A large amount of recent literature on global health makes the suggestion that a major barrier to improved health outcomes in low income countries is the general weakness of their healthcare systems.  Some literature suggests that some targeted programs to control certain diseases may even be acting to weaken health care systems in low-income countries through creating superfluous parallel delivery systems and disrupting the day to day activities of the health system.  
These observations parallel concerns sometimes expressed to Giving What We Can that the organisation’s recommendation of charities that target specific diseases such as the Against Malaria Foundation and Schistosomiasis Control Initiative shows a lack of attention to the need to strengthen healthcare systems in the developing world.
In this report I will assess the current literature on this subject and relate it to our recommended charities.
The concern with strengthening health systems in low income countries can be seen in a debate in the aid literature over the comparative benefits of “vertical” and “horizontal” interventions. A stereotypical horizontal programme would seek to tackle overall health problems on a wide front and on a long-term basis through the creation of a system of permanent institutions. This is one thing people have in mind when they talk about the need to “strengthen health systems”. By contrast a stereotypical vertical intervention would be a free-standing program focussing on a single goal which it achieves through specific interventions. Smallpox and malaria eradication programmes are examples of vertical interventions. AMF and SCI would best be classified as vertical programmes.
There has been debate about the relative merits of both forms of intervention with some enthusing about the ability of vertical interventions to create large and immediate improvements in health and take pressure off general health services with others claiming that much of the good done by vertical health interventions is outweighed by disruption of general health services and inefficiencies caused through the lack of integration between different health services. 1-4, 
The first thing to note is that the debate over the relative effectiveness of vertical and horizontal programmes takes place at such a high degree of generality that it is difficult to come to any strong conclusions. Another issue is that existing healthcare systems tend to involve a mixture of both horizontal and vertical programmes making it difficult to make assessments of relative effectiveness. In light of these issues a recent report by the WHO states that:
“The available evidence on the relative benefits of vertical versus integrated delivery of health services is limited and too weak to allow for clear conclusions about when vertical approaches are desirable”
The same problems seem to be faced even by more limited studies. A recent literature review of programmes targeting Neglected Tropical Diseases argues that the lack of studies on the impact of targeted NTD interventions on health systems means that:
“Further research is therefore timely and needed to go beyond opinion-based discussions on effects of NTD campaigns on health systems and health care delivery”.
It is worth bearing in mind the lack of a good evidential basis when discussing concerns around strengthening general health-care systems.
Despite these concerns the WHO report suggests that there are circumstances in which it is plausible to think that vertical programmes are more effective than horizontal ones. One of these circumstances is programmes operating in countries which have very poor health infrastructure. This is the case for a variety of reasons, in some countries health infrastructure may be so poor as to make vertical interventions the only feasible option for providing high levels of coverage. Additionally the ability of vertical interventions to achieve results more quickly means that they are a better route to immediate improvements in public health that have positive follow on effects. Due to these considerations we might think that worries about the relative effectiveness of vertical interventions are just misplaced when applied to AMF and SCI given that they operate exclusively in some of the poorest countries in the world.
We might still be concerned about the potential of vertical programmes to undermine general health systems, if this is true of a specific programme then we should revise our estimates of it’s effectiveness. The way in which some vertical programmes are thought to weaken general health systems is through the creation of “parallel subsystems” for the delivery of healthcare. These create problems such as
We can look at the work of our recommended charities AMF and SCI in detail to assess to what extent we think that they risk creating these problems.
We can be confident that duplication of service provision is minimised in the case of AMF. In deciding where to distribute nets AMF looks at the Alliance for Malaria Prevention’s list of countries with significant net gaps and discusses where to distribute nets with other malaria-control funders, the relevant National Malaria Control Program and the African Leaders Malaria Alliance. When investigating a country AMF works with partners already present on the ground and requires these partners to carry out pre-distribution surveys to ensure that they are not distributing nets to areas that already have them.
What is more difficult to establish is to what extent the work of AMF might cause distortion, distraction or disruption to provision of other services. Nets are distributed by low level government health workers and staff at the NGOs AMF are working with. In Malawi for instance government workers called Health Surveillance Assistants were used to distribute nets. HSAs are the first line of response to public health issues. They disseminate health related information conduct basic nutrition support, report levels of stunting, wasting and common communicable diseases to clinicians and carry out immunization programs. Distribution of nets would appear to fit well with their other duties although more information is needed to know whether the distribution of bed nets hampers or complements their other responsibilities. It might be that distributing malaria nets means that HSAs do not have time to do other work but it might also be the case that by going into communities and giving out nets increases their standing within the community and provides information to them which facilitates their other work. Regardless the fact that net distribution only takes a couple of days seems to indicate that any negative impact is minimal.
As with AMF duplication of services does not seem to be a problem with SCI. SCI works with governments to roll out treatment programs through financing these programs and providing technical expertise and support if necessary. For the most part they operate in countries that had no treatment program before SCI began working there.
It appears that the capacity for distortion, disruption and distraction through SCI initiatives is limited. Distribution of treatments only takes a couple of days each year. Additionally the distributors of treatment are not trained health professionals but rather teachers and community drug distributors who receive a day’s training in measuring correct dosages. It therefore seems reasonable to suppose that SCI is not causing disruption through using the time of professional health workers in a suboptimal manner.
There are two clear means by which AMF and SCI strengthen health-care systems. The first is that covering some of the cost of malaria control programmes and deworming makes these programmes less expensive for governments to administer leaving more money and resources to be spent elsewhere. The second is that they reduce the burden of diseases which would otherwise be placed on other parts of the health system.Malaria is a huge burden on global health with 3.3 billion people at risk of being infected with malaria and developing the disease and a further 1.2 billion at high risk. Distribution of insecticide treated bednets as carried out by AMF is a highly cost-effective way to combat the disease. AMF supports the work of vertical National Control Programmes for Malaria which have positive effects on other areas of the health system. For example the 2011 WHO World Malaria report states:
“In Rwanda it has been estimated that while it would cost US$ 265 million to sustain the malaria control programme over the next five years, the public health system could avert about US$ 267 million in the costs of diagnosing and treating malaria; and households could avert about US$ 547 million in direct and indirect costs, equivalent to about 7% of household income.”
A recent study looking exclusively at the use of insecticide treated bed-nets concludes that:
“The use of mosquito treated nets can induce a 6.2 percent reduction in outpatient costs, a 6.6 percent reduction in inpatient treatment costs, a 6.3 percent reduction in productivity losses, and a 22.6 percent reduction in disability adjusted life years (DALY), all on an annual basis.”
SCI’s deworming initiatives target a number of parasites including soil transmitted helminths and schistosomiasis. 3 billion people in the world’s poorest countries are infected by STHs and 261 million by Schistosomiasis.  As well as more severe symptoms these diseases cause anemia, malabsorption of nutrients and stunting of growth and development which are general causes of ill-health. Deworming programmes are a highly cost-effective means to combat these problems and it is reasonable to assume that the general improvements in health caused by such programmes must take pressure off other components of the health systems.
It’s important to comparatively assess other options for foreign donors to strengthen general health systems other than supporting charities that are minimally disruptive and massively reduce the burden of disease. One option might be supporting charities that engage in horizontal interventions, Living Goods for instance runs networks of Community Health Promoters who provide health advice and sell basic healthcare goods. However there is no reason to think that NGOs engaging in horizontal interventions are any less likely to face the problem of “parallel subsystems” than vertical programs. We might think that simply giving money directly to the Ministries of Health in low income countries would strengthen health systems. But presently Ministries of Health are often choosing to engage in deworming and malaria control programs themselves, indicating that they see spending in these areas as the most effective use of available resources.  Furthermore recent evidence suggests that whilst health assistance for development directed to NGOs increases government spending on health-care, giving money directly to governments displaces money that would have otherwise been spent on health-care. We might consider spending money on organisations that advocate for strengthening health systems, but we are not aware of any organisations that advocate for this. Further, it is worth noting that all of these efforts to directly strengthen health-care systems may end up being less effective than indirectly strengthening health-care systems by reducing the burden of disease through successful vertical interventions.
Finally we should note that strengthening health-care systems is just one component that should factor in our assessment of charities. Just because one intervention strengthens the general health-care system more than another does not mean that it is necessarily more cost-effective. It may be that at this time the benefits of combating specific diseases outweigh the benefit of any attempt to directly increase the general standard of healthcare.
Ultimately there is a lack of good evidence regarding the cost-effectiveness of taking steps to directly improve the general health-care systems of a country through horizontal rather than vertical programmes. Nevertheless the threat of undermining general health-care systems through the creation of parallel subsystems is something that we ought to account for when estimating the cost-effectiveness of our charitable giving. In the case of Giving What We Can’s top charities we can see that concerns about disrupting health-care systems may simply be misplaced to begin with given that they operate in countries which have extremely weak health-care systems. Furthermore when we examine the model for how parallel subsystems are supposed to emerge and weaken general health-care systems and compare this to the work of AMF and SCI we see that there is little reason to think that they conform to this model. Additionally it is clear that AMF and SCI must strengthen general health-care systems insofar as they reduce the burden of disease that would otherwise be met by the rest of the health-system. Indeed it might be the case that the best way for foreign donors to strengthen health-care systems is to support the work of these charities, particularly given the apparent lack of other options at this time.