Projects per year
Abstract / Description of output
Background
Incomplete reporting of the design, conduct and analysis of early phase dose-finding trials can hinder interpretability and reproducibility, and lead to erroneous conclusions on tolerability and efficacy. This methodological review investigates the reporting quality of published trials using broadly the CONSORT 2010 checklist with added items unique to dose-finding trials.
Methods
MEDLINE (PubMed) was searched for articles published in 2011-2020. Phase I or I/II trials, with planned interim dosing decisions (de/escalate, stay at the current level or an early stop), which aim to find a recommended dosing regimen(s) for further testing, were included. Data were extracted for 476 randomly selected trials, stratified by cancer/non-cancer settings.
Results
The key items that are frequently not reported include:
n (%) reported
cancer
(n=238) non-cancer (n=238)
Methods
Planned/maximum sample size 69 (29%) 105 (44%)
with justification 35 (15%) 59 (25%)
Recruitment method 19 (8%) 51 (21%)
Oversight committees 39 (16%) 90 (38%)
roles and structure 17 (7%) 40 (17%)
Who makes dose decisions 10 (4%) 39 (16%)
Definition of analysis population:
dose-determination 108 (45%) 111 (47%)
Safety 114 (48%) 129 (54%)
key outcomes 100 (42%) 131 (56%)
Rationale for starting dose 52 (22%) 42 (18%)
Results
Baseline demographic and clinical characteristics by each dose level 70 (29%) 148 (62%)
Settings and locations where data were collected 86 (36%) 149 (63%)
Participant flow diagram/table 85 (36%) 144 (61%)
Losses/exclusions for each dose level 30 (13%) 85 (36%)
Only 1 (0.4%) cancer trial was randomised compared to 180 (75.6%) non-cancer trials. Notably, very few trials (6.3%) provided accessible protocols. Improvement in the reporting over time is evident in some items, including participant flow and sample size justification.
Conclusion
Important methodological features in design, conduct and analysis are frequently omitted. Overall, non-cancer trials appear to be better reported, as mainly randomised, they may have adopted CONSORT 2010 guidance. This review confirms the need for robust consensus-driven guidance for researchers and journals reporting dose-finding trials, to allow accurate assessment of their results to reduce research waste.
Incomplete reporting of the design, conduct and analysis of early phase dose-finding trials can hinder interpretability and reproducibility, and lead to erroneous conclusions on tolerability and efficacy. This methodological review investigates the reporting quality of published trials using broadly the CONSORT 2010 checklist with added items unique to dose-finding trials.
Methods
MEDLINE (PubMed) was searched for articles published in 2011-2020. Phase I or I/II trials, with planned interim dosing decisions (de/escalate, stay at the current level or an early stop), which aim to find a recommended dosing regimen(s) for further testing, were included. Data were extracted for 476 randomly selected trials, stratified by cancer/non-cancer settings.
Results
The key items that are frequently not reported include:
n (%) reported
cancer
(n=238) non-cancer (n=238)
Methods
Planned/maximum sample size 69 (29%) 105 (44%)
with justification 35 (15%) 59 (25%)
Recruitment method 19 (8%) 51 (21%)
Oversight committees 39 (16%) 90 (38%)
roles and structure 17 (7%) 40 (17%)
Who makes dose decisions 10 (4%) 39 (16%)
Definition of analysis population:
dose-determination 108 (45%) 111 (47%)
Safety 114 (48%) 129 (54%)
key outcomes 100 (42%) 131 (56%)
Rationale for starting dose 52 (22%) 42 (18%)
Results
Baseline demographic and clinical characteristics by each dose level 70 (29%) 148 (62%)
Settings and locations where data were collected 86 (36%) 149 (63%)
Participant flow diagram/table 85 (36%) 144 (61%)
Losses/exclusions for each dose level 30 (13%) 85 (36%)
Only 1 (0.4%) cancer trial was randomised compared to 180 (75.6%) non-cancer trials. Notably, very few trials (6.3%) provided accessible protocols. Improvement in the reporting over time is evident in some items, including participant flow and sample size justification.
Conclusion
Important methodological features in design, conduct and analysis are frequently omitted. Overall, non-cancer trials appear to be better reported, as mainly randomised, they may have adopted CONSORT 2010 guidance. This review confirms the need for robust consensus-driven guidance for researchers and journals reporting dose-finding trials, to allow accurate assessment of their results to reduce research waste.
| Original language | English |
|---|---|
| Pages (from-to) | S24-S24 |
| Number of pages | 1 |
| Journal | Annals of Oncology |
| Volume | 33 |
| Issue number | S1 |
| DOIs | |
| Publication status | Published - 1 Mar 2022 |
Fingerprint
Dive into the research topics of 'Assessing the reporting quality of early phase dose-finding trials'. Together they form a unique fingerprint.Projects
- 1 Finished
-
Dose-finding Trials: Development of guidance in reporting, analysis and design to accelerate knowledge transfer
1/02/21 → 31/01/23
Project: Research