‘The phone is my boss and my helper’ – A gender analysis of an mHealth intervention with Health Extension Workers in Southern Ethiopia

ABSTRACT Background There is considerable optimism in mHealth’s potential to overcome health system deficiencies, yet gender inequalities can weaken attempts to scale-up mHealth initiatives. We report on the gendered experiences of an mHealth intervention, in Southern Ethiopia, realised by the all-female cadre of Health Extension Workers (HEWs). Methodology Following the introduction of the mHealth intervention, in-depth interviews (n = 19) and focus group discussions (n = 8) with HEWs, supervisors and community leaders were undertaken to understand whether technology acted as an empowering tool for HEWs. Data was analysed iteratively using thematic analysis informed by a socio-ecological model, then assessed against the World Health Organisation’s gender responsive assessment scale. Results HEWs reported experiencing: improved status after the intervention; respect from community members and were smartphone gatekeepers in their households. HEWs working alone at health posts felt smartphones provided additional support. Conversely, smartphones introduced new power dynamics between HEWs, impacting the distribution of labour. There were also negative cost implications for the HEWs, which warrant further exploration. Conclusion MHealth has the potential to improve community health service delivery and the experiences of HEWs who deliver it. The introduction of this technology requires exploration to ensure that new gender and power relations transform, rather than disadvantage, women. Keywords communities, e-health, gender


Introduction
Mobile health (mHealth) provides health services and information via mobile technologies, including mobile phones. 1 There is considerable optimism in mHealth's potential to overcome health systems' deficiencies to ensure access to safe, effective and affordable health services. 2 This has led to an 'explosion of mHealth activities' 3 and 'large-scale adoption and deployment of mobile phones' 4 by Community Health Worker (CHW) programmes. MHealth innovation in relation to CHWs, on which low-and middle-income countries (LMICs) disproportionately depend, has been reported to be 'particularly promising'. 5 CHWs' use of mHealth has the potential to improve their motivation; decision-making; training; adherence to guidelines; data entry and quality; planning and efficiency; and communication and health promotion; while also enhancing coverage and timeliness of services and reducing costs. 1,2,5-8 MHealth also allows the monitoring and tracking of health indicators in real time, providing crucial insights to policy makers and enabling CHWs to better serve communities. 9 Research on CHWs' use of mHealth focuses on the formal, work-related aspects, such as health outcomes and/or health system benefits. Recent systematic reviews 1,4,10,11 note CHWs acceptance of mHealth's potential to enhance health outcomes and health systems and benefit CHWs.
Adoption however, is hindered by infrastructural limitations (e.g. electricity and internet), security issues and a lack of sustainability given that most interventions are pilots. With a few exceptions, 5,12 CHWs' perceptions of mHealth, and their positionality as gendered subjects and workers is overlooked. As Lupton (2014) has argued, examination of how CHWs have used mHealth has 'received almost no attention from critical scholars'. 13 Sustainable Development Goal 5 calls for 'the use of enabling technology, in particular ICTs [Information Communication Technologies], to promote the empowerment of women'. 14 Aligned with this, pilot mHealth interventions provide women with greater access to health care information and more autonomy in health decision-making. MHealth technologies for CHWs, an often-feminized and sometimes volunteer cadre, offer a unique opportunity to explore the potential for empowerment. CHWs are crucial lynchpins in LMICs' health systems, yet experience limited training, heavy workloads and undertake a range of service delivery tasks. 5,15 Situated within broader socio-cultural and gendered contexts, they are burdened both by their workloads and by gendered roles and responsibilities to kin and communities. 16,17 Gender inequalities can weaken attempts to scale-up mHealth and mHealth initiatives do not always lead to women's empowerment; 2,4,12 gender transformative initiatives that promote equality and transform gender norms are needed. 12 Several tools encourage a more gender transformative approach assessing gender norms and power relationships. This includes the World Health Organisation's (WHO) gender responsive assessment framework, 18 which helps position a project from gender blind to gender transformative by setting out basic criteria to be met in each category (see Fig. 1). This framework identifies necessary milestones/actions for interventions seeking to achieve gender transformation.
This study explores the impact of an mHealth pilot intervention, a smartphone-based digital health management information system (HMIS), 19 on Ethiopia's female HEWs. Framed by a socio-ecological model, and using the WHO's gender framework, it assesses whether technology acts as an empowering tool for HEWs and how the intended and unintended consequences influence gender and power dynamics.

Methods
In Ethiopia, the Health Extension Programme (HEP), initiated in 2004, is a free primary health care package in which 38 000 female HEWs offer 16 essential health packages. [20][21][22] HEWs are salaried government employees who have completed at least grade ten. They are selected by their communities to complete one year of training in basic health service delivery. A health post serves a population of about 5000 and is staffed by two HEWs accountable to the kebele (lowest administrative unit). HEWs are supported by female volunteers, known as the 'Health Development Army' 18 and supervised by health professionals from health centres. Health centres in turn, are overseen by the woreda (district) health office (Fig. 2).
In spite of low ICT access and usage compared with other African countries, the Ethiopian Federal Ministry of Health has embraced mHealth in its national strategic health plan. 23 Ethiopia prioritizes maternal health services and calls for improved HEW performance on maternal health-related tasks. 20,21,24

The intervention
An mHealth intervention that focussed on the priority areas of TB and maternal health services 19 and linked to the Ethiopian Ministry of Health's mHealth strategic framework was conducted in Sidama zone, Southern Ethiopia, with a population of about 3.7 million. Our research, undertaken in six Primary Health Care Units across six districts, worked closely with and was realized by HEWs, their supervisors, health workers based at the catchment health centres and policy makers at woreda health office and zonal health department.
One smartphone, assigned to each health post, was shared between two HEWs, who used the phone to input data on expectant mothers and TB. The data was uploaded to the HMIS where it was instantly available to other levels of the health system. Reminder messages prompted HEWs to follow-up on expectant mothers' due dates and sputum examination for TB symptomatic cases. Ninety-seven smartphones and eight computers were distributed to HEWs, their supervisors, health centre staff and focal persons from district and zonal levels. Ongoing theoretical and practical training was conducted and a monthly airtime allowance of 100 birr (3.64 USD) was provided for the first five months. Subsequent top-ups were paid for by HEWs.

Ethics statement
Ethics was approved by the Liverpool School of Tropical Medicine 16-22 and by the Ethiopian Ministry for Science and Technology in June 2016, and supported by the Regional Health Bureau. All participants gave written informed consent.

Data collection process
Qualitative methods were used to generate rich insights into participants' experiences of the intervention. 25 They included face-to-face semi-structured in-depth interviews (IDIs, n = 19) and single sex focus group discussions (FGDs, n = 8) with HEWs, supervisors and community leaders (Table 1). (In the study districts, all HEWs are female and all community leaders male. Supervisors are predominantly male. Disaggregating by gender and district would breech confidentiality.) Interview topic guides explored the gendered elements of the intervention; ways in which the mobile phones helped or hindered HEWs' roles, how HEWs used the phones outside of work and the impact on their relationships. Analysis, informed by an adapted socio-ecological model, was designed to evaluate how the intervention impacted the interface position of the HEWs and to establish how the intervention fared along the WHO's gender transformative scale. Interviews were conducted in four districts purposively selected for variation in geographic location and performance.
In interviews, a local trained female research assistant, fluent in Sidamigna (the local dialect), ensured HEWs felt comfortable, and used topic guides to facilitate conversation. The lead researcher (RS) was on hand to clarify any questions or concerns. Interviews were conducted at health posts, health centres and woreda health offices, scheduled in private spaces, and recorded. These were transcribed and translated into English. Translation quality was reviewed (AZK). Qualitative analysis was done by reading and rereading transcripts to identify iterative themes 26 and select appropriate quotes (RS with inputs from AZK and DGD). Software NVivo was used to code and run queries on the data. Attention was paid to give voice to the majority and minority views.

Results
In line with the WHO framework, our results explore HEWs' experience of the inequalities generated by unequal  Fig. 3, informs our analysis and frames our results. 27 The introduction of mHealth shapes these relationships in new and complex ways. Relationships and behaviour, although illustrated in Fig. 3 as distinct from one another, can span more than one sphere. The results are presented starting from the relationships among the HEWs, and then on to different forms of interpersonal relationships, organizational relationships, and finally community level relationships.

Impact at individual level
Participants spoke of the limited access women in their communities had to mobile phones, which they attributed to their lack of decision-making power. Although all HEWs come from the communities they serve, when they speak about women in their communities, they often set themselves apart in terms of their decision-making ability. This assertion was influenced by HEWs' education, paid employment and the phone.
A woman living in our community is uneducated and she cannot decide everything like I can. Our culture even does not allow her to decide. Moreover, there is still a negative attitude, that is woman cannot participate in meetings or discussion. [HEW, IDI] The intervention helps HEWs be more accurate in their data collection and reporting and reminds them to follow-up with patients and clients. HEWs see it as a helping hand and a means to upgrade their skills and knowledge. Many reported increased confidence, more participation in woreda and kebele meetings, greater oversight of the data and improved quality.
[O]ur participation rate increased… [due to] the presence of this mobile service…[T]here is great variation in the participation rate from HEWs who work in Kebeles where there is no mobile service program. [HEW, IDI] My commitment to my work is improved. The data quality is improved, I develop self-confidence in my work, and the skill of using the technology is improved. My data handling is also improved. I can say the phone is my friend. [HEW, IDI] Several HEWs perceived the mobile as an additional person, friend or boss who helped them in their work. This was especially pronounced for HEWs working alone at health posts.
HEWs unable to partake in training, due to maternity/ educational leave, felt that their trained counterparts had more opportunities to input into meetings. This suggests that training on the smartphone, rather than the presence of, a Merged with participants from District 3 due to geographical proximity and convenience of participants.
b Merged with participants from District 4 due to geographical proximity and convenience of participants.  ii20 JOURNAL OF PUBLIC HEALTH or primary access to, the smartphone is the defining factor for increased participation in meetings. Drawbacks of the intervention were also identified. Some HEWs reported that they spent extra time inputting data, as paper-based reporting was still required. This was partly because the intervention had not been scaled-up and partly because it focused only on TB and maternal health, excluding the other 14 essential packages that HEWs deliver. The intervention added to HEWs' work burden and some HEWs, and their supervisors, reported that it also changed the workload divisions and thus the relationships between HEWs.
It is not helpful. Because the one who is trained to handle the mobile is not performing other jobs, once she starts to fill information in mobile…It is not only the mobile text that support the mothers…, rather [it is] our knowledge and skill.
[HEW, FGD] The one who doesn't have the mobile phone may consider herself inferior to her colleague. [Supervisor, FGD] Financially, the phones may also inadvertently burden HEWs as they continue paying for the airtime charges. Primarily the HEWs seemed happy to do this as they considered the phones to be their personal property, using the airtime for personal calls and data. Though HEWs also reported being fearful that, in the case of a lost or stolen phone, they would have to bear this cost; which could put additional pressure on their limited financial resources.
I take great care of this mobile…Additionally, if it is lost, it is said that there should be paid by the one who lost it and this is practiced in somewhere. For this reason I feel discomfort. However, I felt great satisfaction getting it, currently I am using it to call. [HEW, IDI]

Impact on interpersonal relationships
Most HEWs reported an intrinsic desire to help their communities as their reason for joining the HEP, though many also desired 'an occupation', 'a monthly salary' and noted that HEW 'occupation is good especially for females'. Many respondents lacked information about the HEW role and workload when joining, with some suggesting they would not have joined had they known the work burden. In Ethiopia, HEW work is in addition to a domestic work burden.
[A woman] starts her routine early in the morning and continues working till night and no one understands her problem. …She prepares food, feeds their children, caring for her children and cleans the compound. She cannot feed herself properly even. [HEW, IDI] HEWs find this dual burden onerous but stress that they are motivated by the positive results in their communities, the introduction of mhealth and their paid employment.
HEWs serving in rural areas report that male household heads commonly own phones, as men have their own incomes. HEWs are uniquely positioned: as women and wives, with their mainly subordinate status, and as paid workers, with a slightly higher status. The smartphones elevated the HEWs' status becausealthough most HEWs owned a phone prior to the interventionit was not usually a smartphone.
HEWs [were] delivered with better quality of mobile compared to their own [personal phones]… for this reason HEWs working in non-project areas…are saying to HEW working at project area 'you own a special quality mobile, why it is delivered only for you?'…Our HEWs are glad to have this mobile, because they feel as if this increased their social status…[Project Supervisor, IDI] One supervisor reported a husband's appropriation of the phone, however overall HEWs controlled these phones. Thus, when asked about smartphone access, the power dynamic of men as mobile phone gatekeepers was invertedat home HEWs managed these phones and they reported not allowing their husbands or children access. They were proud of the phones, seeing them as their own, but also as an extension of government property. This indicates some gender transformative attitudes and enhanced HEW status.

Impact on relationships with supervisors
The introduction of mobile phones may have resulted in HEWs' increased collaboration with supervisors. There were reports of a pragmatic, 'teamwork' approach: when phones were lost, or network connectivity was down, supervisors would collect and upload the data.
It was lost… I informed to supervisor and currently he is feeding data. [HEW, IDI] HEWs without primary access to the smartphones also relied on supervisors to address technology skill gaps. Supervisors, in turn, noted the improved quality of HEWs' data and seemed impressed by HEWs' ability to adapt to the mHealth system. I observe that they have positive attitude and the motivation of the HEWs towards their work is improved. They have good attitude towards us too. We are helping them, and we have friendly relationship. [Project Supervisor, IDI] Supervisors also took more responsibility for monitoring pregnancies and motherhood in the community: When the message alarm [for a] particular woman comes to them after entering data in mobile, they feel great satisfaction. This is common not only for HEWs but also us. Sometimes the message comes to myself and it feels good so I go to community to follow up those pregnant women. [Project Supervisor, FGD] Government supervisors, without smartphone access, felt their lack of skills made it harder to appropriately support HEWs who needed help. This, combined with supervisors' desire to assert their superior skill, knowledge and status, may have a negative impact on their relationships.
HEWs have good skills using this technology, but supervisors do not have skill. Supervisors should have at least one step better skills on this technology compared to HEWs. It is better to give the mobile phone to supervisors and the rest of HEWs to exchange information….

Impact on organizational relationships
Additional intra-cadre power differentials arose after the introduction of one smartphone per health post. HEWs perceived the smartphone as gifted to the health post for good performance and it was generally 'adopted' by the more senior HEW, as a signal of status. The senior HEW treated the smartphone as her own property, taking it home each evening. One HEW was said to have retained the phone during maternity leave, demonstrating strong personal ownership. This meant that HEWs 'without' smartphones were unable to upgrade their skills. Despite initial training, they had fewer opportunities to use the smartphones.
Yes, for example she [the senior HEW] increased her knowledge and skill about the mobile and also she increased her own work performance. Additionally, she can use Facebook as well as she can take a picture. [ HEWs believed that the Smartphones increased communities' recognition of their status and that this was not dependent on phone ownership.
… They are happy while we follow-up with mothers after feeding their data in this mobile. And mothers give great respect to us. [HEW, IDI] Q: Does the community give special respect for your colleague after she got phone?
A: The communities are giving respect for both of us in the same way. Only I and she have information about the mobile phone including its purpose. [HEW, IDI] Although HEWs stressed the shared status, their own accounts of smartphone usage demonstrate that status is linked to the smartphones. A kebele leader similarly suggested that the HEW in possession of the mobile is perceived by the community to have a higher status.
Additionally, one HEW reported that smartphones led to greater community expectation which could inadvertently place additional stress on the HEWs. Q: What is the feeling of the community on the mobile phone?
A: They are happy, encourage us to work hard more and they are giving positive comments since we get this chance, so they expect more things. [HEW, IDI] Discussion Main finding of this study HEWs' lives are embedded within the communities they serve. As individuals they have a unique role and agency to shape health outcomes. Like other women health workers, they juggle multiple workloads; undertaking employed health work alongside household and childcare responsibilities. 17,28,29 Adapting the McLeroy et al. 1998 socio-ecological model, this study explored the experience of HEWs interface role with regard to mHealth's impact on their work and on relationships within households, communities and in the health system. In parallel, applying the WHO framework demonstrates that while the mHealth intervention did change gender and power relations, it did not address the underlying causes of gendered health inequalities. In this sense, it did not achieve the WHO milestone for gender transformative interventions. Rather, the intervention intentionally addressed gender-based health inequalities through remedial action. This, in accordance with the WHO framework, is a 'gender specific' approach and interventiona step towards gender transformative research, but not yet there. Incremental improvements nonetheless result in better healthcare provision for populations. As demonstrated here, the intervention equipped HEWs with knowledge and tools to perform effectively and provide more equitable care to communities; reflecting mHealth's positive potential for health outcomes and health system strengthening. 2,5,12 These findings demonstrate the increased social status and agency felt by the HEWs, who as government employees, already experience superior status; they acted as gatekeepers to the phones within households inverting the traditional patriarchal norms where men are the primary keepers of technology. In most cases, HEWs felt the smartphones improved their skills; gave them opportunities to share their knowledge in meetings; and aided their workloads by acting as a 'boss' or a 'friend', serving helpful reminders to follow-up with patients. However, the unequal distribution of the smartphones also changed power dynamics between HEWs, impacting on workload distribution. Inadvertent financial burdens linked to the running costs and potential loss or theft of the mobiles introduced new pressures on the HEWs. These drawbacks require consideration for scale-up.

What is already known on this topic
Literature on mHealth interventions and patient empowerment shows mHealth has the potential to empower communities and transform harmful gender norms but access and use may also reflect and extend current gender inequities. 12 A 2013 review looking at studies from Nigeria, India, Tanzania, Uganda and the Congo found that in cases where husbands did not have access to phones, female community members would render phones to husbands. 12 However, our findings show a shift in patriarchal norms -HEWs became the gatekeepers of the technology, not allowing their husbands access. Similarly, an Indian intervention 30 which registered CHWs phones in women's names, led to male household members requiring permission to handle the phones.
Our findings echo concerns in the literature about mHealth financing of HEWs' smartphones. By not providing unlimited airtime, health extension programmes risk transferring the financial burden to those least able to afford it. Hampshire et al. argue that CHWs subsidize health care from their own pockets when expected to pay mobile airtime. 5 This links to wider debates in moral economies of care, which sees women undertake a large proportion of the unpaid care workforce. Maes (2015) describes the institutional rhetoric of urban Ethiopian CHWs as 'priceless'. 31 This rhetoric is internalized by CHWs, who feel a strong moral obligation to care for the sick or pregnant in their communities at their own expense, 5,31 at times further impoverishing themselves and their families. 32 Conversely, financing smartphone usage could further limit informal use. Although no restrictions were issued, some HEWs refrained from personal gain as the smartphones were government property. Given Ethiopia's political climate during data collection, which had seen social media use restricted in the wake of anti-government protests, 33 it's possible that HEWs felt uncomfortable using government devices beyond their official capacity. While there is no evidence as to whether women are more likely to be affected by the political climate, our results demonstrate proud women over-burdened in their work and limited in their choices. HEWs are low in health system hierarchy 34 and they saw their jobs as 'good for females', but arguably the litmus test for gender transformative programmes is that HEW employment becomes acceptable as men's work with attractive employment conditions. 17,35 Medhanyie and colleagues found, in Tigray, Northern Ethiopia, that HEWs' unrestricted smartphone usage helped familiarize and motivate HEWs. 36 Unrestricted use may have multiple benefits to healthcarein Kenya CHWs' Whatsapp groups disseminated health information at times of outbreak, built morale, improved supervision and documented the quality of services delivered. 37 Moreover, HEWs in Tigray used smartphones for accessing the internet and social media, thus independently gaining information and resources. 36

What this study adds
Many studies have focused on communities' gendered access to mobile technology 12 whereas this paper examines mHealth's impact on the experiences of female HEWs. Unlike in Bangladesh, where technology was appropriated by husbands or seniors as a consequence of gender dynamics, 2,12 our findings show signs of change in household dynamics in a context where, although phone ownership was a male norm, HEWs have become the gatekeepers of the phones accorded to them.
Application of the WHO framework for gender transformative research demonstrates positive HEW empowerment in skill building and data handling. The intervention increased HEWs' expectations of themselves and communities' expectations of HEWs. This is not unique to Ethiopia. In Malawi and Ghana, CHWs' mobile phones led to additional time and emotional burdens, with CHWs often responding to out-ofhours calls. 5 These burdens, along with the risk of theft and lossalso reported as a concern of Mozambique's CHWs 7show smartphones add status and help HEWs perform well on the one hand and increase risks on the other. HEWs, as women, may have diminished ability to endure such risks as they have fewer resources and networks.
HEWs face large workloads and technological interventions should support, rather than undermine. In this study, changed power dynamics and some tensions were reported between HEWs as one of the HEWs adopted the phone for her personal use. While tensions over phone ownership and phonerelated activities affecting work burdens have been reported between couples, 12 to our knowledge this has not been reported within health system cadres. This is an example of how, even with the best of intentions, there is still opportunity for gender and power dynamics to play out in unexpected ways.

Limitations of this study
First, our study focused on a relatively small subset of HEWs in one region of Ethiopia and this context may differ from other parts of the country. 38 Secondly, as this intervention was a pilot, we could not explore how mHealth technology played out in gendered ways across all health packages. Thirdly, we must consider our positionality as researchers. While every effort was made to ensure participants understood the confidentiality and were able to speak openly, it may be that they saw the (local and otherwise) data collectors as project staff, government workers or 'outsiders' and tailored their answers accordingly. 39 It is also important to situate our findings within the political context of Ethiopia, which may limit freedom of speech during the study period. 40

Conclusion and recommendations
Although intended to enhance female HEWs' role in Ethiopia's health system, introducing technology without addressing power relations or other dimensions of their work can bring challenges. Specific actions could make the intervention more gender transformative: distribute smartphones to all HEWs to avoid creating inequalities; ensure workloads are equally shared and that all HEWs are given opportunity to upgrade their skills. Additionally, scale-up of the intervention may alleviate HEWs' workload and build their skills as data collection is streamlined across all 16 health packages. However, this will require further research and technical support for troubleshooting that, while manageable in a pilot, may cause delays at scale.
Supportive policy change that fosters progressive changes in the underlying power relations and in the structure of the health system should challenge patriarchy in the household, community and health system. It should recognize women's rights as individuals; challenge norms that equate household and reproductive work to women's work; create opportunities for HEWs to engage in policy-making processes; 34 and enable progression to more senior positions, such as that of supervisor. Failure to address these dimensions may mean that HEWs' mobile phones reproduce a cultural gender imbalance that may be holding this cadre back. -If so, probe further, have you ever felt threatened or in danger due to the phone? ▪ If not, why not?
Thank you so much for time. Is there anything else you want to share with me?
Topic guide -In-depth interview guides for Supervisors in Ethiopia Introduction -Supervisors: Thank you for taking the time to speak to me. I would like to understand a little bit more about the mobile phone programme that has recently been implemented for HEWs in your district and how this has impacted your work and life outside of work. If you don't want to answer any of the questions in the interview please let me know and we can move on to something different. If any of the questions are unclear please tell me and I am happy to ask them in a different way.

Topic Guide Focus Group Discussions with Community Leaders
Introduction -Community leaders Thank you for taking the time to speak to me. I would like to understand a little bit more about the mobile phone programme that has recently been implemented for HEWs in your district and how this has their role and status in society. If you don't want to answer any of the questions in the interview please let me know and we can move on to something different. If any of the questions are unclear please tell me and I am happy to ask them in a different way.
Perception of HEWs Firstly can you tell me a bit about how women become HEWs in your kebele?
ii30 JOURNAL OF PUBLIC HEALTH -Probe: selected from community, self-selecting? Do you think it is seen as a desirable job? -Probes-why?
Can you describe the perception of HEWs in this community? -Probes: Do you think that everyone views them in this way, or could men/ /women/children perceive them differently? If so, in what waysplease describe? -How does this compare to other women in the community? -How does this compare to men in community? -Are HEWs valued in community?
o Probe -Why?
Are HEWs supported in their work? by community members by health system -Or in other ways?

Impact on work
Are you aware of the mobile phones that HEWs in your kabele are now using in their work? (If not briefly explain project and ask can ask the questions hypothetically) What impact do you think this is having on their work, if any? -Probe on community perception of the work they are doing and how this may have changed since using the phones