Extensions to CONSORT

Document Sample
Extensions to CONSORT Powered By Docstoc
					                       Extensions to the CONSORT Statement:
                         a proposal from the CONSORT Group
                                [A CONSORT policy document?]



 Note: this document started as a think piece for the eyes of the CONSORT Executive only,
 but I have come to believe that we should make this a publication, as a minimum on the
CONSORT web site but quite possibly also in a journal. The text thus evolved in style while I
  was writing it and so it now incorporates elements of both ideas. I have highlighted in
           yellow the material I believe should be only for “internal” consumption.

                                 Draft by DGA: 18 March 2010



The CONSORT Statement [ref] provides recommendations for reporting the findings of
randomised controlled trials. It provides a general framework that others have found helpful
both when considering specific trial contexts or methods and also in some applications
outside medical research.

The value of CONSORT has been demonstrated by the increasing number of “extensions” or
implementations of the Statement that have appeared since the revised Statement was
published in 2001. Several of these have been produced with input from the CONSORT
Group [refs], while others were done independently[refs].

The proliferation of variants of CONSORT is gratifying but it is increasingly causing concern
that they can cause confusion, especially as they do not address mutually exclusive
contexts. The idea has been considered of a web tool to produce a customised checklist for
each trial according to that trials scope – e.g. cluster trial in surgery. Most of the various
existing extensions modify and/or add to the generic CONSORT text, making it hard to
integrate the different versions of the same item. The proposal made here should make
combinations easier to handle.

If we consider that the generic CONSORT checklist includes items that should be addressed
in every report of an RCT (allowing for occasional qualification as “when relevant”), then we
can see the extensions as indicating not different but additional information. It follows that
extensions can be expressed as separate requirements rather than revising the basic
CONSORT item text. If we do this, then it will be very much easier to see what is different in
each extension, and to combine different extensions when relevant. This is the approach
taken in the CONSORT extensions on reporting harms [], but not in the other extensions.

For example, items 1 and 2 from the CONSORT extension to cluster RCTs are shown in
Figure 1a, where the extra text is shown in italics. These items could be reformatted as
shown in Figure 1b. Even when the new text is embedded within the existing text separation
is quite simple, as shown for item


                                                                                                 1
We can characterise the two approaches as blending and augmentation. For some published
extensions all the recommended changed comprised additions to the generic text, such as
the extension to trials of herbal interventions []. Reformatting these guidelines is a trivial
exercise.

The augmentation approach has the advantage that it is much easier to see what is required
in addition to the generic CONSORT. Also, it will be easier to update the extensions when
the generic CONSORT checklist changes, as is likely every few years. Unlike with the
blended system, when the wording of the generic item is modified the extended guidance
may not need to be changed.

We thus recommend that the augmentation approach is used in future extensions of the
CONSORT Statement. We note that the CONSORT checklist is currently under revision and
those extensions already published will need to be updated. That would be an opportunity
to put all extensions into the recommended format.

I have not checked all extensions but I believe that this approach can be adopted for them
all without difficulty.

If we all agree with this policy we still have the opportunity to impose this
recommendation/requirement on the CONSORT extensions currently unpublished – NPT,
pragmatic, and TCM. NPT is in submission so we need to act fast, should we wish to do so.

Occasionally, a group developing an extension to CONSORT has felt the need to address an
issue not covered in the generic CONSORT checklist. If the new information cannot be linked
to an existing item then a new item is needed, as for example in the NPT extension [ref].
Such a change causes problems under the current blending system too.

In summary, we the recommend that future extensions to CONSORT (and revisions to
existing extensions) should adopt the augmentation approach. Also, we believe iut is
valuable for all publications of the CONSORT checklist to indicate the date of the version
being used.

?? If we go with this idea, should we say that reformatted versions of existing guidelines will
be available on CONSORT website??



Acknowledgements

We thank Iveta Simera for helpful discussions.



References




                                                                                                 2
Figure 1. Items from the CONSORT extension for cluster randomised trials [ref] (a) as
published, and (b) reformatted as suggested.

(a) Original format

Item         No.   Describe

Design       1     How participants were allocated to interventions (eg random allocation,
                   randomised, or randomly assigned), specifying that allocation was based on
                   clusters
Background   2     Scientific background and explanation of rationale, including the rationale for
                   using a cluster design



(b) Reformatted as proposed (augmentation)

                      Standard CONSORT item                    Extension for cluster trials
             No.      Describe                                 In addition:
Design       1        How participants were allocated to       Specify that allocation was based on
                      interventions (e.g. random allocation,   clusters
                      randomised, or randomly assigned).
Background   2        Scientific background and explanation    Give rationale for using a cluster
                      of rationale.                            design




                                                                                                      3
Figure 2. Items from the CONSORT extensions for cluster randomised trials [] and
noninferiority and equivalence randomised trials [ref] (a) as published, and (b) reformatted
as suggested.

(a) Original format

Item            No.   Describe

CLUSTER TRIALS
Outcomes        17    For each primary and secondary outcome, a summary of results for each
and
                      group for the individual or cluster level as applicable, and the estimated
estimation
                      effect size and its precision (eg 95% confidence interval) and a coefficient of
                      intracluster correlation (ICC or k) for each primary outcome.
NONINFERIORITY AND EQUIVALENCE TRIALS
Interventions   4     Precise details of the interventions intended for each group, detailing whether the
                      reference treatment in the noninferiority or equivalence trial is identical (or very
                      similar) to that in any trial(s) that established efficacy, and how and when they
                      were actually administered.
Sample size     7     How sample size was determined, detailing whether it was calculated using a
                      noninferiority or equivalence criterion and specifying the margin of equivalence
                      with the rationale for its choice. When applicable, explanation of any interim
                      analyses and stopping rules (and whether related to a noninferiority or
                      equivalence hypothesis).



(b) Reformatted as proposed (augmentation)

                      Standard CONSORT item                       Extension for cluster trials
                No.   Describe                                    In addition:
CLUSTER TRIALS
Outcomes        17    For each primary and secondary              (a) Provide results for the individual or
and                   outcome, a summary of results for           cluster level as applicable.
estimation            each group and the estimated effect
                                                                  (b) Provide a coefficient of intracluster
                      size and its precision (eg 95%
                                                                  correlation (ICC or k) for each primary
                      confidence interval).
                                                                  outcome
NONINFERIORITY AND EQUIVALENCE TRIALS
Interventions   4     Precise details of the interventions        Detail whether the reference treatment in
                      intended for each group, and how and        the noninferiority or equivalence trial is
                      when they were actually administered.       identical (or very similar) to that in any
                                                                  trial(s) that established efficacy.
Sample size     7     How sample size was determined.             (a) Detail whether it was calculated using a
                      When applicable, explanation of any         noninferiority or equivalence criterion and
                      interim analyses and stopping rules.        specifying the margin of equivalence with
                                                                  the rationale for its choice.
                                                                  (b) If relevant, indicate whether stopping
                                                                  rules related to a noninferiority or
                                                                  equivalence hypothesis.
[Unsure if need to keep italics for additional text in last column of reformatted versions]


                                                                                                         4

				
DOCUMENT INFO