Editor's choice: February 2021

There has never been a time when evidence was more important or under such intense scrutiny. A every key study is eagerly awaited and devoured on arrival as though the health hopes of global community rest its shoulders. And yet, we know, frequently flawed, equivocal in findings open to interpretation, even it forms critical element most guidelines decision support tools. Over recent years concept continuous eco-system that comprises different but mutually dependent processes, broad range actors stakeholders, increased exploitation systems technology developed [[1]Vandvik P.O. Brandt L. Future ecosystem series: ecosystems learning systems: why bother?.J Clin Epidemiol. 2020; 123: 166-170Abstract Full Text PDF PubMed Scopus (10) Google Scholar]. In this month's issue JCE present articles address challenges have identified at three points eco-system, namely primary research, synthesis translation into [2Boutron I. Créquit P. Williams H. Meerpohl J. Craig J.C. Ravaud 1. Introduction — needs dramatic change.J 135-142Abstract Scholar, 3Nguyen T.-L. Xie Incomparability treatment groups often blindly ignored randomised controlled trials – post hoc analysis baseline characteristics tables.J 130: 161-168Abstract (2) 4Ghannad M. Yang B. Leeflang Aldcroft A. Bossuyt P.M. Schroter S. et al.A randomized trial an editorial intervention reduce spin abstract’s conclusion manuscripts showed no significant effect.J 69-77Abstract (5) 5Piggott T. Baldeh Akl E. Junek Wiercioch W. Schneider R. al.Supporting effective participation development groups: guideline tool.J 42-48Abstract (4) 6Piggott Brozek Nowak Dietl Saz-Parkinson Z. al.Using GRADE frameworks choose from multiple interventions.J 117-124Abstract (3) Randomised continue represent best research tools evaluate effects healthcare interventions. However, Nguyen show, randomisation processes are adequate, substantial proportion studies (about third their study) include least one variable might, by chance, be substantially unbalanced across [3]. The authors report clinical reports fail describe assess measures comparability. Even where tests for comparability undertaken, they flawed. They potential solution uses measurement standard mean difference between difference. This technique enables initially expressed units viewed similar scale. not simply dry methodological issue: context COVID-19 pandemic, source concern those tasked with interpreting data. can span usual parameters age, sex, setting, stage disease co-morbidity, also vexed co-interventions, unknown given purely compassionate grounds. perfectly conducted risk bias scrupulously minimised undone reporting does fairly accurately reflect findings. Spin, ‘misrepresentation, overinterpretation inappropriate extrapolation results’, less well studied than bias, frequency importance both should underestimated. Ghannad colleagues undertaken what believe may first RCT aimed reducing [4]. identify four elements spin: selective reporting, information supported evidence, interpretation inconsistent design results, recommendations practice justified study's implemented simple compared practice. Perhaps unsurprisingly, none results demonstrated statistically benefit, although confidence intervals were wide, reflecting uncertainty rather effect. Thus, being, task policing will sit teams. We hope further follow, perhaps extending far press releases studies, which currently assuming unwarranted our discipline. At end process work organisations factor current alongside other issues develop guidelines. Contributing part group heavy responsibility, so seems surprising until paper Piggott included issue, there little guidance members [5]. describes mixed methods iterative tool. international researchers who extensive experience Grading Recommendations, Assessment, Development Evaluation (GRADE) working group. tool helpful all participants, particularly limited whom must seem challenging. Importantly, incorporated training Guidelines International Network. result better prepared members, greater inclusiveness engagement meetings improved outputs. his second addresses emerging relation application comparison analyses Evidence Decision (EtD) framework, use [6]. Network meta-analysis become preferred approach ranking options, utilising does, indirect direct evidence. describe, only dimensions benefits harms, whereas EtD framework typically incorporates certainty, values, resources, equity, acceptability feasibility. presented here represents output distinct scenarios, includes incorporation module based developments Pro application. structured supports full transparency making comparisons. begins evaluation each relevant pairwise comparisons, assessment transitivity interventions being studied. facilitates criteria, used inform recommendations. here, others demonstrate how addressing undermine eco-system. By empirical solutions presented, policy care populations move forward.