Science, Finance and Innovation

Making research more useful: experiences in East Africa

7 min


Opiyo Makoude

A five-year donor-funded program that began in 2006 aimed to reduce HIV risks and vulnerabilities for mobile populations within the Lake Victoria Basin in East Africa. This column, written by a manager involved in the regional research program between 2008 and 2011, outlines what happened, and explores some of the general issues around linking research to evidence to policy.

The EAC-AMREF Lake Victoria Partnership was funded by the Swedish International Development Cooperation Agency (Sida) over five years, from 2007 to 2012, with an initial inception phase in 2006 to allow for program design. The program aimed to reduce the mobility-induced HIV risks and vulnerabilities for populations within the Lake Victoria Basin in East Africa.

The Program was a partnership between AMREF (an African NGO) and the East African Community (EAC), the regional economic bloc then comprising Kenya, Tanzania, and Uganda, and which later added Burundi and Rwanda.

The research was not in the original design of the program. The idea was to conduct desk reviews to inform the program design. But we came across a major challenge: the data available were general population-based. There was no data on HIV prevalence among specific categories of mobile populations. Data on the specific behaviors driving HIV risk among these populations were also lacking.

Consultations with the EAC, AMREF, and the donor led to the inevitable conclusion that to have a meaningful regional intervention, we would have to carry out a regional survey of HIV prevalence and map behavioral indicators of HIV risk among selected mobile population categories in the region.

The hurdle was that this type of survey had never before been carried out regionally. We knew what we wanted, but we were unsure of how to go about it. I was then tasked, as the manager, with the duty of exploring how to conduct the study.

The process

I was confronted by one major question: how do we ensure that we carry out a regional study that enjoys the support of key stakeholders, whose involvement would transcend mere words to the actual use of the study’s findings? I consulted with knowledgeable insiders within the EAC system who provided useful insights not only on the politics of policy-making within the EAC secretariat but also on how best to link regional efforts with national-level policy processes.

These insights led to the development of a list of key departments across Kenya, Tanzania, and Uganda, and individuals in these departments whose involvement would be critical. Based on this list, I requested and obtained the support of a senior executive within the EAC who went around with me to these departments.

We held consultations with senior officials and technical staff across the three countries. At the end of these consultations, we compiled a report, which was considered by the AMREF and EAC management, and our suggestion that we constitute technical teams in each country was endorsed. We named these groups the National Technical Teams, mirroring the Technical Working Groups at the EAC and the sectoral working groups at the national level.

The National Technical Teams were drawn from key HIV-AIDS coordinating agencies in each country, the national bureaus of statistics, and the national HIV research laboratories. The National Technical Teams were useful in delimiting the sub-population groups to be studied, based on the available national data, on their knowledge as practitioners. The sub-groups selected were university students, agricultural plantations, and fishing communities.

Two further processes were necessary. The first was getting the buy-in of universities and agricultural plantations. Universities in East Africa comprise public, private, and faith-based institutions. There was initial reluctance by nearly all the universities, but more particularly by the private and faith-based ones: they were clear that they were not interested in the study.

We held a regional meeting for the vice-chancellors and senior administrators of 25 universities in East Africa. This meeting addressed most concerns that the universities had, and we obtained buy-in from them to conduct the study.

Similarly, the agricultural plantations that we approached were all private enterprises and expressly refused to be involved in the study. We visited each of the plantations and held elaborate discussions with their management and clinicians.

The second key process was to have a study design that would meet the rigors of scientific scrutiny while also being ‘politically’ acceptable.

The National Technical Teams adopted an approach that would ‘blind’ the studies by not making public the names of institutions involved by using unique identifiers. This option was acceptable to both the universities and the agricultural plantations.

We also decided to create lower-level technical committees at the university and plantation levels.

Finally, we identified study consultants drawn from leading universities in East Africa, who worked with statisticians from the national bureaus of statistics from each of the three countries to come up with an acceptable sampling design for the study.

The findings – and how they were used

The findings were striking. HIV prevalence was much lower among university students than in the average population. Across the three countries, national HIV prevalence averaged around 6%. By contrast, HIV prevalence across the 18 universities averaged around 0.5%, with the highest HIV prevalence being 2.21% at just one university.

As expected, plantations and fisheries had HIV prevalence higher than the normal population estimates: on average about 8% for plantations and 20% for fisheries.

Based on the findings of the study, several research-policy links were made. From each of the population categories, sexual behaviors tended to be similar. There was evidence of concurrent multiple sexual partnerships among university students, fishing communities, and agricultural plantation workers.

The major difference was that university students had higher levels of consistent condom use compared with the other two groups. Similarly, university students possessed correct information on HIV infection and how it is spread.

Lack of information on HIV and low condom use among plantations and fisheries populations, coupled with multiple concurrent sexual partnerships, explained the difference in HIV prevalence across these groups.

Most universities in East Africa responded by including training on HIV-AIDS as a compulsory and examinable subject in all undergraduate course work. The reasoning was that if students had correct information on HIV, they would adjust their behavior accordingly and reduce their risks of HIV infection and spread. Also, most public and private universities adopted or scaled up their condom distribution policies.

Similarly, agricultural plantations acted on their enhanced understanding of drivers of HIV risk by creating workplace policies that addressed the vulnerabilities identified in the study. But cost considerations meant that the plantations could only intervene at the workplace level, and minimally at the surrounding community levels where the drivers of the epidemic seemed to emanate.

Plantation workers tend to be migrant laborers who have left their spouses behind in rural villages. Working in agricultural plantations gives them disposable income that tends to heighten their vulnerability to sexually risky behaviors. The plantations are unable to address this aspect of their vulnerability, other than by providing information.

National governments used the findings to adjust their responses to these populations, particularly fisheries and agricultural plantations. Both groups were included within a broad category comprising mobile populations, seen by governments as key populations at risk of HIV that require targeted responses.

Policy change tends to be gradual and many of the potential outcomes identified at the program design stage could not be realized within the project’s lifespan. But the findings set in motion many policy discussions and HIV programming, which have kept the spotlight on agricultural plantation workers and fishing communities in East Africa.

The EAC and its various institutional organs have used the findings to structure their response to HIV-AIDS. The study also contributed to the development of regional policy and legal frameworks that address the drivers of cross-border HIV risk and vulnerability: the EAC HIV and AIDS Prevention and Management Act, 2012.

What made the research-policy link successful?

Broad-based consultation and collaboration

Users of research have interests, fears, and needs that must be addressed for research to be meaningful. Each of the parties involved in the studies – the EAC, national governments, universities, agricultural plantations, ministries of health, and fisheries departments – were implementing HIV-AIDS interventions that were not informed by sound evidence, especially concerning these population categories.

Making the research and its findings less threatening, especially for those who expressed initial rejection or reluctance towards the research, was critical. Consultations enabled fears and reservations to be addressed. Giving specific universities and agricultural plantations findings specific to their institutions worked as a good strategy, and they used the findings.

Working with knowledgeable insiders in the policy process

Knowledgeable insiders often provide invaluable insights into the politics of policy-making, and how the policy process works. The initial step is to identify such individuals, to get their buy-in and commitment, and to keep them informed and constructively involved in the entire research process, from design to execution to implementation.

In this case, working with senior officials at the EAC and later the National Technical Teams proved invaluable.

Scientific rigor

While working on politics, the scientific rigor in terms of study validity and reliability must be kept under constant check. A badly designed study or one that does not meet the rigors of good research is bad for everyone.



Opiyo Makoude
Evaluation researcher and policy analyst