Provider incentives as a strategy to improve health services and outcomes. Harsha Thirumurthy DIME Workshop, Cape Town December 2012

Transcription

1 Provider incentives as a strategy to improve health services and outcomes Harsha Thirumurthy DIME Workshop, Cape Town December 2012

2

3 Overview Are health investments made by your country producing desired health results? Reduced maternal mortality Reduced infant mortality Are supply side factors to blame? Poor quality, lack of innovation, low productivity Can we find a better way to link financial investment in health to health results? Often, payment to facilities and providers may be based on inputs, such as hours worked, # of salaried personnel, fuel, maintenance If services are not delivered or if population health does not improve, payment continues to be made

4 Maternal mortality example Few countries likely to reach MDG 5 (reduce MMR by 75%) Reasons? In SSA, from 2003 to the present, 78% of births among the poorest women occurred at home, of which 56% were unattended (Montagu PLoS One 2011) Results in high rate of obstetric complications Why the high rate of unattended home births? Identifying the barrier is useful for finding effective approaches

5 Provider incentives Innovative approach involving provision of incentives that are conditional upon a contracted party s performance of predetermined, measurable actions Primary objective is to increase quantity and quality of services that health-care providers deliver Incentive can be bonus salary or budget Increased interest among donors and funding mechanisms for its potential to link funding with results and enhance accountability of recipients

6 Definitions Performance based financing (PBF) or pay for performance (P4P) Any program that rewards the delivery of one or more outputs or outcomes by one or more incentives, financial or otherwise, upon verification that the agreed-upon result has actually been delivered Payment can be to facilities OR to individual practitioners (relative efficiency an issue to consider) Part of larger family of incentive-based approaches to improving health outcomes

7

8 Overview Rationale for provider incentives Design considerations Example from Rwanda & China Lessons and additional considerations

9 Rationale for provider incentives PBF has its origins in agency theory; more specifically, what is referred to as the principal-agent problem Principal is entity that provides funding; agent is entity hired or contracted to pursue the principal s interest Interests of the agent may not necessarily be in line with the interests of the principal Principal may not have complete information on extent to which agents act in accordance with goals of their agreement (asymmetrical information) Staff whose pay is not linked to their performance may not be motivated to improve quality of care, productivity, or be present at work PBF may resolve the principal-agent problem

10 How will it affect healthcare delivery? Two possible ways that improvement can happen: Giving incentives for providers to put more effort into specific activities Increasing amount of resources/money available to deliver services What could go wrong? Schemes could have detrimental effect on health services if P4P requires completion of reports and other things that take up providers time

11 Mutual benefits Funders, providers, and patients can benefit Funders achieve desired outcomes, get increased accountability on the part of funding recipients Providers and health facilities benefit because they receive more money for services To some extent, they can control how much money they make: the better they perform, the greater the reward they receive Population/patients benefit from improvements in quantity and quality of health services

12 Design: Setting appropriate targets Setting measurable indicators/targets is essential % of infants who are fully immunized to measure primary health care delivery Score on standardized surveys/exit interviews to measure consumer satisfaction % of TB patients completing treatment to measure health outcomes Providers should have direct influence over the performance measure of interest Avoid reduction in child mortality rates because there are many determinants of health beyond influence of providers; better to have # children who are fully immunized In some cases, make it relative to each recipient s baseline Tournament model or uniform threshold may fail

13 Design: Payment mechanism Incentives imply some amount of risk No payment if results not verified Need to consider how much risk enough to motivate behavior change; how much is too much Majority of provider funding will be regular and reliable with only a small portion conditional on attaining performance targets In many cases, roughly 10% of financial risk to providers Fee-for-service may generate excessive health spending Should payment be adjusted for quality Especially relevant if coverage high but quality low

14 Design: Individual vs. team incentives Targeting the individual health worker makes more sense if individual action (i.e. working harder, doing more of what they are already doing) is all that is needed Targeting the institution (facility) level makes more sense if team work is warranted to improve performance or if system -wide changes are needed For example, an individual health worker may not be able to change clinic hours or implement community outreach strategies Incentives at the team level will motivate team members to pressure other members to increase productivity

15 Case studies and examples Relatively few rigorous evaluations of large-scale provider incentive schemes Often there is no control group in evaluation Evaluations done in Haiti and Cambodia Rwanda s experience with PBF is most extensively documented; recognized as a global innovator in PBF Experimental evaluation of national rollout Another study conducted in China to reduce anaemia prevalence Many more underway

16 Evaluation of Rwanda s P4P scheme Assessed effect of P4P for health care providers on use and quality of child and maternal care services in health-care facilities (Basinga Lancet 2011) Evaluation done in parallel with national rollout of P4P Randomization by coin toss: 166 facilities randomly assigned at district level either to begin P4P funding in 2006 (intervention group; n=80) or to continue with traditional input-based funding until 23 months after study baseline (control group; n=86) Surveyed facilities and 2158 households at baseline and after 23 months

17 How did P4P work in Rwanda? Payments made directly to facilities and used at each facility s discretion 14 key maternal and child health-care output indicators for which P4P payments given Reasons for visit (prenatal care, delivery, etc) and also services provided (tetanus vaccination during prenatal care) Based on national priorities & available budget

18 How did P4P work in Rwanda? Facilities submit monthly activity reports and quarterly requests for payment to district steering committee Auditors sent to facilities every 3 months on unannounced, randomly chosen days MOH also did one-off tracking survey and interviewed roughly 1000 patients to verify accuracy Lastly, overall quality of facility calculated (0-1) Based on structural & process measures

19 What was incentivized?

20 Quality score calculation

21 Improvements due to P4P in Rwanda

22 Main results Incentive effects were identified because control facilities received equal amount of additional resources P4P facilities had 23% increase in number of institutional deliveries Increase in number of preventive care visits by children aged 23 months (56%) and months (132%) No improvements in number of women completing four prenatal care visits or children receiving full immunisation schedule Increase in prenatal quality

23 Interpretation Increase in facility delivery Providers encouraged women to deliver in facilities during prenatal care encounters Some also partnered with community health workers to promote institutional delivery Lack of effect on immunisations Completion of a full immunisation programme for a child takes a lot of effort Per-unit payments were not large enough; providers chose not to expend necessary effort

24 Provider incentives in China Effectiveness of rewarding school principals for better health outcomes of children (Miller BMJ 2012) School-based randomization Control group, no intervention (27 schools) Intervention 1: principals received information about anaemia (15) Intervention 2: principals received information and unconditional subsidies (15) Intervention 3: principals received information, subsidies, and financial incentives for reducing anaemia among students (15) Link between anaemia and school performance (part of the information provided)

25 Provider incentives in China: results and interpretation Mean student haemoglobin concentration rose by 2.4 g/l (0 to 4.9) in incentive schools compared with control group Reduction in prevalence of anaemia (Hb <115 g/l) of 24% in incentive schools School principals used subsidies to pursue iron supplementation strategies that increased only students multimicronutrient intake Less likely to pursue broad feeding strategies that increased both multimicronutrient intake & energy intake (meat and other food added to school lunches) - consistent with incentive to raise iron levels Other things happening in schools are also important Principals with incentives to achieve good academic performance focused relatively more on feeding Principals in the information arm with academic performance incentives did so most intensively

26

27 Some general lessons Higher payments generally provide stronger incentives Larger effect on services in which providers have more control over delivery, such as prenatal care quality Larger payments not only for services more important for improvement of health outcomes But also for those in which more provider effort is needed Pay more for verifiable indicators; closely related to outcomes, measurable, and within control of providers Incentives to patients more useful when uptake of services depends on patients care-seeking behavior Another option is to give CHWs incentive to identify patients and encourage them to visit clinics

28 Key design elements Quarterly verification of performance and payment Facilities/networks have autonomy to manage funds and services within regional oversight Financial incentives reach all health workers (% of salaries to be incentivized is unclear) Payment = Quantity of services x Quality score Quality Checklist/Balanced Scorecard Transparent, independent data quality audit Cheating more expensive than improving performance Some independent verification via community client surveys

29 Community/patient participation Via independent verification, payment for patient satisfaction surveys and public reporting Indicator for client satisfaction surveys Public reporting on performance increases transparency and credibility Community knows the quality and quantity scores of their providers and hold them accountable Payments for quantity encourages providers to attract patients, improving access for vulnerable

30 Criticisms of this approach May disproportionately benefit regions with stronger health systems; already in better position to reach targets than regions with weaker health systems PBF may reward health facilities in higher-income regions for doing what they would normally do even in absence of PBF Health-care providers could be discouraged from working in lower-income areas due to the challenges associated with meeting targets What happens when incentives removed? Does the intrinsic motivation of providers go down?

31 Mistakes/problems to avoid Failure to consult with stakeholders to gain input to design, to maximize support and minimize resistance Failure to explain rules (or rules that are too complex) Fuzzy definition of performance indicators and targets, too many performance indicators and targets, and targets for improvement that are unreachable Too much or too little financial risk for providers Tying hands of managers so that they are not able to fully respond Insufficient attention to systems and capacities needed to administer programs Inadequate attention to verification of indicators (deters deliberate misreporting) Failure to monitor unintended consequences Inability to meet increased demand

32 Way forward Identify top five performance problems that PBF can address Determine recipients and how to select them Determine indicators, targets and how to measure them Determine payment mechanisms, sources of funding and how funds will flow Determine who will manage PBF and how to make it operational Develop an evaluation and learning strategy