Smartphones for community health in rural Cambodia: A feasibility study

Background: Village Malaria Workers (VMWs) are lay people trained to provide a valuable role in frontline testing and treatment of malaria in rural villages in Cambodia. Emergence of artemisinin-resistant malaria highlights the essential role of such VMWs in surveillance and early treatment of malaria. Smartphone technology offers huge potential to support VMWs in isolated and resource-poor settings. Methods: We investigated the feasibility of issuing established VMWs with a smartphone, bespoke Android application and solar charger to support their role. 27 VMWs in Kampong Cham and Kratie provinces participated. Results: 26/27 of the smartphones deployed were working well at study completion twelve months later. Interviews with VMWs using quantitative and qualitative methods revealed pride, ease of use and reports of faster communication with the smartphone. VMWs also expressed a strong wish to help people presenting with non-malarial fever, for which further potential supportive smartphone applications are increasingly available. Conclusions: As a result of this pilot study, two smartphone based reporting systems for malaria have been developed at the Cambodian National Malaria Center, and the programme is now being extended nationwide. The full code for the smartphone application is made available to other researchers and healthcare providers with this article. Smartphones represent a feasible platform for developing the VMW role to include other health conditions, thus maintaining the relevance of these important community health workers.


Introduction
More than 3 billion people, including 70% of the world's poorest people, live in rural areas 1 , where access to skilled medical care may be difficult 2 , especially during the rainy season. Trained lay people in target communities can offer, via supported schemes, a valuable contribution to diagnosis and management of medical conditions such as malaria, childbirth and childhood diarrhoea 3 . However, giving sufficient support to such community workers, including training updates, diagnostic support, stock control and data collection is challenging. Mobile phone technology provides the potential to improve the range and quality of services delivered by community health workers 4 , but the majority of published literature on feasibility in developing countries is focussed on standard mobile phones ("feature phones") rather than the next generation of smartphones.
The Cambodian National Malaria Center (CNM) has an established network of Village Malaria Workers (VMWs) across 19 provinces covering a population of approximately 1.5 million people 5 . VMWs review villagers with fever, test them for malaria with a rapid test (SD Bioline Malaria Ag Pf/Pan; Standard Diagnostics Inc, Gyeonggi-do, Republic of Korea) and treat those with positive results with artemesinin-based antimalarial therapy. Rapid case reporting by coded text message (short message service; SMS) from feature phones has been established by CNM. We set out to evaluate the feasibility of replacing the feature phone part of this reporting system with smartphones and a bespoke Android-powered app (smartphone application).

Methods
A new smartphone application ("app") was designed and built for Android phones using the Open Data Kit platform version 1.3 6,7 . All instructions for installation, configuration, implementation and operation are explained in detail with examples on the Open Data Kit website. The minimum system requirements for the smartphone are Android devices 4.0 and above, and the phone should be connected to the Google App Engine hosting platform.
Code is available 18 . The app was designed to be simple to use, not require current network access to be operational, and allowed the VMW to enter case report data in an intuitive way using Khmer language ( Figure 1). The information recorded by the app and sent to CNM was as shown in the screenshots in Figure 1, namely gender, age, malaria rapid test result, and residency status (permanent villager or mobile migrant). Experienced VMWs in Kampong Cham and Kratie provinces ( Figure 2) were trained to use the new smartphone system. The numbers were limited by the cost of the phones. This was a service improvement project for the VMW network and ethical permission was not sought by the authors.
VMWs were called to a meeting and the proposed study explained by CMN staff. All VMWs gave verbal consent to participate in the study, and then VMWs received one smartphone (Acer Liquid Z3, Acer, Virginia, USA) and one solar power charger (

Results
Twenty-seven experienced VMWs in Kampong Cham and Kratie provinces were trained to use the new smartphone system ( Figure 3). VMWs in Kampong Cham province had previous experience with feature phones, while those in Kratie province had no experience of using phones in their work. Populations   were broadly similar in the two provinces in demographics and socio-economic factors. The average distance of each village to the nearest health centre was 7.53km (range 2.9-10.3 km) for Kampong Cham province, and 16.9km (range 4.57-39.1 km) for Kratie province. Road access to the health centre during the rainy season was typically considered impossible or extremely difficult.
The results of the pre-implementation in-person questionnaire at the outset of the study are shown in Table 1 and Dataset 2.
15/27 (56%) of the VMWs were female and the mean age was 39 years (range 20 -62). 100% were literate (a requirement for being a VMW) with the highest level of education completed being primary school level only for 81%, and all had been in post for more than three years. 85% already owned a feature phone, but only three VMWs (11%) owned a smartphone and the majority (85%) had never used one. The baseline survey of expectations showed 78% were expecting the smartphones to make their work easier and 70% were excited/proud to be using them. Motivations for being a VMW were to help others (78%), stop malaria (63%), Education beyond primary, n (%) 1 (7) 4 (31)
At study completion 12 months later, a follow-up survey (Dataset 3) showed 26/27 smartphones (96%) and 27/27 solar chargers (100%) issued to VMWs were working well. One USB connector on the smartphone was broken so it could not be charged. In addition, one smartphone was stolen during the training session prior to issue. CNM reported that training was quicker and easier to conduct than for paper or feature phone data collection methods. 20/27 (74%) VMWs reported cases using smartphones -14/14 (100%) VMWs from Kratie province (no previous feature phone experience) and 6/13 (46%) from Kampong Cham province (with previous feature phone experience) reported data. For the 7 VMWs who did not report cases using smartphone, 2 recorded malaria cases by paper records, while for the remaining 5 there were no malaria cases reported. Negative test results were not routinely reported by SMS in accordance to CNM policy (due to funding limits on the number of SMSs allocated to the scheme).
All 27 VMWs were interviewed by CNM health workers at the end of the study using a semi-structured format with a topic guide, to evaluate experience and attitudes (Supplementary Data 3). Three main themes emerged: enjoyment and pride of owning the smartphone, finding it easier to use than a feature phone for their role (Kampong Cham province), and thinking that it allowed faster communication.
VMWs were asked about what people in their village said about the smartphone and the key themes were curiosity, envy, surprise and being aware of the phones attracting a lot of attention in the village.
The pre-implementation questionnaire found evidence that in villages with much reduced malaria incidence, many villagers who have fever are no longer consulting VMWs. 59% of VMWs reported that only 'some' villagers would see them with a fever. In a 3-month period (Sep-Nov 2014) the information captured by the smartphone app showed tests on 89 febrile villagers were reported in the 13 villages in Kampong Cham province (mean 7 per village, range 3-10) including 19 malaria positive cases. In contrast, in the same period in Kratie, 688 tests on febrile villagers were reported (mean 49 per village, range 25-75) including 371 malaria positive cases. We also found evidence of frustration amongst VMWs that they were unable to help those with non-malarial fever, with 17/27 (63%) wanting to learn more about healthcare.

Discussion
This pilot study demonstrates the feasibility of using a smartphone with a bespoke app to support community health workers in rural villages in low-income countries. VMWs reported pride and enjoyment at having use of a smartphone for their role, and felt it allowed faster communication. The smartphone raised the profile of the VMW in their village.
A previous systematic review of 42 studies of using mobile technology in developing countries 9 has shown the feasibility and some descriptive evidence of effectiveness of delivering healthcare using predominantly feature phones by professional front line workers such as midwives, pharmacists, nurses, doctors and some Community Health Workers (CHWs). Innovative studies are emerging demonstrating the role of smartphone apps delivered by professional healthcare workers in developing countries for health such as hearing screening in South Africa 10 and visual acuity testing in Kenya 11 . In addition, use of a smartphone-based electronic decision support system by CHWs has shown promise for cardiovascular management in rural Tibet and India 12 . The current study expands the literature to demonstrate feasibility and acceptability in an impoverished population of Cambodian CHWs, to support malaria control strategies.
The emergence and spread of artemisinin resistant malaria in Cambodia 13-15 is a major threat to global health, thus control of malaria in rural Cambodia is of the utmost importance. We identified low numbers of malaria cases detected by the VMWs in this study, with many VMWs feeling frustrated at being unable to help those with non-malaria fever under the current scheme. After a number of 'negative' tests people might not consult the VMW with a fever anymore, thus the sentinel role of the VMW scheme in surveillance for malaria is reduced. This undermines malaria control programs because future outbreaks of malaria could go unnoticed and then spread to surrounding regions.
Based on the evidence of this study and the national capacity built in the relevant coding skills, the co-authors at the national control program in Cambodia are now developing two smartphone based reporting systems for malaria. At the VMW level, the application developed in this study has been extended to record all the case data recorded by VMW in paper records. This information is then uploaded directly to the national malaria database after entry. There is a plan to install this application on 3300 smartphones for introduction in 14 districts, 22 districts, 9 districts in 2018, 2019, 2020, respectively, with a view to paperless reporting in the future. An additional feature of an SMS alerting system has been included in selected pilot areas, which immediately reports every new case to every level of the surveillance system. This system has been integrated into a reactive case detection intervention in the piloted areas. At the health centre level, a similar application has been developed which focuses on case reports and stock-out alerts for treatments and diagnostics. This application has been installed on tablets in 816 health centres covering the entire malaria endemic area of Cambodia.
The limitations of this study are the small numbers of VMW involved, the short duration of follow up and the restricted amount of information collected on follow-up. These restrictions are due to the project being conducted with no extra resources available. In addition, the extension of the project by the CNM due to excellent informal feedback meant longer follow up was no longer feasible or deemed necessary.
The VMW scheme in Cambodia urgently needs to be enhanced and extended to be the front line in monitoring for drug resistance and future epidemics. To achieve this, and to maintain its relevance to local communities, it should also become a service for other health problems, transforming VMWs into CHWs. These CHWs would treat or refer patients with other health concerns making them the first point of contact in a village. We have demonstrated smartphone technology to be a robust platform for delivery of these services in the local language, enabling use of voice calls, SMS, photography, video, audio and GPS location tracking. As telecommunication network coverage and mobile use expand globally, exploitation of smartphone applications hold growing promise to tackle the world's greatest health issues. This highly successful pilot project of introducing smartphone-based reporting for malaria has allowed the CNM to develop the platform for implementation.

Competing information
No competing interests were disclosed. Overall, I felt the article was well-written and succinct, identifying a relevant improvement to support the work of an important niche health service provider (the VMW). Data collection for the study involved administration of a pre-study questionnaire, and in person questionnaire to collect qualitative and quantitative data on volunteers' experiences. Outcome measures were (1) Survival of smart phones and chargers over 12 months post-study commencement (as determined by inspection of the equipment by CNM staff) and (2) acceptability of smart phone use by VMWs (emerging themes from the in-person questionnaires were collated).

Grant information
The study made the assertion "the current study expands the literature to demonstrate feasibility and acceptability in an impoverished population of Cambodian CHWs, to support malaria ." In my view, however, the methods were not sufficient to capture some key feasibility control strategies and accessibility information. Furthermore, it appears that in some communities, the role of VMW is becoming obsolete in some areas "In the context of their current role in villages with reduced The authors malaria, villagers presenting with fever are frequently not consulting VMWs." acknowledge the limitations posed by the small number of VMWs in the pilot study, the short duration of follow-up, and the restricted amount of information collected on follow-up, and also suggest the necessity to enlarge the role of VMWs to include more CHW responsibilities. the authors might consider demonstrating how readily the VMW role could be Recommendation: expanded to a more general CHW role, which to me would ably make the case for feasibility. Most developing countries are plagued by a range of infectious diseases and finding broader public health use of technology would be more beneficial.

Assessing feasibility of the proposed implementation
The study was conducted in two health areas in communities that had different malaria profiles, with VMWs who had differing levels of prior exposure to smart phones, which makes the results feel more compelling. Literacy is a requirement for being a VMW, which may not be the case for community health workers in other settings.
: the authors might consider discussing the ramifications of deploying the app more Recommendation generally in the context of the literacy requirement.

Assessing technological and system capabilities/requirements
1.

3.
Implementation required provision of a smart phone, solar-powered smart phone charger and three hours of instruction.
This amount of training seemed to be adequate and effective, even in the cohort which had little prior exposure to smart phones, which suggests good feasibility in terms of skills transfer/capacity building. The study was carried out over a 12-month period, with the majority phones and chargers being found to be in good condition after that interval, further making the argument for feasibility. The authors acknowledged the short time frame (given the high cost, it would be expected that the lifetime of the phone would need to be significantly higher than one year). However, usage and storage conditions were not specifically addressed, and it was not clear to what extent smart phone durability was tested in real world conditions. If the expectation is that the phone systems are kept in a locked cabinet in the clinic the majority of the time, this may affect feasibility.
the authors might consider addressing how and when phones were accessible to the Recommendation: VMWs.
The mobile app is simple to use, intuitive and does not require current network access to be Advantages: operational. The software code open source and available online, making the study intervention accessible and replicable in that respect. The availability of 99% mobile data coverage in the study setting made it feasible for this study, but would perhaps be a concern for other remote health areas interested in trying this model. the authors might consider discussing the practicalities of implementation in settings Recommendation: with less robust mobile date coverage (however, this may be considered to be outside of the scope of the study)

Concerns
The cost of USD175 per VMW seemed prohibitive to scaling. The article reports that "the a challenge which lies at the heart of numbers were limited by the cost of the phones" feasibility.
: the authors might consider addressing how this cost would be managed (the Recommendation use of "supported schemes" is mentioned early in the text but not elaborated upon when referring to the plan for staged introduction of phones).
The authors chose to combine the findings of pride and excitement/proud to be using the phone. However, this concealed the fact that when disaggregated, the numbers for each category were actually quite low: My concern was that the study may not have adequately captured concerns relating to the use of the smart phone (given the closed nature of the questionnaire), and that problems might therefore be replicated. I would be interested to know why the 2 VMWs used paper records rather than the phones, and whether, if the other 5 VMWs had no cases to report, is introducing an expensive cellphone really feasible? I would further be interested in gaining more specific information about the uptake and use of the technology in the village with high numbers of reporting, and to what extent paper reports were generated in that setting, which is not elaborated in the text.
I felt that the study did not capture how the use of phones would/did make the work easier: I assume it would be related to data reporting, but this is not specified in the article. While the study was able to record that VMWs the smart phones would allow faster communication, it is not felt specified if that was actually the case. The methods used did not capture this question, and it is not 3.
specified if that was actually the case. The methods used did not capture this question, and it is not clear if that was assessed or determined in any way.
The authors make the assertion: "we have demonstrated smartphone technology to be a robust platform for delivery of these services in the local language, enabling use of voice ." Recommendation: the authors calls, sms, photography, videos, audio and gps tracking might consider disaggregating the data from the group which reported a large number of malaria cases using the smartphones to more strongly support this assertion.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Partly

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Partly
No competing interests were disclosed.

Competing Interests:
Referee Expertise: Field epidemiology, infectious disease and child vaccines I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. 05  Overall, a very well written and useful paper delivering useful and practically-relevant evidence.

Major comment
The need to shift from VMWs to CHWs is clearly spelled out in the final paragraph of the 1,2 1 2 1. 1.
The need to shift from VMWs to CHWs is clearly spelled out in the final paragraph of the discussion, but needs to be more clearly and explicitly stated in the abstract.

Minor comments
The following sentence from the second paragraph of the results section is vague: "100% were literate (a requirement for being a VMW) with the highest level of education completed being primary school level only for 81%". Does that mean that 81% had not completed primary school, yet were literate? Or that 81% had primary education and the remainder has secondary or above? I'm guessing the word "only may have been misplaced in this sentence.
Third paragraph of the results section: It is a little unclear whether the smart phone with the damaged USB port is the one counted as not working well. If so, I think it would be worthwhile to make it clear that all 27 were still in place.
The first sentence of the second paragraph of the discussion is too long. Please split in two and make it read easier.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed. Competing Interests: Referee Expertise: Malaria transmission and control I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.