Guidelines

1
Immunization coverage data Collection,
interpretation and disseminationPresentation to
the GAVI Joint Board Meeting28 Nov 2006Dr Rudi
Eggers, WHO/IVB
2
Outline

• Challenges and examples
• WHO/UNICEF Joint Reporting Form
• WHO/UNICEF Routine coverage estimation
• Use and dissemination of coverage data
• Other data available

3
The need of global partners

• Getting high quality data
• Reasonable accuracy
• Reasonable completeness
• Reasonable timeliness
• Confidence in the estimation system
• Validation by occasional surveys
• Participation in the process

4
Challenges
Periphery
5
What does it take to do monitoring?
AFP case investigation ? Include measles
6
Challenges
National
7
Outline

• Challenges and examples
• WHO/UNICEF Joint Reporting Form
• WHO/UNICEF Routine coverage estimation
• Use and dissemination of coverage data
• Other data available

8
Overview
9
Overview
10
Joint Reporting Form 2006

• 25 pages including instructions
• 1057 data elements, including
• routine coverage
• annual incidence of selected VPDs
• nationally recommended immunization schedules
• sources of vaccines
• immunization coverage
• immunization system indicators
• supplementary immunization activities.

11
Timeline of JRF data for 2005 received in HQ
1 June 147 countries
20 September 190 (99) countries
Deadline for Data in HQs 15 May 100 countries

• Countries not reported
• USA
• Japan

Data in HQ as of 24 November 2006
WHO/UNICEF data reconciliation and estimation
process
12
When was JRF received in HQ for GAVI countries?

• By deadline - 15 May 54 countries
• By end of May 14 countries
• Bolivia, Cuba, Haiti, Djibouti, Pakistan,
  Somalia, Sudan, Yemen, Bosnia Herzegovina,
  Honduras, Afghanistan, Nicaragua, Uzbekistan,
  Albania
• Between 1 and 30 June 4 countries
• Laos, Mongolia, Armenia, Guyana
• After 1 July 2 countries
• Azerbaijan, Solomon Islands

GAVI countries better than average
13
Overview
14
Initial data Believe it or not?

• DTP3 coverage increased from 28 to 68 in Sierra
  Leone between 1997-8
• OPV3 coverage in Kenya from 1996-1998 77 -
  36 - 64
• No data available from Norway, Denmark?
• Or 98 measles coverage in Iraq in 1998?
• OPV3 dropped from 82 to 33 between 1996-7 in
  Togo
• 96 DTP3 coverage in Bangladesh in 1999?
• 92 measles coverage in China in 1999?

15
Why do we need coverage estimates?

• Service delivery
• How well are routine immunization services
  functioning?
• Characterize cause of poor performance (low
  access? poor management? Both?) and appropriate
  remedial action
• Use data to strengthen immunization services
  where they are weak
• Epidemiologic indicator
• What proportion of infants are vaccinated through
  routine immunization service delivery?
• Equity
• How equitable of health service deliveries (by
  district)

16
Background WHO/UNICEF Estimates

• Begun in 1999, methods were reviewed, approved,
  and first released in 2001, updated annually
  since 2001.
• WHO UNICEF joint endeavour. Produced as part
  of the WHO UNICEF annual review of national
  immunization coverage.
• Country-specific for 193 countries and
  territories.
• Estimates are made for routine coverage
• 1980 2005 - BCG, DTP1, DTP3, Polio3 measles
• 1985 2005 - HepB3
• 1990 2005 - Hib3
• Constitute an independent assessment - In many
  cases uses data officially reported but estimates
  are not necessarily approved by member states.

17
Annual review of coverage data

• National reports to WHO UNICEF - JRF
• Administrative coverage data
• Country official estimates
• Published and grey literature
• DHS, MICS, other surveys
• Additional information
• Stock out information
• Data quality audit results
• UNICEF supply division data
• Expert opinion/local knowledge (consultation with
  RO/country focal point...)

WHO/UNICEF estimates of national immunization
coverage
18
WHO/UNICEF Estimates of National Immunization
Coverage is used

• MDG 4 reduction of child mortality, indicator
  15 Percent measles immunization coverage.
• 2001 UN Special Assembly on Children to report on
  progress established by the 1990 World Summit for
  children.
• UNICEF State of the World Children.
• WHO World Health Statistics / World Health
  Reports.
• GAVI Progress report
• Other uses ? data are publicly available.

19
WHO UNICEF Estimates Process

• Prepare draft estimates
• Working group (WHO UNICEF) meet for 1-2 weeks
  in June.
• External guest invites beginning in 2005
• To broaden the consultation
• To encourage better use among partners
• Consolidate and reconcile reported data.
• Review all data (reported, published surveys,
  etc) country by country.
• Make draft estimates, update data base.
• Prepare document for national review.

20
Overview
21
WHO UNICEF Estimates Process

• National consultation and review
• 2 months - June-July
• Allows countries to elaborate on the data
• Not an "approval" process

22
WHO UNICEF Estimates Process

• Finalization
• 1 week, working group.
• Review national replies. If necessary, revise and
  update data base.
• Prepare estimates for publication dissemination
• Global regional estimates
• Spreadsheets web pages
• Publications
• Release
• End of August

23
Rules of Estimation

• Evaluation of data no statistical model no
  formula
• Look at patterns of across years and antigens
• Changes from year to year
• Differences between antigens
• Country specific no "borrowing" from other
  countries

24
Rules of Estimation

• Estimate reported data if reported data are
• Consistent with quality survey results (/- 5
  points)
• Consistent across years (no sudden, unexplained
  changes).
• Consistent between vaccines (DTP3 OPV3).
• No other data are available.
• If data are inconsistent select most "likely"
  value.
• 100 vaccination coverage not achievable
• Include private sector (becoming increasingly
  important)

25
Example Burkina Faso Before / After survey
results
26
Example Burkina Faso Before / After survey
results
27
Reported / estimated coverage DTP3
1995
2005
2000
Data source WHO/UNICEF coverage estimates
1980-2005, as of August 2006
28
Relation between reported and estimated DTP3
coverage1990,1995, 2000,and 2005
Data source WHO / UNICEF coverage estimates
1980-2005, as of August 2006
29
Overview
30
Data dissemination
http//www.who.int/immunization_monitoring/ http/
/www.childinfo.org/areas/immunization/
31
Slide Date September 06
Global DTP3 Immunization 1980-2005
Source WHO/UNICEF estimates, 1980-2005, as of
August 2006 192 WHO Member States.
32
Slide Date November 06
Global Immunization 1980-2005, DTP3 coverage
33
Number of countries introduced HepB vaccine and
global infant HepB3 coverage, 1989-2005
Slide Date August 06
excluding countries where HepB administered
for adolescence
Source WHO/UNICEF estimates 1980-2005, as of
August 2006 and WHO/IVB database, 2006 192 WHO
Member States.
34
Countries with most unvaccinated children DTP3,
2003-2005 - in millions ( coverage)
Source WHO/UNICEF coverage estimates 1980-2005,
August 2006 Date of slide 11 October 2006
35
Outline

• Challenges and examples
• WHO/UNICEF Joint Reporting Form
• WHO/UNICEF Routine coverage estimation
• Use and dissemination of coverage data
• Other data available

36
Immunization data collected in WHO

• Regular systematic programme monitoring data
  collection
• Annual (WHO-UNICEF Joint reporting Form on
  Immunization (JRF))
• Beyond coverage
• Disease surveillance data
• From case-based rapid AFP and measles data to
  annual aggregate data on diphtheria etc
• Modelling disease burden and mortality estimates
  
• Ad Hoc data collection
• By special request
• From reports / other sources
• Secondary data use
• Population data from UNPD
• Development indicators (WB)

37
WHO Disease Burden Estimation Process

• Hib / Pneumo disease burden estimate
• 2.5 years
• 2 Expert Review panels
• 58 people
• gt400,000

• Database of evidence
• Systematically collected
• Publicly available
• Methods for estimation
• Transparent methods very complicated
• Manuscript prepared for peer-review
• Communication of uncertainty of estimates
• Independent expert group
• Clearance through WHO-EIP
• Compatibility with other disease burden estimates
• Country consultation prior to release of
  country-level estimates

38
Where would we be without the data?

• Inefficient use of financial/human resources
• Inability to fine-tune programme direction
  ("flying blind")
• Inadequate measurement of vaccine impact
  (excessive reliance on modeling)
• Inadequate data to estimate burden of VPDs
• Lost opportunity to build national and regional
  capacity (eg, bacterial diseases)
• Inadequate detection of VPD outbreaks
• Inadequate capacity for detection of new and
  emerging agents

39
What could be done to improve the estimates?

• High quality data could improve the estimates.
• Many countries have not validated their routine
  systems for several years. Surveys or data
  quality self-assessments (DQS) could provide
  important information on the validity of national
  coverage data.
• Improved consultation.
• While we have had several opportunities to
  conduct detailed reviews of the data and
  estimates with national officials the major
  mechanism for consultation remains through
  correspondence. We would like to conduct
  regular reviews workshops with national and
  partner staff.

40
End