Download Collection and Analysis of Drug Safety Data in

Collection and Analysis of Drug
Safety Data in Pregnancy
Myla Moretti
Motherisk Program
The Hospital for Sick Children
Toronto
Aims
• to discover possible adverse pregnancy
events related to maternal exposure
during pregnancy
• to identify signals and establish risks early
• to disseminate this information back to
patients and health care providers so that
rationale choices about disease
management in pregnancy can be made
But why?
• drugs are RARELY evaluated in human
pregnancy prior to market release
– pregnant women are actively excluded
• 50% of women will take at least 1 drug in
pregnancy
• patients may require continued
pharmacotherapy in pregnancy
1
Where are we now?
• currently in Canada there are no
regulated, standardized methods of
collecting or reporting drug safety data in
pregnancy (mostly voluntary)
• worldwide methods vary, but can include:
– manufacturer registries
– observational studies from research facilities
– database linkage (often governmental)
Issues
• events can occur long after exposure (at
least the duration of a pregnancy)
• even for teratogens, adverse outcomes
are rare
• physiologic changes in pregnancy and
concomitant exposures make evaluating
the data complex
Strategies: The Registry
• usually initiated by manufacturers (can be
done in house or outsourced)
• spontaneous reporting
• often retrospective (reporting bias)
• variable data quality
• possible duplication of results
• no consistent methodology
2
US FDA/CDER Document
•
•
•
•
•
•
•
very comprehensive
when, where, how
active recruitment
sources of follow-up data
how to select comparison groups
data collection and minimum data sets
sample size
Manufacturer’s Registries
• since the guidance documents few NEW
registries were voluntarily initiated
• most still do not generate/collect/report
comparator groups
• very little data analysis or interpretation
– limited to summary reports and case counts
• linkage to experts in the field?
• rarely published in peer review literature
– not readily accessible by clinicians
3
Strategies: Database Linkage
• populations where health care and
medications are paid for (government/
insurance)
• duplication
– can occur if there are errors in linkage
•
•
•
•
•
cases can be unlinked if patients move
exposure data not verified (ie.Rx records)
huge data sets
good data interpretation
accessible, peer reviewed
Strategies: Observational Study
• research groups or industry
– cohorts or case control studies
• good data quality within a single centre
• some standardization across different
research centres
• can be arduous and cumbersome without
electronic collection
• smaller sample sizes
• good interpretation of data
• accessible, peer reviewed
What is really needed?
4
Work with existing infrastructure
• a number of centres (research, pharma
and government) have the capability to
provide the expertise to achieve this goal
• Teratogen Information Services and large
obstetric units are ideally situated to
collect the patient data as they encounter
the patients as part of day-to-day
operations
Tool: Electronic Data Collection
• Move towards electronic health records
• Computer usage increasing – storage,
capture, analysis
• Reliance on automation is increasing
BUT
• do it right
• make sure its working
• ensure system is secure, validated, tested
• funding for creation and maintenance needed
Uniform Data Standards
• the existing FDA document and currently
published studies suggest good clear
guidelines for a minimum data set
• patients can be good reporters of their
own health information
• uniform data collected electronically can
more easily be linked to existing
databases (prescription, medicare)
5
The Bottom Line
• the situation is complex
– data that is convincing to one person is
unconvincing to another
• standardization and a defined data set will
promote good data collection
• electronic capture software can be
implemented across various sites, thus
increasing sample size
• link human expertise to interpret data
– statisticians, clinicians, epidemiologists,
teratologists
6