Data and safety monitoring

This section contains the following:

Introduction
Things to consider
Things to consider when reporting to the Data Monitoring Committee
Additional resources
Further reading

Introduction

It is crucial to ensure the safety and integrity of individual patients participating in the trial. As such, accruing data should be monitored for any evidence of of adverse events. Its is important that this monitoring of accruing data is undertaken by individuals who are independent of the trial team as knowledge of the unblinded data or emerging trends may make the investigator in charge of patient recruitment reluctant to enroll new patients or to continue treating those already enrolled if the data suggests one intervention is better than the other.

A Data Monitoring Committee (DMC) should be established and meet regularly to give independent advice on whether the trial should a) continue as planned b) stop further recruitment or c) discontinue the intervention if there is evidence that the latter is substantially better (or worse) than the alternatives. Sometimes it is recommended that the intervention be modified (e.g. treatment dosage reduced or one treatment arm stopped in a multiple arm trial). Such data and safety monitoring committees often use statistical stopping rules to help with their decision making.

The data should ideally be unblinded by the intervention group so that the DMC can fully assess efficacy and safety.

Things to consider

Establish a DMC early in the trial.
Establish with the DMC how often they should meet and what data they would like to review at each meeting.
Ensure that adverse events reporting procedures are in place.
Ensure that the trial statistician knows the timing of the planned DMC meetings as they will need time to prepare the DMC reports.

Things to consider when reporting to the Data Monitoring Committee

Grant et al (2005) have highlighted several issues

Provide comprehensive, but digestible information, summary tables and statistics in graphical form are helpful.
If possible agree the template for the interim report at the start or soon after the beginning of the trial.
Two reports are often prepared one for the open session and one for the closed session.
The length of the report depends on the context of the trial.
It is very important to give the DMC up-to-date information on the outcome measures. This will reduce the risk that subsequent data updates substantatively alter the analysis conclusions.

Additional resources

Data monitoring committees

Picture of Carl Lombard Carl Lombard, MRC Biostatistics Unit, South Africa, explains the purpose of a data monitoring committee (MP4, 36 seconds). Can't see the audio controller? Play it in your media player.

What is the ideal composition of a data monitoring committee? (MP4, 45 seconds) Can't see the audio controller? Play it in your media player.

What are the key things to think about with regard to a data monitoring committee? (MP4, 36 seconds). Can't see the audio controller? Play it in your media player.

Checklist of possible evaluations to be made to the DMC.

This is a list of the basic generic tables that should be made available to the DMC from Grant AM et al Issues in data monitoring an Interim analysis of trials. HTA 2005;9(7).