To ensure that the integrity of clinical research
data is maintained and that there is total agreement
between the data recorded on CRFs, the data
entered in the computer, the data recorded in data
listings and cross-tabulations, the data entered into
statistical and clinical study reports and finally the
data in the sponsor/CRO and investigator archives,
it is essential that the data must only be changed by
following a formal procedure. Thus, requests for
data clarification and all resolution of queries must
be documented. All data changes must be author-
ized by the investigator ultimately. Obviously, theTable 12.9 Source data verification
For all study subjects, source data verification requires a review of the following items:
Existence of medical records/files at the study site. There must be a medical file, separate from the CRF, which
forms a normal part of the clinical record for the study subjects. The medical file should clearly indicate the full
name, birth date, and hospital/clinic/health service number of the study subject
Eligibility of study subjects. The medical file must show compliance with the inclusion and exclusion criteria. At a
minimum, demographic characteristics, for example sex, weight and height, diagnoses, for example major
condition for which subject was being treated, and other ‘hard’ data, for example laboratory results within a
specified range or normal chest X-ray, should be clearly indicated. All required baseline assessments must be
evident. If the medical file has little or no information concerning medical history, it would not support selection
of the subject
Indication of participation in the study. The medical file should clearly show that the subject was in a clinical study
in case the information is necessary for future clinical care
Consent procedures. The original signed consent form should be maintained with the subject’s medical files
or in the investigator files and an indication that consent was obtained (with the date specified) should be noted
in the medical files. Signatures and dates must be checked carefully to ensure that the correct individuals
were involved in the consent procedure and that consent was obtained prior to any study intervention
Record of exposure to study medication/device. The medical file should clearly indicate when treatment began,
when treatment finished, and all intervening treatment dates
Record of concomitant medications/devices. All notations of previous and concomitant medication/device
use must be examined. All entries in the CRF should be verifiable in the medical file by name,
date(s) of administration, dose and reason (or indication). All entries in the medical file during the time
period specified by the protocol must be noted in the CRF. Concomitant medication/device use must be
explicable by an appropriate indication and must be consistent from visit to visit. The reasons (indications)
for use of concomitant medications/devices, newly prescribed during the study period, must be noted as AEs.
The medical history should be reviewed to determine whether medical conditions arising during the study
already existed at baseline. The dispensing records, which are normally separate from the medical file,
must also be examined to determine consistency
Visit dates. All visit dates should be recorded in the medical file. Interim visit dates recorded in the
medical file, but not in the CRF, should be noted by the monitor in case they signify occurrence of AEs or
protocol violations. The final visit date should be so indicated, for example ‘study finished’ or ‘withdrew from study’
AEs. All AEs noted in the medical file during the time period specified by the protocol must be recorded
in the CRF. The monitor must also carefully check other documents (e.g. diary cards, quality of life forms)
for sources of information about AEs. Occurrence of out-of-range laboratory values, which are considered
to be clinically significant by the investigator, must be reported and assessed as AEs
Major safety and efficacy variables (to be decided and documented in advance). It is not necessary for all
measured variables to be recorded in the medical file. Present and future clinical care of the study subject is the most
important factor in determining whether or not measured variables should be recorded in the medical file. The
investigator should record what he/she would normally record to care for the study subject, but also take into
account any recording needed because of the special circumstances of a clinical study. The entire medical file
should be reviewed to ensure that no additional information exists in the medical file that should have been
recorded in the CRF
12.5 COLLECTING DATA WITH INTEGRITY 151