class: center, middle, inverse, title-slide # Reporting Clinical Trial Registration ## Responsible Metrics ### Maia Salholz-Hillel ### QUEST, BIH, Charité ### 2020-06-23 (last updated: 2020-06-23) --- class: section-header <style type="text/css"> .xaringan-extra-logo { width: 10em; height: 10em; z-index: 0; background-image: url('https://www.bihealth.org/typo3conf/ext/dreipc_bih/Resources/Public/Frontend/assets/Image/logo_bihealth_en.png'); background-size: contain; background-repeat: no-repeat; position: absolute; bottom:-6em;left:2em; } </style> # Background --- # From Registration to Reporting .pull-left[ ### **Registration** of clinical trials (CTs) is necessary but not sufficient for CT transparency and findability   ] -- .pull-right[ ### Clinical trial registration numbers (TRNs) should be **reported** with associated publications and meta-data   ] --- # .small[Recommended by CONSORT and ICMJE Guidelines] > In an abstract reporting a trial, authors should provide details of the trial registration number and name of trial register. Hopewell, S., Clarke, M., Moher, D., Wager, E., Middleton, P., … Altman, D. G. (2008). CONSORT for Reporting Randomized Controlled Trials in Journal and Conference Abstracts: Explanation and Elaboration. PLoS Medicine, 5(1), e20. doi:10.1371/journal.pmed.0050020 > The ICMJE recommends that journals publish the trial registration number at the end of the abstract. International Committee of Medical Journal Editors (ICMJE). (2019). Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. http://www.icmje.org/icmje-recommendations.pdf --- # Where should the TRN be provided? .left[**Abstract**] .left[] -- .center[**Secondary Identifier**] .center[**(Meta-Data)**] .center[] -- .right[**Full Text**] .right[] --- # Research Aims - Evalutate the **prevalence and characteristics of TRN reporting** for German UMCs (and globally) - Evaluate which **registries** are reported, and co-reported (i.e., cross-registration) - Evaluate **sources (meta-data, abstract, full-text)** of TRN reporting # Applied Aims - Develop software (R) and a tool (ShinyApp) which researchers, universities, and other stakeholders can use to evaluate their TRN reporting practices --- class: section-header # Pilot --- background-image: url(images/2020-06-23_2018_inclusion-criteria.png) background-position: center background-size: contain # Record Selection --- # Preliminary Manual Check - Random sample of 50 with TRN detected and 50 with no TRN detected - 100% agreement with manual check by blind coder (thanks, Martin!) - More rigorous validity checks to follow --- class: section-header # Preliminary Findings --- # What proportion of clinical trials report trial registrations, and where? Of 607 PubMed Clinical Trial records for 3 UMCs (Charité, Hannover, LMU) in 2018: -- - 226 (37%) reported one (n = 215) or two (n = 11) TRNs in the PubMed Secondary Identifiers field. - 212 (35%) reported one (n = 179), two (n = 30), or three or more (n = 3) TRNs in the abstract. - 249 (41%) reported one (n = 213), two (n = 32), or three or more (n = 4) TRNs in either the Secondary Identifiers field or the abstract. --- class: hide-logo # Which registries are reported in the Secondary Identifiers field? .center[] --- class: hide-logo # Which registries are reported in abstracts? .center[] --- class: hide-logo # Which registries are reported in either Secondary Identifiers field or abstract? .center[] --- # Next steps ## Classifier - Analyse full-text (XML and PDF) - Evaluate whether TRNs resolve in registry - Train classifier to predict incorrect TRNs ## Analysis - Explore covariates of registration reporting practices (e.g., ICMJE Journals, country, etc.) - Evaluate TRN errors (e.g., typos, different CT, etc.) --- class: section-header # Thank you!