Electronic databases (MEDLINE, EMBASE, and Cochrane Database of Systematic Reviews) were searched to identify peer-reviewed studies of mHealth interventions for TB. The systematic search was supplemented by reviewing relevant review papers identified from the initial search. The search strategy for the study population includes key terms describing LMICs such as “resource poor” or “developing country.” The search strategy for the mHealth intervention combined multiple keywords such as “mHealth” and “text-messaging.” As for the target disease, “tuberculosis,” “TB,” “multi-drug resistant tuberculosis,” and “MDR-TB” were used as search terms. No restrictions were applied to the publication type or publication date, but the language filter was applied to identify studies published in English. The search included articles published up to May 2018. The full search strategy is available in Multimedia Appendix 1.

Furthermore, 3 authors (ShL, YL, and SmL) independently reviewed the retrieved studies throughout the selection process. Each study identified from the databases was screened by 2 reviewers and then a full-text review was conducted for the potentially eligible studies. The disagreement on the selection process was resolved by the other authors who were not involved in the review of the specific study under discussion.

The inclusion criteria for this systematic review were as follows: First, studies conducted in the context of LMICs, as defined by the World Bank’s income cutoffs [16]; second, studies involving an intervention using mobile devices (ie, mHealth intervention); third, the target disease of the study should be TB or MDR-TB; fourth, studies designed to evaluate the effectiveness or benefits of mHealth interventions for TB (eg, observational study, mixed-methods study or implementation project, randomized controlled trial [RCT]); fifth, studies reporting more than one type of outcome; finally, only full-text studies published in English were considered eligible. In addition, the authors attempted to identify individual studies from reviews or systematic reviews, which were included in this study.

The qualitative synthesis of evidence focused on the outcome measures reported in each type of mHealth intervention. Information about the study setting, mHealth technologies used, target populations, and types of outcome measures was extracted. To classify the diverse types of detailed outcome measures, the following categories were used: (1) treatment outcome; (2) treatment outcome as defined by WHO; (3) adherence; (4) process measure; (5) perception; (6) technical outcome; (7) health outcome; (8) quality of life; (9) knowledge; (10) cost-effectiveness; (11) cost; (12) psychosocial outcome; and (13) mortality. Some explanations about these measures are provided in the next few paragraphs.

The treatment outcome includes any outcome measure that deals with the result of TB treatment, such as sputum smear conversion or microscopy test result. The treatment outcome following the WHO definition was separately categorized [17]. The WHO definition was developed to make a distinction for treatment outcomes between the drug-susceptible TB and drug-resistant TB, which are mutually exclusive groups. According to the WHO definition, any patient with TB should belong to either group and then 1 of the 7 treatment outcome cohorts: (1) cured; (2) treatment completed; (3) treatment failed; (4) died; (5) lost to follow-up; (6) not evaluated; and (7) treatment success. The WHO definitions for each of these 7 categories differ between the drug-susceptible TB and drug-resistant TB as described in Multimedia Appendix 2.

The adherence outcome includes medication adherence or treatment adherence. The process measure is any outcome measure related to treatment process, including the receipt of diagnostic test, attendance to appointments, or reporting of adverse events. The perception indicates any outcome measure that captures the user’s thoughts on mHealth for TB management. The technical outcome relates to the outcome measures that investigate the technical feasibility such as processing times or system installation. Health outcome, quality of life, knowledge (eg, patients’ understanding of the disease or the technology), cost-effectiveness (ie, the extent to which an alternative provides value for money), cost (ie, costs associated with an intervention from different perspectives), psychosocial outcome, and mortality outcomes are additional categories which are self-explanatory.

To evaluate the quality of individual studies included for our review, risk of bias was assessed with the existing tools. As this systematic review includes various types of studies, it is important to have a coherent set of quality assessment tools for different study designs. Therefore, we used the modified version of the Critical Appraisal Skills Program (CASP) that provides the checklists specific to various types of studies ranging from RCTs and qualitative studies to economic evaluation [18]. In case of mixed-methods studies whose CASP tool has not been developed yet, the quality assessment criteria for mixed-methods studies from the previous study were employed [19]. The quality assessment of studies is presented in Multimedia Appendix 3.

Among the 312 studies identified after removing duplicates, 260 articles were excluded during the screening process based on the titles and abstracts. Therefore, 52 articles were assessed for eligibility through a full-text review. Of those, 27 studies were included for the qualitative synthesis of evidence. A flow diagram for the selection process, based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, is provided in Figure 1 [20].

Table 1 presents the results of qualitative synthesis of the mHealth studies for TB management. Approximately, half (15 out of 27) of the studies were conducted in African countries. The study designs were diverse, including 6 RCTs, 5 mixed-methods studies, 1 cohort study, 3 qualitative studies, 4 observational studies, 6 implementation projects, and 2 economic evaluation studies. The types of mHealth technologies utilized were diverse as well: 12 studies employed SMS, 6 studies used mobile app, 5 studies used SMS plus other technology, 3 studies utilized phone calls, and only 1 study applied mHealth for mobile data collection. With regard to the study population, a majority (20 out of 27) of the studies were targeted for TB patients or TB/HIV patients, but there were a few studies that examined the experience of health care workers or general patients at the clinic for TB test results notification.

In terms of the outcome measures, there was a wide range of variants in their definitions even within each category of the outcomes. For instance, both Mohammed et al and Bediang et al defined treatment success as the primary outcome measure of SMS intervention for TB medication adherence in their RCT, but they defined treatment success differently [15,22]. Mohammed et al defined it as “the sum of patients clinically reported as cured (ie, a patient whose sputum smear or culture was positive at the beginning of treatment but who was smear- or culture-negative in the last month of treatment and on at least one previous occasion) or treatment completed (ie, a patient who completed treatment but who does not have a negative sputum smear or culture result in the last month of treatment and on at least one previous occasion)” [22]. On the contrary, Bediang et al defined treatment success as “having completed 6 months treatment and having negative sputum smears at 5 months” [15].

Table 2 summarizes the types of outcome measures reported in the included studies, using the categories defined in this review. The most frequently reported outcome type was treatment success. Approximately half of the studies reviewed (14 out of 27) included treatment outcome but only 6 studies among them followed the WHO definition for the treatment outcome (Multimedia Appendix 2). The second most frequently reported category was perception on the mHealth intervention (13 out of 27). Other categories reported were diverse and included technical outcome, medication or treatment adherence, process measure, etc. However, there was substantial variability within each category of outcome, as shown in Table 1. For example, acceptability and satisfaction within the perception category were defined differently from one study to another [15,28-30,32,35,36,40,44]. Also, SMS-only intervention studies did not focus on technical outcome [12,15,22,29,30,32,35, 36,42,43,45,46] whereas studies involving other mHealth technologies such as an app or mobile data collection did so [21,28,33,37,39-41]. On the contrary, outcomes related to cost or cost-effectiveness were reported only via studies involving SMS [29,45,46].

This systematic review critically evaluated the outcomes reported in mHealth studies for TB management in LMICs. The reason why rigorous evidence synthesis is warranted is that recent literature for TB reports mixed results despite the rapid implementation of mHealth technology for TB management. The fragmented pieces of evidence on the effectiveness partly resulted from the wide variations in the definitions of outcome measures in TB mHealth interventions. Even though treatment outcome has been reported by many studies, they often did not adopt the standard definition recommended by the WHO [17].

The WHO definition of TB treatment outcome is part of an effort to standardize outcome measures for TB at the global level. To promote the use of standardized sets of outcome measures for TB, WHO provided the standard definitions and classifications of TB in terms of diagnosis or treatment outcomes [17]. The intention for this standardization effort was to coordinate international comparison of TB treatment outcomes through health information systems. However, the findings of our review revealed that mHealth studies for TB have not comprehensively adopted this standardized approach for TB treatment. Only 6 out of 27 interventions chose to report the treatment outcome according to the WHO definition. Interventions involving phone calls, interactive voice response, or mobile data collection did not consider the WHO definition.

Our findings also suggest that, to rigorously evaluate the effectiveness of mHealth interventions for TB, future studies should be carefully designed with regard to the selection of outcome measures. Indeed, using standard definitions for outcome measures within some commonly reported categories can improve comparability across different studies. As assessed in this study, examples of such categories include treatment outcome (preferably using the WHO definition), perception, process measure, adherence, and technical outcome.

The value of this systematic review can be found in its potential to motivate and facilitate consensus on standard definitions of outcome measures used in mHealth interventions for TB so that such effort can guide more effective mHealth intervention designs for improving TB management. Although current literature shows considerable variability in the definition of outcome measures, discussion and coordination among researchers can promote standardized methods in measuring outcomes. Specifically, the outcomes should be comparable, promote transparent communication, and maintain consistency in terminology. Coordination at the global level is necessary to develop a core set of outcome measures for TB mHealth interventions by study design and technology type utilized.

When choosing a core set of outcome measures with standard definitions, there are additional issues to consider. First, the time point for reporting outcome measures should be clinically meaningful and feasible [47]. Second, a detailed description of the measure should be provided, such as calculation method or definitions. Third, a clear explanation on the target population for each outcome measure can be useful. For example, some outcome measures may be more appropriate for MDR-TB patients rather than TB patients on their first-line therapy course. Finally, long-term outcome measures should be considered to establish fundamental evidence for TB mHealth interventions. The long-term outcomes can be related to physical, psychosocial, or mental health. Despite its importance, our review showed that only 2 out of 27 studies reported long-term health outcomes; Mohammed et al, reported self-reported psychological and physical health measures [22] and Hunchangsith et al, reported DALYs averted [46].

Another issue to consider is related to evidence for cost. As this study suggested, insufficient evidence exists in terms of cost-effectiveness or cost of mHealth interventions for TB management. Those previous studies that have evaluated the cost-effectiveness or cost of mHealth interventions only considered SMS as mHealth channels and did not consider or evaluate other mHealth channels and technologies. However, other mHealth channels and technologies such as mobile apps or global positioning system are now available for TB patients [48]. Therefore, future studies need to assess the cost-effectiveness of such channels and technologies for improving TB management in LMICs.

This study has some limitations. First, the database used for identifying relevant studies is limited to the 3 most frequently cited sources, namely MEDLINE, EMBASE, and Cochrane Database of Systematic Reviews. Grey literature or other sources of information can supplement our findings. To complement this limitation, we attempted to identify additional related studies from relevant systematic reviews searched from our study. Second, the effectiveness of the mHealth interventions for TB was not quantitatively evaluated because of the heterogeneity of the outcomes reported.

Despite these limitations, this study provides an overview of the currently reported outcome measures for mHealth interventions intended to improve TB management in the context of LMICs. The results from this review can be used as a starting point for discussion to adopt standardized definitions within different categories of outcome measures for future mHealth interventions for TB management in LMICs.

This systematic review of mHealth studies for TB suggests that substantial variability exists with regard to the definitions of outcome measures across studies. Our review highlights that a standardized method for measuring the different outcomes is warranted to improve comparability of outcome measures across studies for a more rigorous and reliable evaluation of the effectiveness of mHealth interventions for TB. In doing so, the coordination among researchers and the development of a core set of outcome measures based on standardized methods would be necessary. Our study provides useful information for researchers to better assess the effectiveness of mHealth interventions for TB. In addition, the study provides insights into the possibility of developing a core set of outcome measures by intervention type and study design based on a standardized or coordinated set of methods.