Trials Sample Size

Project Title: Reported sample size for phase III trials currently registered on the ClinicalTrials.gov registry.

Start Date: 01.06.2016
End Date: Ongoing

Authors:

  1. Nikola Cihoric,
  2. Ivan Soldatovic

Incentives and description of a project:

It seems that reported sample size is smaller than expected. Further examination are warranted.

Material and Methods

We evaluated the reported sample sizes of phase III (phase 3) clinical trials reported within the clinicaltrials.gov registry. We

Results

In total 17409 phase III trials were evaluated (start date 2005-2015, < 4 Arms). Median reported sample size classified by primary sponsor were (n, 25-75 perc.):

  • National Institute of Health was 318 (122-738)
  • Industry - 300 (121-592)
  • U.S. Fed - 158 (62-323)
  • “Others“, mainly investigator-initiated trials - 135 (28-326)

Discussion

The reported sample size is dependent on the type of primary sponsor (Industry, National Institute of Health, Academy, US Government) and on the presence (or not) of a Data Monitoring Committee (DMC).
This strongly indicates that trials without DMC academy initiated trials are of lower quality. It is not plausible that phase III trials have such low number of patients.
Whan was the last time that someone has reported pivotal trial with 20% difference in survival? How effective should be a treatment for cancer, diabetes, hypertension, depression or schizophrenia to show such effect on this sample size?
Keeping in mind that the sample size determination is one of the central parts of a clinical trial protocol, we believe that this finding is of interest for the broader medical community, as well as for all involved parties in the clinical trials enterprise.

Conclusion

Our findings show that the median reported sample size for all phase III trials (actual or anticipated) is lower than expected.


Further directions

Currently, we are classifying the data saved in several data fields/tags, such as short or scientific summary, interventions, eligibility criteria, and endpoints. Our ultimate goal is to manually categorize all clinical trials according to the discipline and type of intervention and eventually pinpointing the most important determinants of the sample size. An undertaking that requires the allocation of significant human and technical resources.

Keywords

clinical trial, phase 3, sample size