Affordability of comprehensive community health worker programmes in rural sub-Saharan Africa

Introduction

Community health workers (CHWs) provide basic health promotion and healthcare within the communities in which they live. They are supported by the health system but are not necessarily a formal part of it.1 CHWs are often cited as part of the solution to the shortage of health workers and lack of universal access to healthcare in low-income settings2 3 and feature prominently in the WHO’s Workforce 2030 strategy for Human Resources for Health.4 They have two distinguishing features when compared with other potential providers of healthcare: being from the community in which they serve engenders trust from their clients, particularly when the community have been involved with their selection5 and, by travelling to clients’ homes, they reduce the need for long, difficult and/or expensive journeys to healthcare facilities. There is accumulating evidence that CHWs can provide individual health interventions effectively, for example, screening for and treatment of tuberculosis and promoting immunisation uptake.6 7 Many CHW programme providers are now starting to increase the scope of CHWs’ activities, for example, with CHWs providing integrated community case management, which includes screening for malnutrition and detection and treatment of malaria, pneumonia and diarrhoea in children.8 9

There is some evidence that these more formal, integrated CHW programmes can be a cost-effective component of health services, particularly for child health outcomes,10 11 and modelling shows the potential high long-run economic returns from investing in CHW programmes.8 However, by September 2016, the campaign to have one million CHWs in rural sub-Saharan Africa by the end of 2015 had counted just 332 000 (data available for 37 countries).12 There are three potential economic reasons for the shortfall. First, the total cost of initial scale-up and maintenance of CHW programmes is high: McCord and colleagues estimate annual maintenance costs of US$3.4 billion in 2012 prices per year across sub-Saharan Africa.13 Second, CHW programmes have to compete with other priorities for limited government resources, such as other cadres of health personnel. Third, even where financial resources are available, there are inevitable delays before CHWs can start working including the time taken to design appropriate programmes, build an effective supply chain and recruit and train CHWs.

One particularly malleable component of CHW programme costs is the incentive or reward paid to CHWs for their time and effort. Rewards vary widely across programmes. Of the CHWs enumerated in the one million CHWs campaign, 13% were reported as being salaried, 67% received a combination of monetary and non-monetary incentives, 18% received non-monetary incentives and 2% received no incentives at all.12 There can also be different ‘cadres’ of CHW within a health system: in Ethiopia, for example, salaried health extension workers are supported by a volunteer health development army. Although paying CHWs has generally become the accepted standard,4 8 14 it has also been argued that ‘financial incentives that are too low, irregularly paid or discontinued due to a lack of sustainable programme financing may result in more of a disincentive to CHWs than no payment at all’.15 A properly set-up CHW programme staffed by volunteer CHWs could therefore be appropriate where the alternative is no CHW programme at all given the local health and economic returns that could be realised.

The resources available to fund CHW programmes depend on total government expenditure, the proportion of such expenditure allocated to healthcare in general (and a CHW programme in particular) and the availability of external funding, including from donors, low-cost loans and human capital bonds repaid over an extended period based on future contingent savings.8 However, external sources of finance may not be reliable and may divert national resources away from healthcare, rather than supplementing it.16–19 There is also the possibility that aid is misused and does not reach its intended destination. The long-term optimal position would therefore be that nations can fund their own CHW programmes without requiring external aid.

The cost of a CHW programme and the resources available to pay for it therefore influence, at least in part, the scale of the programme in any particular country. The purpose of this paper is to analyse how the affordability of comprehensive CHW programmes would be affected by varying CHWs’ salaries and the resources available to pay for them. We also examine the relationship between the level of economic development in a country (proxied by gross domestic product (GDP) per capita) and the affordability of a CHW programme. Our work complements that of Bossert and Ono,17 who challenge the affordability of the WHO target of 2.3 physicians, nurses or midwives per 1000 population, and extends that of Oxford Policy Management,20 which estimated that a CHW programme in Pakistan providing 74% population coverage would require 27% of total government healthcare spending to be devoted to CHWs. As in this study in Pakistan, we provide a relative analysis to show CHW programme costs as a proportion of the public funding available for healthcare as a whole. By doing so across all mainland countries of sub-Saharan Africa, we are able to illustrate how the challenge of financing a CHW programme varies across countries. This multi-national approach enables us to explore the extent to which this variation can be explained by the level of economic development in each country.

Methods

Results

Variation in affordability as a function of per-capita GDP

Figure 1 shows the relative cost of a CHW programme financed from actual central government healthcare spending (budget 1) with CHWs paid the equivalent of US$80 per month in relation to GDP per capita. Values of both variables are shown on a log scale given their large range and positive skew. There is a strong negative linear relationship (R²=0.83, p<0.001) between the natural logs of GDP per capita and affordability. The five countries with a GDP in excess of around US$5000/8.5 on the log scale (Botswana, Equatorial Guinea, Gabon, Namibia and South Africa) should be able to afford a comprehensive CHW programme without any external aid, with relative programme costs of up to around 2.5% of healthcare spending (0.9 on the log scale). Equatorial Guinea is an outlier on the chart (to the bottom right); among countries with a similar GDP per capita, it spends a relatively low amount on healthcare, which increases the relative cost of a CHW programme. There is no clear maximum GDP per capita at which a CHW programme is unaffordable, although relative programme costs are at least 12% of actual healthcare spending (2.5 on the log scale) in all 23 countries with a GDP per capita of less than US$1200/7.1 on the log scale.

Figure 1

Request permissions

Relationship between GDP per capita and CHW programme affordability (percentage of actual central government spending on healthcare required to fund a CHW programme with CHWs paid US$80 per month), both on a log scale. CHW, community health worker; GDP, gross domestic product.

Discussion

Even the original WHO goal for the ratio of 2.3 skilled physicians, nurses or midwives per 1000 population is unrealistic for many countries,17 and the latest target has increased this ratio to 4.45.26 Expanded CHW programmes appear to offer the least expensive and a relatively quick route to the goal of universal access to healthcare in low-income and middle-income countries. However, our results suggest that CHW programmes could only be afforded by a small number of countries without significant increases in central government healthcare spending or the availability of external funding. As might be expected, the countries that appear to be able to afford CHW programmes without having to rely on external funding tend to be those with higher levels of economic development based on our proxy measure of per-capita GDP. Even in these countries, affordability may not mean provision, as doing so may require resources to be reallocated from elsewhere in the health system: the total healthcare budget must pay for hospitals, out-patient care (which would include any CHW programme), medical goods, devices and pharmaceuticals, public health, administration and research, education and training. Affordability is least likely in countries with a lower level of economic development (GDP per capita below US$1200). Thus, in the poorest countries, solely central-government-funded, full-scale comprehensive CHW programmes for the rural population are not yet a realistic solution to the human resources for health crisis. Such disparities in affordability for such a basic level of healthcare inevitably raise questions regarding equity: our work is intended to highlight such issues rather than to provide a justification for them.

Given general views that CHW programmes are inexpensive, it is important to consider why the relative cost of CHW programmes is high in relation to the funding available for healthcare. One such reason is the sheer scale of a CHW programme that provides universal healthcare access for the rural population of a country: while per inhabitant cost is low, the total cost of CHW programmes is high. There is a similar issue with primary care systems (which incorporate CHW programmes) in general. McCord et al suggest that such a system would cost around US$55 per capita,13 but of the 37 countries included in our analysis, only 10 spent this much on any form of healthcare in 2012.21 The significant challenge of meeting the global need for human resources for health (including CHWs) has been recognised by WHO, who state that ‘the scale-up required… to meet increasing demand, address existing gaps and counter expected turnover is greater than all previous estimates’.4

What, therefore, are the options in countries where affordability is currently low? The first option is to increase the funding available for CHW programmes, although raising additional domestic funds may not be possible in the poorest countries, which have very small tax bases. Some reallocation of funds within the health service (or across government departments) or efficiency savings could be achieved, but as noted above, this may not be easy in practice.27 28 Any redistribution also needs to be careful not to ‘rob Peter to pay Paul’ because different components of the health service are partly complementary to each other. Innovative methods of financing,8 increased publicity and awareness among donors regarding the positive impact of CHW programmes could raise additional external funds. However, a CHW programme provider needs to know that such funding will be sustained until the country has reached the level of economic development required to take over financing if potential health gains are to be realised.18 Changing attitudes to investments in human resources so that investment in the health workforce is seen as investment in a productive sector4 may also help to encourage investment.

The second option is to reduce the cost of CHW programmes. Costs can be reduced by limiting the scope of CHWs’ roles (ie, focusing on one or two key tasks/disease areas). This would enable CHWs to work at a lower coverage ratio, such as 1 to 1000 population rather than 1 to 650, but requires a comparative economic analysis of where CHWs would add most value to population health. Alternatively, models in which CHWs earn commission on the sale of medical or other goods reduce net programme costs to the provider/payer, but such models could distort CHW behaviour from what is optimal for their clients.29 We included CHWs as volunteers as one of our CHW salary options, and others would question the wisdom and long-term feasibility of this option.8 14 We support the payment of salaries but appreciate in some countries that CHWs are willing to volunteer as the alternative is no CHW programme at all (eg, Uganda15). Other approaches to cost-cutting, such as reducing monitoring and supervision, are unlikely to be an appropriate solution given the potential detrimental impact on the quality of the care provided.30 Total costs could also be reduced by not introducing CHW programmes in areas currently without any CHW provision. In the longer term, this would be challenging to justify as any existing inequality of provision, even within a country, would remain. However, in the short term, this option provides an opportunity to test out various approaches to CHW deployment before undertaking large-scale roll-out that would be difficult and expensive to alter. Of course, the optimal solution in one country may not be optimal in another, for a variety of reasons, including local norms and customs as well as access to finance.

Our work is limited by its reliance on secondary data, which is mostly from 2012. Lu et al 16 also highlight discrepancies between WHO and International Monetary Fund data on public spending on healthcare. Although we have used a published and robust source of CHW programme costs,13 this focuses on CHW programmes for the rural population. The cost of such programmes may be higher than for those in urban areas because of the time taken for CHWs to travel between households. Furthermore, the CHW programme as described is not a ‘one size fits all’ solution (as was clearly noted by McCord and colleagues13). In practice, there is a wide variety of CHW programme designs—in terms of tasks assigned and CHW coverage ratios—both within and between countries. More flexible costing tools that can be tailored to local contexts are available for programme providers, such as that produced by Management Sciences for Health.31 CHWs’ roles are also likely to evolve over time as economies grow, populations become more able to self-manage some aspects of their own health, as the burden of disease changes and perhaps as CHWs begin to be professionalised with formal training and even accreditation requirements. As in Ethiopia, new volunteer roles could emerge as the original CHW cadre starts to take on more responsibility and get paid. On the other hand, it is also plausible that CHWs as a provider of ‘first-line’ healthcare are a temporary phenomenon, bridging the gap until a full-scale professional workforce can be supported. We recognise that the labour inputs provided by CHWs (particularly in terms of hours worked but also the effort made during those hours29 30 32) may vary across the three salary structures included in our analysis and this would cause the non-salary costs of a CHW programme (as well as its outcomes) to vary correspondingly. For example, CHWs working longer hours are likely to require more medical supplies.

Our analysis does not incorporate the cost savings to the health system, microeconomic benefits to CHWs’ clients or macroeconomic growth that may be realised from investing in a CHW programme that has a beneficial effect on population health. Valuation of such effects would be requirement for a full, societal-level cost–benefit analysis of CHW programmes, which is beyond the scope of this paper. In addition, we did not attempt to apportion external funding for healthcare to different programmes: CHW programmes may well attract a disproportionately high share of such funding given their focus on improving access to healthcare.

With careful and ongoing consideration of how CHWs’ contributions can be maximised in relation to current and expected future social norms and disease burdens, CHWs can play a vital role in improving health outcomes. However, they cannot be effective in a vacuum—without links into the formal healthcare system—and they should not be viewed or implemented as healthcare ‘on the cheap’.8 Indeed, for many countries in sub-Saharan Africa, full-scale, comprehensive CHW programmes are currently not cheap at all. However, we hope that our results will initiate national-level discussions about how healthcare budgets can be increased to meet Abuja Declaration targets, best supplemented with sustainable and sensible foreign aid and subsequently allocated across services to maximise health outcomes, while working towards universal access. Such discussions need to recognise a country’s current level of economic development and consider plans for reducing reliance on foreign aid and/or future expansion of CHW programmes as the economy grows.