Research in Sub-Saharan Africa: expensive, distorted and morally questionable?

 

In the last decade in many Sub-Saharan African (SSA) countries health indicators have seen impressive improvements. The basis for these success stories is multifactorial: Domestic economies and capacity have grown while international and national health initiatives have been implemented. Many efforts have been targeted towards achievement of the Millenium Development goals for the year 2015. Over the last decade the recently defined sector of Global Health has benefit from increased funding and a rising public profile. While the remit of Global Health has included intervention, it has also brought a boom in research in developing countries led by government, independent and academic institutions. Indeed, it has become essential for any respected academic centre to have a Centre for Global Health/International health. One that will be judged by the number of successful grant proposals and high impact factor publications.

Figure 1. Research Articles on PubMed, searching under term "tanzania", from 1980-2011. Total in 2011 - 621

In the context of the developing industry of research into health sciences in developing countries or, specifically to this paper, SSA, I would like to ask three questions of the role of this research:

1. Is it money well spent?

2. Is it what SSA needs to know?

3. Is it ethical?

By writing this I am outlining my viewpoint rather than writing a systematic paper, and I aim to stimulate debate by describing perceived inadequacies, rather than suggestions for improvements.  Let me also clarify my background in this debate. With an ongoing interest in healthcare in developing countries, I admit to a relative paucity in “on-the-ground” experience having spent only the 10 months living, studying and working in healthcare in East Africa. I have never been involved in any research, but have had the opportunity to witness it in action, and speak to many of the researchers.

1. Is it money well spent?

To make this question more specific, perhaps it is better to phrase it as “does money spent on research represent an economically sound investment for the country?”. In answering I want to use Tanzania as the case study and use two approaches. Firstly, by comparing current research spending to total healthcare spending in Tanzania, the UK and the US. And secondly, by framing Tanzania as a developing healthcare system in need of rapid improvements, is formal research the best way to get results?

Research and Development and Total Healthcare Expenditure: The facts and figures

Before the discussion, it is important to comment on the figures being used. Ideally, one would compare using data researched using the same methodologies with distinct inclusion/exclusion criteria (e.g. including/excluding capital projects for research institutions etc). Finding this comparison data is impossible and I have had to settle with the most accurate data possible, which means reports from National/International institutions and peer-reviewed publications. Unfortunately, in the absence of more regular data, this has resulted in a range of 9 years between some estimates.

National research and development spending is challenging to estimate. For the UK I used a HMRC report from 2006[i], which reports to have included all peer-reviewed research grants, although I suspect may still be an underestimate. If we include private healthcare and pharmaceutical company grants, one healthcare blogger suggested that it may be as high as USD 2.4 million[ii]. In the US, although the figure took into account private investment and therefore a more accurate representation at the time, the figure is now almost 10 years old and therefore may have changed in comparison to total healthcare spending[iii]. The Tanzanian figure is based on a more recent study in 2009 and represents an amalgamation of three areas of funding. The national research institute (partly foreign supported) provided approximately USD 175.5 million, private grants (from the Bill and Melinda Gates foundation, the Rockafeller foundation and the Wellcome trust) contributed USD 10.7 million and donor agency aid contributed USD 15.8 million. Of the three, the national budget is most difficult to estimate. Over recent years the government has committed to increased public spending on R&D, to total of 1% of GDP by 2015. In 2007 this equated to USD 234 million, of which 38% was from external budget support[iv]. Over the years 1995-2004, an average 73.95% of R&D spending went on health sciences[v]. Using these figures, an estimated USD 173.04 million of the national budget was spent on health sciences research in 2007, giving a total of USD 202 million. 

Country	Source	Year	Total Funding(USD/million)
UK	HMRC Report	2006	1,492
US	Medical Research – State of Science (JAMA)	2003	94,000
Tanzania	COHRED report + Science in Africa: A view from the frontline	2007	202

Table 1. Research and Development funding for Tanzania, UK and USA (x1 million USD)

In estimates of total healthcare spending, data for the UK and USA is likely to be relatively well-evidenced coming from respected national institutions; Tanzanian healthcare spending is more difficult. I have chosen to use WHO National Statistical data for total spending in 2008 (private – 40.8% and public – 59.2%) converted at the USD exchange rate at the time[vi]. The alternatives would be to adjust for purchasing power – which I have not done for research spending – or to use Ministry of Health and Welfare data. Interestingly, Ministry data for 2009/10 suggests public national spending of USD 861.46 million USD, significantly higher than compared to WHO data on public spending. This may well lead us to an assumption, which would correlate with experience in other African countries that, due to informal fees and traditional healers, total spending is higher than expected, although by quite how much, we cannot be sure. It is important to note however, that Ministry data includes social welfare spending as well as healthcare. For the purpose of this essay, I will continue to use the WHO data, although on the suspicion that it may be an underestimate.

Country	Source	Year	Total Spending (USD/billion)
UK	HM Treasury Public Expenditure Statistical Analyses[vii]	2006 2008-9 2012-13	89.9 155 196
USA	National Healthcare Expenditure, Office of the Actuary in Centre for Medicare and Medicaid[viii]	2003 2007 2010	1,678 2,297 2,594
Tanzania	WHO National Statistics	2007-8	0.874

Table 2. Total Healthcare Expenditure in Tanzania, UK and USA (x billion USD)

Despite the already described shortcomings of the data, the difference between total healthcare spending and total research spending is striking. In short, in comparison to total healthcare spending, research spending in Tanzania far outstrips that of the UK and the USA (see figure 1). At USD 202 million, the Tanzanian health science research budget is almost 1/5 (23.1%) compared to the total healthcare expenditure for the country, compared to the UK and USA, in which research spending is only a small fraction (1.7% and 5.6% respectively) compared to expenditure on care. It should be noted that this comparative difference is not present in real terms, in comparison to Tanzania, the UK spends 7x as much (1492 vs 202 million USD) while the US spends 460x as much (94 billion USD)

How much would you spend?

Both the UK and the USA pride themselves on their healthcare research credentials and see it as an essential, internationally competitive industry. In the 2011 budget, healthcare research was one of only 6 areas to have above inflation increases in spending. The USA is widely accepted as an academic powerhouse in basic science and medical research. For both countires, exporting research and the financial benefits of housing research is well understood and, as a result research is actively encouraged, often with financial incentives. Taking these facts into consideration one might expect research spending to outstrip patient focused healthcare spending, especially in comparison to the academic minnow of Tanzania. But it doesn’t. Tanzania fights its corner and comparatively spends almost 13x and 4x as much on research when compared to the UK and USA respectively.

So, in answer to the first part of this question, in a country where many fail to receive even the basics of access to healthcare – an ECG in a suspected MI, or Labetalol for pre-eclampsia – does spending 13x or even 4x what a developed country does on research represent an economically sound investment? Personally, I think not and based on what we have seen from the UK and USA, at least an allocation of a similar proportion would seem more appropriate.

But then again would it? Imagine you are given responsibility for a complex task, say setting up a chain of supermarkets in Tanzania. You have limited resources and you are told that, for the moment, your aim is to have a functional, rather than flamboyant product. What is the most efficient way about this? Do you start from the beginning by inviting foreign consultants (at significantly higher cost than local or regional consultants) to carefully researching consumer tastes – what colours they like, how they might shop and what brand of peanuts they like. No, you start by building a few stores in the busiest cities based on your past (or your competitors!) experience and see what works. Essentially, you bring about change through a rapid process of trial and error while learning from your and others mistakes, quickly ending up with a relevant and cost-effective outcome. Monitoring and Evaluation is central to this, but large volumes of academic research is not.

In defence of research spending?

There are several likely responses to the research spending data I have presented. I think one of the most likely defences is what I will call “the blank slate”: There is so little research in any areas of medicine in SSA that we urgently need to know more. For example, the prevalence of cytochrome 1A1 and 1B1 gene diversity in the Zanzibar Islands (its low and so there are relatively few side-effects to Amodiaquine – so I’m unsure what prompted the initial research question, but I’m sure they had a great holiday)[ix]? The real inadequacy of this argument struck me in a clinic last week. At the end of the clinic I reviewed our patients and those of the colleague next to me, approximately 25 in total. Every single one had a well-defined, well-researched disease; hypertension, heart failure (sometimes secondary to Rheumatic Heart Disease), chronic kidney disease and the odd epilepsy. The tragedy of their management was not the absence of research, it was in the failure of the health system in streamlining their care, the laboratory unable to process their renal function tests and the inability to provide evidence-based interventions. This dichotomy between knowledge and practice is already well documented in the “know-do” gap, the closure of which is already a WHO priority[x].

One journal article also highlights the experiences that local staff may get from working with research projects as a potential benefit¹. Whether this is technical, managerial or educational, one might argue that intellectual capital is a benefit to the country. Maybe it is, but if the alternative is for the healthcare worker to be working in their health system, I don’t believe that their experience there would be any less valuable to their future career. In fact, many academics would suggest that after a career in research, its very hard to go back to clinical work.

Unfortunately I could not find up to date and congruent data sets between countries, especially on research spending. In collecting the values, I have chose quality over currency, a virtue I hope is clear through the discussion of the values. I appreciate that research spending in the US has changed over the past 9 years, but so has healthcare spending (as shown in Table 2.). And although there are likely to be inaccuracies in the estimates, I do not believe they would change the final outcome: that there is a large disparity between proportional research funding in Tanzania and the UK and US. 

Differences in relative research spending and what to make of it

As I reviewed this data I found myself asking, why does Tanzania spend proportionally so much more on health sciences research than the UK and USA? I have no clear answer to it, but I suspect the answer may lie in two areas. Firstly, the emphasis placed on research and development, rather than direct care, by external donors dictates an essential industry in the area. And secondly, that for ministries and donors choosing to understand the problem is easier and cheaper than fixing it. When facing the problem of Chronic Kidney Disease in Tanzania it is easier to map epidemiology and identify the aetiology than implement a Kidney Care Network that aims to prevent and manage (including dialysis and transplant) the problem. For foreign donors and ministers, a manageable and achievable aim is a much more attractive prospect than one mired with the prospect of failure. I choose to believe these options in the face of a much more sinister viewpoint, that foreign research institutions use SSA populations to produce cheap and impact factor busting publications which win prestige for their institution.

Finally, this analysis does not suggest the right answer is to take from research and give to care. We can already see that in real terms Tanzanian research in under-funded compared to the UK and USA. Rather that it would be beneficial to spend more on both, and given the current situation just outlined, proportionally more on care to bring it in-line with developed countries. A proportional change would necessitate a redefining of what research is essential and should be funded.

2. Is it what SSA needs to know?

Let me start this section with an example of a study I recently read but will choose not to quote, on the aetiology of febrile illness in two large hospitals. The study has an interesting outcome, that febrile presentations have a much wider set of pathogens than initially suspected. The study was led by foreign investigators who used their own standards, equipment and laboratories. This varied between blood cultures grown locally in a specially installed private lab, and flying samples back to the research team’s country for serological testing. This study is suggested to have use for local physicians (of whom only a couple had even heard of the study, let alone knew the results) by proposing that they explore a wider set of differentials when reviewing febrile patients and in certain cases think of an antibiotic change here or there.

But, is it of any use to the healthcare staff, or research population? I would argue not, in fact the research is inappropriate. Despite being the larger of the two hospitals being a tertiary referral centre, and therefore one of the best equipped public hospitals, it does not have the serological tests while laboratory blood cultures are expensive and notoriously insensitive. Furthermore, they lack a set of antimicrobial guidelines due to unknown antibiotic sensitivities. The authors also suggest the results add weight to arguments of creating cheap point of care tests for diseases such as Q fever. To me, this is a precocious and far-fetched justification; after all, malaria rapid tests are relatively cheap and well-evidenced but not yet in use in any of these hospitals.

The alternative could be to work with clinical staff to perform relevant in-house research on their febrile patients. Necessarily, this would be complex, protocols may change, standards in the laboratory would have to controlled and available parameters extended. Local healthcare staff would need to be sensitized and involved in framing the research question. But in the end, you might not only come out with a publishable study of febrile illness, but one that is relevant to and supportive of the research population and feasibly reproducible.

Clinically orientated research must have its roots in the experiences of those on the frontline of medicine. There is little utility in parachuting studies into populations using unavailable tools and following an external agenda. Furthermore, because the results should intend to create an effect in the research population, the results must be disseminated within the frontline staff.

3. Is it ethical?

I find this question the most emotive and struggle to deal with its concerning implications. To highlight some of the issues, I would like to tell the story of a 13 year-old girl I met a few months ago in a National Referral Hospital. She came in with the worst headache she had ever had. She lay on the bed and her mother held her head as she cried. She had recently completed treatment for TB. Although both mother and daughter said her HIV status was negative, a test today found that she was positive. Given all of these pieces of her history, the local team admitting her astutely felt the most likely cause was Cryptococcal meningitis.

She was clearly very unwell: Cryptococcal meningitis in high resource settings has a mortality of approximately 25%. In the best study settings in low resource countries this is around 40%. A team from a large foreign research program running an ongoing trial into Crytpococcal meningitis management came to see her and agreed with the diagnosis. Unfortunately, being under the age of 18, she did not fit their study criteria and therefore could not receive the “best available treatment”: A CT head (externally performed as the hospital one was broken), serial lumbar punctures (LPs) from pre-prepared packs, a strictly adhered to follow-up from the study team using a treatment protocol and either an Amphotericin based or high dose fluconazole regime. All at no cost to the patient (the normal practice being user fees).

Instead, and despite being in the best equipped national hospital in the country, she would be looked after by an overstretched team of Doctors and Nurses without easy availability of equipment (an LP would be charged to her and her family, and even then equipment is not easy to find), medications (Amphoterecin B was available at a high cost outside the hospital, Fluconazole is free), expensive lab tests and no available CT (if it were available it would be at a cost). The contrast is care is obvious and goes a long way to explaining the sadness of her situation. It’s also probably why the research team offered an LP pack to her family. When she duly received her LP her cerebrospinal fluid hit the wall as the needle went in, and within a few minutes her headache was better. She will need a full course of antifungal treatment, serial LP’s and careful follow up to start antiretrovirals (ARV) but, for the moment, she was much more comfortable.

For me, it’s an incredibly sad and frustration story to recount. Although I will never know, on the basis of what I have seen since, there is a significant probability that she will have died. The case pitches a stark contrast between the “lucky” study participants and patients like this young girl, who failed their enrollment criteria, or had their disease before or after the study period. Despite being cared for in a National Referral Hospital, there is still a deep chasm between research and standard clinical practice. The research protocol is so divorced from reality that it barely resembles standard treatment and the results will be of little use to future patients. Nevertheless, the study will no doubt be hailed as a success when it is presented internationally as a new dawn for the treatment of Cryptococcus. This 13-year old girl will be presented in the “Background” slide, as one of the 1.8 million who died that year as a result of their HIV infection. One could go on to describe how the research fails to satisfy the Declaration of Helsinki – a constantly updated guide to performing morally sound research. However, I think for the case of this 13-year old girl and other patients in her position there is no need, the injustice is plain to see.

The failure to reflect the standard of care in the research population was clear in the HPTN052 study. An international study that was widely accepted as a “game changer” for HIV showed that in sero-discordant couples, early treatment of the infected partner with ART reduced transmission by 96%, when compared to standard treatment. At the end of the study researchers were in a sticky situation. Whilst in developed countries, resources were available to start all trial members on ART, in developing countries (where the burden of the disease is) it wasn’t so simple: How could they universalize the results? They couldn’t, and in the face of this betrayal the study team create a weakly defined “demonstration project” for those already in the study to carry on their ART’s. Whilst this study is clearly fascinating, how long it will be until their results affect all of the population only time will tell.

To put this issue in another context, imagine you work in the emergency department of a western hospital. In the hospital diagnosis of an acute myocardial infarction is made by ECG and just recently you have started thromolysis as management. The study arrives, complete with their own team, ECG and also troponin assays. They will compare outcomes for their patients when treated with either a drug eluting or bare metal stent. The hospital is completely broke and can barely support the current, relatively inexpensive treatments let alone shoulder the burden of newer, more expensive alternatives. I can’t help but feel that, given a situation like this, there would be local uproar – why are we supporting this study when it will have no benefit for our patients? And even this blatant example fails to encompass the awkward question of the withdrawal of treatment from the chronic disease patient, as evidenced by the HPTN052 study. 

What crisis?

This essay aims to stimulate debate around the subject of SSA healthcare research, it is opinionated and based on experiences. It does not aim to exclude the essential role of research in healthcare. Indeed, the successes of many research projects in SSA cannot be doubted, for example studies of scaling up ART with the aim of advocating for universalizing treatment or, evaluating the efficacy of artemesin based regimes for severe malaria. I would argue however that they satisfied two criteria. Firstly, they were implementation based, converting already gathered knowledge into action. The importance of implementation research is appreciated but as yet, its real potential is unharnessed. Secondly, they have been followed by a commitment of resources to tackle the ongoing and wider issue.

The field of Global Health is dominated by professionals from developed countries well-versed in the utilitarian approach to health systems that pervades vertically themed health programs. They may point to a crisis in knowledge in SSA healthcare, the gaping holes in our understanding of diseases or potential treatments that we haven’t fully evaluated. Go to a local health centre, district hospital or referral hospital in SSA and ask them what their problems are and you will find a different answer. The research crisis is far removed from the crisis in care they face on a daily basis. Human, technical and economic resources constantly fail, the result being that their patients with treatable conditions die unnecessarily. The families of these patients return home, feeling betrayed by western medicine, let down by a service that should care for their relatives. The utility of research is lost on them. In the face of this, many of these Global Health professionals will cling to a belief that more research can change this picture. To an extent they may be right, high quality and applicable research undoubtedly has a role, but currently this role is relatively over-financed, distorted by externally driven ideas and the result is morally unsound research at the expense of improvement in care.

---

[i] Sir David Cooksey. A Review of UK Healthcare Research Funding. HMRC December 2009 accessed at www.hm-treasury.gov.uk/d/pbr06_cooksey_final_report_636.pdf on 24/6/2012

[ii] http://really.zonky.org/?p=556 accessed on 24/6/12

[iii] Fontanarose P et al. Medical Research – State of Science. JAMA. 2005;294(11):1424-1425.

[iv] Viviene Irefeke et al. Science in Africa: A view from the frontline Nature 474, 556-559 29/6/2011 accessed at http://www.nature.com/news/2011/110629/full/474556a.html

[v] Montorzi G et al. An Assessment of the Health Research Systems: A country report for the AHA series. 2009 Council on Health Research for Development http://www.cohred.org/downloads/cohred_publications/Tanzania_low_res.pdf

[vi] Total Health Spending (USD) WHO Statistical Information System, accessed at http://www.gapminder.org/data/ on 23/6/12

[vii] HM Treasury Public Expenditure Statistical Analyses accessed at http://www.ukpublicspending.co.uk/year_spending_2006UKbn_11bc1n_10#ukgs302 on 19/6/12

[viii] National Healthcare Expenditure, Office of the Actuary in Centre for Medicare and Medicaid http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/index.html?redirect=/nationalhealthexpenddata/, 2008

[ix] Cavaco I et al. cytochrome 1A1 and 1B1 gene diversity in the Zanzibar Islands Trop Med Int Health. 2012 May 18.

[x] Bridging the Know Do Gap. Meeting on Knowledge Translation in Global Health 2005 accessed at www.who.int/kms/WHO_EIP_KMS_2006_2.pdf