitatively evaluating patient support without comparison to a control group. Of the nine studies that quantitatively evaluated patient support without comparison to a control group [40,41,[43][44][45][49][50][51]68], seven evaluated different forms of DOT [40,41,44,45,49,50,68], one evaluated SE support [43] and one combined HE support with PE support and reminders [51] (Table 3). Under the different forms of DOT, treatment Patient support for tuberculosis patients in low-incidence countries: A systematic review  3 Adjusted OR provided in the publication 4 OR provided in the publication 5 Adjusted OR provided in the publication 6 Adjusted RR provided in the publication https://doi.org/10.1371/journal.pone.0205433.t002 Table 3. Quantitative outcomes of tuberculosis patient support interventions in low incidence countries described in studies not allowing for comparisons to a control group. Patient support for tuberculosis patients in low-incidence countries: A systematic review Blogging about MDR-TB treatment Patients mentioned blogging about MDR-TB treatment was supportive for adherence, considered blogging a tool to receive and provide peer support, a platform to express themselves and mean to record their achievement of which they can be proud; Stakeholders considered blogging a tool to provide treatment support to patients and to empower patients; Project staff and stakeholders considered blogging a tool to enhance patient practitioner relationships and to improve the understanding of the patient's experience with the disease; One blogger mentioned expectations of financial gain as a result of blogging (Continued) completion/ cure rates ranged from 70% to 89%. Treatment completion was 78% in one study in which SE support was provided to patients [43] and 61% in one study in which HE support combined with PE support and reminders was provided to patients [51].

Acceptability of patient support described
Thirteen studies qualitatively described how acceptable patient support was to patients and providers (Table 4). Patients appreciated individualized support by nurses [56,60,62] which included DOT at the patients home [60,62] or the health care centre [56] but also various forms of psycho-emotional support, such as nurses providing food and building rapport.
Interventions appreciated by patients as well as providers included VOT [39,44,65], a "TB link worker" [43], blogging about TB [55], and Somali health care workers providing support, together with specialized TB nurses, for people of Somali origin living in the UK [61]. Patients appreciated VOT due to its convenience, privacy and flexibility [39,65]. Providers were satisfied with VOT due to travel time saved [44] and the ability to easily assess patients repeatedly [39]. Some obstacles were encountered by providers, which were mainly of a technical nature

Study Intervention Outcomes
Kawatsu et al., 2013 [56], Japan Nurses provide DOT at the public health centre, and food and drinks when patients come for DOT; Nurses consult social welfare offices and other organizations; Nurses build rapport, address concerns and organize congratulation ceremony for successfully completed treatment Patient empowerment was achieved comprising the fulfilment of emotional needs, improved mental health, improved health behaviour, improvement of living environment, improved interpersonal relationships and improved attitudes towards society Mtui and Spence, 2014 [57], UK DOT is provided by the general practitioner; TB nurses provide DOT in pharmacies for substance-abusers, on the streets / public bars by TB nurses for homeless patients; Nurses talk about TB and provide leaflets; Nurses assist in accessing social care while delivering DOT, provide incentives for some cases and bring people to the clinic; Nurses build rapport with patients, support in coping with the treatment and perform home visits Nurses reported that lengthy time for travel, long duration of visits, and high number in receipt of DOT were challenging; One National Health Service board reported threats to nurses from people known to patients while visiting; One National Health Service board reported problematic treatment adherence with the immigrant population due to fears of deportation Synchronous VOT Patients valued the convenience, flexibility and reliability and privacy of VOT, developed rapport with the nurses via VOT and found the technology was easy to use; Nurses reported that many more patients could be seen in a shift than with a drive-around service, more convenient scheduling was regarded as improving patient adherence, absent patients could be readily called back repeatedly and patients who had difficulty taking all their tablets at once could be called in stages; Frustrating, substantial and ongoing problems with video call quality were reported as well as the potential to not swallow the pills correctly Patient support for tuberculosis patients in low-incidence countries: A systematic review and patient-related challenges such as patients not following observation protocols [39,65].
The "TB link-worker" was generally appreciated by patients as they had to go only to one "stop" for both social and emotional support. After the introduction of the TB link worker, 88% of the patients achieved jointly set goals such as temporary housing and secured income. Stakeholders appreciated additional time and increased information and knowledge exchange with the clinical teams [43]. Blogging was perceived, both by patients and providers, as a tool to empower patients [55]. For both patients and providers challenges were reported related to DOT provision. Lengthy travel time was mentioned as an obstacle for patients when DOT was provided in the health facility [54] and for providers when DOT was provided at the patient's home or the community [39,44,57,65]. Also long duration of the home-DOT visits themselves and threats from the vicinity while visiting patients were reported to be obstacles for HCW [57]. Homebased DOT by a general nursing service implemented in Norway was perceived as humiliating and discriminating by some patients. They felt that there was little room to negotiate whether they consent to DOT. Some patients perceived inflexible DOT appointments as frustrating as it restricted their daily activities. Frequent changes in health care workers proving DOT hampered the establishment of a trustful relationship between the patients and providers [58].
Risk of bias within studies. Risk of bias was assessed for RCT (n = 2) [63,64] as well as for NRS with a control arm (n = 6) [30,35,39,48,65,66]. For both RCT there was an unclear risk of bias, i.e. insufficient information provided to determine the risk of bias, in more than one key domain (S3 Appendix). For all NRS a risk of bias was identified in one or more domains (S4 Appendix, S5 Appendix).

Discussion
Our systematic review included 40 publications describing a variety of support interventions for TB patients in low-incidence countries; interventions that we categorized into DOT, SE support, PE support, HE support and other support. Only eight studies compared the intervention with a control group: six treatment supervision options (with reminders) and two HE support with reminders. In four of these studies the intervention(s) significantly increased treatment adherence and completion. Providing health education and reminders for medication intake and appointments [63,66], DOT by peers [64] and DOT at any location convenient to the patient [35] improved treatment adherence. DOT at the health care centre or the dispensary [48] and VOT [39,65] did not improve adherence. Qualitative and descriptive evaluations of patients' experiences with treatment support showed that patients appreciate individualized support by nurses [56,60,62] or a "TB link worker" [43] including DOT, psycho-emotional as well as socio-economic support, and digital health interventions such as VOT as opposed to traditional forms of DOT [39,44,65] and blogging about TB [55]. These interventions may be beneficial in TB patient support in low-incidence countries.
Based on the evidence found in this review, combining health education and appointment/ medication reminders might be considered a best practice in TB patient support. The combined effect of health education and reminders has not been systematically reviewed yet. The importance of patient education in TB case management, however, was highlighted in another recent systematic review [69]; the success rate of treatment as well as the confirmed cure rate was significantly higher in the group receiving an intensive triad model (health education combined with support) compared to the control group. A review on reminder systems showed that appointment-reminders had a significant effect on treatment success and can be a valuable addition to other interventions [18]. A systematic review on digital technologies in TB treatment found that SMS reminders had no statistically significant effect on treatment adherence, while electronic medication monitors, i.e. medication boxes that record when the box is opened, increased the probability of cure in one observational study and reduced missed treatment doses in one trial [70].
Individualized support by nurses including PE support, SE support and possibly DOT might also be a beneficial intervention in low-incidence countries. An important finding from our Europe-wide survey among policy makers and nurse representatives also was that appointing a TB nurse to coordinate TB patient support (TB case management) may be considered a best practice in European low-incidence countries [71]. A systematic review on SE support and PE support of various forms found that most studies were conducted in in high-incidence settings, and that both SE and PE support improved treatment results [16]. SE support comprised direct enablers (food supplements, dietary advice, travel reimbursement, vouchers for local shops, clothing and hygiene kits), incentives (financial support, board games, newspapers and household goods) as well as legal support and assistance in procuring documentation for access to health care and social services. PE support included self-help groups, counselling, home visits, community groups and psychotherapy combined. Studies conducted in America before 2006 show that incentives and enablers, such as housing, monetary support, grocery coupons and transport reimbursement, improve treatment adherence [72][73][74][75].
Ideally, TB case management should be tailored to the individual patient and aligned with his or her expectations, i.e. patient-centred. Patient-centred care is much recommended in TB care in both low-and high-incidence countries [76]. A patient-centred approach includes exploring the patients' needs for information, emotional needs, and life issues, finding common ground concerning the nature and the management of disease, enhancing prevention and health promotion as well as the relationship between the patient and the provider [77]. Patients receiving support as described in the publications included in this systematic review were frequently patients belonging to hard-to-reach groups such as immigrants, substance abusers, homeless and prisoners. A risk-group-focussed approach has been recommended by the WHO for countries progressing towards elimination [1]. The type and intensity of support needed is, however, not only dependent on the presence or absence of certain risk factors but also on individual patient characteristics [6]. Even within risk groups the need for support may differ between individuals and it is not one size fits all. In general, to allow a patient-centred approach, the patient's needs are required to be assessed before implementation of treatment and support [77]. It may be essential to train health care workers in how to assess these needs and how to involve patients in the decision process.
Our review provided inconclusive evidence on the effectiveness of DOT in low-incidence countries but suggests that VOT might prioritized over traditional forms of DOT to improve patient and provider acceptance. A previous systematic review on DOT in both high and lowincidence settings concluded that TB cure and treatment completion did not improve substantially with DOT versus self-administered treatment [17]. Additionally, it showed that treatment completion or cure did not differ between patients who received DOT at home by family member or community health workers and patients who received DOT provided at the health facility by health workers [17]. In our review there was also no clear benefit of DOT for risk group patients. While peer DOT for substance abusers lead to positive effects on treatment adherence [64] while there was no correlation between DOT and treatment adherence in a population, in which DOT was given with priority to patients at risk of non-adherence [30]. A previous systematic review by Heuvelings et al. (2017) on TB treatment adherence interventions in hard-to reach populations in countries with low and medium TB incidence similarly concluded that DOT may improve adherence in only certain risk populations including homeless populations as well as migrants, prisoners and people living with HIV [20].
Similar to previously published literature [17,[78][79][80], our review found that provision of traditional forms of DOT is surrounded with challenges for patients and providers, indicating a need for innovations. Only two studies included in this review found that patients appreciate DOT [56,58]; DOT combined with SE-, PE-and HE support for homeless patients reported patient-empowerment [56]. Migrants experienced DOT as an expression of care, especially when they were living very isolated [58]. Several other publications included in this review reported obstacles related to DOT such as patients feeling humiliated because DOT restricted their social life as they had to be at home during certain time slots for the DOT [58]. Also financial barriers through travel time and costs were reported, either for the patient when DOT is provided at the health facility or for the health care provider when DOT is provided at the patient's home [39,44,54,57,65]. A substantial body of literature has already described obstacles and ethical concerns related to conventional DOT and questioned its' effectiveness [17,[78][79][80], and the need for new more flexible forms of DOT approaches has been expressed [80]. Few approaches addressing these obstacles related to more conventional DOT are identified by this review: only VOT and the use of peers. Use of peers increased adherence, which may be explained by reduced social distance between patient and provider [64]. VOT is described as an alternative to reduce travel time and costs for both patients and providers and to increase privacy [39,44,65]. As VOT can be provided at any place, if there is internet connection, it may also be easier to incorporate it into the patients' everyday life [39,44,65]. The VOT studies conducted showed that adherence is similar to health facility-and home-based DOT [39,65]. The 2017 update of the WHO Guideline [76] for treatment of drug-susceptible TB and patient care states that VOT may replace DOT if the video communication technology is available and can be appropriately organized and operated by health care providers and patients. Further research will be needed to determine how VOT can best be combined with other treatment support resulting in satisfaction by both patients and health care workers, and in a positive treatment result.
Integrating digital health interventions in TB patients support may be beneficial for patients and providers, especially in low TB incidence countries. Both interventions identified in our review, VOT and blogging about MDR-TB, were appreciated by patients as well as providers. A recently published systematic review on the impact of digital health technologies on TB treatment showed that provision of SMS reminders and electronic medication monitors until now have been evaluated only in high-incidence countries [70] The implementation and evaluations of digital health in TB patient support interventions in low-incidence countries yet appear to be scarce. but. Especially for low-incidence, high-resource countries it may be useful to conduct further research on the value of digital health interventions in TB patient support. Digital health is expected to be increasingly implemented in TB patient support as information and communication technology is becoming more widely available and affordable, especially in high-income countries [11,81]. Digital health could be supportive to the need for more flexibility in patient support in general and in treatment supervision specifically. In addition to VOT, there are substantially more opportunities to use digital health solutions in patient support interventions such as a one-stop internet hub that links up to different services of relevance to TB care and health education [11]. These innovations may contribute to making support not only more flexible and more patient-friendly but possibly also more efficient and future-proof taking into account the decline of both patients and specialized TB health care providers in low incidence countries.
There were some limitations to this study. The literature review yielded only a small number of studies, especially studies including comparative quantitative outcomes of patient support interventions were scarce. Even fewer studies measure clinical outcomes such as cure and treatment failure which would be of greater importance to patient and providers compared to adherence measures only. Further, the studies were of low quality and studies and interventions described in this review were highly variable due to the design of the review aiming to provide an overview of patient support interventions implemented. Due to the small number of studies measuring quantitative outcomes and a large variation in interventions applied, outcomes measured and study populations among these studies could not be quantitatively synthesized and analysed.
Although the finding of this study are insufficient to provide recommendations on effective patients support in low incidence countries, our study identified some interventions that were effective and/ or appreciated by patients which would be useful to subject to further research. Furthermore, the framework we applied, with different patient support intervention categories (DOT, HE support, SE support, PE support and other), provides the opportunity to assess for the individual patients more systematically what categories of support are needed and to decide on the level of support e.g. on a scale of 1 (low), 2 (medium) and 3 (high). A standardised approach of defining the type and level of support to be provided will allow evaluating the complex nature of patient support. Further research is needed to determine for low-incidence countries how patients can best be involved in the decision-making process and to determine which support is most effective and efficient for these countries. This should be done taking into account the individual needs of the patients, risk group policy, the resources of the health care system and digital innovations.

Conclusion
We provided an overview of support interventions currently implemented for TB patients in low-incidence countries and the evidence on its effects on treatment-related outcomes. Our review identified limited published evidence on effectiveness of patient support interventions in low-incidence countries, although we did identify a number of successful patient support interventions and possible best practices. Further research will be needed to verify these best practices and to determine which patient support is most effective and efficient.