Table 2 Summary of study findings

Research evidence

High quality evidence from RCTs that supported the effectiveness of MgSO₄ as first line treatment.

Head-to-head comparisons in RCTs did not find an important difference in effectiveness.

- Timeliness, perceived relevance and quality of the research

- Trust in the research and researchers

- Availability of research summaries with clear recommendations

- Use of jargon and only publishing for a scientific audience

- The perceived relevance and quality of the research on MgSO₄ was high, whereas the evidence for bed nets was not perceived to be relevant. More weight was given to local experience than to evidence from RCTs for malaria control.

- Availability and timeliness of the research were not identified as important for either case.

Involvement of local researchers in evidence production

Leading obstetricians in all three countries were involved in the trials.

Researchers in South Africa involved in one comparative trial.

Researchers in Mozambique were involved in bed net research.

This was not identified as an important factor in either of the reviews cited above

Involvement in trials may be an important influence directly or through other routes. These other routes include champions, international networks, prior experience and the promotion of a culture of evidence-based health care.

Prior clinical or public health experience

Prior experience in the use of specific drugs for eclampsia may explain, in part, differences in policy between the three countries.

Prior experience heavily influenced support for spraying and inhibited policies favouring the use of bed nets.

Policy confirmatory research

The extent to which the research confirmed or challenged prior experience was important for both cases.

Research and policy champions

Obstetrician researchers championed MgSO₄ in all three countries, but in Zimbabwe the key champion emigrated prior to development of a policy supporting MgSO₄ as the first line drug.

Researchers regionally, particularly in South Africa, championed spraying, whereas researchers in Mozambique championed bed nets.

Neither review specifically identified the role of champions in promoting the use of research evidence, although both found interactions and trust between policymakers and researchers to be the most commonly identified factor supporting research use.

In both cases senior researchers actively advocated for specific policies. In the malaria case, researchers in South Africa and Zimbabwe advocated spraying rather than bed nets. In Mozambique, they advocated nets.

International networks, organisations and other stakeholders

International networks that influenced MgSO4 policy were largely evidence-based, such as the Cochrane Collaboration.

A wide range of stakeholders and international organizations with differing interests, including bilateral donors, (e.g. DFID, JICA), and multilateral agencies, (e.g. WHO, UNICEF and the Roll Back Malaria Partnership), influenced malaria policy in Mozambique and Zimbabwe. This was partially due to donor dependence.

Neither review identified stakeholder interests or international networks as being important factors that affect the use of research, although community pressure was identified as an important factor.

It is likely that international organizations play a more important role in policy development in many LMICs compared to high-income countries, due to donor dependence.

Regional networks of policy makers and researchers

Regional networks of policy makers and researchers did not emerge as an important factor.

Regional networks of policy makers and researchers emerged as a key factor influencing malaria policy.

Existence of policy networks

This difference between the two cases may reflect differences between public health policies, particularly for vector-borne diseases that cross borders, and policies for clinical interventions. In addition, powerful champions promoted regional networks for malaria policy and control.

Involvement of researchers in policy making

Researchers played an important role in policy development in all three countries.

Researchers played an important role in policy development in all three countries.

Interactions and trust between policymakers and researchers.

There was extensive interaction between researchers and policy makers in both cases, and researchers also moved between the research and policy environments. This level of interchange may be more common in LMICs than in high- income countries.

Culture of evidence-based health care within specific health domains

This emerged as an important factor supporting the uptake of research findings for MgSO⁴, due to the strong culture of evidence-based health care in obstetrics.

The culture in relationship to evidence varied for malaria, with greater emphasis on local observational evidence. Differences in malaria epidemiology contributed to this emphasis.

Skill and attitudes of those receiving the research

The greater focus on local conditions and evidence may be more typical for public health and communicable diseases management particularly.

Political and bureaucratic processes

Bureaucratic processes can in part explain the failure to include MgSO₄ in the national formulary in Mozambique even though it was recommended [30].

Political processes at national, regional and international levels may have contributed to the continuation of policies that failed to promote the use of bed nets.

- Bureaucratic process including power and budget struggles and conflicts

- Management support

Bureaucratic processes emerged as being potentially important for MgSO₄, whereas political processes appeared more important for malaria policies.

Events within the wider political environment

Although this did not emerge as an important factor, political and economic instability may have influenced policy in Zimbabwe.

Political and economic changes influenced policy in several ways: through South Africa becoming influential in regional politics; through lobbying by interest groups; and with regard to ideological and political perceptions of spraying and bed nets.

The political environment including political stability and community pressure.

External political events can be a limiting factor, but are perhaps more important in public health than in clinical medicine.