MgSO4 in the treatment of eclampsia and severe pre-eclampsia
The research evidence
Landmark trials, subsequently incorporated into Cochrane reviews, have demonstrated the efficacy of MgSO4 in the treatment of eclampsia and pre-eclampsia in pregnancy [23–27]. South Africa and Zimbabwe participated in these trials, which presented high quality evidence for the use of this drug as a first line therapeutic.
Local involvement in evidence production
In all three countries, leading obstetric departments were involved in trials of MgSO4, and the obstetricians concerned participated in policy formulation, influencing uptake of these research findings. In particular, senior obstetricians and other researchers in South Africa and Zimbabwe were involved in the Collaborative Eclampsia and Magpie trials [26, 27]. This enhanced the credibility of the trial results within these settings:
"I think it [the Collaborative Eclampsia Trial] was presented at a number of local meetings.... the leader in South Africa of that trial,... he presented it at many meetings and so it was a well known research in this country and in Zimbabwe." (Clinician/researcher, South Africa)
Involvement in these trials allowed local researchers to gain further experience in the use of MgSO4 for the treatment of eclampsia and pre-eclampsia. However, respondents reported that adverse outcomes experienced by patients receiving MgSO4 in the Collaborative Eclampsia Trial in Zimbabwe raised concerns amongst clinicians regarding its safety. This was thought to have inhibited its uptake as first line treatment for eclampsia in the country.
Respondents in Mozambique reported that a trial conducted locally in 1989 showed MgSO4 to be superior to diazepam for the treatment of eclampsia. Although this study was never published, it was considered important in convincing obstetricians to use the drug:
"We developed the trial up to a certain point, but it was obvious that the patients treated with magnesium sulphate had a lower mortality, and they awoke more rapidly from comas and convulsions..." (Clinician/researcher, Mozambique)
Interactions between research - both local and international - and experiential knowledge therefore influenced researcher-clinicians' acceptance of trial findings on the effectiveness of MgSO4; the value attributed to them; and the uptake of these findings into policies.
Prior practices and beliefs
While research knowledge was important in shaping policy and practice, experiential knowledge also played a key role. Based on their clinical training, obstetricians in South Africa and Mozambique tended to follow the obstetric schools that promoted the use of MgSO4 for the management of eclampsia:
"I think it depended on where people studied. It depended on what people read and it depended on what people believed." (Clinician/researcher, South Africa)
Because practice prior to the landmark trials was often based on MgSO4, many obstetricians noted that they had experienced the clinical success of the drug long before it was shown to be effective in trials:
"...even in the absence of a randomised study, I think the empirical results were very convincing." (Clinician/researcher, Mozambique).
Because the evidence from the later Collaborative Eclampsia Trial  was congruent with existing practice, its uptake into policy was easier.
In Zimbabwe, on the other hand, obstetricians tended, before MgSO4 was demonstrated to be effective, to follow the British tradition of using diazepam for eclampsia management:
"...the older consultants have been trained in Britain and came with what we were doing in Britain, and practising what we were doing in Britain and this has been passed on to the others in the country...." (Clinician/researcher, Zimbabwe)
"... diazepam has been there for as long as I have been there, even pre-independence [before 1980], diazepam has been used all along...." (Clinician/researcher, Zimbabwe)
Champions and lobby groups
The role of, and need for, local champions in placing issues on the policy agenda was highlighted. In all three countries, clinical champions lobbied for MgSO4 to be included as the first line treatment for eclampsia:
"Yes, both [clinicians] in private and in public practice [lobbied], some as groups of researchers and others in their individual capacities...." (Clinician/researcher, Zimbabwe)
Respondents in South Africa suggested that individual and organizational lobbying drew attention initially to the causes of maternal mortality, helping to place maternal and child health on the policy agenda [39, 40]. This helped to ensure that the Department of Health prioritised the development of evidence-based policies to improve maternal health. In Zimbabwe, clinical champions were also important in developing guidelines for the treatment of eclampsia. Their importance was illustrated by the fact that the emigration of the local champion for MgSO4, before the full acceptance of the drug was realised, was regarded by some as a cause for slowing the process of getting MgSO4 into policy and practice for the management of eclampsia.
Lobbying and championing was not limited to the national level. Groups such as the Cochrane Collaboration influenced how local researchers and policy makers thought about evidence-based practice and policy, as we discuss below.
Involvement in national and international research networks
In all three countries, academic obstetricians who were key to local policy development were involved in national, regional and international research networks. These networks were influential in building a culture of research and evidence-based medicine through exposing local clinicians to these ideas as they developed internationally. Directly and indirectly, these networks therefore shaped the translation of evidence into policies. For example, our data indicated that proponents of evidence-based obstetrics internationally, including several linked to the then developing Cochrane Collaboration, participated in conferences in South Africa:
"We actually invited [researchers].... attached to the Oxford Database [the precursor to the Cochrane Library].... So we were sort of, I think from the word go, when the Oxford Database became available for use, we were part of it, we were aware of it, we were using it, and I think quite a few South Africans became involved on their editorial board and as editors or reviewers, or whatever." (Clinician/researcher, South Africa)
Leading international researchers also spent sabbatical periods, or developed research units, in the study countries. In Mozambique, international researchers working in the capital facilitated access to the international literature and organised scientific exchanges:
"...studying abroad, having scientific interchanges not just based on the diseases here, is important. The doctors who come to work here help in training the Mozambican doctors who work with them". (Clinician/researcher, Mozambique)
These networks built links to key research taking place internationally and to international policy debates on treatment. Later these networks were also important in recruiting researchers from the study countries into international trials of MgSO4.
Involvement of researchers in policy making
Key to knowledge translation in the case of MgSO4 was the interface between researchers and policy-making. In all three countries, academic obstetricians who were also active as researchers played important roles in policy development. These obstetricians worked with government officials to draft and review policies, often through expert groups. In Mozambique and Zimbabwe, they were often given policy-making responsibilities, as a researcher in Zimbabwe notes:
"The policy of management of eclampsia is enunciated in the EDLIZ [Essential Drugs List of Zimbabwe] and is a process of consultation within affiliated disciplines to arrive at compilation of the document. Obstetricians and gynaecologists contributed to this through a nominated point person who worked with the Ministry Steering Committee to draw up this document." (Clinician/researcher, Zimbabwe)
Government officials also gave researchers other tasks, such as drawing up operational plans, training and supervision, which contributed to close working relationships. In Mozambique, some academic obstetricians also occupied key positions in the Ministry of Health.
In contrast, limited opportunities existed for academics to engage with government in South Africa prior to the change of government in 1994. The new government, however, prioritized maternal and child health and employed into its ranks key members of the national academic obstetric community. This created opportunities for academic engagement in this sector with key policy-making committees being chaired by academic obstetricians. A clinician researcher described this change in researcher involvement in policy-making:
"...I've worked in obstetrics before the new government and after, and... before there was never the ability to talk about national policies or anything like that. Certainly afterwards [after 1994] there's been a great movement to be able to do that, to participate and to make guidelines." (Clinician/researcher, South Africa)
In all three countries, obstetricians thus had ready access to policy makers and were part of tight-knit policy communities. The closeness of academic obstetricians in each country to the policy making process suggests that they were potentially key conduits for knowledge translation.
A culture of evidence based obstetrics
Attempts have been made internationally to develop a culture of evidence-based research and practice within obstetrics . In the study settings, a culture of basing practice on research findings preceded the availability of evidence from RCTs. We have already described how local researchers became linked into international evidence-based medicine networks. The success of these international initiatives in the study countries was clear from our interviews, with most respondents expressing strong views on the importance of using and generating evidence:
"...so evidence, that sort of thing was really grasped with both hands. I think a lot of our research is clinical. So trials are our - if you want to do research -is our bread and butter... I don't think there's any...O & G [obstetric and gynaecology] academic institution which doesn't use Cochrane extensively." (Clinician/researcher, South Africa)
Most respondents embraced strongly a culture of evidence-based medicine, believing that policies should be based on RCT findings. Since many of these clinicians were key to the policy making process, their definitions of evidence influenced strongly the sorts of information considered during policy development. Respondents placed high value on evidence from RCTs, including the international collaborative trials [26, 27]. Respondents noted that before evidence from these trials became available, other forms of research information such as the Pritchard case series on the treatment of eclampsia [42, 43] were relied upon:
"... in a very famous series of cases of Pritchard, they had 300 consecutive cases of eclampsia without any maternal deaths...with magnesium sulphate. That's clinical proof..." (Clinician/researcher, Mozambique)
Insecticide treated bed nets compared with indoor residual household spraying for malaria vector control
Indoor residual spraying and insecticide-treated nets have been demonstrated to be effective across a wide range of settings. However, few randomised trials have compared directly the efficacy of bed nets and indoor residual insecticide spraying and their comparative cost-effectiveness depends on the context in which they are implemented. It is therefore difficult to justify one approach over another based on the available evidence [44, 45]. At the time of the study, there had also been no trials directly comparing different insecticides. In addition, long-term impregnated nets had not been widely distributed in the study settings.
The stakeholders and international agencies
A large and diverse group of stakeholders was involved in decisions on malaria policy in the three countries. Important players included government officials; multilateral agencies (particularly the WHO and UNICEF); partnerships such as the Roll Back Malaria Partnership (which includes WHO, UNICEF, UNDP, the World Bank and a wide range of other NGOs); foundations; donors (such as the UK Department for International Development (DFID); the Japan International Co-operation Agency (JICA); and the Global Fund against AIDS, Tuberculosis and Malaria); academic and private sector institutions; NGOs (such as environmental organisations); political actors; and commercial actors (such as insecticide manufacturers). These stakeholders expressed varied and contested interests, differed in their use and interpretation of evidence, and promoted different malaria control policies, as we describe elsewhere . Contested issues included which insecticide to use (all three countries) and whether to use nets or spraying (Mozambique). Political and commercial interests were also evident, and these interests attempted to influence both policy makers and researchers. For example, the tobacco lobby in Zimbabwe was important in the decision to stop using DDT because of fears that the pesticide would contaminate the tobacco crop.
" [The] tobacco commercial farming sector lobbied government against use of DDT because the buyers of Zimbabwe tobacco abroad were saying that, if they found traces of DDT in tobacco, they would not buy Zimbabwe tobacco." ( International agency, Zimbabwe)
In Mozambique, younger researchers mostly favoured the introduction of bed nets. This was because they had been drawn into international bed net research networks through their post-graduate studies overseas and through their contacts with international researchers visiting and working in Mozambique. Because of its higher malaria burden, Mozambique was more integrated than were South Africa and Zimbabwe into the international malaria research networks that had undertaken key RCTs on ITNs.
Bilateral donors, multilateral agencies, and international NGOs were important players in shaping malaria control policies, particularly through funding ITN programmes in Mozambique and Zimbabwe:
"...but of course the insecticide treated nets agenda is also pushed, as you may be aware, very strongly by the bilateral donors, and other players and the UN family and so on. And so the government has accepted their advice as it were, of course to actually use nets...Nets, to begin with, I think, were an outside sort of influence" (International agency, Zimbabwe)
"We can see that nets have become an international fashion... there is a lot of pressure regarding them..." (Government health official, Mozambique)
Pesticide and net manufacturers also lobbied for their interests and contributed to developing and shaping evidence. For example, pesticide companies had an interest in promoting insecticides other than DDT for spraying, as greater profits were to be made from newer insecticides. These companies therefore sponsored trials in both Mozambique and Zimbabwe [47–49].
A range of NGOs also worked to influence malaria control policies. Environmental lobby groups, both local and international, played important roles in supporting moves away from DDT in South Africa and Zimbabwe in the 1990s and in mobilising evidence for this:
"The international community, as you probably know, the Greenpeace people and the environmental lobbying groups started putting pressure...we agreed in principle to try reducing the reliance on DDT mainly because of all the things that were published. So those are more or less the reasons why we moved towards reducing DDT." (Government official, South Africa)
Later, the NGO 'Africa Fighting Malaria' took on the task of lobbying for DDT http://www.fightingmalaria.org. This organization, apparently linked to international neoconservative groups , fiercely criticized environmentalists who had pushed for a ban on DDT .
The role of regional networks of policy makers and researchers
Researchers and policy makers were organised into strong regional networks that were important in sharing ideas and approaches to malaria control. These networks played a key role in maintaining the emphasis on spraying in the region.
Historically, South Africa and Zimbabwe had close links and several key South African researchers and malaria control officers had been trained, and formerly held positions, at the Blair Research Institute in Harare - an important regional malaria research institution. Strong relationships later developed between South African and Mozambican researchers and implementers around the Lubombo Spatial Development Initiative (LSDI), which aimed to improve malaria control in South Africa, Swaziland and neighbouring areas of Mozambique.
Involvement of researchers in policy making
Although policy makers may have become aware of research evidence through their own reading, respondents also identified a number of interfaces through which research entered the policy making process. Firstly, researchers were co-opted into formal government advisory committees in all three settings:
"...certainly we [in the government department] might have the background knowledge into it but we're not currently working in researching malaria all the time... So we felt that we needed to bring and call on expertise from the country to advise us on policy... and that is the main reason why the decision was taken to put an advisory group together." (Government official, South Africa)
Consequently, close relationships developed between researchers and health officials responsible for implementing malaria control. These researchers were regarded as experts and evidence uptake was mediated through them.
Secondly, researchers acted in some instances not only as advisors but also as implementers. For example, a senior researcher in South Africa was integrally involved in implementing malaria control initiatives within the LSDI. This initiative, which integrated research and implementation, helped to bring researchers and policy makers together, as one researcher involved in this noted:
"...we brought service and research together. We thought, 'What's happening in the world? There's a lot of research going on but is it translating into implementation?'. I think that you have to bring those two communities together because the one needs to feed into the other... Over the last ten years, maybe fifteen, we've really worked towards trying to bridge that gap." (Researcher, Mozambique)
Thirdly, national research bodies had close links with government policy-making bodies in all three countries:
"I think major advantages that Blair [the Blair Research Institute] has made, or Blair research findings had in influencing policy, is that Blair themselves are part of the national malaria control programme, and Blair do sit in national malaria control programme committees. So the work that they do and their research findings find [their] way almost automatically, naturally into policy and decision making." (International agency, Zimbabwe)
Knowledge translation for malaria control was facilitated through these different avenues.
Defining 'evidence' in the context of malaria control
In contrast to the MgSO4 case study, there seemed to be a less visible culture of decision-making for malaria control being based on RCT evidence. Respondents were clear that, for them, what constituted research evidence went beyond RCTs and included experience, surveillance data and expert opinion. Considerable research was undertaken locally in all three countries. This research - broad in its methodological approach - was regarded as important to decision making. For example, the malaria surveillance data collected for many years in Zimbabwe and South Africa were seen as important evidence for shaping policy (see, for example, ), particularly in showing the effectiveness of the spraying approaches used in these settings. The weight of this evidence meant that RCTs were not seen as necessary to demonstrate the effectiveness of spraying:
"But in terms of indoor residual spraying, I think the evidence is not disputable - it's there to see! In the countries where indoor residual spraying is done, the number of malaria deaths are very low, but in countries where indoor residual spraying is not being done, the number of malaria deaths are so high." ( Government health official, Zimbabwe)
In Zimbabwe and South Africa, the international evidence from ITN trials was also seen as distant and not necessarily relevant as it was felt that the local epidemiology of the disease was different to that of the countries in which the trials were conducted:
"Well I know with nets, even up to now there are a lot of controversies. The epidemiology of malaria in Zimbabwe is quite different from Tanzania, is quite different from the Gambia, or where there is high para-endemicity in those areas" (International agency, Zimbabwe)
Although a national trial comparing bed nets and spraying was conducted in South Africa in the 1990s [53–55], this was stopped when a malaria epidemic demanded quick action that undermined the trial randomisation.
In addition to concerns about the local applicability of RCT evidence, many policy makers were also clear that ease of implementation and sustainability were key to their decision-making regarding malaria control options. Simply showing success in a RCT was not considered sufficient:
"Several randomised and controlled studies have already been undertaken which prove effectiveness in reducing mortality, but how can this ideal situation be translated into practical terms, while maintaining effectiveness?... In practical terms and under real conditions,... this is where questions arise. Why are there so many nets distributed, and why are there so many sprayings, and there is no positive impact?...Research should always be done; otherwise we will not be able to know when an intervention is no longer working." (Researcher, Mozambique)
"... after five years of investment in the Gambia (teams, money, very high cost), when they [the researchers] left, it all fell apart. People need to learn the advantage of nets...because the nets give more work....it is a long-term thing to create a habit. (Policy maker, Mozambique)
Some respondents within the malaria control programme also differentiated between research, and surveillance and outbreak investigation:
" We have very little time for actual research. [...] It's basic problem solving. If there's a small outbreak or a rise in cases in a particular area, you go there and try and evaluate and see what the reason for it is. So you can't study or work in one area for a particular time, which virtually rules out any kind of research." (Government official, South Africa)
While research evidence was regarded as influential, many respondents saw local experience with varied approaches to malaria control as even more important. Thus, the long history of spraying in the region contributed to a preference for its continued use:
"Historical evidence - there was plenty...50 years of spraying in South Africa with large areas free from malaria that previously had been malarious areas." (Researcher, Mozambique)
This contributed to the delayed acceptance of bed nets in Mozambique and Zimbabwe and to scepticism towards this control approach in South Africa.