Improving sample size calculation and reporting for randomized trials with DELTA² guidance

A key step in designing a randomized trial is choosing the sample size. Too small and the study is underpowered, too large and you may have wasted time and resources in needlessly recruiting patients. The new DELTA2 guidance aims to assist researchers in sample size calculations and improve trial quality and reporting.

The size of a research study is based on a calculation of the minimum number of participants needed for the analysis to be able to answer the question the study has been set up to address. This is called the sample size.

We do not want a study to involve more people than necessary as it would mean people are potentially needlessly being recruited. This would also incur additional expenses. Conversely, we do not want the sample size to be too small as the study may not be informative, patients again may have been needlessly recruited for a financial cost that could be saved.

The new DELTA2 guidance aims to improve the choice of approaches when determining the sample size; more particularly for a key driver, called the “target difference”. If we are planning a study to compare an old treatment and a new treatment, the target difference is the level of improvement in the key outcome of the trial which we want to detect in order to enable us to say that a new treatment is better than an old one.

The target difference in the primary outcome is probably the most critical piece of information used in the sample size calculation.

To determine what target difference to use, we need to think about what we are going to measure in our study to help us decide whether a new treatment is better. An “outcome” is how we measure the specific effect of the treatments upon the study participants. While studies will measure multiple outcomes, typically, one outcome is chosen to be the key outcome for the study. This is called the primary outcome.

The target difference in the primary outcome is probably the most critical piece of information used in the sample size calculation. It has received increased focus over recent years as researchers and users of research (for example health professionals and funders of healthcare) have grappled with the interpretation and implications of trial findings.

The DELTA2 guidance was developed by a group of over 20 researchers and funders of research with representatives from the UK and North America. Further information on the development of the guidance is available here. The full guidance is available on the study webpage here.

To further increase the visibility and accessibility of the DELTA2 guidance we are co-publishing a summary article simultaneously in The BMJ and BMC Trials. Alongside this, a set of related articles are available in Trials in an article collection on “Sample size calculations and specifying the target difference for randomized trials”.

The DELTA2 guidance includes recommendations on how to conduct and report the sample size calculation to aid researchers designing a new randomized trial. Additionally, a list of reporting items for key study documents have been provided to help ensure the sample size calculations performed are reported in a transparent and informative way. This is to aid other researchers and interested readers in understanding what the study was aiming to achieve and the thinking behind it.

We hope that the DELTA2 guidance will help researchers when designing randomized trials and will play a role in improving their quality and reporting. Furthermore, we hope that it will also assist users of research when they are evaluating studies. The vision is that funders of randomized trials will recommend the DELTA2 guidance to those who are developing funding proposals for trials.

View the latest posts on the On Medicine homepage