Issues regarding the sharing of interim results by the Data Safety Monitoring Board of a trial with those responsible for the conduct of the trial.
Bibliographic record
Abstract
Background and Objectives: Sharing of interim results by the Data Safety Monitoring Board (DSMB) with non-DSMB members is an important issue that can affect trial integrity. The objective of this dissertation was to determine the views of the stakeholders on what kind of interim results can or should be shared by the DSMB, why, and with whom among those responsible for the conduct of a trial. Methods: We first conducted a systematic search of the literature to assess views and current evidence on sharing interim results. Secondly, we conducted two cross-sectional surveys aimed at those involved in trials to solicit their views on what type of interim results should be shared by the DSMB with non-DSMB members, with whom and under what circumstances. Thirdly, we assessed for any potential association of demographic factors with the sharing of certain interim results and their perceived usefulness, using regression analysis. Results: Mixed views exist in the literature on interim result sharing practices. Evidence from the surveys conducted resulted in the following findings. What to share: Based upon the survey results from our cross-sectional survey (Chapter 4), the interim control event rate (IControlER), the adaptive conditional power (ACP) and the unconditional conditional power (UCP) should not be shared. Most respondents from this survey thought the interim combined event rate (ICombinedER) could be shared provided proper conditions and provisions are in place. However, based on our cross-sectional scenario-based survey (Chapter 3), it was demonstrated that the ICombinedER, when shared at interim, is compatible with three possible interim results (Drug X doing better than placebo, worse than placebo or performing the same as placebo). Why share or not share: Respondents indicate that the ICombinedER can be shared because it does not unmask relative effects between groups, and keeps the steering committee (SC) informed about the trial’s progress; however, with the condition that sharing this type of result should be specified a priori including for what purpose and be at the DSMB’s discretion, especially if the control group rate is known from the literature. However, it is important to note that the ICombinedER, demonstrated with evidence from our cross-sectional scenario-based survey (Chapter 3), is compatible with three possible interim results and should not be shared because it has low usefulness and is flawed due to multiple interpretations. The IControlER and the ACP should not be shared because they are unmasking of interim results. It was mentioned that ICombinedER is usually known by the SC and sponsor making it easy to determine group rates if the IControlER is known. The UCP should not be shared because it is a technical measure that is potentially misleading of interim results. With whom to share: Survey results from Chapter 4 indicated that the ICombinedER can be shared with the SC and that the IControlER, the ACP, and the UCP should not be shared with any non-DSMB members by the DSMB. However, evidence from Chapter 3 also indicates that the ICombinedER should not be shared with any non-DSMB member. Factors associated with sharing: Having experience with greater than 15 trials with private industry sponsorship was found to be associated with not sharing the IControlER and an increase in perceived usefulness in sharing the ACP. Though some other demographic factors were found to be associated with sharing the ICombinedER and the UCP, they were sensitive to missing data upon our sensitivity analysis and will require more validation. Conclusions: Though mixed views exist within an extensive literature review on interim result sharing practices, survey evidence from this dissertation suggests that the ICombinedER, IControlER, the ACP and the UCP should not be shared with any non-DSMB member. The IControlER and ACP can be unmasking of interim results and the UCP is a technical measure that is potentially misleading. We agree with this reasoning. The majority of respondents from the survey in Chapter 4 indicated that the ICombinedER can be shared with the SC because it does not unmask relative effects between groups, however it was also stipulated that sharing this measure should be specified a priori and for what purpose and be at the DSMB’s discretion, especially if the control group rate is known from the literature. Even though the majority from our second survey in Chapter 4 indicate sharing the ICombinedER with the SC, we do not recommend sharing the ICombinedER at interim with any non-DSMB member because, as demonstrated with evidence from our cross-sectional scenario-based survey in Chapter 3, this measure is compatible with three possible interim results potentially leading to the introduction of trial bias at interim by those privy this interim measure and their interpretation. Based on the findings from the survey from Chapter 4, there appears to be a lack of awareness in how sharing the ICombinedER is flawed, of low usefulness, and potentially dangerous. The perceived desire to have this measure shared seems misguided. Experience with greater than 15 trials with private industry sponsorship was found to be associated with not endorsing the sharing the IControlER and an increase in perceived usefulness in sharing the ACP by the DSMB at interim. In regards to implications for future research, this characteristic should be further evaluated to see if this subgroup has insight into interim trial management practices that protect from trial bias. Results from this research have implications for practice and guidelines concerning trial design and protocols, and DSMB charters. These results can also help assess the need for proper safeguards around sharing an interim result when deemed appropriate by the DSMB and under their discretion, that prevent the introduction of bias that could alter the final trial results generated.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.672 | 0.818 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.008 | 0.020 |
| Scholarly communication | 0.013 | 0.020 |
| Open science | 0.006 | 0.010 |
| Research integrity | 0.013 | 0.019 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".