Question 19

College answer

Syllabus topic/section:

2.5.1 Research and evidence-based practice in intensive care / Research methods for randomised trials and aspects of statistics.

Discussion:  

Most candidate had a reasonable grasp of the topic, however answers suffered from a lack of content. The better answers gave succinct explanations of unplanned vs planned reasons including bias, underpowering, predefined criteria for conclusion of a study, interim review and analysis.
Many answers to Section b) were not of an acceptable standard due to failure to demonstrate knowledge and content. This is an area fundamental to the understanding of trials. Explanations would have included, but not limited to, harm vs. benefit, future curtailment of research and clinical opportunities as well as limitations of various types of data collected.

Discussion

"Research methods for randomised trials and aspects of statistics" is not actually a syllabus topic, but the reader familiar with the second edition of the CICM Syllabus for the Second Part Examination will immediately agree that "Aspects of statistics" and "Research methods for each study type including study design, populations, methods, data collection, analysis and outcomes" sounds vaguely the same.

Trials can have a planned early end or an unplanned end:

A planned early stop occurs when interim analysis demonstrates:

• Evidence of harm
• Evidence of benefit 
• Futility (EOLIA, 2018)
• Lack of trial viability  (eg. recruitment goal not reached within timeframe)

An unplanned early stop can occur when:

• Trial is bankrupt, out of funding
• Trial researchers move on to other things ()
• Unexpectedly difficult slow recruitment ()
• Too many protocol violations ()
• Unethical: new information has removed equipoise (PHARLAP)
• Reprioritisation: support for the trial has diminished because of changes in the political climate

Trials can also be stopped strategically to achieve a goal of the sponsor, eg. where the trial drug is performing poorly. This is viewed as an unethical practice.

Patient consequences of stopping the trial early:

• Harm or wasteful futile treatment is prevented 
• Results are reported sooner, which may be beneficial to society if the results are practice-changing
• A smaller enrolled group means less safety data is collected
• Percived futility may lead to a viable therapy being dismissed as ineffective 

Resource consequences of stopping a trial early:

• Reallocates resources to other studies
• Practically, unused funding is often lost

Consequences of early stopping for data analysis:

• Reduced loss to follow-up by shortening the period of follow-up
• The resulting study may end up underpowered
• Subgroup analysis becomes difficult, as the subgroups will be smaller
• Secondary endpoints may never reach statistical significance
• Multiple interim analysis and a smaller-than-expected group has the effect of amplifying the treatment effect by increasing the likelihood of a Type 1 error.
• If a composite endpoint was used, only few patient-centred outcomes might have accumulated