(145) Co-Producing a Pharmacoepidemiology Online Knowledge Hub

Publication Number: 476

Background: The use of routinely collected health and prescriptions data for pharmacoepidemiology is a growing domain, which can overcome some of the challenges of feasibility and generalisability faced by randomised-controlled trials. However, there is a lack of consistency between researchers in the use of codelists to identify cases, interventions and outcomes, data management and analytic methods, which can lead to inconsistent findings.

Objectives: To address issues of consistency, transparency and reproducibility in the use of routine data for pharmacoepidemiology, we are co-producing an online knowledge hub. The purpose of the hub will be to promote open science, share learning and facilitate collaborations. The hub will be the first of its kind in the pharmacoepidemiology domain. In the hub's first phase of development, described here, we aimed to consult multidisciplinary pharmacoepidemiology researchers on the following objectives:
i - Is there demand for an online knowledge hub?
ii - What features would the online knowledge hub need to have to be of use to the research community?

Methods: Over the space of 12 months, we promoted a 3 online events via the UCL e-Research domain, with attendees from 12 universities across 5 countries. The events consisted of seminars and problem solving workshops. In the first two events, we surveyed 120 delegates to establish whether there was demand for an online knowledge hub and to identify features that would be of use to the research community. In the third event, we held a workshop to build on the features for the online hub identified in the surveys, and to co-produce a blueprint for the online pages.

Results: 100% of event delegates surveyed agreed there would be demand for the online knowledge hub.
Three main themes for proposed features of the online knowledge hub were agreed upon during the workshop:
1) An open science platform to share and review codelists/data management coding templates/statistical analysis code, used in pharmacoepidemiology studies;
2) A resource library with information on data sources, training on data management and analysis methods, and papers show casing examples of relevant methods;
3) An online collaboration platform for methodological support and networking
General additional requirements also included:
- Ability for community to input, but with expert curation and rules for appropriate engagement
- Signposting to existing resources

Conclusions: An online knowledge hub for promoting open science, sharing learning, and encouraging collaborations would be a valuable resource for the use of routinely collected data in pharmacoepidemiology. The Pharmaco-Epi Data Collaborative has begun development of the proposed online pages, that will be community led and expert curated.