Capacity-building for health research in developing countries: No quick fix, but efforts could be boosted by greater efficiency

Ironically, capacity-building efforts were of most benefit to industrialised countries, as scientists followed the money and glamour that is associated with cutting-edge research in richer countries. The United Nations Education, Science and Cultural Organization (UNESCO) reports, for instance, that more African PhD graduates now live outside of Africa (about 30 000) than on the continent 2 . In hindsight, the limited success of capacity-building efforts could have been predicted. These initiatives tried to harness the attributes of science that typically drive research and development (R&D), namely:

• increasing scientific and economic rewards when R&D production is scaled up 
• high rates of return to society for investments in R&D and
• a production process driven by demand for product applications.

Not surprisingly, attempts to scale up R&D in less developed countries (LDCs) by building a critical mass of researchers helped strengthen the research enterprise in established market economies. The benefit of investments by LDCs often failed to materialise. And, increasingly, the end products that drove R&D production directed energies away from the health problems of the poor 3 . In response, organisations such as the United Nations Development Programme (UNDP) have rejected the market-driven paradigm implicit in early efforts, and integrated capacity-building designs into a framework for sustainable human development4 . Underpinning this approach is the view that advances in human development are achieved by maximising each persons capability. 

This shift in thinking has strengthened capacity building and further clarified its objectives equity, for example, has been pushed to the forefront.

Our view is that todays efforts at capacity building for health research could be boosted by a more efficient design - drawing on the attributes of science that shaped earlier strategies, but this time around ensuring that less developed countries are the principal beneficiaries. We propose three components of a more efficient design.

First, as a counterbalance to a market-driven lock-in phenomenon, we propose the nurturing and support of multi-stakeholder problem-oriented learning and research networks.

Second, investments in research should explicitly reduce the high costs associated with sharing knowledge and its application in poorer countries. And third, we propose a dramatic shift to the stimulation of demand for new knowledge, while continuing to improve supply.

Strategy one: Support national leaders to forge research and learning networks

The first typical attribute of science that stimulates R&D is increasing scientific (and economic) reward, when the production of new knowledge is scaled up. Unlike many industries, adding more capital and scientists to a research venture tends to increase productivity hence the success of R&D powerhouses like Silicon Valley.

Capacity-building efforts that try to create a critical mass of researchers, build institutional strength and centres of excellence are all attempts to capture the benefits of increasing returns which are proportional to investment. A downside of this characteristic of science is that researchers tend to cluster around particular topics and ignore others. This clustering behavior is reinforced by the reward systems of science, that leads too many researchers to enter research races with particularly lucrative prices, and too few to enter little known, but socially beneficial research endeavours. These rewards include the promotion process, peer recognition, competitive funding, and direct personal financial gain from the commercialisation of research. Lock-in favours the rich. Not only is there intellectual flight to wealthier countries, but even the work of researchers in less developed countries is shaped largely by the dictates of the international market.

So what would a new critical mass look like, that would resist the tendency to cluster massive research efforts on the health problems of the rich? In essence, problem-focused research and learning networks constitute the new critical mass, drawing on expertise and knowledge from diverse sources to improve the health of a countrys people. These networks involve recognition and support for multiple sites of knowledge-generation (many of which are outside of traditional institutions likeuniversities), enabling these sites to communicate effectively with one another, and designing initiatives around problems, not institutions 5 .

In our view, enabling research leaders in less developed countries to cultivate new ways of knowledge production and sharing is the key to freeing up global research, so that it responds more to the health needs of the poor. In practice, this means enabling national research leaders to develop strategic responses to priority health problems. 

The substance of country-focused research and learning initiatives would be shaped by national priority setting processes, broadly representative of a wide range of interests. Priority setting can be a powerful way of improving efficiency of allocation by better revealing aggregate levels of social demand, although there is a risk that one cause of inefficiency (lock-in), will be replaced by another usually political interests. And care needs to be taken to preserve the incentive and reward structures that are intrinsic to science. However, country experiences supported by the Council on Health Research for Development (COHRED) suggest that greater representation can be incorporated into priority setting designs without antagonising the research community and jeopardising long-run scientific endeavour 6 . Priority setting in Uganda, for example, identified tuberculosis as a major concern for both community groups, programme managers and researchers. The next step was to establish an inventory of tuberculosis research in the country, as the basis for concerted interdisciplinary action 7 .

Strategy two: Reduce the costs of communication
The second typical attribute of science that can lead to greater country support of R&D, is high social benefit from substantial investment. These benefits are reflected either in growth of GNP or in better health outcomes. Less developed countries are often hard pressed to demonstrate positive returns to society on investments in R&D. However, even those benefits that can be anticipated are eroded unnecessarily by higher transaction costs in less developed countries. Dasgupta and David (1994) assert that the major cost associated with research is incurred in making information available to users, who are either other researchers or the public 8 . They argue that the relative costs and benefits of information disclosure lie at the very root of knowledge diffusion.

Where the costs associated with making information public are high,knowledge will remain restricted to the world of researchers. Where researchers stand to gain more from sharing their results with the commercial enterprise than social programmes, the benefits of research will tend to be captured as profits by private industry. In poorer countries, the relative benefit to researchers who sell their findings privately, as opposed to public disclosure, is even greater than in wealthier ones. And the relative costs associated with sharing information with both fellow researchers and the public are higher. Perhaps non-economists would prefer to think of these as interaction costs. Obvious causes of higher interaction costs in poorer countries are the limited communication infrastructure and poor access to international discourse. In this regard, the growing commercialisation of health research is pushing up the costs of acquiring knowledge a situation aggravated by the Trade-Related Aspects of Intellectual Property Rights (TRIPS).

This international agreement, enforced by the World Trade Organization, compels poor countries already strapped for foreign currency, to pay global market rates for new information. Less explicit are the high political costs associated with public disclosure of research results in some countries. Not only does undue political
resistance to public disclosure push up the costs of individual interactions, but it reates feedback that discourages future researchers from testing prevailing assumptions, and dampens the long-term effectiveness of R&D in that country.

Strategies aimed at reducing interaction costs associated with knowledge diffusion will need to focus largely on facilitating communication. Facilitating access to new computer hardware and software seems to be a fundamental first step. Related to this is securing electronic access for researchers to international journals and other information resources. As access to resources is enhanced, so the linkages among researchers and between researchers and potential user groups should be developed. In many instances, national leaders will need to encourage researchers to make connections with colleagues outside of their traditional typically intra-disciplinary networks, and user groups that they may never have dreamt of approaching! An expanding function-oriented web of connections (hyperarchy) will put researchers in regular contact with the media, advocacy groups, legislators, international agencies, the private sector, and public health officials.

In time, these linkages will go far beyond e-mail contact. For example, EQUINET is a Harare-based southern African initiative to develop a regional research programme that promotes equity in health 9 . This collaboration lowers the barriers to information and makes best use of limited human resources in the region. In a move that helps reduce the interaction costs of policy-related research, the South African Health Systems Trust shares employment of a journalist with a national newspaper. Research findings get written and published in plain English, and the newspaper gets easy access to the latest studies. With respect to high political costs, international agencies may have a particular responsibility in helping to overcome them. For instance, up-front negotiations with health ministries can minimise the backlash against researchers whose evaluations find unfavourable outcomes of government programs. In sum, many opportunities exist to reduce the high interaction costs that still pertain to researcher-researcher and researcher-user linkages in developing countries.

Reducing these costs may prove to be a powerful way of boosting the efficiency of existing capacity building efforts. Strategy three: Focus on demand-side capacity development.

The third attribute of science listed in our introduction is that its production process is driven by demand for new applications. Market-driven incentives provide much of the impetus for innovation, and whether intentional or not - an implicit assumption of supply-side capacity-building strategies, like basic research training, is that the market will do the rest. However, the demand for research that is expected to meet an enhanced supply of new knowledge often fails to materialise. Public officials, the media, industry, community groups and other potential users rarely seize the opportunities to capitalise on available new knowledge. This weak demand is reflected in low national investments in R&D, low salaries for researchers, and limited use of research findings.

Explicitly stimulating a demand for research may sometimes mean that user groups, as opposed to researchers, are the primary targets for capacity-building efforts. For example, national legislators in South Africa helped design a country wide survey of health facilities that will help monitor progress towards equity in service provision over time 10 . Not only has this enhanced the legitimacy of the studies, but has enabled legislators to see how research can help answer the questions that they choose to ask. The Tanzania Essential Health Intervention project (TEHIP) has clearly demonstrated that district development committees can learn to do dramatic evidence-based planning based on alarming increases in malaria fatalities, the district budget for malaria was substantially enhanced 11 . Similarly, media editors have participated in seminars at rural research sites in South Africa to explore ways of improving the coverage of issues related to the health of the rural poor.

These seminars have established continuing contact with researchers and journalists on a health or development beat. Promoting links with private industry as a way of strengthening market demand will become increasingly important for developing countries over time. However, national ministries of health, and science and technology, will need to provide clear policy and direction to publicly funded research efforts so that growing demand by the private sector does not deviate resources from the national prioritised research agenda.

Conclusion
Implementing these strategies requires a new mindset. Success will be measured not by the number of scientists trained, or the volume of academic publications, but rather by the ability of countries to apply global and country specific knowledge to local problems. In turn, country-specific health research will contribute to global knowledge through innovation, adaptation and occasionally, new discovery.