Deficiencies in the reporting of early phase dose-finding trials: findings from a rapid methodological review

(preliminary results)

​

ICR Clinical Trials & Statistics Unit (ICR-CTSU)

20^th September 2021

Olga Solovyeva, Christina Yap

and on behalf of DF-CONSORT Methodological Review Working Group

Background

• Incomplete or unclear reporting of the design, conduct and analysis of early phase dose-finding trials can hinder interpretability, reproducibility and impact on timely clinical development and lead to erroneous conclusions on tolerability and efficacy.

​

• The CONSORT 2010 statement is the recognised quality standard for reporting randomised trials and its adoption by many journals has contributed to an increase in reporting quality and completeness.

​

• Existing reporting guidelines do not fully cover features specific to dose-finding trials, e.g. starting dose and justification, escalation/de-escalation strategy.

​

• There currently exists no work that comprehensively assesses the reporting quality in this setting - this rapid methodological review addresses this gap by investigating the reporting quality of published dose-finding trials.

2

Dose-Finding CONSORT Extension

3

Project overview:

https://www icr ac uk/dosefindingconsort​

1/ METHODOLOGICAL REVIEW

The overall aim of this research is to develop and disseminate to stakeholders an extension to the CONSORT 2010 statement (DF-CONSORT) tailored to the specific requirements of early phase dose-finding clinical trials across all disease areas.

​

First step: Methodological Review

• Identify gaps and assess quality of reporting,
• Make recommendations towards DF-CONSORT extension

Methodological Review

4

The overarching research question is:

​

​

​

​

​

​

​

​

​

This is broken down into several sub-questions regarding whether the quality differs by:

• time of publication;

• journal endorsement of CONSORT;

• funding source;

• setting (oncology/non-oncology);

• geographical region of corresponding author, and;

• trial design

What are the GAPs and Reporting Quality of

Early Phase Dose-finding trials?

Methods

.

​

​

​

5

Data Extraction Items

​

Drawn from guidance documents (including CONSORT 2010, Adaptive Designs CONSORT Extension (ACE), SPIRIT 2013) with added items specific to early phase dose-finding trials from relevant published literature

Selection of Clinical Trials Papers

• MEDLINE via PubMed was searched for articles published in English, from 2011 to 2020.

​

Key Inclusion Criteria

• Phase I or I/II clinical trials, where interim dosing decisions have to be undertaken using accumulating data to either escalate, de-escalate, stay at the current level or stop a trial early, with the aim of identifying a recommended dosing regimen(s) for further testing.
• Reported main analysis of a trial

​

Key Exclusion Criteria

• Trials that were not planned with interim dose decisions were excluded.

Study Flow Chart

6

• 5,839 articles have been retrieved from PubMed in April 2021;
• The papers were randomly permutated and stratified into the oncology and non-oncology subsets;
• Screening until a total of 476 dose-finding trials which meet the inclusion/exclusion criteria have been selected (238 oncology & 238 non-oncology papers);
• Full data extraction for 476 papers, 10% to be independently checked;
• The sample size of 476 papers will provide a two-sided 95% confidence interval for the reporting proportion of an item which has a width of at most 9% (±4.5%).

Preliminary Results*: Study Design

7

* Henceforward the results provided for the first 260 papers

High Reporting, >80%

8

* Included in CONSORT 2010 checklist

^A Included in Adaptive designs CONSORT Extension (ACE) checklist

^S SPIRIT 2013 checklist

Low Reporting, <25%

9

* Included in CONSORT 2010 checklist

^A Included in Adaptive designs CONSORT Extension (ACE) checklist

^S SPIRIT 2013 checklist

Participant Flow

From CONSORT 2010 Statement:

“Item 13. Participant flow (a diagram is strongly recommended)”

10

* Both Flow Diagram and Flow Table were presented in 14 (5.4%) articles.

Key Differences Between Oncology / Non-oncology

11

Trends

Planned/maximum sample size justified

12

- cancer; - non-cancer; - overall

Planned/maximum sample size specified

A participant flow chart provided by a flow diagram

Interim results reported

Remarks

Reports of dose-finding trials frequently omit important methodological features in design, conduct and analysis, including items highlighted by the CONSORT 2010 statement.

​

Many (particularly those that are non-randomized) may not have used the CONSORT 2010 statement, though many of the checklist items may apply.

​

Early dose-finding trials have specific features not covered in the CONSORT 2010 statement.

​

This methodological review further confirmed the need for a robust and comprehensive consensus-driven reporting guidance for authors and journals reporting dose-finding trials, to enhance transparency, completeness, reproducibility, and interpretation to reduce research waste.

13

Participate in Delphi Survey!

The DF-CONSORT Executive Committee would like to invite interested clinical research stakeholders to register their interest in taking part in the Delphi Survey process by email at:

​

DFCONSORT-icrctsu@icr ac uk

14

Acknowledgements�

15

MRWG (Methodological Review Working Group)

​

​

• Christina Yap, Institute of Cancer Research
• Olga Solovyeva, Institute of Cancer Research
• Christopher Weir, University of Edinburgh
• Shing Lee, Columbia University
• Munyaradzi Dimairo, University of Sheffield
• Aude Espinasse, Institute of Cancer Research
• Jonathan W B Martin, Institute of Cancer Research
• Thubeena Manickavasagar, Institute of Cancer Research
• Rong Liu, Bristol-Myers Squibb
• Andrew Kightley, Patient and Public Involvement lead
• Johann de Bono, Institute of Cancer Research
• Zhulin Yin, Institute of Cancer Research

and

• Emily Alger

DF-CONSORT Executive Group

Methodological Review Protocol

Funded by: MRC-NIHR Methodology

Research Programme grant MR/T044934/1

THANK YOU!