This section describes key results as they relate to the research questions.
What was the experience implementing the DFS program?
Facilitators and Barriers to Successful Implementation
Respondents from both case studies highlighted a variety of common factors that contributed to successful DFS solution implementation. These included technology factors such as: existing mobile/internet network infrastructure and mobile money network operators; strong software development teams; high penetration of mobile phone use across all target clients (see Fig. 1. Mobile Phone Penetration in Rwanda and Kenya); the existence of web-enabled systems facilitating automatic, real-time verification using national ID numbers to ensure the correctness of data and secure electronic financial transactions; a population open to using mobile money for financial transactions - both at the individual and business level (healthcare providers).
People factors were also highlighted. The programs were implemented through initiatives designed to bring together multidisciplinary teams of stakeholders (technology, policy/government administration, health, and finance) from public, private and NGO sectors and building trust between stakeholders (a crucial but sometimes slow process) to enable data sharing and interoperability between independently managed systems.
Both programs benefited from the existence of trusted and functional community service systems. In Kenya, the programs were able to leverage the trusted community health volunteers and workers networks that had established relationships in the community - especially in rural Kakamega County where mothers feared going to hospitals but trusted local health workers from their own community. In Rwanda, the program was able to build upon the network of district- and community-based representatives from Irembo (a public private partnership to enable digital payment for government services) and mobile money agents in remote areas, helping less digitally literate citizens complete their transactions and overcome the digital divide.
The health financing policy environment was also a key enabler in DFS implementation. In both case studies, the DFS schemes were built on top of pre-established functioning ecosystems of health financing (e.g., CBHI in Rwanda; National Hospital Insurance Fund [NHIF] in Kenya). In Rwanda, respondents noted that the government’s policy and vision of digital transformation for government financial services, with the motto “zero paper, zero trips,” drove the change with strong political support. In Kenya, the DFS innovations were driven more by nongovernmental organizations (PharmAccess and AMREF) with social enterprise missions who effectively engaged public and private sector stakeholders.
Both programs also encountered barriers and challenges along the way that required mitigation. Key among these on the technology side were: difficulty negotiating data sharing agreements to enable interoperability between systems managed by different stakeholders (government – national identification (ID) and household/birth registration; insurers - NHIF and CBHI enrollment systems; mobile operators - gateways for financial transactions) and issues related to software platforms themselves (some systems lacked APIs, computer interfaces needed to be updated as systems changed, and some insurance schemes were digitized, which prevented program managers from identifying gaps or overlaps in coverage).
When the Rwanda 3MS development started, there was no interoperable payment gateway available for government programs. Further, some health facilities had spotty internet connectivity which led RSSB staff to attempt an offline system that failed due to challenges synchronizing with the centralized databases.
On the user/beneficiary side, clients and community level agents with low levels of digital literacy were challenged in using some of the DFS services that required familiarity with smartphones. Some of the Rwandan CBHI staff noted that training was insufficient – as the cascaded orientation sessions often focused more on policy changes to the CBHI scheme than on practical exercises using the software. Users in both countries faced financial hardship, exacerbated by COVID-19, and had difficulty making household financial decisions with their limited resources. One i-PUSH program respondent demonstrated the competing priorities they faced: “Do I buy food? Do I pay premium? Do I pay school fees?”
Just as traditional banks have been cautious about lending to private healthcare small and medium enterprises (SMEs) with no credit history or obvious collateral for loans, some health facility leaders were concerned about participating in the mobile lending programs. “Health facility leadership were careful, they were wary, they were even fearful that you want to tap into the M-PESA events, virtual accounts” [when they agreed to share tier financial data]. - MCF program respondent
Table 3, below, summarizes key barriers and facilitators:
Summary of key implementation facilitators and barriers
● Existing mobile/internet network infrastructure and mobile money network operators
● Strong software development teams
● High penetration of mobile phone use and providers and beneficiaries open to or using mobile money for financial transactions
● Existence of web-enabled systems enabling automatic, real-time verification using national ID numbers
● Existing mobile money agent network to extend reach of DFS services (CBHI 3MS)
● Strong community of Kenyan software developers available to manage and improve the M-TIBA platform (i-PUSH and MCF)
● Government’s digital vision: “Zero paper, zero trips” catalyzed change (CBHI 3MS)
● Large network of CHWs to assist with household-level enrollment and bridge the digital literacy divide
● Old infrastructure not up to peak demand; spotty internet in some remote areas (CBHI 3MS)
● Lack of electronic payment gateways and APIs to reliably connect systems managed by different actors (CBHI 3MS)
● Poor quality smart phones produce inadequate images of required certificates for registration (i-PUSH)
● Government health facilities could not apply for mobile credit funds (MCF)
● Low-income households had too many competing demands on their limited income and no experience with savings
● Inadequate onboarding of CBHI and facility staff using 3MS (CBHI 3MS)
|*Facilitators and barriers in bold were identified across both cases. Where a factor was only identified in one case, it is specified in parentheses.|
Barriers identified during program implementation needed to be mitigated in order for the programs to succeed. Program managers from both case studies highlighted examples of how online platforms built to support DFS enabled the programs to be more agile and quick to implement policy changes (e.g., changing insurance coverage wait times, increasing loan limits), tweak the system to make incremental changes and improvements to the user interface, fix bugs, and introduce new features requested by clients. This helped to respond to the challenge of telecommunications companies that were also innovating constantly, so the DFS platforms could evolve and adapt quickly to their innovations.
The COVID-19 pandemic resulted in challenges and opportunities for the DFS programs in both countries. In general, when it came to savings and paying insurance premiums, people had more difficulty paying because of loss of revenue - and job losses hit hardest in the poorer communities that were served by the insurance schemes.
On the other hand, COVID-19's distancing mandates, suggested precautions, and heightened health awareness helped to accelerate uptake of digital health platforms. Both the Rwandan and Kenyan governments encouraged more electronic payments to reduce in-person interactions. In Rwanda, citizens were motivated to enroll in CBHI because they perceived a greater likelihood of getting sick.
Rwandan CBHI program managers felt that the DFS system contributed to resilience during COVID-19, as payments could be made from home during lockdown, unlike through traditional channels that handled cash payments such as banks that reduced work hours and were less geographically accessible. Citizens’ ability to access these services remotely also reduced potential exposure to COVID-19 virus.
In Kenya, the i-PUSH enrollment work faced challenges throughout the pandemic. As some of this required in-person interface with households, the program had to cease activity when surges resulted in lockdowns. Further, requirements around social distancing, wearing proper personal protective equipment, and limiting large group gatherings forced the need for additional trainings with smaller groups at higher cost. Loan agents also faced challenges initially, as they were forced to adapt their marketing efforts to health facilities remotely; however, they adapted by doing more virtual phone calls. Businesses, like hospitals, turned more to MCF’s loans to cover losses in revenue and to purchase specialized equipment required to treat COVID-19 patients (e.g., ventilators).
How was the program perceived to influence health systems performance?
Respondents from both case studies perceived that the DFS for health programs contributed to improved health system performance, including aspects of data quality and use, and improved quality of care, responsiveness, and efficiency. The DFS initiatives also supported national eGovernment initiatives to move from manual to automated management for greater efficiency, transparency, and control.
Data Quality and Use
Both systems promoted use of data by clients and providers and contributed to a heightened awareness of the importance of data quality. The system developers incorporated features to enhance data quality - such as linking to national ID databases to validate ID numbers and retrieve accurate identification data and implementing artificial intelligence image recognition controls in mobile apps to ensure that appropriate documents are being uploaded during the registration process. Effective multi-sectoral coordination is needed to establish relationships between different ministries who manage health services, financial technology (FINTECH) and civil registration in order to set up data sharing agreements. Respondents perceived that the availability of information through mobile phone interfaces enabled citizens and health facility managers to be more engaged in the process, including for checking coverage status, making payments, knowing exact payment requirements, and confirming household members covered in real-time.
In different ways, all programs contributed to enabling different stakeholders to use data more effectively for evidence-based decision-making. For example, 3MS provided more accurate, disaggregated, and timely data on premium revenues and membership campaign coverage than earlier aggregate reporting tools.
Efficiency and Financial Considerations
The DFS for health programs supported national eGovernment initiatives to move from manual to automated management with the potential for greater efficiency and transparency. Some respondents observed that digital systems made it possible to quickly implement changes to their services based on new laws or client-proposed features, making systems more responsive and adaptable. They also noted that citizens have more trust in financial transactions placed through the private sector DFS ecosystem. There was no intermediary (e.g., insurance agent counting on commissions from enrollments) and a perception by enrollees that there was less chance of fraud. Similarly, third party insurance payments that are common in Rwanda’s CBHI – such as relatives in urban areas paying the CBHI premiums for unemployed family members in rural areas, or small businesses paying the premiums for families of their day laborers - were sure to be used for intended purpose as they went directly to RSSB.
The programs also contributed to potential cost efficiencies. In Rwanda, the introduction of 3MS reduced the number of cash financial transactions managed at the facility level and digitized the labor-intensive work of managing paper household record systems thereby requiring fewer staff. While new posts were created at district CBHI sections and at the national level to manage a heavier workload at higher levels - and there were costs associated with orienting staff to the new technologies - the net result was reported to be cost savings. The system also enabled new types of facilities with no CBHI staff (e.g., health posts and telemedicine providers) to check eligibility with a simple SMS message before providing services. Digital membership verification was much more efficient and protected the providers from the risk of not being reimbursed.
The CBHI scheme was also more efficient. Under the old system, Savings and Credit Cooperatives (SACCOs) were collecting money daily and then transferred funds at the end of the month. In fact, there were often substantial delays in this process that worked to the benefit of the SACCOs – who maintained these funds to give loans to their members and earn additional revenues – in addition to the commission they charged on each premium received. Now money is transferred immediately to RSSB. One CBHI manager estimated that “at the end of the year the amount of interest generated by getting the funds into the CBHI pool quickly is bigger than the commissions that are paid to SACCOs and mobile money agents for the transactions.”
Facilities praised the simplicity of using the MCF systems, especially digital cash advance. The ease and speed of the process helped to ensure that SMEs could not only access financial resources, but could do so when they needed it, making the service much more responsive and improving their ability to provide services to clients.
Quality of Care
Some of the efficiencies enabled by digitizing financial and membership management services were perceived to improve quality of care. In both case studies, respondents shared that they were motivated to seek care earlier because they no longer risked paying high out-of-pocket expenses for care. A CBHI program manager in Rwanda noted, we “don’t expect that we will be able to show measurable change in quality of care [through this qualitative study], but clients receive better service - less time waiting in queues, and immediate triage for life saving care”.
The mobile phone applications that support M-TIBA and 3MS enabled citizens to be more engaged with the program. Using a simple mobile feature phone, they could send an SMS to the 3MS system to check the status of their coverage, make payments, know how much to pay, and confirm household member coverage in real-time. Clients of CBHI expressed that service had improved, as prior to the use of digital payments there was a one-month processing delay between paying for the CBHI cover and accessing services. When paying digitally, the client gains access immediately.
In Kenya, the digital financing services were accompanied by significant effort to make sure facilities were working toward quality improvement. The SafeCare program’s digital quality checklists played a role in this, by enabling regular quality assessments in hospitals and linking them to resources through CA/MAF to help make improvements.
Patient registration and eligibility checking also goes more smoothly because verification is done in real time by phone or computer. In Rwanda, this responsibility was transferred from dedicated staff hired by CBHI in each facility to frontline health workers who checked this during triage without patients having to pass through a separate insurance queue as in the past. Many respondents indicated that the introduction of the digital systems has made paying for and accessing health services quicker. “When you get to the hospital, you don’t need to queue; you just use your mobile phone to activate your account, bring out your name and you are quickly attended to”. – i-PUSH participant
Similarly, i-PUSH beneficiaries could choose from a wider range of health providers and could choose to go where more services were offered or providers were perceived to offer better quality of service. This led to increased income in preferred facilities that in turn enabled providers to increase the scope of services they provided, creating a virtuous cycle. “We've seen facilities who had earlier no laboratory services, but they were able to generate more income and [were] able to increase the services by opening labs in their facility”. – i-PUSH NHIF program manager.
Interviews demonstrated that DFS programs have other unintended benefits for the participating healthcare providers. For instance, the record-keeping and data reporting required within the programs encouraged additional rigor in accounting, bookkeeping, and other administrative management practices. The programs inspired a culture of quality program and health facility management in a way that was not planned but was welcomed by all. “It became so easy to inculcate that culture of quality management in terms of healthcare because of this program. Many hospitals were [came] up with quality improvement teams and … there was a lot of training for both medical and non-medical staff through this program. This of course improved their efficiency and productivity [leading to improved] quality of services rendered to the patients.” – i-PUSH NHIF program manager
What was the client/beneficiary experience of the program?
The study respondents identified a wide range of benefits that they perceived to accrue from the programs documented by these case studies. The chart in annex (Fig. 2 Perceived benefits of DFS programs by stakeholder group) highlights some of the key benefits across different actors engaged in DFS for health:
The following sections provide more detailed evidence of some of these benefits as they relate to key sub-components of the beneficiary experience.
Key informants perceived that both programs (3MS and i-PUSH) contributed to increasing insurance coverage (though the DFS tools were only one of many changes in market dynamics that influenced this change). Many interviewees praised the ease of use of all functions of the system—enrollment, payment, and accessing services. Clients could easily make payments from home, helping them to make payment on time, retain coverage, and remain able to access services. “When we used to pay for example NHIF, we were to go to Kiambu so as to pay but for now that we are paying via the phone you can pay at any time even at night.” – i-PUSH participant
The i-PUSH program’s digital enrollment tools enabled registration of many beneficiaries in a short period. “[i-PUSH] managed to enroll more than 35,000 women and their households who had never been on insurance. Giving them a platform, where based on the frequency of their incomes they can put in money in small bits and save for their healthcare.” – PharmAccess program manager. This supports the finding in a previous systematic review that “mobile money accounts help people smooth health and non-health expenditures when faced with a health shock.”3
DFS systems in both countries relied on socioeconomic mapping to identify poor households to target for subsidies and enroll for access to health services. In Rwanda, linking 3MS to the Ubudehe income classification database enabled RSSB to identify those households that fell into the indigent category and enroll them automatically in CBHI with premiums fully paid by the government – while those who could afford to pay were charged on a sliding scale. In fact, the introduction of the progressive premium structure – rather than the simple standard premium for all in the original CBHI scheme – proved to be impossible to implement without the digital platform to lookup a household’s income category in order to determine how much they should pay.
Similarly, i-PUSH’s socioeconomic mapping gave the “government information about who they should subsidize, especially the poorest of the poor, and encouraged other partners/NGOs to contribute by subsidizing NHIF enrollments for households that couldn't afford the whole enrollment fee.” – i-PUSH program staff. Because many clients never had any health coverage before, i-PUSH represented their first opportunity to feel secure in accessing health services without risking a bill they would not be able to pay. Beyond practical financial considerations, respondents described an improved state of mind and confidence. For some clients “the shift from the one-time large payment of NHIF to gradual saving made paying for health coverage easier.” i-PUSH client.
These examples highlight how DFS programs can contribute to increased and more equitable access to health services.
On the provider side, facilities participating in MCF’s Cash Advance and Mobile Asset Financing programs indicated that access to credit improved the ability of facilities to weather dips in funding and consistently pay expenses, maintaining a more stable supply of medications and ensuring health workers are paid. Some respondents indicated that the loans were especially critical in helping them remain solvent through the worst points of the COVID-19 pandemic.
A major advantage to loans through CA and MAF was predictability. Clients reported a greater predictability about whether or not they would receive a loan and how much they qualified for, which allowed them to plan more effectively. Facilities praised the repayment directly from the till, with many reporting that the gradual repayment linked to revenues eased the financial management burden and avoided difficult situations at the end of the month. “I can say the fact that the money is being deducted from the till, is what is making us feel like we don’t have that weight of repayment.” - MCF client facility
Rwanda has seen significant increases in service utilization rates at health facilities over the past decade, but they do not appear to correlate well with temporal trends in CBHI coverage (see Fig. 3 Comparison of CBHI coverage and utilisation rates in Rwanda). There were potentially other confounding interventions in the health sector during the same period that may have impacted service utilization rates. For example, digital tools helped to reduce perceived opportunity cost in traveling to the health facility while they also improved patient flow and wait times. Other interventions such as the establishment of new public and private health facilities, the national hospital accreditation program, improved supply chains for essential medicines/health commodities, and wide-ranging capacity-building initiatives for health workers have also contributed to increases in access to and quality of care that may have also driven up utilization rates.
In Kenya, several respondents from i-PUSH facilities reported rapid increases in the number of clients accessing their services. The clients seemed to be seeking comprehensive health care services in situations where they might otherwise only visit a pharmacist previously. Likewise, the majority of beneficiaries of the i-PUSH program who participated in the study had no access to any form of health insurance coverage prior to the program and affirmed that their service utilization had increased once they were enrolled.