Digital Performance Based Incentives CHWs

Despite the significant investments already made in Health Information Systems in Kenya, community health continues to lag behind in this regard, with many of its reporting systems still paper based and therefore error-prone and inefficient. Deficiencies in basic community-based data collection and reporting tools hinder the optimal use of community health data to inform public health response and resource allocation at county and national levels. As a result, resource allocation across geographic and disease areas are inequitable, and result in poor utilization of primary health care by those who need it most.  

Kenya has a devolved health system, with 47 counties responsible for resource allocation and improvement of primary health facilities. Recent legislature allows for gathering revenues at the primary health facility level, as well as decentralized procurement, increasing autonomy for primary health facilities at expense of county and national government control. Without a clear view on the issues that are at hand at the community level, decision making at primary health facility and sub-county level becomes problematic. As a result, stockouts of common drugs and commodities are frequent, resulting in people skipping the primary care level and seeking care at higher levels of care, including expensive specialized hospitals, resulting into inefficiencies due to overstretched resources and over-burdened health care workers at the hospitals. Ultimately households often end up incurring catastrophic health expenditures not because of their wish, but due to lack of access to functional primary care systems built on a foundation of well-coordinated and functional community units.   

In Kenya, over 86,000 community health workers (CHWs) are responsible for gathering household level data and reporting to the formal system. There is a general lack of government funding for CHWs as part of the primary health care system; no efficient ways to track and trace their activities; and lack of enforcement of community health bills that were passed in legislation in the past two years.  

Research has shown that timely, consistent and predictable rewards to CHWs are a driver for high return on investment on community health worker programs. Thus, next to attrition amongst CHWs, resources from donors and governments are being wasted by not adequately rewarding and supporting community health workers. 

Our solution seeks to improve data quality and availability at the primary health care level by providing community health workers (CHWs) with an integrated platform that provides them with the adequate tools, training and incentivization to perform their duties of health promotion, referrals and community monitoring. The platform integrates the existing digital solutions that Amref deployed in the last decade within the community and primary health space across sub-Saharan Africa: 

• M-Jali, is a smartphone-based data gathering and analysis platform, that is used by CHWs to gather household level data. It uses standardized, ministry of health approved forms, while integrating with the Kenya national health information system (KHIS), that is used at health facilities and hospitals both at the county and national levels. Our digital platform enables tracking of CHW performance as visualized from various metrics including geoinformation of households visited. Other metrics that can be obtained include quantification of work done per disease area (mainly household sensitization), any screening and referral efforts and any interventions provided by the CHW.

• Leap: This is a mobile learning solution, that equips CHWs with the adequate knowledge to perform their roles. Leap delivers bit-sized content via IVR, text or a smartphone-based application to CHWs, according to a tested learning pedagogy. The platform has been used in 8 countries for covid response and separately in 3 for basic curriculum CHW training. Leap is a solution that does not require any internet connection, and works on basic mobile phones.  

• Integration of Leap and M-Jali with m-Pesa, the mobile payment platform in Kenya, so that CHWs can be paid based on their learning results (Leap) and reporting activities (M-Jali). Through this integration, the learning and reporting activities of a community health worker can be tracked, and payment is automated so that any delays at the county level may be prevented.
  

Additional CHW incentivizes that will be explored includes:

• Subsidized registration of CHWs into National Health Insurance Fund (NHIF) by the counties, and; 

• Possible payment of commissions to CHWs for households that they newly register into NHIF and routinely track for adherence of payment, as routine part of their work using the M-Jali platform; 

• and we look at the financial sustainability of the remuneration system for community health workers 

Our solution will entail strengthening the performance measurement and improvement of primary health care services through the use of community health workers (CHWs) in Kakamega, Kilifi, Migori, Taita Taveta, and Meru counties in Kenya. In our socio-economic mapping exercise of 2021, we identified over 168 thousand poor and vulnerable households that were unable to pay for their health services unless supported by their respective governments. They were uninsured, and heavily relying on out-of-pocket expenditures to pay for their health needs. Out of Pocket expenditures have a high chance of pushing already poor households to extreme poverty levels; not considering the current pandemic situation which has as well adversely disrupted the economic well-being for majority of the households. In addition, realizing a 5-kilometer radius as defined by WHO is still a major challenge for the majority of these counties prompting households to incur both medical and non-medical costs when seeking health care services.  

Consequently, primary health care access and health outcomes for this population are lagging behind;

• In Kakamega county, only 52.5% of home deliveries are referred for Post Natal Care (PNC) service in 2021, and only 15.5% households had up to date health insurance.
• In Kilifi County, 66.5% households did not have refuse disposal facilities, and only 9.3% primary care level households had up to date health insurance. In addition, 56.1% of pregnant women were not referred for Ante Natal Care (ANC) services.

• For Taita Taveta county; only 7.2% of households had up to date health insurance, and 69.4% pregnant women were not referred for the critical ANC service as required in the context of community health service.
  

• In Meru county, uptake of maternal health services was poor where 53.2% of the pregnant women were referred for ANC. Consequently, only 11.1% of households reported having up to date health insurance.
  

• Migori, a county that lies within the Malaria endemic zone reported that only 18.5% households had up to date health insurance, and that 52.7% pregnant women were not referred for ANC services. 

For this particular solution, we will strengthen community service system by empowering community health workers through trainings on primary health care indicators relevant to them, health issues to be addressed, use of digital tools for continuous learning and reporting, and delivery of health promotion and awareness creation. Through our solution, that links the activities of the CHWs, to adequate knowledge and appropriate incentives, we can tap into the full potential of CHWs, who are then much better positioned to service their communities.  

We have build a strong track record in the primary health care data gathering and improvement domain in sub-Saharan Africa:

• Previous experience deploying M-Jali for data collection and reporting using CHWs in 10 Counties in Kenya. 

• Amref has strong government relations to various interventions; existing signed MoU’s and project agreements with all counties in Kenya.  

• Strong presence throughout sub-Saharan Africa through COVID-related work with Africa CDC in 50 countries  

• Part of different entities focused on improvement of primary health care and universal health coverage, such as Strategic Purchasing Africa Resource Center (SPARC); knowledge management repository on implementing UHC in SSA 

• Doing UHC delivery hub for Africa, looking at strengthening Technical Assistance to counties and governments on areas surrounding health financing (SEMA) 

• Scaling up community health systems, to pass CHW bills in counties; advocating for facility improvement fund to ringfence funds allocated to PHC (the counties that we are proposing to work in have adopted this) 

• Technologies have been used by the different counties already, and are adopted for further use in selected counties, such as Kakamega. 

Key staff:

Dr Catherine Kanari is a medical doctor with public health specialization currently the Universal Health Coverage Lead at Amref Health Africa. Universal health coverage stream focuses on bringing to scale replicable models of health solutions infused with innovations and leveraging on Amref Health Africa’s evidence-based solutions gained from 60 years of implementing programs and health solutions for Africa. Catherine has garnered more than 16 years of skills and expertise in implementation of sustainable scalable programs including managing several grants with portfolios ranging up to USD 20 million annually and overseeing project implementation across several health areas, incorporating social determinants of health including behavior change communication, social franchising and social marketing of health-related products. The portfolios include but are not limited to funding from various notable donors including USAID, CDC, CIFF, BMGF, FCDO and several private sector players including big pharmaceutical companies.

Peter Otieno: Peter has a background in Innovation and Social Entrepreneurship, his experience involves scaling up and adoption of Technology Driven Solutions in different markets such as Kenya, South Africa, Zambia, Malawi, Rwanda, Ethiopia and Kenya, through different digital platforms Leap and M-jali. His different roles in Amref have seen him managing operations, delivery and technology driven projects, where he has dealt with entities from the public sector such as Ministries of Health and Government institutions, Corporate and Regulatory bodies. He is a Professional Member of the Chartered Institute for IT (MBCS) and has contributed in development of the National Digital Health Strategy in Kenya, Expert Panel for Digital Health for NCDs in LMICS’ through a study conducted by Novartis Foundation

Bill Olwenda is currently Leading Government Relations & Public Private Partnerships for Primary Healthcare, at Amref Health Innovations. He has over 8 years’ experience in Social Enterprise Business and brings on board expertise in managing cross-sector partnerships; new approaches to hybrid value chains, and shared value. A holder of an MBA and an ardent social enterprise intrapreneur, Bill has held leadership positions in Health Financing Advisory and Execution of Innovative and Sustainable Business Models across different industries in Africa. Most recently he was seconded as a member of the Amref Global Health Financing Community of Practice (HFCoP) by Amref Health Innovations and Amref Netherlands.

Eric is a Monitoring and Evaluation professional working at Amref Health Innovations (AHI) as the lead M&E lead. His experience spans several thematic areas including; primary care, health financing, disease control & prevention, Non-communicable diseases, Assistive Technology, Reproductive, Maternal and Child Health.  He also supports program development and implementation of m-health platforms across Kenya and the Eastern Africa region. He is currently pursuing post-graduate studies in M&E and holds an Actuarial Science degree.

Raymond holds a Bachelor’s degree in applied mathematics and currently pursuing a masters degree in Monitoring and Evaluation. I support Monitoring and Evaluation at Amref Health Innovations (AHI) in Non communicable Diseases and Primary care. I have previously worked in Supporting Monitoring and Evaluation and Research on Social Behavior Change and Reproductive, Maternal and Child Health in Sudan and Ethiopia.

Aranka Hetyey is a hands-on venture builder, applying her business and technology acumen gained at Accenture Strategy Consulting to develop financially sustainable health solutions at Amref. Aranka was part of a startup that provided investment services and software to private and corporate equity players in Europe. Aranka majored in business information management at the Erasmus University, Rotterdam school of Business, with an exchange to University of Washington in Seattle.

Sadiq is a software developer with over five years experience and expertise in the design, implementation, and management of information systems. He serves as a software developer at Amref Health Innovations. Sadiq is passionate about providing high quality technical digital solutions to everyday problems. He previously worked for the Nairobi West Hospital and was among the key team that designed and implemented their Covid-19 testing, vaccination and certificate verification digital platform. He also consulted for Laborex East Africa, Amref Enterprises and Amref Health Africa. Sadiq holds a Bachelor of Science in Informatics degree from Moi University, Kenya where he majored in software engineering.

Jochoniah Nzomo is Research Data Analyst working with Amref Health Africa to realize a lasting health change in Africa through the use of Research and Data Science methodologies that draw insights and exploit opportunities centred around addressing Primary Health Care solutions suitable for realizing Universal Health Coverage in community context.

To achieve universal health coverage the driving engine is primary health care. In many African countries community health systems mainly resourced by community health workers are essential to this system. Remuneration of these key resources should be prioritized and done in a comprehensive manner to motivate and ensure retention of this resource. In the past 8 years, we have refined our data gathering tool. The technology is ready, yet challenges remain with the human resource aspect of data gathering and reporting for performance improvement in the primary health care space. The CHW Digital Performance based incentives platform has been designed in close collaboration with the end users and county government to address this challenge. Now we need to test our model to confirm the assumptions that we have made and to ensure the integration between the different platforms works as intended.

As such, the key barriers we want to address with this challenge are:

• Technical: Leap and M-Jali are mature solutions that have been tested at scale already. The integration with M-Pesa has not yet been tested. With this funding, we would like to test the integrated technical solution with approximately 6,000 CHWs
• Financial: a key challenge is the consistent and transparent remuneration of CHWs, with poor economic conditions and motivation as a result. Our assumption is by digitizing this process, and designing an remuneration structure together with the CHWs, we should see an improvement transparency and consistency of rewards of CHWs. This needs to be pressure tested.
• Cultural: although county governments have passed the community health bill that formalizes the payments to CHWs, and have allocated budget for the same, they are lacking behind in execution. Historically, CHWs have been under-recognized, -utilized and -acknowledged for their impact, and they hold very little bargaining power as CHWs are not organized in formal bodies to represent them. It is very difficult for CHWs to hold county governments accountable. With this project we are aiming to change this, however this requires a cultural shift in a domain that governments sometimes prefer to keep opaque. The litmus test is not designing the framework, but actually getting it implemented. And for this, we need all the force we can get, including the support of MIT Solve.

Our platform is the first of its kind, delivering performance based incentives to CHWs, linked to reporting, referrals and learning performance.

The way that we combine our proven solutions in a new value proposition is what makes our solution stand out. Our training of trainers (ToT) approach ensures buy-in and active participation of the supervisors of the CHWs, the community health assistants. Our platform provides real-time visibility on the performance of CHWs, linked to the geolocation of the households that the CHWs are reaching.

We leverage on: the high uptake of mobile money services in Kenya (60 % of the population), increasing smartphone ownership (30% of the population) and the progressive legislation that several county assemblies have passed with regards to recognizing CHWs as part of the health care workforce who also deserve to be trained, equipped and remunerated in order to effectively deliver on their mandate. Through our solution, we strengthen community health units (CHUs) by training and equipping CHWs to effectively gather household level health data in an accurate, timely and cost-effective manner. By automating rewards for performance output, we hope to develop a predictable, and reliable model that can be replicated and scaled. Through this project, we will develop best practices that can subsequently be adopted as a best practice for counties. Historically, rewards for CHWs have been inconsistent because they were part of donor funded projects that came and went or were fickle due to governments not meeting their promises.

By integrating the findings from our socio-economic mapping work during the pandemic into this new proposition, we ensure that the communities and their health workers support our approach.

Our impact goal is to improve community health service delivery through performance based incentivization for CHWs.

In short, our goals are:

1. Improved community health service delivery
2. Improved financial resilience of CHWs
3. Evidence-informed policy formulation about renumeration of CHWs

Our aim is to strengthen primary health care through improved community health service delivery as a result of digitized community based information systems; and automated performance based incentives for CHWs for data gathering, health promotions and referrals. We will achieve our goals by piloting our proposed solution in six counties in Kenya, and using our evidence to conduct advocacy for performance renumeration of CHWs and scaling up of our digital platforms.

The intervention’s progress towards impact will be undertaken through a longitudinal study that will deploy impact evaluation methods for policy research - the quasi-experimental approach to provide learnings from intervention and control groups. 

Specifically, we will assess:

• Learning performance on basic CHW modules (such as RMNCH, NCDs etc)
• Change in data collection and reporting ratesby CHWs
• Proportion CHWs receiving incentives and are subscribed to NHIF

And how these will link together.

Our  theory of change is that primary health care can be strengthened through improved community health service delivery as a result of digitized community based information systems; and automated performance based incentives for CHWs for data gathering, health promotions and referrals. Our integrated digital platform supports these activities, so that performance improvement can be both driven and delivered by the communities themselves.

Our three pillars are:

1. Community based service delivery (prevention, treatment, control and referrals) - Communities and households receive appropriate and effective prevention and treatment services of communicable and non-communicable diseases and are referred to relevant service providers as appropriate at community level.
2. Health promotion -  Health workers are trained with mobile learning in basic concepts, modules, mentored, and supervised to undertake their role adequately.
3. Incentives - Improve the performance and effectiveness of community health units by adequate recognition and renumeration of CHWs, predictable and on a timely basis.

Improve financial resilience of CHWs by increasing NHIF coverage of CHWs by subsidizing their premiums based on their actual output as CHWs (performance based).

Digital tools are core enablers throughout the three aspects to ensure consistency and transparency of the intervention.

M-Jali

“Mobile-Jamii Afya Link”, shortened as M-Jali, incorporates a mobile application for capturing data from the household level and transmitting it online to a web-based database. The device-side application runs on an Android platform, which transfers data via a GPRS/3G network to the back-end database which is supported by a web application enabling data validation, report generation and information sharing. It incorporates alerts and flags to possible intervention and action areas to improve health care and health care related problems based on data as and when it is collected.

M-Jali is a flexible system, which allows for tailoring according to the specific data management needs in different contexts.

Front  end - M-jali Mobile Application: Built on Android Native programming language

• M-jali Administrative Web Framework: built on Python (Django Framework)
• M-jali Dashboard Web Framework: Built on Python (Flask Framework)

M-jali API (Application Programming Interface):

• Built on Python (Flask Framework)

DATABASE MANAGEMENT

• PostgreSQL
•  Reddis

Leap

Leap is a mobile learning solution for training health care workers, that is based on an SMS and IVR gateway that is integrated to a web-based learning management system.

• Java EE 8
• Glassfish
• PostgreSQL
• Linux
• Antlr

Attributes of Leap

• Multi-tennancy
• Point-in-time recovery (PITR) 
• SMS enabled Identity Management
• Stand-alone desktop command line compiler for developers developing apps on the platform
• SOA architecture
• ISO/IEC 27001:2013 compliant
• Supports Mobile Network Operator APIs for mobile money and USSD
• Supports inbound and outbound SMS
• Scheduler to execute functions at intervals.
• Logic and data objects can be very easily exposed as RESTful APIs

Trained community health workers collect data using M-Jali. Data is reported to health facility level, Community health assistants check the data and amend if needed, and then upload to KHIS. It then rolls up to the level of the county, where the Health Information Records Officer reviews the data and creates reports for the county and national MoH decision makers.

Amref’s promise is to improve the lives of disadvantaged people in Africa and bridge the gap between communities, health systems and governments. With over 65 years of experience working in the most remote African communities, we have built a wealth of knowledge to reach the poorest of the poor.

We incorporate diversity, equity and inclusion through a three-steps approach: identification of vulnerable groups; implementation of agreeable social inclusion strategies and monitoring of impact on the specific groups.

Identification of vulnerable groups in this project includes both the CHWs and the population of the six selected counties. In our previous project, we identified vulnerable Bottom-of-the-pyramid households that are in need to improved access to primary health care services. In alignment with the counties, we can leverage these lists to target our continued data gathering activities to this group. This will bring out specific needs that the CHWs and health facilities can use to target their service delivery towards.

With regards to the CHWs, we can build on our experience on identifying disabled and marginalized CHWs and supporting their reporting activities through our projects.

The social inclusion strategies will be aligned with the community leaders as well as county governments.

The monitoring frameworks will include standardized tools to measure inclusivity of our approach, for example incorporating our socio-economic status mapping questionnaire to identify and track poor households, as well as segmenting our impact measurements across the different vulnerable groups including people with disability and youth.

M-Jali is developed by Amref Health Innovations (AHI), which is an integral part of Amref Health Africa. AHI covers the expenses for the development and deployment of M-Jali primarily through donor funding as well private partner sponsored projects. Any surplus remaining at year-end is reploughed into the non-profit (Amref Health Africa) and utilized for high-impact projects.

In the projects, we charge a license fee for M-Jali (4.5$ per CHW / month); software development costs for customization; as well as consultancy / project implementation staff costs.

We charge at-cost rate internally, e.g towards other Amref entities such as country offices, and at a slight surplus for external customers, such as private sector players.

Eventually, we will target the government directly as paying customer. Although there is an increase in the number of counties that have adopted digital solutions for health, only a limited number of counties allocate their budget to support the same. For the sustainability of the solutions, payment by the governments is favorable. However, from our experience there are many challenges getting commitments from the government to pay for digital solutions. Therefore, funding by donors is still the most viable option in the short to medium term, while the governments make progress in institutionalizing digital solutions.

The path towards financial sustainability is achieved in three steps: establishment of a strong core platform; adding value added services; and integration into the systems. In the past 8 years, we have focused on the first step, which is building the core platform. Through donor funding, the platform has been sustained and developed further. Currently, we are in the process of designing and building value added services. After pressure testing these services, integration within the broader ICT systems can take place. Through these phases, our intention is to shift from primarily donor funded projects towards a blend of donor funded, private sector and government funded projects.

Through the initial donor funding, we have grown the base of CHWs that are onboarded on the platform. This CHW base can be deployed to new projects in the same counties. With each donor-funded project, the costs per household visited declines, due to lower CHW training expenses. As the different counties are passing the community health strategy bills, their need for a platform that facilities data gathering at the community level with the incentivization of CHWs grows. We are looking to play into this development by supporting the counties in the execution of the bill through our platform.

The core of M-Jali is household level data gathering, focused on health data, as well as tracking referrals and service delivery. This functionality has been tested in different projects in Kenya, for example the Innovative Partnership for Universal Sustainable Healthcare project (I-Push), focused on maternal, neonatal and child health and Blueprint project, focused on non-communicable diseases. These projects were donor funded. As a result of the I-Push project, Kakamega county is currently embedding M-Jali in its processes. As a result of the Blueprint project, Meru county is now allocating budget towards primary health care, specifically for NCDs. NCDs were first not allocated in the annual development budgets. By showing that investments in NCDs actually creates impact, we were able to advocate for inclusion of NCDs in the primary health care budget. We also conducted advocacy to get stipends for 2000 out of 3500 CHWs in Meru county through Leap completion and use of Mjali to complete household mapping. The county were able to supply CHWs with phones to download M-jali and use it for their work. Also M-Jali is used to input screening data for NCDs as well as referral data.

In 2020, we started to develop value added services, specifically, the poverty mapping we conducted in 2021. Between November 2020 and July 2021, we partnered with the Bill and Melinda Gates Foundation (BMGF) to implement a project dubbed, “Expanding Financial Resilience” during COVID-19 Project (EFIRE). The goal of the Project was to provide accurate data and evidence to help monitor and mitigate the social and economic effects of the health crisis by expanding the financial resilience of poor and vulnerable households during the pandemic. EFIRE project was successfully implemented in 6 counties: namely Kakamega, Kilifi, Makueni, Meru, Migori and Taita Taveta. The Project managed to reach over 480,000 households across the 6 counties by the time it closed. Over 158 thousand indigent-households were identified on an accurate, timely, non-biased and affordable manner.  

In total, through donor-funded projects, we have reached 11,690 CHWs who have mapped a total population of 2,555,674 across 10 counties in Kenya. We have reached a further 105 CHWs in Zambia. With our digital learning platform Leap we reached over 162,000 CHWs. 

Key partners that supported this work are the Bill & Melinda Gates Foundation, Africa CDC, Takeda, Roche, GSK, Achmea Foundation, Pepfar, amongst others, around 300k$ in grant funding for developing and deploying the different in-scope technologies.