Case Report Form Completion and Data Quality

Part of Good Clinical Practice (GCP) — ICH E6(R3): Data Collection, Source Documentation and CRFs.

About 8 minutes. Free, in your language, no account needed.

Why this matters

The case report form (CRF) is how your site's work reaches the sponsor and the regulator. Data that arrive late, incomplete or inconsistent with the source create queries, delay the analysis and raise doubt about the site. A pattern of poor data quality can lead a sponsor to increase monitoring at your site or to exclude your data.

The CRF is now one kind of data acquisition tool

This is a headline change in the International Council for Harmonisation (ICH) Good Clinical Practice (GCP) guideline E6(R3). Older training treated the CRF as the data collection instrument. R3 introduces a broader term: the data acquisition tool (DAT).

The E6(R3) Glossary defines a DAT as:

"A paper or electronic tool designed to collect data and associated metadata from a data originator in a clinical trial according to the protocol and to report the data to the sponsor."

The Glossary explains that the data originator may be a human, such as the participant or trial staff. It may also be a machine, such as a wearable or a sensor. It may be another computer system, for example where data are extracted from an electronic health record or a laboratory system.

The Glossary lists examples of DATs: CRFs, interactive response technologies (IRTs), clinical outcome assessments (COAs) including patient-reported outcomes (PROs), and wearable devices, irrespective of the media used. The Glossary entry for Case Report Form now says plainly that a CRF is "a data acquisition tool".

Why this reframing matters to you at the site:

• A participant's electronic diary is a DAT, with the same data quality expectations as the CRF.
• The randomization system is a DAT. Its entries are trial data, not just logistics.
• A wearable sensor is a DAT. Device assignment, charging and data upload become data tasks.
• Training and oversight cannot stop at the CRF. Every tool that feeds the sponsor is in scope.

Reviewing the data before it is reported

Data cannot be reported well if nobody has looked at it. E6(R3) §2.12.3 gives the investigator both a right and a duty. The sponsor should provide the investigator with timely access to data (see §3.16.1(k)), and the investigator is responsible for the timely review of that data — including relevant data from external sources that can affect participant eligibility, treatment or safety. The guideline's own examples are central laboratory data, centrally read imaging data, another institution's records and, where appropriate, electronic patient-reported outcome (ePRO) data. The protocol may provide exceptions for access, for instance to protect blinding.

Central laboratory results are the everyday case. A grade 3 value that arrives in the portal and that nobody at the site opens is a safety failure, not a data-entry one.

Completing the tool correctly

E6(R3) §2.12.5 places the duty on the investigator to ensure "the accuracy, completeness, legibility and timeliness" of the data reported to the sponsor in the data acquisition tools completed by the site, and in other required reports such as serious adverse event reports. The investigator also reviews and endorses reported data at important milestones agreed with the sponsor.

E6(R3) §2.12.6 requires that data reported to the sponsor are consistent with the source records, or that the discrepancies are explained. Changes in reported data must be traceable, explained where necessary, and must not obscure the original entry.

E6(R3) §2.12.4 requires the investigator to ensure that data acquisition tools and other systems deployed by the sponsor are used as specified in the protocol or trial-related instructions.

On the sponsor side, E6(R3) §3.16.1(d) requires the sponsor to ensure that data acquisition tools are fit for purpose, designed to capture what the protocol requires, and validated and ready before they are needed.

For data capture itself, E6(R3) §4.2.1 states that where data captured on paper or in an electronic health record are transcribed manually into a computerized system, "the need for and the extent of data verification should take the criticality of the data into account". Acquired data from any source should be accompanied by relevant metadata. Automated validation checks that raise queries at the point of capture should be considered, based on risk.

Remember also §2.12.2: unnecessary transcription steps between the source record and the data acquisition tool should be avoided.

Corrections and queries

E6(R3) §4.2.4 sets the rule for data corrections. Corrections should be:

• Attributed — to the person or the computerized system making the correction.
• Justified — the reason is recorded.
• Supported — by source records from around the time of the original entry.
• Timely — made without unnecessary delay.

Two related sponsor rules protect the site. Under §3.16.1(i), the sponsor should not change data entered by the investigator or by participants unless the change is justified, agreed in advance with the investigator and documented. Under §3.16.1(l), the sponsor should not have exclusive control of data captured in data acquisition tools, so that undetectable changes are prevented.

Answer queries with what the source shows. Never change a value in the system to make a query close if the source does not support the new value.

What monitoring checks — and why 100% verification is not required

E6(R3) §3.11.4.5.4 describes monitoring of clinical trial data. The monitor checks "the accuracy, completeness and consistency of the reported trial data against the source records and other trial-related records", and whether the data were reported in a timely manner.

The guideline then states this directly:

"This can be done on the basis of using samples and supported by data analytics, as appropriate."

So E6(R3) does not require 100% source data verification. The sample size and the types of records may be adjusted based on earlier monitoring results, or on other signs of insufficient data quality. Monitoring focuses on data identified as being of higher criticality in the monitoring plan.

Section 3.11.4.5.4 also asks monitoring to identify "missing data, inconsistent data, data outliers, unexpected lack of variability and protocol deviations", and to examine data trends within and across sites.

In practice

Rafael is a monitor for a multi-site trial. Centralized monitoring flags one site for review.

At that site, three consecutive participants have identical vital signs recorded at every visit: blood pressure 120/80, heart rate 72, temperature 36.5. There is no variation at all.

Real physiology varies. Identical repeated data often suggest copy-and-paste behavior, or estimation instead of actual measurement. This is exactly the "unexpected lack of variability" that E6(R3) §3.11.4.5.4 asks monitoring to detect.

Rafael does not accuse anyone. He asks to see the source. He asks who takes the measurements, with which device, and when the values are entered. He finds that a new coordinator records vitals at the end of the clinic day, from memory, and copies the previous visit when unsure.

The finding is a data integrity issue, not a typing error. The site retrains the coordinator and moves recording to the point of measurement. The sponsor increases source data checking at that site and assesses the affected data.

Good data is not data that looks tidy. It is data that matches the source.

Key takeaways

• E6(R3) reframes the case report form as one kind of data acquisition tool, alongside interactive response technologies, clinical outcome assessments, patient-reported outcomes and wearables.
• The data originator behind a data acquisition tool may be a person, a machine or another computer system, and all of them are subject to the same data quality expectations.
• Under §2.12.3 the investigator must be given timely access to trial data and is responsible for reviewing it in a timely way, including data from external sources such as central laboratories, centrally read imaging and ePRO, unless the protocol restricts access to protect blinding.
• The investigator is responsible for the accuracy, completeness, legibility and timeliness of data reported to the sponsor, and for endorsing the data at agreed milestones.
• Data corrections under E6(R3) §4.2.4 must be attributed, justified, supported by source records from around the time of original entry, and made in a timely manner.
• E6(R3) does not require 100% source data verification; §3.11.4.5.4 explicitly allows sampling supported by data analytics.
• Unexpected lack of variability in data is a recognized monitoring signal, because identical repeated values often mean the measurement was estimated rather than taken.