Building an evidence base to support and enhance community health workers' (informal) use of mobile phones in Ghana, Malawi and Ethiopia

Africa's recent communications 'revolution' has generated optimism that using mobile phones for health (mhealth) can help bridge healthcare gaps, particularly for rural, hard-to-reach populations. However, while scale-up of mhealth pilots remains limited, community health workers (CHWs) across the continent possess mobile phones. CHWs form a vital part of healthcare delivery across Africa and many countries are scaling up their deployment (target of 1 million CHWs).

There is thus enormous untapped potential here. While much research has focussed on formal mhealth initiatives (small-scale successes, challenges of scale-up), we know almost nothing about what CHWs across Africa are doing with their own mobile phones, and with what implications for healthcare. Last year, we published what we believe to be the first study to have looked seriously at this issue (Hampshire et al, 2016, Health Policy & Planning) Our very small-scale preliminary study, based on interviews with 16 health-workers in Ghana and 18 in Malawi, revealed that CHWs used their own mobile phones regularly (often on a daily basis) to communicate with colleagues and patients, obtain help in emergencies, seek information, organise work logistics such as drug supplies, etc. By contrast, only one had ever participated in a (now defunct) formal mhealth programme.

The potential for what we have called 'informal mhealth' to enhance primary care in resource-limited settings could thus be very significant. The CHWs we interviewed were innovative and adept at harnessing new features on their phones to help manage their work. However, our study also pointed towards challenges, both for patient care and health-worker wellbeing, including: the financial burden of phone costs and emotional burden of being available to patients 24/7; decreased face-to-face interaction with patients; concerns about patient confidentiality when communicating through personal phones; and difficulties in accessing reliable online information, etc.

Our aim for this Foundation Grant is to build a strong evidence base of current mobile phone use among community health workers in Ghana, Malawi and Ethiopia (three countries committed to major CHW programmes), in order to enhance the effectiveness of 'informal mhealth' and address challenges. If our hypothesis is correct, and CHWs' 'informal' mobile phone use is widespread, this study could provide crucial evidence to support innovative ways to strengthen health systems in resource-limited settings.

Methods:
1) Comprehensive policy reviews of CHW programmes and mhealth initiatives in each country, plus analysis of health systems and contextual factors affecting implementation.

2) Questionnaire survey of CHWs in each country to estimate levels of work-related mobile phone usage and collect comparable data on:
(a)Split between formal and informal mhealth usage, and function/purpose (e.g. communicating with patients, colleagues, logistics, information seeking, etc.);
(b) Estimated financial costs of phone use and who meets these costs;
(c) Perceived benefits and challenges arising from this 'informal mhealth' for CHWs and patients.

500 CHWs in Ghana and Malawi and 1000 in Ethiopia (where the total number of CHWs is much higher) will be sampled across multiple sites to cover a range of urban, semi-rural, rural settlement types.

3) Two sets focus groups of CHWs (minimum 14/country) and patients (minimum 6/country) will be convened before and after the survey, to reflect on current practices and experiences (incorporating survey findings), and to identify possible ways of supporting, enhancing and sharing good practice, and addressing challenges.

4) Meetings and on-going discussions with national stake-holders throughout the project to feed into policy/practice (see impact summary).

CHW Questionnaire: 500 CHWs will be sampled in Ghana and in Malawi, using a multi-stage cluster sampling approach across all three regions in Malawi (Northern, Central, Southern) and three contrasting regions in Ghana (Central, Brong Ahafo, Northern), across urban, semi-rural and rural locations. In Ethiopia, with its much larger number of CHWs (38,000), 500 CHWs will be sampled in each of two regions (Oromia and Amhara). The sample size of 500 CHWs per sampling unit was determined envisaging 5% margin of error in estimation of level of work-related mobile phone usage and survey design effect of 25%, including missing data. In Amhara region in Ethiopia, for example, the first stage of our sampling scheme will be a random sample of 4 out of the 14 zones in the region. Then a random sample of 12 districts will be selected in each selected zone with an average of 10 CHWs per district: all CHWs in the selected district will be surveyed. A similar scheme will be used Ghana and Malawi, adapted to specific structure of each country's health system.

Quantitative data will be analysed using survey methods in R, SAS9.4 and STATA. Descriptive analyses will be performed for each country to characterise 'informal mhealth'. To estimate level of CHW mobile usage, a survey logistic regression model will be used to account for survey design effects resulting from the multistage cluster sampling, with survey linear models used to analyse continuous data. A flexible mixed effects model will be used to perform "participant level" meta-analysis of data across countries.

Focus group discussions with CHWs and patients in a sub-set of sampled clusters: each FDG will be facilitated by 2 RAs, one to manage the discussion and one to record. FGDs will be transcribed as soon as possible in the original language then translated into English, with a sample independently back-translated to ensure accuracy. Coding and thematic analysis will be performed using NVivo software or equivalent

This Foundation grant will provide a crucial evidence base for developing sustainable strategies in each country (Ghana, Malawi, Ethiopia) to support Community Health Worker (CHW) mobile phone use, as a potentially innovative and important means of strengthening health systems and healthcare delivery.

Throughout the project, we will engage stake-holders at multiple levels to disseminate and discuss our findings, in order to inform policy and support/develop good practice.

1) Local level:

At the local level, and in keeping with our commitment to a bottom-up approach, focus groups of CHWs and patients will be convened before and after the survey, to reflect on current practices and experiences (incorporating survey findings), and to identify possible ways of supporting, enhancing and sharing good practice, and addressing challenges. It would be pre-emptive to pre-specify what form these might take but, from our preliminary work in Ghana and Malawi, possibilities might include:
- establishing locally-appropriate guidelines for use of personal/shared phones when dealing with confidential patient information,
- guidelines for sharing phones numbers with patients;
- managing patient expectations re CHW availability,
- managing phones costs, network availability, phone charging, etc.

2) National level:

Crucially, these reflections and proposals from those 'on the ground' will be fed upwards to national-level stake-holders, who will be engaged throughout the project via two high-level meetings and ongoing dialogue. In each country a national consultative group will be convened, to include: Ministries of Health and associated government agencies, phone network providers, associations/unions representing health-workers and patients, NGOs/iNGOs associated with CHW and/or mhealth initiatives (e.g. 1 million CHWs), etc. We have already begun work to constitute this group in each country, via preliminary discussions with Ministries of Health (or equivalent), all of who have expressed their eagerness to work in partnership with us.

A first national stake-holder meeting, to take place in each country within the first four months of the project, will enable stake-holders' experiences and concerns to be fed directly into the study design/refinement, and will help maximise buy-in and ownership. A second meeting in each country (scheduled for the last four months of the project) will entail dissemination and identifying areas of appropriate policy intervention. In between these two meetings, we will maintain regular contact with the stake-holders to discuss progress with the projects and emerging findings.

Again, it would be pre-emptive to specify now what areas of policy focus might emerge, but possibilities might include, for example
- Drawing up national guidelines or codes of practice to support CHWs' work-related phone use;
- Discussing with phone network providers the possibility of provided free/subsidised phone credit for CHWs (as part of corporate social responsibility);
- Ensuring phone charging points are available for CHW use at health centres;
- Developing e-resources to enable CHWs have access to accurate information sources and guidance on patient care.

3) Engaging international audiences:

While the study findings will be immediately applicable just in the three focus countries (Ghana, Malawi, Ethiopia), they will have relevance too for widely in under-resourced contexts that have yet to achieve universal coverage (UHC) and where formal mhealth programmes continue to have limited reach. Communication/dissemination to international audiences will be via the project website (with links to all papers and meta-data) open-access journal publications, and conference presentations.