Clinical Trials and Safety Data

Download Report

Transcript Clinical Trials and Safety Data

Clinical Trials and Safety Data
–
Different View of the same
Reality
Uwe Trinks, Ph.D.
PRISM Forum, San Diego
Clinical Data Management vs. Clinical Safety












Goal: Profile Drug for Use
Longitudinal Data Collection
Patient and Visit driven
Large Part of Analysis after
LPLV
Mostly placebo-controlled,
blinded
Combined with other studies
and analyzed
No expedited reporting
No annual reporting
Goal: Prevent Harm
Situative Data Collection
Event driven
Immediate Follow up and
Analysis
 Mostly unblinded data
 Expedited Reporting
 Annual Reporting
 Reporting to Ethics Committees
and Investigators
 Reports on paper or as ICH
E2BM dataset
 Data reported at the end of
product development in a
CTD as CDISC SDTM
dataset
2
Efficacy vs. Safety
Efficacy
Safety
3
Looking for the Outlayer
Come on! It can‘t go
wrong every time...
4
Technical Hurdles
5
A Study
Patient Recruitment
Randomization
FPFV
LPLV
Data Cleaning, QC, Data Management
SAE
6
SAE
SAE
Reconciliation
SAE
Data Pooling
Analysis
Documentation
Data
Unblinding
Data Cleaning
Reconciliation
Visits
SAE
DB
Lock
The old clinical trials paradigm
CRF
Study Sites
SAE
CTMS
CTS
Reconciliation
7
Safety
The brave new world of EDC
Study Sites
EDC
EDC
EDC
EDC
EDC
EDC
CTMS
CTS
Safety
8
Patient View vs. Event View
9
SDTM Clinical Domains by Class
Interventions
Findings
Events
Other
Exposure
AE
Labs
InclExcl
Demog
ConMeds
Disposition
Vitals
SubjChar
RELREC
Supp Qual
Subst Use
MedHist
PhysExam
QS
ECG
Other
Domains most relevant to safety review
10
Comments
Trial Design
ICH ICSR DTD-XML
 Message Header
 M.1. Message Identifier

Type, Format, Version, Number, Sender ID, Receiver ID, Date
 Case Admin Data
 A.1. Identification of the case safety report
 A.2. Primary source(s) of information
 A.3. Information on the sender and receiver of the case
safety report
 Case Information
 B.1. Patient Characteristics
 B.2. Reaction(s)/Event(s)
 B.3. Results of tests and procedures relevant to the
investigation
 B.4. Drug(s) information
 B.5. Narrative Case Summary and further information
11
SDTM Clinical Domains by Class
Findings
Events
Interventions
B4
B2
B3
Other
B1
Exposure
AE
Labs
InclExcl
Demog
ConMeds
Disposition
Vitals
SubjChar
RELREC
B4Supp Qual
B4
Subst Use
MedHist
PhysExam
QS
ECG
Other
Domains most relevant to safety review
12
Comments
Trial Design
B3
What suits one system might not suit the next...
13
The ideal Process
Dictionaries
Codelists
Studies
Query
Trigger
E2B
Plus
file
Inform
Argus
Interchange
SAE
form
Diff
Report
14
Argus
Safety
The real Problems
15
The Problem
 Trigger Points
 EDC data collection is continuous




Safety Collection is a “Snapshot”





Study Sites update information almost daily
Mostly without thinking about SAE
Most prominent updates: Death Dates
What fields and when should trigger a follow up
Avoid having up to 40 FU reports
Which labs and vital signs to transfer?
Update of non-significant information?
Feedback Loop



EDC System must mark transferred data as such
Requires redesign of SAE form and SAE process
Requires standardization of certain form design elements
16
The Problem
 CDISC SDTM to E2BM Conversion difficult
 Field inconsistencies


Need to create a mapping routine (ETL tool)
Easier to create the E2B file directly
 E2B Intake
 Incomplete E2B form



Requires separate profile and DTD
Need E2B difference report
Harmonize Configuration



Need to be able to automatically find Study Information in
Argus
Population of Suspect Drug etc. from Safety System not from
EDC System
Requires quick configuration of Studies
17
Coding
 Medical Semantics
 Vary widely
 Are usually not harmonized
between Safety and Clinical
 Coding in Safety is
 Real-time
 Most current MedDRA version
 Important for Reporting
 Coding in Clinical is
 Harmonized at the end
 Used only for Analysis
 Two Possibilities
 Harmonize Coding Strategies or
 Only transport Verbatim
18
Major Changes in Inform
 Standardized SAE form
 Zero deviation allowed
 Additional fields for




Lab Data
Vital Signs
Event Assessments
Dechallenge/Rechallenge
 Selection Tables for Data that are uploaded by
relevance





Concomitant medications
Medical History
Vital Signs
Patient Demographics
Treatments
19
Major Changes in Inform
 Pop ups in other Inform Sections
 “Is this an SAE/Update to SAE?”
 Need to be “cloned” for EVERY database
 Death Date





Other Seriousness Criteria
Change in Relatedness
Change in Drug Exposure


Cause of Death
Relatedness -> Update SAE form
Dechallenge/Rechallenge
Pregnancy
 System Down: Fax to Safety, but enter into EDC
System
20
Major Changes in Inform
 Standardized SAE Query form
 Assure that answers end up in SAE form
 Lab Data Integration difficult
 Standardized and harmonized Codelists
 Biggest source of errors
 Need full alignment of CRF writers
 Can probably build an E2B check into Inform
 Trigger Mechanism
 Difficult balance between




Regulatory Requirements
Operational Efficiency
Basic rule: No more than one FU per day
Scheduled upload of non significant information
21
Major Changes in Argus
 Study Information
 Needs to be configured before FPFV
 Needs to be harmonized with Clinical
 E2B Import Module
 Needs to be configured for partial E2B files
 Added non E2B fields





Sequence numbers
Event Assessments
Configure for Always Accept
Configure to not overwrite certain fields (e.g unblinded
product information)
Configure not to upload certain events (e.g entry criteria)
 No changes to Argus SAEs
 Except Narrative and Unblinding
 All Changes/Queries through EDC System
22
Communication
 Most difficult task
 Lots of Clinical and IT lingo
needs to be translated into
Safety Lingo
 Lots of Safety lingo needs
to be translated into Clinical
and IT Lingo
 Very few people are bi- or
tri-lingual
23
E2B Transport
 Non native DTD
 Partial E2B form
 Header of non-standard format
 No study product information
 Only Study # information
 No Narrative Information
 Added fields
 Sequence numbers
 Event Assessments/Causality
 Non overwrite information
 No real acknowledgement
 Send Argus Case number back to Inform for
sequencing and FU
24
The good news
 It works very well at some companies
 GSK
 Technical problems are mostly solvable
 It saves time and money once it works
25
The bad news
 It takes an effort to harmonize things
 SAE forms
 Codelists
 Dictionaries and Coding Conventions
 It takes a while to train people
 Safety to use Inform for Queries and Updates
 Investigators to “think safety”
 It requires adherence to agreed standards
 The “extra codelist item” will break the process
 Safety has to conform with not updating data in Argus
 Savings will happen late in the process
 Once the majority of the studies is in Inform and using
the standard SAE process
 Can’t change studies midway
26
Don’t forget
27