Final consensus statements |
Agreement (%)* |
Agreement (%)* |
Agreement (%)* |
Underpinning information source** |
|
Delphi 1st round
(threshold 76.5%)
|
Delphi 2nd round
(threshold 68.4%)
|
Consensus meeting
|
|
General |
|
|
|
|
Clinical trial integrity guidelines and policies must be explicit,
visible, and prospectively enforceable at all levels through an
implementation plan.
|
82.7a
|
|
|
SPS
|
Trialists, ethics committee members, journals editors and
peer-reviewers should receive appropriate methodological and integrity
training.
|
80.8a
|
|
|
SPS,1-7
|
Trial ethics committees should have accreditation and regional,
national and international harmonisation of ethics assessment criteria
and review process.
|
92.3a
|
|
|
8,9
|
There should be continuous public documentation of trials during the
entire study lifecycle.
|
61.5
|
61.5
|
80.0
|
SPS
|
Journals should support adoption of responsible research practices in
the design, conduct, analysis, reporting and archiving of trials.
|
88.5
|
|
|
SPS
|
Institutions should avoid excessive publication pressure.
|
76.9
|
|
|
SPS
|
Design and approval |
|
|
|
|
Ethics approval should be obtained for all trials, including those
using de-identified data.
|
67.3a
|
65.5a
|
100
|
10,11,20,21
|
Informed consent should be developed with patient (or their
representative) and public involvement.
|
80.8
|
|
|
12,13,14,15,16
|
Informed consent should be examined and approved by the ethics
committee.
|
96.2
|
|
|
1,12,14
|
Informed consent should include explicitly how the de-identified data
will be shared at the time of publication or used for future analysis
|
73.1
|
65.4
|
96.4
|
17
|
Trials should be prioritised and resourced according to local health
care needs, strategy, and culture, especially in multi-country trials
including low-resource settings.
|
69.2
|
69.2f
|
|
1,12,18
|
Trials should be approved according to local ethics and regulatory
framework, especially in multi-country trials including low-resource
settings.
|
76.9
|
|
|
1,12,18
|
Translations of patient reported outcomes should be culturally
sensitive in multi-country trials including low-resource settings.
|
84.6
|
|
|
19
|
Equality, diversity and inclusion should be embedded in trial design
to maximize generalisability of findings.
|
76.9
|
|
|
SPS
|
Sample size estimation should be sufficiently detailed to permit
replication.
|
92.3
|
|
|
24
|
Primary and secondary outcomes should follow the internationally
agreed core outcomes whenever available.
|
80.8
|
|
|
SPS
|
The trial protocol, including ethics approval, should be prospectively
registered with an open-access trial registry prior to participant
recruitment. This policy should be included in research institutions‘
and sponsors‘ regulations, and researcher employment and funding
contracts.
|
78.9a,b,f
|
|
|
SPS, 30,32,35
|
Conduct and monitoring |
|
|
|
|
Trial site assessment should put in place measures to mitigate
integrity breaches with the support of local research governance
departments.
|
88.5a
|
|
|
SPS
|
There should be promotion of admission of honest or unintentional
errors in the conduct of the trial without fear of blame. A part of
this policy should be training.
|
94.2a
|
|
|
SPS
|
Innovative recruitment strategies should be participant-driven and
should comply with ethics principles.
|
88.5
|
|
|
15,25,26e
|
Routinely collected data should be validated before analysis and
reporting.
|
69.2
|
84.6
|
|
SPS, 20,27
|
Informed consent oversight should be part of trial audit.
|
92.3
|
|
|
10,13
|
The membership of independent trial steering and data monitoring
committees should declare any potential conflict of interests.
|
100
|
|
|
SPS
|
The membership of independent trial steering committees should include
patient and public stakeholders.
|
69.2
|
65.4
|
79.3
|
SPS
|
Minutes of the independent trial steering and data monitoring
committees should be available when required.
|
69.2
|
61.5
|
83.0
|
SPS
|
Data monitoring committee charter should include responsibility for
data integrity.
|
92.3
|
|
|
SPS,28
|
Centralized monitoring and selective source data verification should
be deployed for ensuring data integrity.
|
80.8
|
|
|
29
|
There should be transparency in the method(s) of handling missing data
at all stages of monitoring and reporting.
|
96.2
|
|
|
SPS
|
Early termination of a trial should be undertaken with the input of
the independent trial steering and data monitoring committees.
|
96.0
|
|
|
SPS
|
Any amendment to study protocol should be reported to the trial
registry (with dates). Major changes also require ethics approval.
|
100
|
|
|
SPS
|
The statistical analysis plan should be developed and published at the
start or during the early stages of the trial before the data is made
available to the investigators.
|
88.5
|
|
|
SPS
|
All analyses should be pre-specified from the outset (the analysis of
the primary outcome and secondary outcomes, sub-group analyses, and
sensitivity analyses).
|
84.6
|
|
|
SPS
|
There should be a single primary outcome pre-specified; when there are
multiple key outcomes, valid testing strategies should be considered
for maintaining familywise type-1 error within the acceptable limit of
5 %.
|
65.4
|
69.2f
|
|
SPS
|
Trial funders should mandate in their contract with researchers that
outcomes are analysed and reported according to preregistration.
|
42.3
|
57.7
|
88.0
|
SPS
|
Databases for trials should include auditable access logs and
permission management systems to prevent illicit access to data or
editing of data.
|
n/ad
|
n/ad
|
100
|
SPS
|
Trial integrity and quality evidence synthesis both require the
avoidance or minimisation of bias in trial conduct.
|
n/ad
|
84.6
|
|
SPS
|
Reporting of protocols and findings |
|
|
|
|
Trialists are strongly encouraged not to submit to a predatory
journal, avoiding journals without transparency and integrity.
|
69.2
|
65.4a
|
83.3
|
30
|
Journals’ authors’ instructions should explicitly and comprehensively
cover the requirements for openness and transparency.
|
84.6a
|
|
|
SPS, 31,32,33,34
|
Journals´ electronic submission system should facilitate compliance
with the integrity-related authors‘ instructions.
|
73.1
|
92.3
|
|
SPS
|
Professional medical writing could help in reporting more clearly and
succinctly to meet the integrity requirements. Its contribution should
be reported.
|
61.5
|
69.2f
|
|
36
|
The speed with which editorial and peer-review decisions are made
should be balanced against the possibility of future complaints and
retraction.
|
65.4
|
65.4
|
83.3
|
37
|
Reporting of ethics approval and informed consent details should be
obligatory part of reporting guidelines and authors’ instructions.
|
84.6a
|
|
|
10,13, 14,17,38
|
Ethics or independent data monitoring committee should provide
confirmation that the trial was conducted as planned.
|
61.6a
|
69.5a
|
|
SPS
|
Authorship contribution (credit according to international guidelines)
should be made explicit in the manuscript.
|
94.3a
|
|
|
SPS,22,23
|
Trial protocol and statistical analysis plan should be submitted in
unredacted form along with data set, statistical syntax and analytical
outputs.
|
69.2
|
88.5
|
|
SPS,7,33
|
Reporting of conflict of interests, funding sources and payments
received by all authors should be standardised.
|
78.9a
|
|
|
SPS,23,34,39,40,41
|
Declaration of conflict of interest, funding sources and payments
should be mandatory for peer-reviewers and editors.
|
88.5
|
|
|
SPS
|
Reporting of patient and public involvement in the trial should be
mandatory.
|
76.9
|
|
|
SPS
|
Manuscripts should be prepared according to standard reporting
guidelines (e.g SPIRIT, CONSORT, GRIPP-2, etc) and their specific
extensions for particular trial types (e.g. human challenge trials,
trials of social and psychological interventions, etc.).
|
76.9a,c,f
|
|
|
SPS,42,43, 47
|
Plagiarism checks should be routinely carried out on the article main
text.
|
84.6
|
|
|
44
|
Errors, deviations from protocol, losses to follow-up, missing outcome
data and solutions applied should be transparently reported.
|
92.3
|
|
|
45,46,54
|
Reporting the use of data monitoring committees, its responsibilities
and its membership should be mandatory.
|
73.1
|
96.2
|
|
28
|
Among trials conducted in various languages use of translations in
patient reported outcomes should be explicit.
|
53.8
|
53.8
|
91.6
|
19
|
Primary and secondary outcomes should be mandatorily linked to
prospectively registered outcomes.
|
76.9
|
|
|
35
|
Spin in writing to misrepresent, overinflate or distort the methods,
findings, results and conclusions should be eliminated.
|
82.7a
|
|
|
SPS
|
The strengths and limitations of the integrity-related issues, as well
as any flaws in terms of less-than-ideal method implementation that
was unavoidable, should be discussed in the manuscript.
|
73.1
|
96.2
|
|
SPS
|
Post-publication |
|
|
|
|
When a post-publication review detects integrity breaches, the
implication is that the scientific process failed, so the focus should
be on correction and learning lessons openly and collectively.
|
76.9
|
|
|
SPS
|
Journals have the responsibility to conduct their pre-publication
assessments and peer-review in a manner so as to minimise the risk of
post-publication dishonesty allegations.
|
92.3
|
|
|
SPS
|
Any guidance concerning post-publication integrity concerns (e.g, COPE
https://publicationethics.org,
https://doi.org/10.24318/o1VgCAih,
https://doi.org/10.24318/cope.2019.2.4) should explicitly
emphasise the investigators‘ responsibility to evaluate the integrity
of the complaint and to support the trialists.
|
73.1
|
88.5
|
|
SPS
|
Institutions and journals should be equally supportive to the
complainant(s) and author(s) in handling such complaints. There is a
responsibility to protect honest trialists against harassment.
|
84.6a
|
|
|
SPS
|
Trialists must engage with any request for an explanation for apparent
data discrepancy if required by the journal during both peer-review
and post publication stages, or by systematic reviewers during
evidence synthesis.
|
92.3
|
|
|
SPS
|
Trialists have the responsibility to keep detailed records of their
trial including original protocol (with any subsequent amendments),
ethics approval, details of the trial registration, de-identified raw
data set, randomisation sequence employed, statistical plan, syntax
and outputs of all the statistical analyses in case these are required
to address any post-publication complaints.
|
80.8
|
|
|
SPS
|
Declaration of conflicts of interest, funding sources and payments
should be mandatory for complainants.
|
84.6
|
|
|
SPS
|
Journals should act in an unbiased fashion transparently managing the
conflict of interest of their own editors and advisors handling
complaints.
|
80.8a
|
|
|
SPS
|
Trialists, with their institutional input, should be permitted to
provide independent expert reports to the journal investigating a
complaint.
|
76.9
|
|
|
SPS
|
If honest mistakes are identified in post publication, an erratum
should be published.
|
96.2
|
|
|
SPS
|
Retraction notices should be clear and interpretable.
|
88.5
|
|
|
48
|
Post-retraction management of trials with proven misconduct should be
based on a system that avoids continued citation and data misuse.
|
96.2
|
|
|
48
|
Future research and development |
|
|
|
|
Educational effectiveness of integrity training should be evaluated.
|
69.2
|
84.6
|
|
53e
|
The factors influencing participant willingness to give consent for
data sharing should be evaluated.
|
61.5
|
76.9
|
|
51,52
|
The minimum requirement for adequate informed consent should be
established.
|
61.5
|
69.2
|
|
49
|
The criteria for and level of data auditing required during conduct of
trial should be delineated.
|
61.5
|
65.4
|
100
|
10,49
|
The integrity remit of data monitoring committees should be clarified.
|
69.2
|
80.8
|
|
28
|
The best method(s) for publication credit (authorship contribution)
should be determined.
|
65.4
|
88.5
|
|
50
|
Effective peer review models should be developed for evaluation of
trials.
|
84.6
|
|
|
55
|
Automated checks for compliance with reporting guidelines items (e.g
CONSORT, SPIRIT, GRIPP-2) should be developed.
|
80.8
|
|
|
SPS
|
For the raw data to be shared, journals should clarify the
requirements, e.g. randomisation sequence, cleaned or original
de-identified dataset, statistical codes, etc.
|
69.3a
|
92.3
|
|
SPS
|
The validity of early post-submission and post-publication integrity
tests should be evaluated.
|
65.4
|
84.6
|
|
44
|
A common research terminology should be developed for prevention of
selective reporting.
|
57.7
|
53.8
|
86.9
|
54
|
Evidence syntheses of trials using reported study-level (not raw) data
should develop methods (e.g. subgroup meta-analyses or
meta-regression) to evaluate integrity concerns.
|
n/ad
|
69.2f
|
|
SPS
|
Evidence syntheses of trials should develop methods to access
patient-level (raw) data to maximize transparency.
|
n/ad
|
76.9
|
|
SPS
|