Idea Combinations in Biomedical Science and Clinical Impact: The case of Alzheimer’s Disease Research

Scholars of science and biomedical researchers both argue that combining ideas from different research streams and different disciplines generates a deeper understanding of human disease, produces breakthroughs in preventing and treating illness, and more rapidly promotes health. This paper examines whether scientific ideas in combination are more likely to lead to therapeutic advances and whether advances in therapy are more common compared to therapeutic advances that follow from a single stream of science, using Alzheimer’s disease (AD) research as a case study. Because Alzheimer’s risk factors are complex and the disease itself is heterogeneous, many researchers in this field especially argue that multiple approaches tried in combination are necessary.

We use the text of published clinical guidelines and systematic reviews in AD research to identify clinically important ideas. Clinical guidelines directly inform physicians of best practices in prevention and therapy based on the scientific literature. Clinical guidelines are often based on systematic reviews, which synthesize and analyze results from the scientific research, summarizing the literature in a form that can be incorporated into actionable guidelines or even directly inform physicians about the efficacy of a certain therapeutic interventions. We use text-analysis to identify the important concepts in the AD literature that end up in systematic reviews and guidelines. We ask if combinations of ideas—co-occurrences of concepts—from different scientific disciplines are more likely to appear as clinically important—mentioned in clinical guidelines and systematic reviews--than important scientific ideas that arise from a single discipline or line of research.

This paper addresses three sets of questions. First, how do co-occurrences of concepts compare to other important concepts in the AD literature in terms of scientific impact (as measured by forward citations, for example) and clinical impact (as measured by mentions in systematic reviews and clinical guidelines)? Second, what is the path to translation for co-occurring concepts compared to important concepts? Does it take longer or less time for co-occurring concepts to produce a clinical impact? Do combinations of concepts produce broader scientific and clinical impact? Third, what are the conditions associated with the interaction of ideas? How do co-occurring concepts and their impact vary by whether they arise from within or outside the AD field? How does this production function compare to the production function of scientifically important stand-alone concepts? For example, do diverse teams generate more important new pairings and more narrowly-representative teams produce more scientifically important concepts? Are the boundaries of gender, race, and ethnicity and age important for the emergence of new combinations?