Article Text

The cost-effectiveness of using results-based financing to reduce maternal and perinatal mortality in Malawi
1. Jobiba Chinkhumba1,2,
2. Manuela De Allegri3,
3. Stephan Brenner3,
4. Adamson Muula4,
5. Bjarne Robberstad2
1. 1Department of Health Systems and Policy, Health Economics and Policy Unit, University of Malawi College of Medicine, Blantyre, Malawi
2. 2Department of Global Public Health and Primary Care, Faculty of Medicine and Dentistry, University of Bergen, Bergen, Norway
3. 3Institute of Public Health, Medical Faculty, University of Heidelberg, Heidelberg, Germany
4. 4School of Public Health and Family Medicine, University of Malawi College of Medicine, Blantyre, Malawi
1. Correspondence to Dr Jobiba Chinkhumba; jchinkhumba{at}mac.medcol.mw

## Abstract

Introduction Results-based financing (RBF) is being promoted to increase coverage and quality of maternal and perinatal healthcare in sub-Saharan Africa (SSA) countries. Evidence on the cost-effectiveness of RBF is limited. We assessed the cost-effectiveness within the context of an RBF intervention, including performance-based financing and conditional cash transfers, in rural Malawi.

Methods We used a decision tree model to estimate expected costs and effects of RBF compared with status quo care during single pregnancy episodes. RBF effects on maternal case fatality rates were modelled based on data from a maternal and perinatal programme evaluation in Zambia and Uganda. We obtained complementary epidemiological information from the published literature. Service utilisation rates for normal and complicated deliveries and associated costs of care were based on the RBF intervention in Malawi. Costs were estimated from a societal perspective. We estimated incremental cost-effectiveness ratios per disability adjusted life year (DALY) averted, death averted and life-year gained (LYG) and conducted sensitivity analyses to how robust results were to variations in key model parameters.

Results Relative to status quo, RBF implied incremental costs of US$1122, US$26 220 and US$987 per additional DALY averted, death averted and LYG, respectively. The share of non-RBF facilities that provide quality care, life expectancy of mothers at time of delivery and the share of births in non-RBF facilities strongly influenced cost-effectiveness values. At a willingness to pay of US$1485 (3 times Malawi gross domestic product per capita) per DALY averted, RBF has a 77% probability of being cost-effective.

Conclusions At high thresholds of wiliness-to-pay, RBF is a cost-effective intervention to improve quality of maternal and perinatal healthcare and outcomes, compared with the non-RBF based approach. More RBF cost-effectiveness analyses are needed in the SSA region to complement the few published studies and narrow the uncertainties surrounding cost-effectiveness estimates.

• maternal health
• child health
• health economics
• public health
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

View Full Text

## Statistics from Altmetric.com

### Key questions

#### What is already known?

• There is increasing evidence that results-based financing (RBF) can promote service use and quality of maternal and neonatal care.

• Little evidence exists ascertaining whether RBF interventions provide value for money in low-income and middle-income countries.

Table 2

Facility characteristics and provider economic costs (US$) ### Cost-effectiveness of RBF relative to non-RBF The model estimated that RBF would avert one additional DALY (maternal and perinatal), one additional death (maternal and perinatal) and gain one additional life year (maternal and perinatal) at an incremental cost of US$1122, US$26 220 and US$987, respectively. Averted perinatal DALYs, deaths and LYG contributed most to the total effects (table 3).

Table 3

Incremental cost-effectiveness ratios of RBF compared with non-RBF by outcomes

### One-way sensitivity analysis

The share of non-RBF facilities providing quality care, the life expectancy of mothers at time of delivery and the share of births in non-RBF facilities had the strongest impact on ICERs for any DALY averted (figure 2). Relative to baseline, increases/decreases in the share of non-RBF facilities providing quality care would lead to more favourable/unfavourable ICERs while high/low life expectancies of mothers at time of delivery would be associated with unfavourable/favourable ICERs. The 10 parameters with most influence on RBF cost-effectiveness are shown in figure 2.

Figure 2

One-way sensitivity analysis showing variations in incremental cost-effectiveness ratios per disability adjusted life year averted. RBF, results-based financing.

### Probabilistic sensitivity analyses

The ICER scatter plot illustrates that all iterations of DALYs for RBF relative to non-RBF consistently have positive costs, reflecting higher RBF costs with certainty. However, some have less (negative) DALYs relative to non-RBF, reflecting that the probability of overall lower health effects in the RBF arm relative to non-RBF cannot be completely ruled out (figure 3A). The probabilities that RBF is cost-effective compared with non-RBF at different levels of willingness to pay are shown in figure 3B. At a willingness to pay of US$495, US$990 and US$1485, (1X, 2X and 3X Malawi GDP per capita, respectively), RBF has a 0%, 35% and 77% probability of being cost-effective. At a willingness to pay of US$1144 per DALY averted, RBF and non-RBF have equal probabilities of being cost-effective. Thus, with a lower willingness to pay for health (<US$1144), non-RBF is most likely to be optimal, while at higher levels of willingness to pay RBF represents the more optimal policy. Figure 3 (A) Incremental cost-effectiveness scatter plot for RBF relative to non-RBF. (B) Cost-effectiveness acceptability curves for RBF compared with non-RBF funding option. RBF, results-based financing. When deaths averted (maternal and perinatal) are used as a measure of health benefit, the threshold is at a willingness to pay of US$26 500 per death averted. For a lower willingness to pay, non-RBF remains optimal, while RBF is the policy choice with highest probability of being optimal when the willingness to pay for a life is higher than this threshold (online supplementary figure 1).

### Model validation

The model structure was informed by data availability and consensus view that it reasonably represents the situation on the ground. It estimates perinatal mortality rate of 47/1000 births in the baseline scenario, which is comparable to 56/1000 live births reported for the SSA region8 and 40/1000 live births estimate for Malawi.73

## Discussion

This study demonstrates that an RBF intervention with a strong quality improvement component is probably cost-effective compared with status quo care. These results were produced in a context characterised by high levels of FDs and in the absence of significant changes in service use. Most of the health benefits resulted from averted perinatal deaths due to improvements in quality of care, underscoring the potential gains for newborn survival if RBF is rolled out.

This is the first full economic evaluation of an RBF (combining PBF and CCT components) intervention for MNH in a low-income country. We therefore have no relevant previous studies against which our results may be compared. When maternal or perinatal outcomes are considered separately, our ICER results are larger compared with findings by Alfonso et al17 for a voucher programme in Uganda, which reported an ICER per maternal death averted of US$20 756, which is 1/3 of the US$61 260 per maternal death averted in this study. This estimate however ignores the benefits in terms of perinatal health. Our ICER results are also larger than those published by Hounton et al74 assessing a health worker surgery training aimed at increasing access to EmOC in Burkina Faso, which reported an ICER per perinatal death averted of US$11 757. We estimated an ICER of US$45 841 per perinatal death averted, although this estimate ignores maternal health benefits. In Zambia, Zeng et al75 have reported an ICER of US$809 per Quality adjusted life year gained for a PBF scheme, but this estimate only considers financial costs. Compared with some specific MNH interventions including quality improvements efforts76 and new-born care,77 RBF has higher ICERs per DALY averted. However, RBF is a broader and more complex intervention, where capturing all effects is challenging. We find substantially lower ICERs when we combine both maternal and perinatal outcomes, confirming assertions that the cost-effectiveness of MNH interventions is underestimated if benefits are assessed separately, rather than jointly, for mothers and perinates.20 Because the RBF4MNH intervention appears more effective and also more costly than providing MNH care under status quo conditions, decisions to adopt it would depend on policy makers’ willingness to pay. At a willingness to pay of US$1485 per DALY averted, policy makers can be 77% certain that RBF is more cost-effective compared with status quo care. This confidence is reduced to 50% if we consider a willingness to pay of US$1144 and further reduces to 35% at willingness to pay of US$990 per DALY averted.

The choice of a new intervention is not based on cost effectiveness analysis (CEA) results alone but also on the capacity of resource-constrained governments to sustain its routine use.78 The implications for low-income countries like Malawi, where coverage gaps in other vital MNH interventions still remain, is that implementing new costly interventions based on the WHO threshold can only occur at the expense of displacing other interventions, thus risking lowering overall population health attainment and increasing health inequalities.79 Going forward, an important line of inquiry is therefore to conduct a budget impact analysis of the RBF4MNH. Besides ascertaining the affordability of RBF and potential dividends from economies of scale in case of scale up, such analysis may also offer needed insight regarding broader impact of RBF on other interventions.

Our model is sensitive to estimates of non-RBF facilities with effective coverage and share of births in non-RBF facilities (both directly influenced by quality of care) and life expectancy of mothers at time of delivery. Quality of care is important in RBF programmes and there are ongoing efforts to improve its measurement and reporting.80 The observed sensitivity underscores the importance of quality as it relates to health outcomes and that better data are needed to improve our model accuracy. The sensitivity to RBF costs is not surprising, given the financial outlays associated with RBF.81 It does nonetheless highlight the imperative to contain operational costs and improve programme efficiencies. Though the model is sensitive to the share of births in both RBF and non-RBF facilities, it is not as sensitive to the percentage of complications among facility deliveries (FD), which depends on FD coverage since women with complication self-select into care, especially when FD rates are less than 40%.82 As FD rates are variable in SSA, ranging from as low as 12% in Ethiopia83 to as high as 91% in Malawi,84 we postulate that cost-effectiveness of RBF may be strongly influenced by the share of obstetric complications at much lower FD rates than those obtained in Malawi. In this regard, CEA studies in different setting are required to contextualise findings.

We did not consider benefits from potential reductions in perinatal morbidity due to lack of quality data. Studies on MNH report a heavier morbidity burden due to disabilities than to mortality per se.85 Inclusion of averted perinatal morbidity would thus increase effectiveness and improve cost-effectiveness of RBF. Future CEA studies should account for the potential of RBF to reduce all disabilities.20

We observed large differences in rewards based on cadre. The size of financial rewards is assumed to positively influence performance86 while perceived unfairness in distribution of rewards may demotivate staff, undermining RBF objectives. Assessment of adequacy of rewards, perceived fairness in how rewards are shared and their impact on health system performance should inform future lines of inquiries.87

This study has limitations. First, estimates of RBF effects on stillbirths and CFRs were based on programme evaluations in Zambia and Uganda due to lack of randomised trial data. Intense monitoring and supervision under programme settings may have improved programme effectiveness, biasing our results downwards. We attempted to minimise this by using mean effectiveness estimates in the baseline scenario. Second, we were not able to account for maternal deaths that occurred before 28 weeks of gestation due to lack of data. The share of maternal deaths before 28 weeks is small and as early maternal deaths are not explicitly targeted by RBF, they may be assumed to be a constant that does not affect our estimation. Thus, we believe that this omission is less likely to substantially bias the effectiveness estimates. Third, the model considered diverse events and use of incidence in DALY estimation would have been more appropriate. We cannot ascertain how use of prevalence based approach affected our results but given use of similar approach by recent global burden of disease estimates,65 we believe our results can be trusted. Fourth, RBF was preceded by one-off infrastructural/equipment support. Disentangling the effects of such investments from the quality effects of RBF is problematic. Our results therefore represent the costs and effects of the combination of infrastructural improvements and RBF rather than RBF alone. Fifth, not all central level start-up costs were captured. This might have led to underestimating RBF costs. We attempted to characterise these uncertainties by using wide ranges (±20%) in probabilistic sensitivity analysis and found that the main results were stable across these ranges. Finally, GDP-based thresholds have been shown to be easily attainable and unconnected to local budget constraints;88 the reader should note this caveat when considering our results.

## Conclusion

The RBF4MNH Initiative is a potentially cost-effective way to fund health facilities to improve quality of maternal and perinatal health, and to increase pregnant women’s access to EmOC, compared with the current non-RBF based funding in Malawi.

Although delivery services at RBF supported facilities are about five times as expensive per delivery compared with services offered at non-RBF facilities, we estimate that the intervention will avert 1.5% of perinatal and 12.1% of maternal deaths that occur with status quo MNH care in Malawi. The cost of US\$1122 per DALY averted is lower than Malawi’s three times GDP per capita, which is one of the decision rules towards implementation.

More RBF CEAs are merited to explore cost-effectiveness of different intervention types, different health systems settings and health services and to reduce uncertainties around RBF CEA estimates related to modelling. Researchers in SSA should take advantage of the numerous RBF studies being implemented in the region to generate needed economic information to support policy decisions. Further, RBF studies should particularly prioritise generating more health outcomes data, while future economic evaluations should focus on identifying optimal RBF designs and implementation models that have lower transactional costs. This would allow to better assess the adequacy of different reward options and aspects of fairness in the allocation mechanisms to maximise individual health worker and team efforts.

## Acknowledgments

The authors would like to thank Professor Don Mathanga and Dr Jacob Mazalale for their support during data collection. The authors are also grateful to Dr Matthias Arnold and the two anonymous reviewers for their helpful comments.

View Abstract

## Footnotes

• Handling editor Lei Si

• Contributors JC and BR made substantial contributions to the development and design of the model. SB, MDA and AM provided substantial input to ensure model structure reflects situation on the ground. SB, AM and MDA made substantial contributions to acquisition of some of the data for the study. JC drafted the initial manuscript. All authors contributed substantially to the manuscript and interpretation of the data. All authors approved the final version of the manuscript to be published. The corresponding author attests that all listed authors meet the authorship criteria and that no one meeting the criteria has been omitted.

• Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

• Competing interests None declared.

• Patient consent for publication Not required.

• Ethics approval Ethical approval for the study was obtained from University of Malawi, College of Medicine Research and Ethics Committee (COMREC) protocol P.02/13/1353.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.