Date Published: August 28, 2007
Publisher: Public Library of Science
Author(s): Mark Tomlinson, Mickey Chopra, David Sanders, Debbie Bradshaw, Michael Hendricks, David Greenfield, Robert E Black, Shams El Arifeen, Igor Rudan
Abstract: Nearly 100,000 children under 5 years die annually in South Africa. This paper defines health research priorities to address this unacceptably high mortality rate.
Partial Text: Research can play a critical role in the response to global health challenges. But when needed to assist decisions on defining the priorities for health research investments. An early attempt at the global level to define health research priorities was made through the Commission on Health Research for Development in 1990 . A number of subsequent initiatives addressed this problem by attempting to set priorities in global health research [2–4]. However, these approaches promoted funding for research predominantly focused upon the generation of new technologies, knowledge, and processes. Research concerned with implementation of already proven technologies and interventions was downplayed .
Activities of the technical working group. The rationale for CHNRI methodology and its conceptual framework, application guidelines, and strategies to address the needs of the stakeholders have all been described in detail elsewhere [5,11]. Box 1 presents the elements of the methodology at a glance. In the first step, a group of six leading South African technical experts in the area of child health formed a technical working group (TWG). All are involved in national policy development and have experience working in both rural and urban settings across South Africa. Four of the experts are child health specialists with three currently engaged mostly in clinical practice, one is a clinical child psychologist, and the sixth is a medical demographer. Only two of the six have qualifications or expertise in public health. Because of valid concerns that the choice of TWG members may significantly affect the outcomes of the process, we discuss how this issue is addressed in the CHNRI methodology under the “limitations” section below.
The final results of the scoring process of the technical experts (top 20 and bottom 10) are shown in Table 1. In this table, the scored research options are ranked by their final RPS multiplied by 100, which gives a range of score values between 0 and 100. This score takes into account the scores from technical experts, based on five criteria relevant to priority setting, and the weights defined by the LRG. The ranks in parenthesis indicate RPSs before weighting by the LRG. The final RPSs for the 63 research options ranged from 88.6 to 45.6, with the lowest score of 30.8 being an outlier. This range shows substantial variation between the research options in their likelihood to address the five criteria, as assessed by the TWG and LRG, and indicates that the methodology has the power to discriminate among many competing research options using a simple conceptual framework with 15 questions.
The results of this prioritization exercise suggest that child health research funding in South Africa should concentrate on HPSR options, especially those related to diarrhoea, pneumonia, and malnutrition. Our results are in line with the findings from Anthony Costello and colleagues’ recent priority-setting exercise with international experts that used the Delphi method . The results probably reflect the fact that, although pneumonia and diarrhoeal disease represent 38% of the global burden of disease in children, only an estimated 0.2% of the total funding spent on research and development is allocated to these conditions . Furthermore, a recent investigation found that 97% of research grants from the largest not-for-profit sources of funds for health research (the US National Institutes of Health and the Bill and Melinda Gates Foundation) were for developing new technologies, which could reduce child mortality by 22%, a reduction one-third of what could be achieved if existing technologies were fully realized .
We do not believe the predominance of HPSR is due to a bias of methodology. In the context of the high remaining burden of child mortality in South Africa  and the presence of cost-effective interventions and sufficient available resources to implement them, we expected that the methodology should highlight the issues of improved delivery and increased coverage of those interventions as an immediate priority. Furthermore, among the 16 research options at the bottom of the list of rankings, four of them (25%) are HPSR options (including the outlier at the bottom). The results may still be biased by the choice of the technical experts and LRG. However, even if there were such a bias, the methodology transparently presents the input from every expert on every criterion, so the input would be reviewed and challenged by other experts, and the feedback loop within the methodology would take those changes into account. This process of review also helps to identify the points of widespread agreement and the points of controversy, thereby directing and focusing the discussion on research prioritization on the key issues.
The main advantages of the CHNRI methodology presented in this article over the alternative approaches can be summarized as: (1) being systematic in listing and scoring competing research options, thus limiting the influence of personal biases on the outcome; (2) being transparent in regard to the input and contribution of everyone involved in deciding the research priorities; (3) preventing one or a few individuals from dominating the process; (4) presenting a simple quantitative outcome (RPS); (5) simultaneously evaluating different domains of health research using the same set of criteria; and (6) incorporating the opinion of stakeholders and wider public. The methodology proved to be a feasible and transparent approach toward setting priorities in child health research investments in South Africa.