Clinical Data Management

Sponsored by
Center for Cancer Research
National Cancer Institute

Introduction
Clinical data management (CDM) consists of various
activities involving the handling of data or information
that is outlined in the protocol to be
collected/analyzed. CDM is a multidisciplinary activity.
This module will provide an overview of clinical data
management and introduce the CCRs clinical
research database. By the end of this module, the
participant will be able to:
Discuss what constitutes data management activities in clinical
research.
Describe regulations and guidelines related to data management
practices.
Describe what a case report form is and how it is developed.
Discuss the traditional data capture process.
Describe how protocols are developed in Cancer Central Clinical
Database (C3D).

Clinical Data Management

A multi-disciplinary activity that includes:

Research nurses
Clinical data managers
Investigators
Support personnel
Biostatisticians
Database programmers

Various activities involving the handling of

information outlined in protocol

Clinical Data Management

Activities

Data acquisition/collection
Data abstraction/extraction
Data processing/coding
Data analysis
Data transmission
Data storage
Data privacy
Data QA

Guidelines and Regulations

Good Clinical Practice (GCP):
Trial management; data handling, record
keeping (2.10, 5.5.3 a-d)
Subject and data confidentiality (2.11; 5.5.3
g)
Safety reporting (4.11)
Quality control (4.9.1; 4.9.3; 5.1.3)
Records and reporting (5.21; 5.22)
Monitoring (5.5.4)

Guidelines and Regulations

21 CFR Part 11
Applies to all data (residing at the institutional site and
the sponsors site) created in an electronic record that
will be submitted to the FDA
Scope includes:

validation of databases
audit trail for corrections in database
accounting for legacy systems/databases
copies of records
record retention

Case
Report
Forms

What is a Case Report Form

(CRF)?...
Data-reporting document used in a clinical
study
Collects study data in a standardized
format:
According to the protocol
Complying with regulatory requirements
Allowing for efficient analysis

What is a Case Report Form

(CRF)?
Allows for efficient and complete data
collection, processing, analysis and
reporting
Facilitates the exchange of data across
projects and organizations especially
through standardization
Types: Paper, electronic/web interface
Accompanied by a completion/instruction
manual

CRF Relationship to Protocol

Protocol determines what data should be
collected on the CRF
All data must be collected on the CRF if
specified in the protocol
Data that will not be analyzed should not
appear on the CRF

General Considerations for CRF

Development
Collect data with all users in mind
Collect data required by the regulatory
agencies
Collect data outlined in the protocol
Be clear and concise with your data
questions

General Considerations for

CRF Development
Avoid duplication
Request minimal free text responses
Collect data in a fashion that:
allows for the most efficient computerization
similar data to be collected across studies

Elements of a CRF
The term CRF indicates a single page
A series of CRF pages makes up a CRF Book
One CRF book is completed for each subject
enrolled in a study
Three major parts:

Header
Safety related modules
Efficacy related modules

Header Information
Key identifying Information

MUST HAVES
Study Number
Site/Center Number
Subject identification number

Safety Modules
Keep safety analysis requirements of the protocol
in mind
Follow the general guidelines for CRF development
Safety Modules include:

Demographic information
Adverse Events
Medical History/Cancer history (e.g., diagnosis, staging)
Physical Exam, including Vital Signs
Concomitant/Concurrent Medications/Measures
Deaths
Drop outs/off-study reasons
Eligibility confirmation

Efficacy Modules
Considered to be unique modules and can be
more difficult to develop
Protocol dictates the elements required in efficacy
modules
Define
Key efficacy endpoints of trial (primary and secondary_
Additional test to measure efficacy (e.g.: QOL)
How lesions will be measured (longest diameter, bidimensional, volumetric)
CR, PR, SD, PD
Required diagnostics

Include appropriate baseline measurements

Repeat same battery of tests

Standard CRFs
Allows rapid data exchange
Removes the need for mapping during data
exchange
Allows for consistent reporting across protocols,
across projects
Promotes monitoring and investigator staff
efficiency
Allows merging of data between studies
Provides increased efficiency in processing and
analysis of clinical data

CRF Development Process

Begins as soon in the study development
process as possible
Responsibility for CRF design can vary
between clinical research organizations (e.g.:
CRA, data manager, Research Nurse,
Database Development, Dictionary Coding,
Standards)
Include all efficacy and safety parameters
specified in the protocol using standard
libraries

CRF Development Process

Collect ONLY data required by the protocol
Work with protocol visit schedule

Interdisciplinary review is necessary

Note:
each organization has its own process for
review/sign-off
Should include relevant members of the project
team involved in conduct, analysis and reporting of
the trial

Properly Designed CRF

Allows components or ALL of the CRF
pages to be reused across studies
Saves time
Saves money

Poorly Designed CRF

Poorly designed CRFs will result in data
deficiencies including:
Data not collected as per protocol
Collecting unnecessary data (i.e.: data not
required to be collected per protocol)
Impeding data entry process
Database requiring modifications throughout
study

Electronic CRFs
The use of Remote Data
Capture (RDC) is increasing
In general, the concepts for the design of
electronic CRFs/RDC screens are the
same as covered for paper

No need to print and distribute paper

CRF Completion

CRF Completion
According to GCP Section 4.9.1, the investigator
should ensure the accuracy, completeness,
legibility, and timeliness of the data reported on
the CRFs and in all required reports. This
includes ensuring:
all sections have been completed, including the
header with identifying items
all alterations have been properly made
all adverse events are fully recorded and that for all
serious adverse events, any specific documentation
has been completed

CRF Completion
Data is taken from the source documents
(e.g.: medical record) and entered onto the
CRFs by study personnel. This is referred
to as data abstraction.
Only designated members of the research
staff should be allowed to record and/or
correct data in the CRFs
Typically this responsibility resides with the
Data Manager/ Research Nurse

Tips: CRF Completion

1. CRF completion/instruction manual should be
observed to ensure the accuracy,
completeness, legibility, and timeliness of the
data reported to the sponsor
2. Make sure appropriate protocol, investigator
and subject identifying information is included
in the Header (for RDC, may be pre-populated)

3. Ensure data is entered in the correct location or

data field

Tips: CRF Completion

4. Use the appropriate units of measurement
(UOM), and be consistent

5. Check to see that data is consistent across

data fields and across CRFs

E.g.:
Make sure visit dates match dates on the
laboratory or other procedure reports;
Make sure the birth date matches the
subjects age;

6. Use only the abbreviations authorized per

completion/instruction manual
7. Double check your spelling

Tips: CRF Completion

8. Watch for transcription errors

E.g.: sodium level should be 135 and entered as

153

9. Do not allow entries to run outside the

indicated data field; this important data might
be missed during data processing
10. Use comments section to elaborate on any
information, but keep to a minimum

Timeliness of CRF Completion

Ideally CRFs should be completed as
soon after the subjects visit as possible
Ensures that information can be retrieved
or followed-up on while the visit is still
fresh in the healthcare providers mind,
and while the subject and/or the
information is still easily accessible

REMEMBER.
Data cannot be entered onto a CRF if it is not in
the medical record or for some documents, in
the research record
If the individual completing the CRF, finds
missing or discrepant source data he/she
should:
Notify the research nurse or health care provider who
then will provide the data
If applicable, contact outside source (i.e.: outside lab
or doctor's office)

Common Errors
Logical
date of the second visit is earlier than the first
visit

Inaccurate information
source document says one thing, the CRF
says another

Omissions
AE is recorded on the CRF but not on the
source document

Transcription errors
date errors, 11-2-59 instead of 2-11-59

Common Errors
Abbreviations
unless an approved list of abbreviations is
distributed and utilized, data entry personnel
often misinterpret abbreviations

Spelling errors
Illegible entries/write-overs
Writing in margins

Correcting Paper CRF

Entries
If corrections are necessary, make the change
as follows:

Draw one horizontal line through the error;

Insert the correct data;
Initial and date the change;
DO NOT ERASE, SCRIBBLE OUT, OR USE
CORRECTION FLUID OR ANY OTHER MEANS
WHICH COULD OBSCURE THE ORIGINAL
ENTRY

These procedures ensure a complete audit

trail exists for all entries.

Correcting Paper CRF Entries

01/JAN/2005

05C1234

NCI
NIC

EN
9/8/05

12345678

03/JAN/1925

80

x
x
1.

Complete each form in black or blue pen to ensure good photocopies.

2.

All dates are to be expressed in day/month/year (dy/mth/yr) format. To

avoid ambiguity,months are to be recorded using a three letter
abbreviation (i.e., Jan, Feb, Mar., etc.). Years are to be recorded as four
digits (i.e. 1998).

Electronic Data Collection

Process
Web-based interface
Sponsor or site dependent

Ensures data integrity:

Controls the ability to delete or alter
previously entered data
Provides an audit trail for data changes
Protects the database from being tampered
with
Ensures data preservation (e.g. automatic
back ups)

Process of Data Transfer to

Sponsor
Traditional (Paper)
Electronic

Traditional Data Transfer

CRF Books developed by sponsor and supplied
to the site for completion along with
completion/instruction manual
Paper CRFs are either 2 or 3 part NCR (No
Carbon Required paper)
Use a black or blue ballpoint pen for permanency
and PRESS HARD

At the time of a monitoring visit, CRFs are

reviewed for adherence to completion
guidelines and verified against source
documents by the Monitor

Traditional Data Transfer

During the monitoring visit, site staff make
required corrections to CRFs
Verified/corrected CRFs are submitted to
the sponsor, leaving a legible copy of the
CRF at the site
e.g.: CRA may hand carry completed
CRFs to the sponsor;

If data is not retrieved at the time of the

monitoring visit, sponsor may want the
CRFs submitted via mail or facsimile

Traditional Data Transfer

Sponsor enters the CRF data into a centralized
database (generally done by 2 separate
individuals, called double data entry) and
reviews the data for errors

If inconsistencies are found, the sponsor

generates data queries (forms may vary slightly
from sponsor to sponsor) and sends to the site
Site staff investigates these queries and
responds to them either directly on the data
query form or on the CRF. The data correction
is then re-submitted to the sponsor for entry into
their database.

Data Transfer:
Electronic CRF (eCRF)
Site records data from source documents to the
electronic database or the web interface
Data periodically electronically transmitted to
Sponsor/CRO or automatically resides in Sponsor
database
Real-time review of data performed by in house
CRAs
Less frequent CRA visits

Electronic queries generated and sent to site

Database lock

Cancer Central Clinical Database

(C3D)
C3D is an integrated clinical trial
information system for the CCR
System is secure, compliant with
regulatory requirements (21 CRF Part 11 )
System is friendly and flexible for user

Cancer Central Clinical

Database (C3D)
Designed to allow integration with the NCI
extramural divisions and the NIH Clinical Center
CRIS (Clinical Research Information System).
Currently this is being done with labs drawn at the
Clinical Center.

Oversight is done by the Control and

Configuration Management Group (CCMG)
whose membership has clinical and IT expertise

C3D Overview
Based on commercial software produced by the
Oracle Corporation called Oracle Clinical (OC)
Allows for Remote Data Capture (RDC) so that
local and remote personnel enter and manage
clinical data over a LAN, intranet, telephone line,
or the Internet
Data can be electronically transferred to
Sponsors (responsibility of DM IT team)

C3D Overview
A template set of master CRFs have been
created to collect the data required by
CCR protocols
Templates are reused and each study will
only use the eCRFs that are appropriate
and required for that study
Confidentiality statement signed at time of
training

J-Review
J-Review is a software product that allows us to
get data out of C3D into a variety of reports
Numerous template reports have been
developed including:
Adverse event summary
Demographics
Drug administration

Also allows for customized reports

C3D eCRFs Resources

C3D Data Entry
Manual for the Completion of the
NCI/CCR/C3D Case Report Forms
Access to J-review is granted once
training occurs.

C3D Protocol Build Process

OCD determines if a protocol will be
built in C3D

Currently the following are built:

All CTEP-sponsored, non-cooperative
group trials
All industry-sponsored trials with company
agreement (if not, sponsor will then
provided paper crfs)
All internal/non-sponsored interventional
trials

 Clinical Analyst (CA)

receives protocol from
IRB
CA identifies standard
eCRFs to be used
CA develops the eCRF
book and identifies if
new eCRFs are needed
CA meets with research
team to confirm eCRF
book

$$

C3D Protocol Build Process

Control & Con

Protocol
Receiving

Requirement
Specification

Initiation
Meeting

Fo

Clinical Analyst Clinical Analyst Clinical Analyst

Research Team P
(PI, RN, DM)

 C3D Protocol Build Process

Control & Configurations Management Group (CCMG)

$$

Clinical Programmers
(CP) build protocol
$
(eCRFs) in C3D
Requirement
Initiation
Protocol
Research
team tests
Specification
Meeting
Receiving
the build/enters data
Modifications made
as needed
Protocol activated in
Clinical
Analystby
Clinical
Analyst Clinical Analyst
C3D
CA/CP
eCRFS available for
data entry

Forms & Rules

Building

Forms & Rules

Testing

Clinical
Programmers

Team

Modification
Activation

Clinical Analyst
Clinical Programmers P
Team

Signoff

 If a protocol
amendment requires
changes in C3D (e.g.
eligibility criteria),
CA/CP will develop
new eCRF
Team will review,
sign-off
CA/CP will activate
new eCRF Book

$$

C3D Protocol Build Process

Control & Configurations M

Protocol
Amendment

Update
New Forms & Rules
Requirement
Building
Specification Activation of New Forms

Clinical Analyst Clinical Analyst

Clinical
Programmers

Training
There is specific training required for use
of C3D and I-review.
See Training Sessions for date, time and
location.

Industry Sponsored Queries

Sponsor generates questions/queries:
During/end of a monitoring visit
After data sent to sponsor and
reviewed/entered in sponsors database

Site corrects CRF:

During/between monitoring visit
May need to also sign-off on query form itself

CTEP Sponsored CTMS

Clarification
These are paper queries generated for
CTEP-sponsored, CTMS-monitored trials
Sent every Monday by Theradex (contractor
for CTEP)

CTEP Sponsored CDS

Rejection/Notification
These are electronic data queries for CTEPsponsored, CDS-monitored clinical trials
CDS submitter receives notice
For studies in C3D, the notification will be sent to the
CCR IT Programmer who transfers the data to CDU

CCR staff corrects data in the database and

resubmits
Process occurs until data is loaded correctly in
CDS

Missing Data at Time of Transfer

Missing data elements
Source Document (SD) not supporting CRF
CRF not supporting SD

Referred to as:
Discrepancies
Queries
Clarifications

Identified by:
Sponsor
Database

Sponsor Queries
Sponsor generates:
During/End of a monitoring visit
After data sent to sponsor and
reviewed/entered in database

Site corrects CRF:

During/between monitoring visit
May need to sign-off on query

Database Discrepancies
Failure of entered data to pass a validation
check as applied by a database
Univariate discrepancy single data
element errors (e.g., not using provided
pick-list, missing data in a field)
Multivariate discrepancy multiple data
element errors (e.g., male patient with +
beta HCG)

Quality Control
According to GCP Section 5.1.3 quality
control should be applied to each stage of
data handling to ensure that all data are
reliable and have been processed correctly.

Assessing the QC/QA Process

Are staff checking their own work?
Are staff relying on others to check their work?
Does the organization have a QA plan for
monitoring protocol adherence and data
collection?
Are there SOPs related to data management?
How soon after a visit is a CRF completed?
Is all data, as defined in the protocol, captured
from the source document to the CRF?

Terminology
Quality Control
Quality Assurance
Quality Improvement

Quality Control (QC)

Ongoing and concurrent review of subject data
Typically 100%
Checking your own work and work of others

Verify that data collected and abstracted:

Correctly entered onto CRF
Able to be found in source document
Follows regulations and guidelines

Individual team member level

Quality Assurance (QA)

Planned, systematic check done at the branch or
organizational level
Verifies:
Trial is performed as per the approved plan
Data generated is accurate

Identifies problems and trends:

Retrospective and involves sampling of subjects and
data
Pulls all the pieces together to gain a picture
(measurement) of compliance
Ensures staff is compliant with internal and external
regulations/guidelines

QA Activities
Internal monitoring/audits
Compile all data components and gain a
measurement of compliance

Clarification monitoring
Assess for trends
Review clarifications responses before they are
submitted to sponsor

Measure data inconsistencies and trends using

a sampling of the data prior to audits/monitoring
visits
Summarize QA findings and report to
management
Identify learning needs

QA Activities for CCR

The following are examples of QA activities
for the CCR:
Office of the Clinical Director (OCD)
Internal monitoring/audits
Conduct audits per upon request, for PI sponsored
studies
Clarification monitoring

Data Management Contractor

Develop QA tools
Summarize QA findings and report to management,
education and training
Identify needs

Quality Improvement (QI)

Result of QC and QA
Developing a plan includes:

Identifying root causes of problems

Intervening to reduce or eliminate these problems
Taking steps to correct the process(es)
Identifying trends and areas for improvement
Identifying solutions:

Assess work flow and time management activities

Develop tools for source documentation
Assess training needs
Involve appropriate staff in resolution

Implementing new/updated solution

QI Activities for CCR

Team Level:
Based on QC activities: identifying trends
Based on audit/monitoring visit results

OCD CCR Level:

Based on audit/monitoring visit results
Guide in implementing processes for making
corrective changes

Responsibilities
Research Team responsibilities
Research Nurse responsibilities
Data Manager responsibilities

Research Team
Ensure that all source data is documented in the
Medical Record/Research Chart with accuracy,
completeness, and consistency
Ensure the overall quality of the research data is
verifiable and acceptable for sponsor
submissions, publications, etc.
Review data discrepancy/clarification resolutions
for accuracy, consistency and timely response

Research Nurse.
Provide accurate and complete source
documentation

Develop, implement, and maintain a team QC

plan:
Establish a schedule of QC activities
Quality check source documentation, data
abstraction, CRFs completion
Quality check of database
Verify function in database

Develop team quality improvement plan, as

needed

.Research Nurse
Lead Team QC meeting:
Provide administrative updates
Provide patient updates
Perform QC on data/resolve issues

Review query/clarification:
Assign to Data Manager(s), if appropriate, to
investigate and resolve or resolve yourself
Review and sign off:
Follow sponsor SOP

Data Manager.
Abstract data onto CRFs according to what is
found in the source documents (Medical Record
or Research Chart) and CRF Instruction Manual
Abstract data in a timely fashion, this includes
entry into database
Code Adverse Events accurately utilizing the
appropriate version of CTCAE, as per protocol

.Data Manager
Apply quality control checks at each stage
of data handling
Ensure that data elements abstracted are
complete and accurate
Contact Research Nurse for missing source data

Resolve discrepant data ongoing

Utilize database report tools to assist with QC
activities

Guiding Principles
Source documents need to be accurate and
complete
Data abstraction should occur in real time
QC/QI is the responsibility of every research
team member
QC/QI should be completed on all protocol data
for all protocols
QC/QI should be proactive and ongoing
Each team member should know and
understand the roles and responsibility of each
team member

Resources
Guidelines for Good Clinical Practice.
International Conference on Harmonisation
(ICH).
[Link]

FDA, Title 21 CFR Part 11

[Link]
r/[Link]?CFRPart=11

Evaluation
Please complete the evaluation form and
fax to Elizabeth Ness at 301-496-9020.

For questions, please

contact Elizabeth Ness
301-451-2179
nesse@[Link]