Can Solving Behavioral Design Issues Help Position the Private Sector for Universal Health Coverage?
by Thierry van Bastelaer
Senior Technical Director Health Financing - Palladium, Washington DC email@example.com
All health systems ultimately include some type of private provision of services, both for clinical care and potentially financial products. The debate over what role the private sector can play in extending coverage of financial protection against catastrophic health expenditures remains a hot topic. While the following blog does not necessarily reflect the P4H Network’s opinion (the Coordination Desk or members), it does provide food for thought and we look forward to a productive debate.
Private sector actors in low- and middle-income settings are increasingly motivated to bring their entrepreneurial drive, out-of-the-box thinking, and technological expertise to expand the reach of affordable health financing solutions, thereby contributing to their countries’ progress toward universal health coverage. These actors coalesce into a variety of interacting or overlapping groups, described below with examples from sub-Saharan Africa:
- Health insurance companies that partner with grassroots organizations to engage the informal sector—one example from Tanzania is the partnership between Kilimanjaro Co-operative Bank Limited and Tanzania Coffee Board;
- Insurance companies that partner with health networks to offer low-cost insurance products—for example, Bima Ya Mwananchi, recently launched by Britam Insurance Company in Kenya;
- Mobile network operators that sell and service insurance via cell phone platforms—one such innovator is CoralPay/Airtel in Nigeria;
- Tech innovators that develop applications to streamline and accelerate transactions and financial flows—one well-known example is M-TIBA in Kenya;
- Private (often church-related) medical facilities that develop their own insurance schemes to attract local residents (and help fill their beds more predictably)—see Kisiizi Hospital in Uganda.
These privately owned entities often work closely with public entities, including national health insurance schemes and public health providers. Given the increasing focus on public-private partnerships in health systems, the growth of the private sector in health financing is worth a closer examination.
Complex sets of human behaviors shape how successfully health financing tools and policies function on the ground. It should therefore not be surprising that many well-intentioned and innovative product designs fall short of their goals. This blog shares lessons from two private initiatives in sub-Saharan Africa—an electronic maternity savings product in Kenya and a partner-agent health microinsurance product in Nigeria. We deliberately leave open the question of whether the challenges encountered by these efforts can be traced mostly to basic design weaknesses or are the result of more critical unresolved issues around the role of the private sector in paying for health.
Electronic Maternity Savings in Kenya
Motivation for the Product
Until Kenya moved in 2013 to provide free maternity care under its Linda Mama program, frequent non-payment for maternity services in Kenya was problematic for facilities and for patients, who were required to stay in the facility until they could settle their bills. In response to this challenge, Nairobi-based Changamka worked with Pumwani Hospital (at the time, the largest public maternity facility in sub-Saharan Africa) to make electronic savings cards available for free at a kiosk inside the hospital. The card, in use between July 2010 and October 2011, could be used at the hospital to pay for antenatal care (ANC) visits, delivery, and postnatal care. It could be topped up at the kiosk in the facility, and on the mobile payment platform M-Pesa. Fees (15 percent of service costs were borne by the hospital). The product’s goal was to increase facility-based deliveries by spreading the cost of ANC and delivery over several months.
The product was initially well received by pregnant women visiting the facility. More than 2,800 cards were distributed and used for first ANC visits. Large majorities of users indicated that the card was convenient to use and safer than carrying cash, that it helped them pay for services so they could be discharged faster, and that it helped users save money by preventing them from spending it on other items (15 percent of users also indicated that the card prevented their families from spending the money on other items).
Massive discontinuation followed this promising launch. After being used to pay for 2,800 first ANC visits, the card was only used to pay for 314 follow-up visits, 181 deliveries, and 10 postnatal visits. The three most common reasons users gave for this discontinuation were limited knowledge about how to use the card, insufficient cash flow to save money on the card to be used for delivery, and uncertainty about card use in other facilities.
The cited issues of how to use the card and where it was accepted could have easily been addressed by Changamka from the outset. But the mention of insufficient cash flow exposed a behavioral issue that had not been considered during the design phase: on average, users would acquire the card at their first ANC visit at about 30 weeks into pregnancy and then start saving money on the card. Aside from the health consequences of such late care-seeking, the compressed savings window of about 10 weeks before delivery meant that users had to save almost four times more per week to pay for a normal delivery than if they had acquired the card at an earlier ANC visit, for example when they discovered they were pregnant.
The fact that users had to trust an unknown tech startup to securely hold their money instead of a card from a trusted financial institution was another design obstacle. But the most consequential design decision was a baked-in selection bias that worked against the product’s goal of increasing facility-based deliveries—because the card was made available only inside the facility, only women who already were considering delivering at the hospital were even offered the chance to access the card. While the card may originally have strengthened the resolution of some pregnant women to deliver at the hospital, it did nothing to attract women who did not visit the hospital but might have considered delivering there if they had known the card existed.
Active marketing of the card to low-income families outside of the hospital would have helped address this design issue. The lack of that marketing probably doomed the idea from the beginning. (Later on, some of these challenges would be addressed successfully and on a larger scale by the M-TIBA mobile health wallet that Safaricom, PharmAccess and CarePay launched in Kenya in 2015.)
Partner-Agent Health Insurance in Nigeria
Motivation for the Product
Seventy-two percent of health expenditures in Lagos, Nigeria, are paid out-of-pocket. On average, the cost of one inpatient episode is equal to one month’s household income. Responding to this issue, a national health maintenance organization (private insurance) teamed up with a state-based microfinance institution (MFI) to provide affordable quality healthcare to the MFI’s clients and their families. Active between 2013 and 2017, the product covered a basic package of health services, including hospitalization; outpatient care; and maternal, newborn, and child healthcare at public and private health facilities in and around Lagos. There was no formal period for this product, which was priced at US$18.50 per year for individuals (US $110 for family), with premiums deducted directly from clients’ microfinance accounts in weekly installments. To ensure scheme sustainability, premiums were not subsidized. Even without subsidy, the product offered good value—the ratio of insurance premium to estimated annual total health expenditures for targeted women was 0.12.
The product rollout took place in three phases, each part of an ultimately unsuccessful effort to balance the interests of the insurance and the MFI. First sold as a voluntary product by the MFI’s loan officers, the insurance suffered from low enrolment and low re-enrolment. In addition, the sale of the insurance product slowed down the main work of the loan officers to issue and service credit and savings accounts. This may help explain why officers focused on enrolling clients who required the least amount of convincing—either because they were pregnant or presented serious pre-existing conditions—injecting high adverse selection in the portfolio. In response to these challenges, the insurer and the MFI agreed to make the product mandatory for MFI clients who had a loan above $600, with voluntary family enrollment. This led to better control of adverse selection, but the clients perceived the compulsory premium as an additional fee, which led many of MFI’s most valued clients to defect to competitors. In reaction, the MFI requested a further loosening of the mandatory insurance requirement, which concurrently led to worsening adverse selection, and reduced enrolment to zero after four years.
On the face of it, selling health insurance products through microfinance institutions makes sense for all parties. The insurer does not have to bear the costs of finding and selling products to individual clients in the informal sector; the MFI protects its portfolio by limiting the impact of health shocks on its clients’ ability to repay their loans, and the clients can be protected from the financial impact of health care through small and regular withdrawals that are already familiar.
In practice, however, balancing the interests of all parties can be challenging. First, insurers prefer to spread risk and lower adverse selection using compulsory products, but clients may perceive these products as an additional cost of credit, even if health benefits are immediately visible (which is rarely the case). The resulting loss of MFIs’ most valuable clients hurt the core of their business.
Second, relying on loan officers to sell insurance products is an apparently attractive proposition: these staff have already built trust with clients, which would help convince them to part with income against the hope that quality healthcare will be available and free when they require it. In reality, however, loan officers—under pressure to enroll insurance clients while still performing their normal credit and saving service duties—tend to seek out clients most likely to be interested in insurance, which would include the sick, the old, and pregnant persons—those people who pose the highest risk for insurers. Further, the fact that loan officers were not policyholders affected their credibility with clients and meant they could not speak from personal experience about the benefits of insurance.
Third, there was no waiting period. This was an important selling point for clients who wanted to access care as soon as they need it—which would also provide confidence in the insurance product—but it only reinforced the risk of adverse selection.
Finally, careful consideration of cultural, religious, and psychological factors is critical at all stages of product design and sales strategy—as exemplified by the words of a woman in a Lagos market who had been offered an insurance product to protect her family and business from the impact of likely health shocks: “So let me make sure I understand: you just cursed me, and now you want me to give you my money?”
The Kenyan and Nigerian health financing initiatives described suffered from critical design weaknesses that led to their relatively rapid demise. Most significantly, they failed to incorporate human behavioral factors to align incentives and expectations of all interested parties. With perfect 20-20 hindsight, it would be tempting to propose design fixes that may have expanded the life and outreach of the two products. Would these fixes, systematically and rigorously implemented, have resulted in products that could be added to the established set of health financing strategies? This question relates to core issues regarding what role, if any, the private sector can play in reducing catastrophic financial burdens on families and in helping a country to achieve universal health coverage. I very much look forward to your reactions to it.
van Bastelaer, T.; B. Woodman, M. Chatterji, and D. Long. 2016. “Saving for Delivery through Technology: An Inquiry into an Electronic Commitment Savings Product for Maternal Healthcare in Kenya.” International Journal of Health Planning and Management 31:254–272.
Peterson, L., A. Comfort, L. Hatt, and T. van Bastelaer. 2018. “Extending Health Insurance Coverage to the Informal Sector: Lessons from a Private Micro Health Insurance Scheme in Lagos, Nigeria.” International Journal of Health Planning and Management 33:662–676.
Photo © Palladium
Disclaimer: The responsibility for opinions expressed in this blog lies solely with its author(s). This disclaimer applies to any means of dissemination, distribution or publication of said blog at any time