2.2 Identify relevant trials

Page last updated: September 2016

2.2.1   Search results

Use a PRISMA flowchart (Figure A3.1, Appendix 3) to present the study selection process for each search.^1,2

Exclude studies on the following bases:

A     not a randomised trial (not relevant when the search is repeated to find nonrandomised studies)

B     incorrect intervention (such as when the intervention is used in combination with another medicine that is outside the use described in the requested restriction)

C     does not include the PBS population (not enough patients are enrolled who would be eligible for the proposed medicine according to the requested restriction or relevant to the proposed patient indication)

D     not compared with the main comparator (or other relevant comparator) as identified in Subsection 1.1 (this is not relevant for submissions based on indirect comparisons of randomised trials via a common reference arm).

Adapt the selection process as indicated for submissions relying on an indirect comparison of randomised trials or nonrandomised evidence.

2.2.2   Annotated search results

List identified trials (eg in a spreadsheet), and indicate which trials were excluded and the basis for their exclusion (use categories A–D from Subsection 2.2.1).

2.2.3   Master list of relevant trials

Present a master list of all included trials and relevant systematic reviews or meta-analyses that meet the inclusion criteria from Subsection 2.2.1 according to Table A3.1 of Appendix 3.

2.2.4   Further selection of trials for an indirect comparison

Where the submission includes an indirect comparison, justify the exclusion of trials that are unsuitable for use in the indirect comparison. Appendix 3 provides an acceptable approach for selecting appropriate trials to include in an indirect comparison. A general approach is summarised here:

1.    Follow the approach outlined in Subsection 2.2.1 to identify all trials involving the proposed medicine (irrespective of comparator arm) and all trials involving the main comparator (irrespective of comparator arm).

2.    Draw a network diagram to show all the possible links.

3.    Where pairwise comparisons are possible, the submission may seek to exclude linkages requiring multiple steps, or include these as a supplementary analysis.

4.    Examine heterogeneity within trial sets and across trial sets, and justify the exclusion of trials with differences in factors that may affect the transitivity of the trials in the indirect comparison.

5.    Examine the event rates in the common reference arms and justify the exclusion of trials.

6.    Present a list of the trials included in the main analysis, the trials included in supplementary or sensitivity analyses, and the trials excluded from all analyses.

2.2.5   Approach taken to support the clinical claim

Describe how the included studies are combined or compared to support the clinical claim.

Example:

The submission is based on a meta-analysis of three trials of [medicine X] compared with [medicine Y], which is then compared indirectly to a single trial of [medicine Z] compared with [medicine Y]. A claim of noninferiority is made on the outcomes of time to progression and quality of life.

If subgroups were used to support the clinical claim, justify this in Subsection 2.6.1.

2.2.6   Copies of included trials

Include sufficient details of the relevant trials (key publications, supplementary data, clinical study reports) as attachments to the submission. Indicate the location of the trial reports in the document table at the beginning of the submission.

Where there is more than one report of a randomised trial (eg one or more published papers, one or more trial reports internal to the sponsor), provide the published paper(s) and the complete internal trial (clinical study) report(s). If the results vary between reports of the same randomised trial, discuss the differences, justify the results used in the base case and cross-reference the source of the extracted results. Provide a copy of each publication that reports data from a listed randomised trial.

For any relevant trial identified from a meta-analysis, include the individual trial report or publication(s) as above.

Provide reputable translations of trial reports that are not published in English.