Characteristics of randomized trials that have been extended by linkage to routinely collected data: results of a scoping review of the published literature

Session: 

Oral session: Innovative solutions to challenges of evidence production (4)

Date: 

Tuesday 18 September 2018 - 14:40 to 14:50

Location: 

Carrick 1

All authors in correct order:

Fitzpatrick T1, Henry D2, Perrier L3, Cairncross Z1, Shakik S1, Tricco A4, Straus S4
1 Dalla Lana School of Public Health, University of Toronto, Canada
2 Centre for Research in Evidence Based Practice, Bond University, Australia
3 University of Toronto Libraries, Canada
4 Li Ka Shing Knowledge Institute, St Michael's Hospital, Toronto, Canada
Presenting author and contact person

Presenting author:

David Henry

Contact person:

David Henry
Abstract text
Background: The value of randomised controlled trials (RCTs) can be enhanced by independent re-analyses and inclusion in meta-analyses. However, these activities are limited to the data collected during the original trial. An additional approach is to extend trials beyond their planned follow-up by linkage of participant information to routinely collected data held in health administrative records, registries and electronic medical records. There has been no systematic attempt to characterize this potentially important literature.
Objectives: To perform a scoping review to characterize the types of extended trials, the clinical topics studied and the value generated by extended follow-up.
Methods: We published the protocol and performed a literature search using Cochrane Register of Controlled Trials, Embase, PubMed, and CINAHL. We tested the sensitivity and specificity of various combinations of search terms, as there is no generally accepted terminology in this field. We performed determination of study eligibility and data extraction in duplicate.
Results: We retrieved 2811 unique abstracts (after removal of duplicates), and reviewed 309 of these in full text form; 113 extended trials met our inclusion criteria and we retrieved reports of the original trials. Half of the trials were conducted in Scandinavia and 58% of the trial extensions were published after 2010. Record linkage extended follow-up by between one and 50 years. Interventions included drugs or vaccines (45%), surgery (17%) and screening (16%). Commonly studied outcomes were mortality, cancer (recurrence or second tumors) and cardiovascular events. Most extended outcomes were reported according to the original randomization and in 29 instances (25%) the long-term intervention effects differed from those reported in the original trial. Comprehensive information on post-trial treatments was rarely reported and not accounted for in most of the statistical analyses.
Conclusions: The number of reports was small in comparison with the total number of trials that have been conducted. Post-trial extension by linkage to routinely collected data appears to be a valuable but underused mechanism for extending participant follow-up in RCTs. Guidelines are needed for the analysis and reporting of post-trial extensions with agreement on terminology.
Patient/healthcare consumer involvement: None.

Relevance to patients and consumers: 

If routinely collected (administrative) health data can be linked regularly to clinical record data from participants in randomized trials this will enable long-term followup, enhance the value created by their participation and reduce the degree of intrusiveness of follow-up data collection.