The first thing to note is that, although we identified over 300 papers on the use of research in health policymaking, the vast majority of these were descriptive

The first thing to note is that, although we identified over 300 papers on the use of research in health policymaking, the vast majority of these were descriptive. Very few – in fact just 14 of 304 articles – actually concerned testing interventions to see whether they worked. There is a serious discrepancy, therefore, between surging interest in this area and the small number of studies actually testing strategies.
The 14 articles we did find (reporting on 13 intervention strategies) tended to be methodologically weak. Only one study used an experimental design, while one other used a pre/post-test design. The others used a range of approaches and were characterised by an absence of control groups, small sample sizes, and self-report data. Most measured outcomes related to factors that influence research use rather than actual research use. — http://blogs.lse.ac.uk/impactofsocialsciences/2018/11/12/less-than-5-of-papers-on-the-use-of-research-in-health-policymaking-tested-interventions-to-see-what-worked-but-those-studies-reveal-a-number-of-strategies-for-improvement/