Barriers and facilitators to sharing of individual patient data: a randomised controlled trial

ID: 

286

Session: 

Poster session 2

Date: 

Monday 17 September 2018 - 12:30 to 14:00

All authors in correct order:

Veroniki AA1, Ashoor H2, Le S2, Rios P2, Stewart L3, Clarke M4, Straus S2, Tricco A2
1 Department of Primary Education, School of Education, University of Ioannina, Greece
2 Li Ka Shing Knowledge Institute of St. Michael’s Hospital, Canada
3 Centre for Reviews and Dissemination, University of York, UK
4 Northern Ireland Hub for Trials Methodology Research, Queen’s University Belfast, UK
Presenting author and contact person

Presenting author:

Areti Angeliki Veroniki

Contact person:

Areti Angeliki Veroniki
Abstract text
Background:
Individual participant data (IPD) meta-analysis is considered the 'gold standard' for exploring the effectiveness of interventions in different subgroups of patients. However, it is time consuming to contact authors of original randomised controlled trials (RCTs) to obtain IPD. To date, there are no studies evaluating strategies to optimise the process for retrieval of IPD from RCTs.

Objectives:
To examine the impact of providing a financial incentive to RCT authors that are eligible for a systematic review versus usual contact strategies to obtain IPD. To describe the potential barriers and facilitators associated with the data sharing process.

Methods:
We used RCTs identified through two systematic reviews and contacted both study authors and sponsors to obtain IPD. We randomised eligible authors to either the intervention or control group using a 1:1 procedure. The intervention group included contacting authors via email, mail and phone, along with a financial incentive. The control group included the same contacting author process with no financial incentive. Primary outcome: the proportion of authors who provided complete IPD. Secondary outcomes: time taken to obtain IPD between initial request and authors' provision, and completeness of IPD received. We conducted descriptive analysis for characteristics abstracted from RCTs or collected through the author/sponsor contacting process.

Results:
Of the total 138 trials, we were unable to locate contact information for eight trials. Of the 130 authors, we were able to obtain 83 (64%) responses (38 (29%) positive and 45 (35%) negative responses). In total, one author provided complete IPD (control group) within 472 days. Of the 138 studies, 107 reported at least one industry-sponsored funder in their publication. We contacted 17 sponsors for 137 studies (83 studies were funded by one sponsor, 23 studies by two sponsors and one study by five sponsors), three (18%) of which did not respond to any of our contacting attempts. To date, we have two complete IPD datasets from a single sponsor. For study sponsors, data sharing agreements were required and the time to clarify the process ranged between 0 and 24 days.

Conclusions:
We encountered important barriers in obtaining study IPD threatening the validity of an IPD meta-analysis. These include study identification, data ownership and limited data access.

Relevance to patients and consumers: 

The joint analysis of data from multiple randomized clinical trials (RCTs) in collaborative efforts may strengthen scientific evidence, and the gold standard approach is pooling individual participant data (IPD). However, sharing IPD has legal, ethical, and logistical constraints, often leading to the conduct of aggregate data (network) meta-analysis that compromises statistical power and accuracy. Since we rely on evidence-based medicine to give us reliable information about the risks and benefits associated with medical interventions, it is of critical importance to make anonymised IPD publicly available.