In a recent World Health Organisation bulletin, Atle et al. draw attention to the complexity surrounding studying and understanding performance-based financing (PBF) models, their effects on health outcomes and their purported impact on the functioning of health systems at large.(2) In spite of the gaps in understanding, PBF models are being adopted in more than 20 African countries, and in some instances, scaled up to involve the national health system as a whole.
Part 1 of this series examined the evolution of performance-based financing within health systems in Africa, the underlying theory supporting these models and the evidence for their effectiveness from research and practice. This final paper considers some of the potential setbacks that could arise with the adoption of performance-based financing models and, based on experience from within various contexts, identifies factors that appear to be instrumental to successful implementation.
The potential perverseness of financial incentives
The introduction of financial incentives for performance can be distorting and could even deter good performance. PBF models assume that financial incentives are a strong motivation for behaviour, stronger at least than intrinsic incentives. Individuals could, however, be motivated by more than financial gain. One multi-country study revealed that in one situation intrinsic factors were found to be even more influential in improving performance than financial incentives.(3) Psychological studies have posited that extrinsic rewards could pose a risk to intrinsic motivation among health workers in particular, which stems from professional ethics, altruism and social status, for example.(4) Introducing financial incentives for performance could lead to ‘crowding out’ of a worker’s internal motivation.(5) PBF contracts may, in their design, even devalue tasks that would otherwise be appreciated for their intrinsic worth.(6)
Financial incentives have also been associated with other negative effects. Health workers spend long hours documenting performance, and attending supervisory meetings which cut into time for clinic work. In addition, observations have been made of instances of ‘cherry-picking’ patients who would have favourable outcomes and cost less to treat.(7) Furthermore, a focus on particular indicators could lead to ‘gaming’, with a neglect of unrewarded aspects of work and distortion of reports.(8) It has even been reported that, to avoid stock-outs, Rwandan hospital pharmacies refused to deliver the last box of medications to wards.(9)
As regards financial incentives to groups, facility-based incentives are only effective to the extent that performance bonus distributions are transparent and equitable. Otherwise, they could also be de-motivating. For instance, depending on the culture of the organisation, rewards may be considered a reflection of seniority rather than performance, and may even be viewed as an instrument used by supervisors to control healthcare providers.(10) A culture of rewarding performance could fragment social networks and breed unhealthy competition.
Provider behaviour, utility-maximisation and anticipated results
When structuring incentives, it is therefore important to recognise that individuals may diverge systematically from behaviours predicted by rational, utility-maximising models such as the principal-agent model. For example, individuals may find contracts in which the financial incentive is a bonus more attractive than one in which losses are incurred for not attaining targets, even if these targets are attached to rare incidents. In addition, though the direction of people’s responses to incentives may be obvious, it may be harder to estimate the magnitude of their reaction and the effect on the rest of the system. It is therefore often necessary to repeatedly adjust targets and rewards.
Even when the agent’s behaviour is in the predetermined direction, the value of these incentives in creating real results in terms of health outcomes is also a concern. In paying for performance, one assumption is that provider behaviour is a key determinant of results in that area. However, numerous extraneous factors influence health and social outcomes.(11) Even if a physician is motivated to work for financial gain, patient non-compliance, organisational constraints, lack of technical skills and time constraints may prevent the physician from doing a good job. Drawing conclusions on the influence of financing incentives on fluctuations in outcomes is thus an extremely controversial issue. PBF in some cases may not be effective in actually improving population health outcomes. As a matter of fact, improvements in indicators may not even actually reflect improvements in health. A target such as ‘number of antenatal visits’ is appropriate only to the extent that it reflects the care that mothers receive. The danger is that this assumption of quality care may not be true, and often all that the target tells us is that the target itself, not the overarching goal, has been achieved.
Performance-based financing, equity and quality of care
The effectiveness of the PBF model in increasing quality of care and promoting equity is also questionable. A potential advantage of the PBF model lies in the fact that whether or not patients recognise quality, rewarding performance can promote quality assurance. However, a major challenge in PBF is the incorporation of quality of care outcomes into the monitoring process. Meessen and colleagues report that in Butare, Rwanda, there was a focus on high impact activities that were easy to deliver and measure.(12) This tendency to focus on activities that are easy to measure can be a drawback with PBF. Indicators of quality, such as attention given to the client during clinical consultations, can be difficult to measure, but can be essential to improving health outcomes. If PBF does not contribute to the overarching goal of improving population health, its whole purpose is arguably defeated. A focus on outcomes and outputs must therefore be accompanied by increased attention to the processes of care and quality of care delivered.
The effect of PBF on equity in health systems can actually be negative. It is often assumed that PBF models will, by definition, increase equity of access to care and financing of health facilities. However, as targets are often non-specific with regard to the population profile, and just set coverage goals, there is a perverse incentive to focus on people that can be reached with little effort. Also, health services usually cluster in well-off areas. Thus, by default, poorer families who live in hard-to-reach areas, have limited financial resources and less understanding of health problems and services, are placed at a disadvantage. It is important, therefore, to structure incentives in such a way that access to care for disadvantaged communities is increased. In Cambodia for example, by conditioning rewards on reaching of poorer families, coverage of the poorest fifth of the population rose to 40% from 15%. This rise was twice as high in facilities that were contracted using financial incentives for this purpose than regular facilities.(13) The combination of supply-side incentives such as PBF with demand-side incentives could also have multiplicative effects and deal with the complex determinants of inequitable access to care. Illustratively, in India a small monetary incentive to clients in addition to supply-side incentives led to an increase in immunisation rates from 17% to 38%.(14)
In the same vein, at facility and national level, the need to set targets that take account of the differences in resource availability is pertinent. Otherwise, reward systems could reinforce inequitable financing by setting unrealistic targets for facilities with low resources. As a matter of fact, with PBF systems used by donor agencies such as GAVI, countries with stronger health systems find it easier to meet targets, and benefit more from financial incentives, than countries that have weaker systems and are arguably more in need of financial help.(15) In some cases, when targets are not met, PBF models involve not only the loss of bonuses but the withholding of funds, which could reinforce inequity of financing when the resource constraints significantly influence poor results.
Is performance-based financing the sustainable alternative?
As the effectiveness of PBF models is presently the subject of controversy, a major issue to consider is the relative cost and resulting implications for financial sustainability of these programmes. PBF models aim to design contracts that pay for performance such that it is in the interest of the agent to align with the principal’s interests. Therefore the cost of monitoring the production process is expected to reduce. However, in many quarters, the cost of monitoring, evaluating and independent verification of performance for contracts is believed to increase transaction costs.(16) Simple cost analyses have revealed that the cost of administration of PBF is high, varying between 15% and 30% of the per capita health expenditure.(17)
The PBF model without a budgetary cap also provides incentive for supplier-induced demand if there is payment for increased enrolment in care, for example. This can lead to unanticipated budgetary pressures. The status quo of paying for inputs usually presents predictable budget schedules such that supplier-induced demand is unlikely to be an issue. It is important to consider, however, that in many of the environments where PBF has been introduced within Africa, utilisation rates for healthcare were already quite low, supplier-induced demand was not discouraged. Unlimited spending is not feasible. Mechanisms to control expenditure growth and ensure quality of outcomes are therefore essential. One multi-country study indicated that in order to maintain predictability of budgets, the implementation of PBF schemes was often accompanied by block grants with regular renegotiation of payments.(18) In the absence of formal comparisons of these costs with input-based alternatives in various contexts, the fear that PBF models are financially unsustainable may, however, be unfounded.
Institutional arrangements for performance-based financing
Implementing PBF, especially at the national level, demands changes in institutional arrangements beyond financing. The separation and delegation of provision of care, purchasing of services, and regulation is considered imperative for transparency and accountability in PBF implementation. The details of this arrangement usually involve navigating the balance between technical capacity to handle such duties by the relevant parties, and the need for Government ownership which is important for ensuring sustainability and scale-up.(19) Purchasing of services should ideally be the duty of an autonomous fund-holding agency. Soeters et al. argue that the in countries such as Zambia, where Government agencies are fund-holding agents in addition to being providers of regulatory oversight, there is conflict of interest and a tendency to engage in rent-seeking behaviour.(20)
Therefore, ideally, a separate regulator would provide oversight and verification of commitments. The engagement of all these actors from the design phase of programmes is considered essential to reaching contextually appropriate agreements on contractual arrangements and to institutionalising PBF within health systems in future.(21) In Sub-Saharan countries with multiple financing streams, the need to engage all actors in the design of a single contract is pertinent. When agents have multiple principals, the effectiveness of incentives established by any one principal could be weakened. If funds are received from multiple streams, such as the federal Government, bilateral agencies and private donors, with competing instructions attached to each resource, this could lead to confusion and detract from the overall goal of improving healthcare.
In what context is performance-based financing likely to succeed?
In drawing to a close, it must be asked: what are the determinants of success of the PBF model? Analysis of the Rwandan success story points to certain factors. There was autonomy at the operational level to allow for tailoring of the healthcare process to suit the context; national ownership of the PBF model from the onset; consensus on performance indicators among stakeholders at all levels; an autonomous local fund holding agency; split of responsibilities for provision, purchasing and regulation of contracts; a functioning monitoring system, which also took quality into account; and an appropriate reward system.(22)
There was also strong political leadership, increased health expenditure, strong management capacity, a financing system that was flexible enough to mobilise funds to the frontlines, and the technical capacity to design the reward system that was effective and equitable.(23) Multiple simultaneous health policy reforms helped to reduce financial risk through health insurance, as well as to increase utilisation rates through demand-side incentives and performance contracts between the republic and the mayors.(24) The PBF model in isolation is, therefore, not sufficient to deal with the complex issues that often plague health systems in Africa. The Rwandan example demonstrates that the effectiveness of PBF is dependent on a supportive socio-economic and political policy environment for system reform at a national scale.
It would be difficult to deal with all the nuances of performance-based financing in this series. There is still some uncertainty regarding the effectiveness of this model in positively changing provider behaviour, promoting equity and improving quality of care and health outcomes. Furthermore, as outlined in part one of this series, PBF models are not without flaws. It is therefore important to anticipate and check negative effects while ensuring that the programme components necessary for success are included in PBF models to be implemented. There is, however, a need to continue to prioritise research as the impact of PBF on health outcomes, equity, access to care, cost and provider behaviour remain unclear. Despite potential drawbacks, in the current global financial climate, a culture of rewarding efficiency is, however, more likely to persist than not.
It is important to note that successful performance-based financing involves a change in culture, with a focus on the results that we consider pertinent to the health of the communities. It therefore appears to be a method of financing that, if properly structured, can facilitate positive changes in institutional arrangements, supply systems and information systems. These changes, if leveraged effectively, could strengthen national health systems and improve population health outcomes in general.
(1) Contact Adanna Chukwuma through Consultancy Africa Intelligence’s Public Health Unit (
(2) Atle, F., et al., 2012. Performance-based financing in low- and middle-income countries: Still more questions than answers. Bulletin of the World Health Organisation, 90, pp. 559-559A.
(4) Savedoff, W.D., 2010, ‘Basic economics of results-based financing in health’, Social Insight, Background paper for Results-Based Financing for Health, http://www.rbfhealth.org.
(5) Kalk, A., 2011. The costs of performance-based financing. Bulletin of the World Health Organisation, 89, p. 319.
(6) Frey, B. S., 2001. Motivation crowding theory. Journal of Economic Surveys, 15(5), pp. 589–611.
(9) Kalk, A., Paul, F.A. and Grabosch, E., 2010. Paying for performance in Rwanda: Does it pay off? Tropical Medicine and International Health, 15, pp. 182-190.
(10) Magrath, P. and Nichter, M., 2012. Paying for performance and the social relations of health care provision:
An anthropological perspective. Social Science & Medicine, 75, pp. 1778 -1785.
(11) ‘Can we demonstrate effectiveness without bankrupting our NGO and/or becoming a randomista?’, Oxfam, http://www.oxfamblogs.org.
(12) Meessen, B., et al., 2006. Reviewing Institutions of rural health centers: The performance initiative in Butare, Rwanda. Tropical Medicine and International Health, 11(8), pp. 1303–13017.
(13) Gwatkin, D.R., 2009, ‘Ensuring that the poor share fully in the benefits of results-based financing programs in health’, The World Bank working paper, http://www.rbfhealth.org.
(14) Sjoblom, M., Beith, A. and Eichler, R., 2012, ‘Performance-based incentives for child health: Taking stock of current programmes and future potentials’, Abt Associates, Health Systems 20/20 report, http://www.healthsystems2020.org.
(15) Heaton, A. and Keith, R., ‘A long way to go: A critique of GAVI’s initial impact’, Save the Children UK briefing analysis, January 2012, http://www.savethechildren.org.uk.
(18) Ettelt, S., et al., 2006, ‘Reimbursing highly specialised hospital services: The experience of activity-based funding in eight countries’, London School of Hygiene and Tropical Medicine, http://www.international-comparisons.org.uk.
(20) Soeters, R., Habineza, C. and Peerenboom, P.B., 2006. Performance-based financing and changing the district
health system: Experience from Rwanda. Bulletin of the World Health Organisation, 84, pp. 884-889.