記住我
Mobile health (mHealth) technologies, defined by the World Health Organization (WHO) as “the use of mobile and wireless technologies to support health objectives,” have the potential to improve health outcomes globally, including in low- and middle-income countries (LMICs) [-]. This is achievable through improving patient education, improving disease self-management, decreasing health care costs, and performing remote monitoring of patients, as reported in a recent systematic review of mHealth in LMICs []. In addition, mHealth can support preventative measures, facilitate disease management, or support health workers to strengthen the delivery of health care [-]. The WHO has highlighted the need to advance national digital health strategies that can facilitate universal health care []. The WHO’s “Recommendations on digital interventions for health system strengthening” highlights that digital technologies, including mHealth, can directly address health system challenges by supporting more widespread coverage across population groups and improving the quality and affordability of health care []. This digital transformation of the health care system has been made possible by the widespread availability of affordable digital technology; currently, 95% of the world population has internet access [].
mHealth interventions targeting infectious diseases care have the potential to greatly transform the health care landscape of LMICs, where infectious diseases still represent a substantial burden [,]. This is particularly important given health system challenges such as low health service utilization, poor adherence to clinical protocols among health workers, and geographic inaccessibility of health facilities [].
The success and sustainability of mHealth interventions require overcoming context-specific barriers and enhancing facilitators of mHealth implementation; these factors must be considered prior to intervention design. The WHO’s “Global Strategy on Digital Health 2020‐2025” acknowledged the need to adapt digital health intervention implementation to unique national contexts, health situations, and trends, as well as a country’s vision, available resources, and core values []. The WHO’s “Recommendations on digital interventions for health system strengthening” further identified key implementation enablers including health content aligning with recommended practices, intervention functionality, and greater leadership and governance []. The WHO has emphasized that recognizing and addressing digital health implementation challenges uniquely faced by the least-developed countries is a large factor influencing the scalability and sustainability of emerging mHealth technologies [].
mHealth interventions in LMICs have had limited success due to a range of health system factors not considered during the development and implementation of interventions [-,-]. Many mHealth interventions in LMICs remain as pilot studies that investigate feasibility, usability, and effectiveness, and they have not been scaled-up for integration within the broader health care system [,]. mHealth initiatives have often been developed for use in higher-resource health systems, with little consideration of differing contexts affecting implementation, such as social norms around a health-promoting behavior or access to resources []. Industry representatives, such as those from mobile phone providers, often push the scale-up of mHealth interventions rather than researchers, governments, or health workers []. This excludes crucial end user perspectives when developing mHealth interventions and risks having market-driven motives unrelated to health care encouraging the scale-up of interventions []. mHealth intervention teams also often fail to understand the relationship between users and mHealth technologies [,-]. Previously reported barriers to the widespread adoption of mobile technologies that uniquely concern LMICs include poor mobile network coverage, limited health care workforce capacity, limited data access, or negative health worker and patient perceptions toward mHealth interventions [-,-].
Among the limited number of published reviews evaluating mHealth implementation, there is a lack of rigorous evaluation regarding the design and implementation of mHealth interventions to aid policymakers [,,,,].
We therefore conducted a narrative review of the existing literature to understand the determinants of mHealth implementation for infectious diseases in LMICs. The review aims to consider the broader context, drawing on a comprehensive health systems analysis framework.
We searched MEDLINE, Embase, PubMed, CINAHL, the Social Sciences Citation Index, and Global Health. We collected studies that were the earliest available indexed in the above databases, up to and including May 31, 2023; the studies were exclusively in English.
Selection CriteriaInclusion CriteriaStudy population: We included individuals (of any age) with infectious diseases in LMICs. We took LMIC search terms from the Cochrane Effective Practice and Organisation of Care LMIC filters, defined according to the World Bank Classification (2022) []. We did not restrict the type of participants in the intervention (ie, we included health workers, patients, carers, general community members, and multiple types of participants).Intervention: We defined mHealth interventions as per the WHO []. These interventions included SMS, electronic decision-support tools, educational tools, apps, and other strategies to improve health care delivery. We included interventions that used either mobile phones, smartphones, or tablet devices, conducted at any level of the health care system.Comparator: We included studies where the current standard of care was a comparator, where applicable, in addition to studies without a control group listed.Outcomes: We included qualitative and mixed methods studies that included a description of the mHealth intervention and implementation processes and reported on factors affecting implementation (eg, acceptability, feasibility, essential resources) based on interviews or discussion groups.Exclusion CriteriaWe excluded formative research studies (ie, studies conducted before fully developing or implementing an intervention); study protocols; interventions involving computers or web-based health care (eg, websites); telehealth interventions (defined as consultation with a health worker via a mobile phone either through SMS or phone calls); and quantitative studies including randomized controlled trials, Likert scale surveys, and impact evaluations, as they did not provide in-depth reporting of qualitative factors affecting implementation. We excluded studies where mHealth was part of a larger complex intervention, studies from high-income countries, and studies that combined analysis of determinants across multiple countries where it was not possible to separate out findings from LMICs versus high-income countries.
Data Extraction and AnalysisOverviewThe full search terms and strategy for the databases are detailed in . Briefly, we included terms pertaining to LMICs, infectious diseases (eg, communicable disease), and mHealth terms (eg, mHealth, text message, mobile app). We did not include additional filters for qualitative versus quantitative studies.
Extracted DataOne reviewer (JGO) screened the titles and abstracts of the search output for relevant studies. As a next step, we conducted full-text screening. Where eligibility criteria were unclear, final consensus on article eligibility was based on discussions with another member of the author team (HMY).
We extracted characteristics of the mHealth intervention including intervention setting (country, LMIC status, health care setting); intervention design; content and purpose; target disease and population; and its quantitative impact on health outcomes originally targeted by the intervention, as reported in the included qualitative study. This was to aid our interpretation of how the qualitative implementation determinants we identified may have affected targeted health outcomes. The original quantitative impact evaluations were not sourced for this review. We also extracted data on details of the qualitative implementation study setting, study population, research question, data collection method, and study size, as well as broad implementation determinants considered by the authors.
Data Analysis and Reporting: Tailored Implementation for Chronic Diseases FrameworkWe performed a framework analysis of determinants affecting implementation based on the Tailored Implementation for Chronic Diseases (TICD) framework []. The TICD framework is a comprehensive checklist of determinants of clinical practice developed to inform implementation research projects that are tailored to local conditions [,]. The framework can be applied beyond contexts of chronic diseases, as the framework broadly focuses on health system components that determine quality of care. It identifies 7 key domains: guideline factors (clinical care guidelines or mHealth as a “guideline”); health worker factors; patient factors; professional interactions; incentives and resources; capacity for organizational change; and social, political, and legal factors. Its strength lies in its emphasis on ensuring tailoring to local conditions, which is valuable to consider for LMIC interventions, and inclusion of contextual (including political and legal) and patient factors affecting implementation. The TICD framework is a comprehensive health systems framework aligned with the systems thinking framework, which considers interactions with the broader context and patient needs in addition to the structural components of a health system [].
The framework was used to initially categorize the reported determinants of mHealth intervention implementation and was used to structure the reported findings in the results. Additional emergent themes, such as intervention design and forgetfulness, were extracted.
A total of 17,041 records were initially identified. After removing 6537 duplicates, 10 non-English studies, and 9625 non–mHealth-related studies, a total of 869 studies underwent abstract screening in 2 rounds, each examining different criteria, as seen in . A total of 20 studies were included for data extraction and final analysis.
Figure 1. PRISMA flow diagram. Screening strategy and PRISMA reported according to flow diagram for systematic reviews, including database searches. mHealth: mobile health; LMIC: low- and middle-income countries; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Characteristics of Included StudiesDetails of the intervention including study setting; intervention type and purpose; target disease and population; and quantitative impact evaluations (as reported in the included qualitative manuscripts) are presented in . details qualitative studies analyzed in this review. Among the 20 included studies, 9 were apps and 11 were SMS interventions. Based on the World Bank 2022 Country Group by Income LMIC status [], 8 were low income [,,-], 6 were lower middle income [,-], and 6 were upper middle income [,,-]. Of the 20 studies, 7 focused on people living with or at risk of acquiring HIV [,,,,,,]; the remaining studies targeted malaria [,,,], tuberculosis [,,,], pneumonia [,], dengue [], a grouping of “respiratory illnesses” [] or “childhood chronic infectious diseases” [], HPV [], or COVID-19 []. There were 2 studies that addressed multiple diseases [,]. Half (10/20) of the interventions targeted health workers [,,,,,,,,,,], while the remainder were for patients, individuals at risk of disease, or general community members [,,,,-,,,]. Most studies (11/20) involved community level health care [,,,,,,,,,,].
Determinants of mHealth implementation are reported according to the 7 TICD domains and additional emergent themes.
Table 1. Characteristics of mHealth intervention among included studies.Country, author, year, referenceWorld Bank country income classificationLevel of health care; health care contextmHealth interventionStudy design of original quantitative study/comparatorContent of the interventionPurpose of the interventionTarget disease(s)Target user(s)Quantitative impact of mHealth intervention on targeted health outcome(s) as reported in the qualitative studyArgentina, Straw et al, 2023 []Upper middlePrimary; primary hospital and primary health care centers consisting of community health workersSMSC-RCT/standard of careReminder messages to women with HPV, and one to community health workers about women with no triage 60 days after positive HPV testImprove Pap triage of HPV-positive womenHPVHPV-positive women and health workersBoth acceptability of the intervention by HPV-tested women and its adoption by health workers were high: 15% increase in percentage of women with triage Pap after HPV result. Statistical significance not reportedGhana, Ginsburg et al, 2016 []Lower middleCommunity; health centers and community-based health planning and service centers in rural GhanaApp (phone or tablet)Not reportedSoftware-based breath counter and a pulse oximeter to count child’s breaths, off-the-shelf reusable pediatric pulse oximeter to detect hypoxiaImprove pneumonia diagnosis and treatment and childhood illnesses in general in accordance with the Integrated Management of Childhood Illness guidelinesPneumoniaHealth workersNot reportedKenya, Jones et al, 2012 []Lower middlePrimary; government dispensaries and health centersSMSC-RCT/control groupPediatric outpatient malaria case management accompanied by “motivating” quotes to health worker’s personal mobile phones. Two messages per day across five working days for the duration of the study (26 weeks)Improve health worker’s malaria case-management practices, specifically drug dispensing and managementMalariaHealth workersIntention-to-treat analysis showed 24% improvement compared to baseline in correct antimalarial drug management immediately after the intervention, sustained effect of 25% six months later. Statistical significance not reportedMalawi, Ide et al, 2019 []LowerCommunity; village clinic in Northern MalawiAppNot reportedApp-directed assessment and management of the visit; the visit was documented in both the app and the village clinic registerImprove assessment, classification, and treatment of seriously ill children, facilitate disease monitoring and surveillance []Childhood infectious disease outbreaksHealth surveillance assistants, caregiversNot reportedMalawi, Kaunda-Khangamwa et al, 2018 []LowerPrimary and tertiary; health facilities operated by government or private care (including tertiary hospitals)SMSC-RCT/control group with no messagesTwice-daily text message reminders on case management of malaria, pneumonia, and diarrhea sent to clinicians and drug dispensersImprove case management of malaria, diarrhea, and pneumoniaMalaria, diarrhea, pneumoniaHealth workersNonsignificant, 4% improvement in correct malaria case management. Statistical significance not reportedMali, Mangam et al, 2016 []LowerCommunity; rural district communitySMS3 pilot intervention villages against 3 nonintervention villagesEducate and instruct households about indoor residential spraying campaignLower malaria prevalence through preventative measuresMalariaHouseholdsSignificantly lower among the mobile-messaging villages than the door-to-door mobilization villages (86% vs 96%, respectively; P=.02) and significantly lower structural preparedness in households mobilized through the mobile-messaging approach compared with the door-to-door approach (household and food items removed; 49% vs 75%, respectively; P=.03)Mozambique, Nhavoto et al, 2017 []LowerPrimary; health care centers providing intense ART and tuberculosis careSMSRCT/no comparatorStructured series of SMS text messages sent automatically based on appointments and scheduled drug pickups. Messages were sent 7 and 2 days before appointment or drug pickupSupport retention in ART and tuberculosis treatmentHIV and tuberculosisPatients and health workersThe majority of HIV patients (61/68, 90%) and the majority of tuberculosis patients (60/68, 88%) reported not having missed any appointments.amHealth: mobile health.
bC-RCT: clustered randomized controlled trial.
cHPV: human papillomavirus.
dART: antiretroviral therapy.
eRCT: randomized controlled trial.
fICMV: integrated community malaria volunteers.
gMDR-TB: multidrug resistant tuberculosis.
Table 2. Characteristics of qualitative study design among included studies.Country, author, year, referenceStudy populationResearch questionData collection method (study size)Reported qualitative factors affecting user uptakeArgentina, Straw et al, 2023 []Health decision makers, health workersStakeholder’s perception about mHealth implementation strategy and factors affecting scale-upSemistructured interviews (n=20)Knowledge of the strategyamHealth: mobile health.
bART: antiretroviral therapy.
cFGD: focus group discussion.
Determinants of mHealth Implementation: TICD FrameworkGuideline FactorsThere were 6 studies that reported on how lack of clarity and understanding of the intervention hindered its uptake. For example, users did not respond to mHealth intervention text message prompts because they were uncertain about how to use the personal identification number, were uncertain about how to respond to the message, or received little to no information about the background of the intervention [,,,,,]. Compatibility of the intervention with users’ past experiences with either the technology and its use in health care, or the intended change in health care practice, was a key facilitator that aided in implementation. Existing familiarity with the intended behavior promoted by the intervention—such as antiretroviral treatment adherence, HIV testing, or general interactions with the health care system—facilitated uptake [,,,,]. The mHealth intervention in Babili et al’s study—WelTel, assessed for COVID-19 case and contact management—was previously implemented for HIV epidemic virtual care, which facilitated its implementation as users were familiar with the platform and its functionality []. Similarly, health workers commented on how an app’s alignment with existing practices of using the village clinic register increased the likelihood of engagement []. Existing mobile phone use or interest in the use of new technologies were additional facilitators, as users were already familiar with making appointments, making calls, or using SMS technology [-,,,,,]. Trustworthiness of the source of the recommendation given by the mHealth intervention facilitated implementation, particularly when the mHealth intervention intended to aid health workers in improving their health care practice or disease management [,,,,].
Individual Health Care Worker Professional FactorsmHealth being perceived as useful by end users in improving existing health care practices facilitated the uptake and integration of the intervention. For instance, after initial use of the intervention, a perception that it might improve existing clinical practices, improve patient engagement with services, or relieve strain on the health system were key for implementation [,-,,,,,]. Health workers in Ide et al’s study perceived the app as advantageous over existing systems, as the intervention aided in conducting more accurate, error-free community case management of malaria, diarrhea, and pneumonia, which facilitated uptake [].
Furthermore, for interventions for which the main users were health workers, attitudes toward the intervention were impacted by the perceived effect of the use of the intervention on the health worker’s reputation. Ellington et al identified that the perceived loss of trust between patients and health workers in the health worker’s ability to diagnose and treat patients due to their reliance on technology to deliver health care was a barrier to using the app []. In contrast, Ide et al commented on how the app facilitated perceived professionalism []. Twimukye et al commented on how the health worker’s use of the app improved how the patient perceived the health worker’s attention to detail and care [].
Patient FactorsPatients perceiving the intervention as beneficial in improving health outcomes through increased convenience, awareness, or reminders facilitated implementation of the intervention [,,,]. Adeagbo et al commented on how the app’s positive messages about HIV testing and adherence promoted users to adopt new behaviors by improving individual competency to make informed, healthy decisions concerning sexual health []. Language literacy was a barrier that limited engagement with 2 SMS intervention studies targeted at patients or community members [,].
Professional InteractionsLimited supervision and follow-up of the intervention implementation by the research team leading the intervention was noted as a barrier, as users requested feedback and confirmation of correct intervention use []. Lack of or limited access to training to provide necessary skills to effectively engage with the mHealth intervention was a barrier to implementation for health workers [,]. Access to professional training was a facilitator of implementation [,,].
Incentives and ResourcesSeveral studies reported that access to resources and essential infrastructure influenced implementation. Specifically, 6 studies reported that poor telecommunications networks, problems with electricity, a lack of phone coverage, limited staffing, and a lack of equipment to implement the behavior change were barriers to implementation [,,,,,]. Network problems were particularly prominent in geographically remote areas. Other barriers included technology that repeatedly malfunctioned (such as periodic freezing and system crashes) and limited access to technology support systems to troubleshoot technological problems [,,-,,,,]. Access to technological resources and support in case of malfunctions facilitated implementation [].
Capacity for Organizational ChangeFinancial instability, existing patient overflow, incompatible technological equipment, and length of appointment time within the clinic hindered the implementation of mHealth interventions [,,,,]. Kaunda-Khangamwa et al reported that 90% of the health worker respondents blamed existing high workloads and drug stockouts as factors discouraging health workers to respond to SMS reminders that promoted infectious disease case management []. Similarly, Ellington et al’s study identified that the existing length of the appointment time was not compatible with mHealth intervention use as the time to complete a health assessment through the app took longer than the appointment duration []. Straw et al commented on the compatibility of the mHealth intervention with the existing organization functionality to facilitate normal workflow as a facilitator of implementation []. A lack of management and ownership of the intervention by health workers further contributed to a limited capacity for embedding the intervention within the health care facility [].
Social, Political, and Legal FactorsOne study commented that the costly nature of airing messages during the daytime and limited funding for the intervention were barriers to successful implementation []. Receiving national-level support on a political level including the Minister of Health or district leadership was a facilitator of implementation [,,]. Babili et al commented on how the Rwandan government’s support of digitization across all governmental sectors by offering resources for implementation aided the adoption of the digital health intervention []. Furthermore, shared responsibility and corporate social responsibility felt among the wider community was an enabler, as the charitable community assisted in reducing costs and improving uptake [].
Emergent ThemesA novel factor not directly addressed in the TICD framework is the importance of considering app or SMS design features. Frequency of messages, language, and integration of local narratives to engage users were reported as affecting ease of use and user uptake [,,,,,,,]. mHealth interventions that adapted content toward the intervention context by using local proverbs, narratives, or language facilitated uptake of the intervention [,,,]. Moreover, features such as a user-friendly interface and a streamlined workflow facilitated implementation, while an intervention design that increased the workload of health workers was a barrier to implementation [].
The study by Mangam et al, whose reminder SMS system replaced the existing door-to-door reminders, uniquely commented on how the absence of face-to-face reminders increased the rate of forgetfulness and patients ignoring the messaging, thereby affecting its implementation [].
mHealth Impacts on Target Health Outcomes and How Implementation Determinants May Have Influenced These OutcomesAs described in , there were 8 studies that identified that mHealth had a positive impact on health outcomes or behavior, whether that was through increased HIV testing, reduced errors in drug and disease management, or improved disease prevention knowledge, results collection, or linkage to care for better clinical practice [,,,,,-]. In 4 studies, the mHealth intervention did not lead to an improvement in the health outcome—responses to SMS prompts were low, preventative measures were worse than in the non-mHealth control group, or user uptake was low [,,,]. There were nonsignificant changes in health outcomes or behavior in 2 studies [,]. Quantitative health outcomes were not reported in 6 of the reviewed studies [,,,,,].
Improved outcomes may be explained by familiarity with the health behavior or technology [,,,,], positive attitudes among health workers toward the technology [,,], or ease of use of the mHealth technology [,,,]. Technological barriers, lack of familiarity with technology, and resource limitations [,,,,,] may have reduced engagement with the intervention or the participants’ ability to implement the behaviors enforced by the intervention, therefore diluting the intervention effect.
Findings from the 20 reviewed studies and categorization into the TICD framework were synthesized to deduce two overarching themes that influenced the successful implementation of mHealth initiatives in LMICs: (1) the acceptance of the intervention by patients and health workers (as well as on a sociopolitical level), regardless of the target user, and (2) the capacity of existing infrastructure and resources to implement the intervention, which was strongly tied to the health system’s capacity for change. This relationship is visually depicted in . The logic flow diagram in further represents these reported factors according to inputs required for mHealth interventions and the required processes for success.
Figure 2. Model of factors influencing implementation of mHealth interventions in LMICs. The two main factors are presented in large boxes: (1) acceptance of the intervention and (2) capacity of the existing infrastructure and resources to accommodate mHealth. Acceptance, in turn, was mainly influenced by perceived usefulness of the intervention, amount of training and communication, and previous experience with the guideline behavior and mHealth or technology. The health care system’s capacity and infrastructure and resources were influenced by funding, network availability, and technological support. mHealth: mobile health. 
Figure 3. Logic diagram depicting determinants of successful mHealth implementation. Overview of mHealth implementation determinants from the mHealth design and resource input stage to the process of implementation and the desired outcomes and impact. mHealth: mobile health. Acceptance of the InterventionOverviewAcceptance of the intervention was strongly linked to uptake of the intervention, which facilitated implementation of the intervention and its potential to improve health outcomes. Patients and health workers were influenced by the perceived benefits of the intervention in improving health outcomes, the extent of training, and previous experience with the guideline behavior or mHealth.
Perceived Benefit of the InterventionPerceived benefit of the intervention was influenced by ease of use and design of the intervention [,,,,,,,,]. Where an app or SMS design engaged users in a way that improved health outcomes, patient engagement with health care services, or health care efficiency, participants were less likely to perceive the use of the mHealth technology as burdensome, facilitating successful implementation [,-,-,-,]. An intervention that provided a clear advantage compared to existing practice facilitated the perception of it being beneficial, as supported by the intervention’s design and ease of use. These findings were consistent with Davis’s Technology Acceptance Model and its application in health care, which suggests that perceived ease of use and usefulness both affect attitudes toward and use of technology [,]. Design considerations was an emergent factor unique to mHealth and technology-related interventions, emphasizing the value of iterative design of the mHealth intervention before implementation [,,,,,,,]. Further reviews reporting on mHealth in the context of antiretroviral therapy adherence and maternal health have identified the importance of intervention design, including tailoring SMS messages and the frequency of reminders [,].
Political support of the intervention ensured effective implementation; this included stakeholder buy-in and approval of the intervention, and support from health administrators on a district level [,,]. The importance of understanding the need for the intervention to reduce disease burdens or improve health care services is supported by a previous review by Opoku, Stephani, and Quentin []. Therefore, regardless of the target user, perceived benefit of the intervention is crucial for implementation of the initiative, emphasizing the community-wide, integrated nature of mHealth interventions.
However, studies by Kaunda-Khangamwa et al and Mangam et al noted that, despite the users’ positive attitudes toward the intervention, factors such as lack of communication regarding intervention use and limited resources to implement the behavioral guideline hindered implementation [,]. Perceived value of the intervention and acceptance alone therefore cannot guarantee successful implementation and positive outcomes from an intervention.
Training and CommunicationUsers who received communication and training on the intervention and its use before or during its implementation engaged well with the intervention and intended guideline practice [,,,]. Lack of awareness and clarity regarding how to engage with the intervention were consistently noted across studies with limited user engagement [,,]. Sufficient training has been previously noted in a review as a contributing factor to mHealth intervention implementation, which is closely linked to the perceived ease of use []. For interventions aimed toward health workers, training was either provided by the intervention research team members or between health workers; follow-up on correct use was identified as being important for encouragement and continued intervention use [,,,,].
Compatibility Wi
留言 (0)