Roundtable: Cutting edge safety systems A global perspective

1
Roundtable Cutting edge safety systems A
global perspective
Vaccine Safety Evaluation Post Marketing
Surveillance Conference Bethesda, MD 10
April,2007 Adwoa Bentsi-Enchill Quality,
Safety and Standards Immunization Vaccines and
Biologicals WHO
2
Outline

• Background
• WHO Programme for International Drug Monitoring
• Overview
• Potential for strengthening vaccine
  pharmacovigilance
• WHO's Global Network for Post-marketing
  Surveillance of New Vaccines
• Selected country examples of new methodologies

3
Background

• Global impact of real or perceived safety
  concerns
• Challenges for ensuring effective safety
  monitoring systems in non-industrialized
  countries
• Relative lack of resources
• Communication, cultural and other local factors
• What would an ideal post licensure vaccine safety
  system be .. globally?
• How can we build more effective systems in
  non-industrialized countries?

4
WHO Programme for International Drug Monitoring

• Network of National Pharmacovigilance Centres,
  WHO HQ, and the WHO Collaborating Centre for
  International Drug Monitoring (the Uppsala
  Monitoring Centre (UMC))
• Established in 1968 and at UMC since 1978
• Jan 2007 82 member countries ( 17 associate
  members)
• Approx. 3.7 million case reports in database
• Only approx. 10 of database relates to vaccine
  reports (82 of vaccine reports from 3 countries)
  M Letourneau, 2005 publication pending

5
Functions of the WHO Programme

• Identification and analysis of new adverse
  reaction signals
• WHO database as reference source for signal
  strengthening and ad hoc investigations.
• Information exchange between WHO and National
  Centres (Vigimed)
• Periodical newsletters, guidelines etc. in PhV
  and risk management
• Tools for management of clinical information (WHO
  Drug Dictionary and the WHO Adverse Reaction
  Terminology WHO-ART to MedDRA mapping)
• Training and consultancy support to National
  Centres
• Software for case report management (VigiFlow)
• Annual meetings for representatives of National
  Centres
• Methodological research

6
Steps in UMC signalling

• Global reporting (from National Centres) of
  safety data to UMC
• Quarterly analysis of data using a Bayesian
  Confidence Propagation Neural Network (BCPNN)
  method
• Generates Combinations Database (all Drug-ADR
  combinations)
• Triage algorithms applied to prioritize
  associations of interest (potential signals) for
  review by clinical "signal reviewers"
• based on an Information Component (IC) value
  unexpected drug-ADR pairs
• based on special terms of interest
• Signal review (including review of literature)
• Publication in Signal document
• Additional pattern recognition to generate
  hypotheses for further study (e.g., specific
  events or ADR risk factors)

7
UMC Signal Detection Follow-up
Combinations.db (reported quarterly)

Triage (filter)
Quarterly analysis BCPNN
Review panel
Vigibase
Pharma Company
Yes
No
Follow-up
SIGNAL
National Centres
Source the Uppsala Monitoring Centre
8
Improving vaccine safety monitoring through the
WHO Programme for International Drug Monitoring

• WHO consultation on improving global AEFI
  monitoring (9-10 Jan 2006)
• http//www.who.int/wer/2006/wer8127.pdf
• Improving AEFI reporting to UMC
• understanding determinants for AEFI reporting
• Improving UMC resources and methods for reporting
  and analysis of vaccine safety data
• Improving advocacy and communication
• Linkages with international vaccine safety
  groups/experts

AEFI adverse events following immunization
9
Global Network for Post-marketing Surveillance
of New Vaccines

• To support WHO vaccine prequalification system
  (supply by UN agencies) with safety data in
  post-marketing phase
• Objectives
• Ensure standardised approach to monitoring
  serious adverse events
• Identify/address potential safety signals in
  timely manner
• Ensure adequate safety information to support
  vaccination policy and recommendations
• may include recommendations for more controlled
  studies

10
Global PMS Network Surveillance methods

• Start with 10-12 countries
• Core/standard PMS guidelines
• adaptation to local resources and systems
• improve use of standardised terminology and case
  definitions (e.g., Brighton Collaboration)
• Case detection
• public AND private sector
• stimulated passive system (based on existing
  systems)
• active sentinel surveillance for selected events
  and vaccines
• monitoring in campaign settings (e.g., JE)
• Strengthen data analysis, investigation
  causality assessment for serious AEs
• reporting to UMC to enhance signal detection

11
Proposed functional structure


UMC (Signal detection)

Network management group Member countries, WHO
(RO HQ), experts/advisors (extended analyses
and specific studies)

Monthly, or ad hoc reporting
Quarterly feed-back
12
A new way to monitor vaccine safety in South
Australia
13
SAVeS study

• First pilot study in Australia to evaluate ?
  technical feasibility ? effectiveness ?
  acceptability
• of linking Commonwealth and State health records
  to monitor vaccine safety.

• Linkage of 1997-2002 data from
• Australian Childhood Immunisation Register
• OACIS
• Womens and Childrens Hospital HAS-ED
• SA passive AEFI reporting system

14
Effective?Demographic data for linking
ACIR Name Date of birth Gender Address Local ID
(eg ACIR 1)
SA passive system Name Date of birth Gender Addres
s Local ID (eg SAVeS 9)
OACIS CRR Name Date of birth Gender Address Local
ID (eg OACIS 5)
WCH HAS-ED Name Date of birth Gender Address Local
ID (eg HAS 31)
Data Linkage Officers in SA Department of Health
15
Effective?Clinical data for analysis
ACIR . . . . . . . Immunisation history Local ID
(eg ACIR 1)
SA passive system Date of admission Date of
discharge . . . . Length of stay Immunisation
history Local ID (eg SAVeS 9)
OACIS CRR Date of admission Date of discharge ICD
principal dx code ICD 2nd ry dx code External
cause code ED presentation code Length of
stay . Local ID (eg OACIS 5)
WCH HAS-ED Date of admission Date of
discharge ICD principal dx code . . . . . Local
ID (eg HAS 31)
Analyst at Curtin University, WA
16
Effective?Sensitivity of SA passive system
Actual events identified through data linkage
Events reported as convulsions identified
through SA passive system
17
Q4. How comfortable are you with data linkage
for general health research?
Omnibus vs providers 2005
18
Q6. Under what circumstances should data linkage
be allowed for vaccine safety research?
Omnibus vs providers 2005 (ranking)
19

• Next steps include
• Plans for assessing national feasibility and
  acceptability of data linkage for vaccine safety
  surveillance.

20
The Italian Hera study on the risk of death due
to unknown or ill defined causes during the first
two years of life

• Centro Nazionale di Epidemiologia, Sorveglianza e
  Promozione della Salute
• Istituto Superiore di Sanità
• Italy

21
Italian Hera study safety of hexavalent vaccine

• 2003 signal of possible association between
  hexavalent product and sudden unexpected deaths
  raised in Germany
• Hexavalent vaccines widely used in Italy since
  2001
• Coverage in Italy for three doses (ICONA 2003, on
  a sample of 4,602 children from the 2001 birth
  cohort Ciofi degli Atti, 2005)
• 78 by 12th month 95 by 24th month
• Case series method used to evaluate risk of death
  due to unknown or ill defined causes during the
  first two years of life following vaccination

22
Objectives

• To evaluate whether the incidence density of
  unexpected deaths during the first two years of
  life is higher in periods of time close to
  vaccine administration (i.e. vaccinations are
  associated with an increased short term risk of
  SIDS-SUD)
• To compare the incidence density following the
  administration of one of the two available
  hexavalent vaccines, and to compare hexavalent
  vaccines with products previously used in Italy

23
Study design and analysis

• Five annual cohorts of newborns in Italy (
  500,000 newborns per year) 1 Jan 1999 - 31 Dec
  2003
• Only vaccinated cases included in analysis
• For each case, the observation period following
  each vaccine dose was divided into
• Three risk periods following vaccination were
  considered 0-1 0-7 and 0-14 days
• Control period between the 14th day and the
  subsequent vaccination or death
• Confounding factors adjusted by age (80 81-100
  101-120 121-180 181-360 360 days)

24
Computation of person-time during at risk- and
control-periods
Classes in days 31-80 81-100 101-120 121-180
181-360 360
P-T stratified by - class of age -
vaccine Other available covariates - mother
citizenships - place of death (home or
not) - calendar year - last dose
Risk period
Risk period
Death
Any vaccine
Control period
Control period
Risk period, Vac Y
Risk period, Vac X
Death
By vaccine
Control period
25
Conclusions

• No increased risk
• for 0-1 days following vaccination
• by dose
• for 2nd year of life
• For the risk periods 0-7 and 0-14 days most RR
  estimates were slightly greater than 1 with
  non-significant confidence intervals
• At 14 days both hexavalent vaccines had the same
  estimated risk
• The RRs concerning concomitant administration of
  six antigens do not differ from the estimates of
  the two hexavalent products
• Manuscript in preparation for publication

26
How can we ensure "ideal post licensure vaccine
safety systems" in the WHO Programme?

• Improving reporting by National Centres to UMC
  and use of UMC resources and methods for
  reporting and analysis of vaccine safety data
• Networks of non-industrialized countries for
  post-marketing surveillance (incl. data linkage,
  epidemiological studies etc.)
• Linkages with international vaccine safety groups
  and experts
• Capitalize on experiences from advanced
  monitoring systems
• Subgroup of Global Advisory Committee on Vaccine
  Safety established in Jun 06 to advise on global
  AEFI monitoring
• CIOMS/WHO Working Group on Vaccine
  Pharmacovigilance

27
Acknowledgements

• the Uppsala Monitoring Centre/WHO Programme for
  International Drug Monitoring
• Mike Gold, Sarah Dugdale et al (South Australia
  Vaccine Safety Study)
• Stefania Salmaso et al (Istituto Superiore di
  Sanità, Italy)