医学科学研究的牵引力和驱动力建模: 健康需求与未知探索

At present, medical research has become one of the most active fields of science and technology innovation in the world, and our country attaches great importance to the scientific and technological innovation oriented to health. The advancement of medical science and technology is propelled by two factors: Scientific and technological innovation and management. In the meanwhile, two forces—Free exploration and demand-oriented innovation—Drive scientific and technical advancement. This article proposed a “three-force” model of medical science and technology innovation, namely pulling, pushing and protecting forces. The protecting force of science and technology management places a strong emphasis on resource allocation (e.g., strategic planning for diseases with high burden and scientific conundrum) and risk management (e.g., ethics governance). The pushing and pulling forces could contribute to a better understanding of innovation process. Unmet health needs, the pulling force, focus more on the health-research priorities for lowering burden of diseases. In particular, unmet health need is also the fundamental motivation for the development of diagnostics, medications, vaccinations, and other health-related treatments. Exploration of the unknown, the pushing power, addresses the existing scientific problems and explores the scientific unknowns. This paper divided knowledge states into four categories, which are known knowns, known unknowns, unknown knowns and unknown unknowns. Through the interpretation of four knowledge states, this paper provides a thorough discussion on how scientific unknowns drive scientific and technological innovation. The approach of big-data analysis provides the methodological basis for further structural optimization of the research landscape based on the pulling and pushing power model. Hence, on the basis of constructing the “three-force” model from the theoretical conceptual level, this paper further proposes that the pulling and pushing force model can be built in a quantitative approach with big data analysis. To be specific, three research objects, namely, health needs, scientific problems and science and technology innovation, could be quantified using various indicators and methods. In terms of measuring health needs, morbidity, mortality, and disability adjusted life years (DALYs) are the widely used and universal quantitative indicators, which can be applied in the “three-force” model. As for the identification of scientific unknowns, we argue that using big data analysis and computational linguistics to identify research topics that are “controversial, conflicting, unverified, unsolved” can reflect scientific problems that are oriented to the scientific frontier and focus on major needs and challenges. For example, terms like “unknown, suspect, unclear, unusual, controversial, consensus, incomplete, conflicting, contrary, debatable, etc”. can be employed as trigger words, and sentences containing such triggers words can be extracted to represent medical claim with uncertainty. In terms of disease mapping, natural language processing tools (such as Medical Text Indexer) can be used to extract Medical Subject Headings (MeSH) involved in various scientific research texts to represent the main content. With the construction of “MeSH-Diseases” concordance table, the mapping of multi-types of scientific research data and disease can be realized. Such research could assist scientific policy in the identification of major scientific issues, the allocation of scientific resources, and the encouragement of coordinated, multidisciplinary research.