Wallace Chigona

University of Cape Town wallace.chigona@uct.ac.za

Mphatso Nyemba-Mudenda

University of Cape Town mphatso.nyemba-mudenda@uct.ac.za

Andile Simphiwe Metfula

University of Cape Town andilesimphiwe.metfula@uct.ac.za

Introduction

Mobile technology affords users ubiquitous access to information and applications, hence providing flexibility with communication, collaboration and information sharing (Kaplan, 2006; Sheng, Nah, & Siau, 2005). Since mobile phones are readily available, even in remote rural areas, governments and development agencies view them as a potential Information and Communication Technology (ICT) tool for developing and improving livelihoods in developing countries. Mobile technology offers innovative opportunities to overcome a wide range of access and resource barriers in healthcare, specific to developing countries. The pervasive social connectivity of mobile phones has imperative implications for healthcare and public health in promoting health, disease prevention and management.

The rapid growth of mobile telephony has encouraged developing countries to apply mobile telephony in solving wide-ranging challenges in the health sector. It is generally agreed that communication and communication technologies are essential for improving the access and quality of health services, and overall health outcomes, by providing informational support and regular care (Klasnja & Pratt, 2012; Noordam, Kuepper, Stekelenburg, & Milen, 2011; Tezcan, Von Rege, Henkson, & Oteng-Ntim, 2011; Kreps & Neuhauser, 2010; Krishna, Boren, & Balas, 2009). Hence, the use of mobile technology for mass communication and disseminating information to the communities, together with other health interventions, seem to have the potential for a wide scale impact in developing countries. However, despite the widespread belief in the usefulness of mobile technology in health, there is still a paucity of study-based evidence.

mHealth, the use of mobile communication devices for health services and information, is a rapidly growing area of practice and research. mHealth interventions are mainly deployed in public health and primary healthcare. Public health is concerned with improving the health of the population, rather than treating the diseases of individual patients; it is an organised response by society to protect and promote health and to prevent illness, injury and disability. On the other hand, primary healthcare refers to a basic level of healthcare that includes programs directed at the promotion of health, early diagnosis of disease or disability, and prevention of disease; these are provided in an ambulatory facility to a community.

Health systems in developing countries, which are often already constrained by resources, have benefited from mHealth through increased access to healthcare and health-related information, especially for the hard-to-reach population (Tezcan et al., 2011; Krishna et al., 2009; Blake, 2008). mHealth is also attributed to have contributed to increased efficiency, cost reduction in service delivery, timely dissemination of health information and more actionable public health information, which have improved the ability to diagnose, treat and track diseases ( De Tolly, Skinner, Nembaware, & Benjamin, 2012; Lim, Xue, Yen, Chang, Chan &Tai, 2011; Chib, Lwin, Ang, Lin, & Santoso, 2008). In addition, it has enhanced the provision and expansion of training and medical education for health workers (Mechael, Batavia, Kaonga, Searle, Kwan, & Goldberger, 2010).

The potential impact of mHealth in health delivery in developing countries is largely untapped, due to technical as well as socio-economic, cultural and regulatory barriers (Kaplan, 2006). The design, implementation and adoption of mHealth in developing countries are beset with wide-ranging challenges. For instance, most of the technology applications used in most electronic health (eHealth) interventions are not designed in the developing countries; rather they are imported from developed countries (Chan & Kaufman, 2010). Consequently, most of these applications are not aligned with the realities of rural settings in developing countries. mHealth design and implementation need to take cognisance of local realities of settings which encompass political, social, cultural, technical and economic factors, under which the intervention would work best (Abimbola, 2011; Chan & Kaufman, 2010). In some countries, there is evidence of limited user acceptance of data collection using SMS and mobile forms (Garai, 2011). Further, low levels of literacy pose a challenge in mHealth utilisation in developing countries; inputting, displaying, transferring and processing of data is challenging for people with low levels of literacy. Moreover, despite increasing network coverage, the mobile network connectivity in most rural areas of developing countries is not reliable; the signal is often weak, which makes it hard to use the phone.

Despite the strides that have been made in mHealth in recent years, there is still a lack of clarity on the status and the direction of research in the field. This paper reviews the progress in both the mHealth field and mHealth research in developing countries. Data for the study is drawn from papers on mHealth in the proceedings of the third Mobile for Development (M4D) Conference, which took place in India between 28th and 29th February 2012. The nature and the focus of the M4D conference made it ideal for our analysis. The proceedings provide insight and understanding of the role of mobile technology on the development landscape, especially in reaching disadvantaged people in rural and poor resource settings. Being a conference, hence having a short turn-around period for publication, it presented us with the current state of affairs in mHealth, allowing us to be up-to-date. The review analysed the different characteristics of mHealth projects deployed in developing countries i.e. the application of mHealth, the aims of intervention and the implementation strategy.

Methodology

Our intention was to analyse papers which focused on mHealth, not those which simply mentioned the subject in passing (e.g. as an example of mobile application). We searched the electronic version of the 2012 M4D conference proceedings, using key words 'mHealth' and 'health'. We manually read the papers to confirm that they fitted the criteria. Studies or papers were included if they were about mobile phone technology usage for health services, health information at primary care level and public health. Out of the total of 62 papers in the proceedings, we identified 13 papers as meeting our criteria.

The analysis was informed by the work of Thomas (2006) in the General Inductive Approach. We read the papers several times to make sure we understood the meanings in them. The analysis is based on the following criteria: research methods; type of project (pilot or full deployment); setting - target population, whether rural or urban; nature of study - implementation, adoption or impact; application of mHealth: primary healthcare, public health; implementation strategy - top-down, bottom-up; and funding. The criteria gave us an understanding of the advancement in mHealth, as indicators that could reveal implementation approaches, evidence of outcomes and broader societal impact.

Results

Of the 13 papers in the corpus, ten were about interventions in Asia (nine from India and one from Bangladesh); the rest were from Africa (see Table 1). The domination of papers from India was most likely due to the location of the conference; this was the pattern with all the papers at the conference (the majority of papers at the conference, 33 out of 62, were from India). The first and second M4D conferences (2008 and 2010) had four and seven mHealth papers respectively. The 13 papers at the 2012 conference could mean that mHealth practice and research is burgeoning. A full list of references for these 13 mHealth papers is provided in the appendix.

To gain an overview of the terms used in the papers, we used a word frequency map constructed by using NVivo 8. The word frequency map is presented in Figure 1. As would be expected, the word frequency map in shows that the terms 'mobile' and 'health' are prominent.

A cross-section view of mHealth research

We evaluated the degree of rigour in the studies in the corpus. A sound conceptual foundation and use of rigorous methods are prerequisites for good research; they ensure that the research is coherent and consistent with what is accepted within a research discipline, and potentially make a longer-term contribution (Duncombe, 2011). The majority of the studies in the corpus employed an interpretive paradigm using qualitative methods; the other studies employed quantitative and mixed methods, using survey techniques and based on small samples. The methodologies employed were detailed, clear and easy to follow. Following Duncombe's (2011, p. 279) categories of conceptual approaches, it was found that none of the studies used a conceptual framework to guide the research and, similarly, no particular conceptual framework seems to guide the planning and implementation of mHealth interventions. This is similar to what Roman (2003) observed in the early years of telecentre research.

In addition, there is no evidence that the research questions of the studies were inspired by a specific theory. This could be due to the overarching challenges burdening health systems of the developing world, such that mHealth interventions are conceived and implemented in an atmosphere of enthusiastic urgency to use ICTs to help solve health problems. Another reason could be the lack of maturity of the mHealth domain; research is still in its infancy and mHealth interventions are in their early years, so that researchers are still trying to gain insight into the field to reach some common understanding. mHealth research, like M4D research in general, still lacks good quality research to inform policy and practice; this poses a major obstacle to effective decision making and implementation of mHealth initiatives on the ground (Webb, 2009 as cited in Duncombe, 2010).

Application areas

To assess the technological advances made in the domain, we assessed the technologies used in the interventions. Most of the interventions were for health promotion, encouraging health behaviour through awareness and counselling, while a few others were for remote monitoring and treatment adherence of patients. These functions are carried out using text messages, multimedia, games and a tailor-made internet-based application. The results show that text messaging still dominates mHealth in the developing context. This is expected, since SMS is supported even on the basic phones, which may be common in the developing country context. The results also show that other technologies are now being exploited as options for provision of mHealth services.

The majority of papers (i.e. six papers) dealt with interventions for improving the health of the population i.e. education/awareness and patient monitoring. Only three interventions used mHealth for connecting health workers, diagnosis and treatment, and treatment adherence (Batra et al., 2012; Khurana et al., 2012). Of interest in the word frequency map (Figure 1) is that the word 'patient' has a higher frequency than 'doctors', 'workers' and 'healthcare'. This could demonstrate that:

• (i) The majority of interventions are concerned with improving the health of the population rather than treating the diseases of individual patients.
• (ii) There is a shift towards citizen-centred healthcare and well-being.
• (iii) Research in mHealth for development is focusing on interventions in the hard-to-reach group of remote and rural areas of the developing world.

These benefits are vital, since they offer a potential that mHealth could work towards, thus reducing the strain on health systems, as more people become empowered to make informed decisions about their health and well-being, leading to reduction in hospital visits.

Connecting health workers

Some of the interventions in the papers enhanced efficiency and reduced costs of service delivery, by improving communication amongst health workers i.e. remote diagnosis and treatment. For instance, the Jiah project in India (Khurana et al., 2012) uses a mobile phone application to assist and connect health workers and trained medical professionals, through a back-end Medical Record System in rural areas, where there are no trained medical professionals such as doctors. The health workers upload the patient's information, including images (scan or X-Ray results) via a mobile phone to a database, from which specialists in urban referral centres download and analyse the patient's information and give advice or treatment, based on the diagnosis.

Patient monitoring

eCompliance, a treatment adherence intervention piloted in a low literacy urban area of Delhi in India for patient monitoring, uses a novel combination of biometric and mobile technology to manage drug resistant tuberculosis (TB) (Batra et al., 2012). The system identifies patients through a fingerprint reader to verify that they are physically present for treatment. Another example of remote patient-monitoring and self-management is AutoCare (Haque et al., 2012), which was piloted in rural Bangladesh to facilitate palliative care in breast cancer patients, using a mobile and sensor-based solution. The system is used to, amongst others, remotely monitor patients for symptoms and danger, in order to reduce the necessity for the patient to visit the medical facilities. The system also benefits doctors by providing up-to-date regular patient data. AutoCare uses wearable sensors and a cell phone for a remote symptom monitoring system. The system is novel "because it does automatic assessment of a patient's emotional states, along with physical states, that have not yet been covered by any cellular phones and sensors based healthcare system" (Haque et al., 2012, p.27). In cancer patients, treatments and interventions concentrating on reducing the suffering and improving the quality of life are essential, and are usually provided through an adaptive feedback-oriented treatment mechanism. AutoCare allowed for patients be assessed regularly.

Education/Awareness

Some interventions were aimed at well-timed dissemination of health information, which has improved the ability to diagnose, treat and track diseases, while providing more actionable public health information. In an intervention for TB and to promote HIV/AIDS awareness in youth in India (Khanna et al., 2012), mobile phone gaming was used to present "serious information by using the challenge and excitement offered by games and the advantages offered by mobile learning technology. … the games were … effective in creating awareness and increasing the knowledge of youth about HIV-AIDS and Tuberculosis" (2012, p.80).

Other mHealth initiatives are targeted at reducing medical errors, by providing correct information online by mobile phone. For instance, the 'drug-drug interactions' application deploys drug interactions information through mobile phone so that doctors can update themselves at any time anywhere, be aware of new drugs and the drug-drug interactions, and promote patient safety (Chaudhury et al., 2012). To provide authenticity to the database, it was refined regularly by the Authentication Committee, consisting of eminent pharmacologists. Both health workers and patients can access the database using mobile phones or the Internet.

The state of mHealth projects

All the mHealth interventions in the papers we reviewed were pilot projects with limited reach, which focused on different specific problems. Interventions for health information dissemination, especially where SMSs were used, were able to reach many people. The campaign for medical male circumcision in Tanzania received requests for information from about 15 000 individuals of the country's northern region. The OpASHA project in India reached 1,400 people with 26 kiosks placed in clinics in urban areas (Batra et al., 2012). However, the other interventions were designed to reach a small number of people. The scarcity of studies on large-scale interventions is a reason for concern. The lack of large scale projects makes it difficult to tell what works and why, to understand the settings under which the interventions work best, and to come up with best practices in the field (Mechael et al., 2010). Large-scale projects would be useful in informing public policy related to mHealth.

The papers express mixed opinions on the success, cost-benefit and practicality of mHealth interventions. There is no study-based evidence yet, from long term or fully deployed projects, on whether mHealth effectively improves health services and health outcomes in developing countries. Only two studies (Batra et al., 2012; Hoefman et al., 2012) presented evaluation results of the pilot studies based on outcomes; however, these studies did not report on the impact of the interventions. The rest of the projects were either at implementation stage or at proof-of-concept testing. To establish the real values and consequences of mHealth interventions, there is a need to heed the calls for further exploration and analysis of the impact of mHealth intervention on the healthcare system or among consumers at institutional and individual level (Lim et al., 2011; Noordam et al., 2011; Chan & Kaufman, 2010).

The targets for most mHealth interventions in the reviewed papers were disadvantaged communities, usually in rural areas, with a few in urban places of high density and low literacy. The underlying principle in the aims and objectives of all these mHealth interventions was to mitigate the delay in communication and other communication barriers to accessing health information and services. Women, due to their place in society in developing countries, were the target population of many mHealth interventions, especially in the field of cancer, family planning and maternal and child health. However, interventions for behaviour change communication targeted men, women and the youth.

mHealth implementation

The interventions in the reviewed papers utilised both bottom-up and top-down implementation methods. Donors or implementers of the system started with a feasibility study to gain an understanding of the local settings, such as socio-political norms, cultural practices and environmental factors that can affect adoption and usage of the technology. The implementers usually worked with a health facility or with non-governmental organisations already working in a particular field of health. The community or participants were involved in the proof-of-concept testing when evaluating whether or not a technology was viable and capable of solving intended problems, and if it had potential for empowerment. While a number of projects involved local communities, there is still a notable lack of bottom-up projects.

Almost all the interventions in the reviewed papers were funded by donor agencies or industry research bodies. The exception was the Medical Scheme Project in India which was funded by the government (Pundir et al., 2012). As such, the mSolution systems were hosted by the facilitating agencies and not integrated into main health systems. This funding arrangement predisposes the systems to sustainability challenges; the likelihood of such systems to survive is low once the funds allocated to the intervention are used up and the donors have shifted their attention to other projects.

Related to being funded by donor agencies, the systems operated in silos, lacking common standards of collecting and reporting patients' data. Integrating the mHealth systems into the national health system would expedite efficient delivery of healthcare services and health information to different stakeholders, for informed decision making and planning purposes. For sustainability and to increase the level of impact, there is a need for future research to consider how the mHealth systems could be integrated into the main health systems of the respective countries.

Discussion and conclusion

This review provides an overview of the progress of the mHealth domain and research progress in developing countries. Our results confirm that contemporary mHealth interventions have enhanced services. The services range from simply focusing on delivering mass health information using text messages, to remote data collection, remote patient monitoring and self-management, appointment booking systems, as well as monitoring systems for diagnosis and treatment. mHealth initiatives seem to have tremendous potential to improve access to and quality of health information and services. However, the body of mHealth literature is still lacking in studies on the effectiveness and impact of these interventions on the healthcare system and among clients. Most of the mHealth studies only focus on the feasibility, implementation, adoption, usage and acceptability of the technology.

Despite the claims of potential benefits that can be realised from mHealth, evidence is yet to be conclusively established on the acceptance and feasibility of this technology for large-scale deployment. Literature on mHealth implementations is still dominated by studies of pilot projects and those implemented to run for a short period, dealing with specific problems. This could possibly be an indication of what is happening on the ground i.e. most mHealth projects are pilots with no clear strategy on how they can be upscaled. From a research perspective, it is imperative to have studies on larger-scale implementations of mHealth. There is a need to analyse implementation strategies, best practices and the effectiveness of mHealth on large-scale implementations. Findings from such studies could be informative for policy makers and implementers of interventions.

Since it was noted that studies were not guided by any conceptual framework, neither were the research questions instigated by existing theories, it is recommended that future research should seek ways to improve on the rigour for studies in the domain. Further, research in the domain has not focused on the impact of the interventions on the various stakeholders. The majority of attempts to evaluate the interventions have focused on health indicators. We argue that the social impact of mHealth interventions may go beyond the health indicators into other social areas. Improved rigour would have a positive effect on the quality of research and hence the impact of research on policy and practice.

For improved effectiveness of mHealth, there is a need for changes in both the practice and the research of mHealth. Practically, the bottom-up innovations of health workers should be encouraged in mHealth implementation, so that all stakeholders, including disadvantaged communities, can contribute to the process of realising desired health outcomes and well-being. Such a process may also empower individuals and communities, leading to social change and improved quality of life. That would substantiate the claim that mHealth could be a tool for human development.

The small sample size is one of the limitations of this study; this review is based on 13 mHealth papers presented at the M4D conference. Moreover, due to the locality of the conference, the majority of the papers reviewed were from India. This may not necessarily represent the overall picture in the field. Again, being an academic-oriented conference, the papers could have been skewed against practitioner-based papers. We recommend that future studies should draw from a wide range of academic research outlets as well as practitioner reports.

References

Appendix 1

References for mHealth Papers at the 2012 M4D Conference