Box 83 Factors that may influence a decision to use limited or no source data verification

♦ The existence of a plan other than source data verification to assure data quality

♦ Trials with a low susceptibility to bias: trials with simple procedures, noncritical eligibility criteria, and readily assessed outcome measures not susceptible to observer bias

♦ Trials in which sample data can readily be compared against supporting records, such as hospital records

♦ Trials conducted by groups with established standard operating procedures and a history of implementing such procedures effectively

♦ The existence of a clear prospectively specified analysis plan that are not industry-sponsored may employ less stringent procedures and may use no source data verification at all (http://www.fda.gov/cber/guidelines.htm). The extent of source data verification should be based on trial-specific factors such as design, complexity, size, and outcome measures. If little or no source data verification is used in a trial, it is essential to ensure close control, review and speedy checking of case report forms, and close contact with collaborating centres to ensure that procedures are understood and are being followed and all queries efficiently dealt with. Factors that may influence a decision to use limited or no source data verification are shown in Box 8.3.

The trial management group should decide for each trial the level of source data verification, if any, that is needed.

When source data verification is used, information provided on case report forms is compared against patients' hospital records at a visit by members of the trial management group to the centre concerned, or sometimes by an independent body. As a minimum, the following checks should be made for all or, more commonly, a random sample of participants.

♦ Check patient eligibility.

♦ Check that informed consent is being obtained.

♦ Check key baseline data.

♦ Check data essential for the trial outcome measures.

♦ Check that important events have been reported.

Inform the centre in plenty of time which patient records or other data will be required for checking.

Some centres may at first find source data verification intrusive, implying, perhaps, that they cannot be trusted to provide reliable data. Explaining the need for checks at all levels, including within the trial coordinating centre, will help to allay any such concerns if it is done sensitively.

0 0

Post a comment