Vaccines in Crisis

1
Vaccines in Crisis

• Challenges in Understanding Efficacy /
  Effectiveness of Vaccination Influenza Vaccine
  Case Study

2
Overview

• What are we trying to prevent?
• How do we know we are preventing it?
• Who should we be targeting for prevention?

3
What are we trying to prevent?

• Cause specific outcomes
• Lab confirmed infection
• Lab confirmed illness
• All cause outcomes
• Clinical illness
• URI
• ILI
• Complications
• P I Hosp / other hosp
• Death

4
Cause Specific vs All Cause Outcomes

• Cause specific
• Does the vaccine work?
• Outcomes specific to influenza
• Cause specific outcomes
• Eg reduction in lab confirmed illness
• Advantage
• Less noise
• Easily understood
• Smaller sample size
• Disadvantage
• Not real world
• Miss some outcomes
• More difficult data to collect

• All cause
• What are the benefits in real life situations?
• Outcomes non-specific but sensitive
• All cause outcomes
• Eg reduction in ILIs / URIs of all causes
• Useful also for CEA analyses
• Advantage
• Include more outcome events
• May be easier data to collect
• May reflect the real world better
• Disadvantage
• Sometimes misunderstood
• Larger sample size
• ?? Acceptable for licensure

5
Influenza Infection , Illness Differing Case
Definitions
Influenza infection
6
Influenza Infection, Illness Differing Case
Definitions
URIs
Influenza Illness
Influenza infection
7
Influenza Infection, Illness Differing Case
Definitions
URIs
Influenza Illness
Influenza infection
ILIs
8
Influenza Infection, Illness Differing Case
Definitions
Influenza Illness
9
Influenza Infection, Illness Differing Case
Definitions
Influenza Illness
Complications (hosp / death)
10

• Influenza Illness
• Only The Tip of the Iceberg

Influenza Illness
- Misery - Absenteeism - Physician Visits -
Antibiotic Use - Hospitalizations - Deaths
Exacerbations of chronic diseases Secondary
infections Other
11
Types of Clinical Studies Relative Strengths

• Studies Using Individual Level Data
• Interventional
• Experimental (trial)
• Quasi experimental
• Observational
• Cohort
• Case control
• Cross sectional
• (Case series)
• Studies Using Population Level Data
• Ecologic Studies

Stronger
Weaker
12
Validity of Inferences

• Measurement
• Reliability validity
• Sampling
• Sample size power
• Sampling scheme sample frame
• Inclusion / participation rate
• Representativeness generalizability to real
  world
• Other
• Bias (eg misclassification, diagnostic suspicion,
  others)
• Confounding / covariates

13
Other Issues In Interpreting Study Results

• The truth is never a specific point estimate or
  found in one study
• Negative studies dont necessarily exclude
  benefit or harm
• Positive studies dont necessarily prove benefit
  or harm
• Sometimes studies are wrong or provide
  conflicting or apparently conflicting results

14
Epidemiologic Criteria for Causality

• Strength of association
• Dose response
• Temporal relation
• Consistency of association across studies
  populations
• Biologic plausibility

15
Outcome / case definition RRR vs ARR

• Typical kinds of outcomes assessed in VE studies
• Cause specific (specific outcomes)
• Infection
• Lab confirmed illness (LC ILI)
• Influenza otitis media
• All cause (sensitive outcomes)
• Clinical illness (ILI)
• Complications
• Otitis media
• Pneumonia
• Death
• Cause specific outcomes provide highest RRR
  because there is less noise
• But this does not mean that the lower RRR seen
  with all cause outcomes means that the vaccine is
  less effective (ie the ARR would be the same if
  it could be measured)

16
Influenza Vaccination Clinical Case Definition,
ARR, RRR (VE of 75)
17
Efficacy vs Clinical Effectiveness of Influenza
Vaccination of Healthy, Working Adults
Clinical Effectiveness 34
Efficacy 89
Adapted from Bridges et al. JAMA 2000 284
1655-63.
18
Examples from the literature

• Cochrane review of influenza vaccination for
  healthy adults, 2004
• Results
• VE lab confirmed flu 70 (56 - 80)
• VE clinical ILI 25 (13 - 35)
• ARR 6
• Work loss decreased by .16 days (.04 - .29)
• Conclusions Influenza vaccines are effective in
  reducing serologically confirmed influenza.
  However they are not as effective in reducing
  cases of clinical influenza number of working
  days lost. Universal immunization of healthy
  adults is not supported by the results of this
  review.

19
US Economic Studies of Influenza Vaccination in
Working Adults
1. NEJM 1995 333 889. 2. JAMA 2000 284 1655.
3. Vaccine 2003 21 2216. 4. JAMA 1983 249
3189. 5. Emerg Infect Dis 1999 5 659. 6. Arch
Intern Med 2001 161 749 7. Clin Infect Dis
2001 33 1879. 8. Ann Intern Med 2002 137 225.
20
Examples from the literature

• Simonsen et al. Impact of Influenza Vaccination
  on Seasonal Mortality in US Elderly Population.
• Ecologic study attempting to make conclusions
  about VE based on aggregate data

21
Effectiveness of Influenza Vaccination for
Reducing All Cause Mortality in the Elderly
22
Disparities in Adult Vaccination by Geography,
Age Race
Disparities by Geography, 2003 flu, elderly
Disparities by Race, 2004 elderly
Disparities by Age, 2003
MMWR 2004 53 1007 NHIS early release Jan Jun
04
23
Death rates among elderly members of 3 US HMOs
by, age and risk subgroups --1999-00
24
Examples from the literature

• Allsup et al. Clinical trial of TIV in 729
  healthy elderly persons 65 74
• No hosp or deaths
• Conclusion influenza vaccination not cost
  effective for this group

25
20th Century Annual US Morbidity vs 2002
Morbidity from VPDs
26
Influenza is 1 Cause of US VPD DeathsVPD Cases
Deaths, US 1989 - 1998
MMWR 2001 48 (RR-53) Thompson et al. JAMA 2003
289 179 Feikin DR, et al. Am J Public Health
2000 90 223-9.
27
VPDs Take the Highest Death Toll Among Adults
Annual VPD Deaths

• Adults 99 of VPD Deaths
• 30,000 to 70,000 deaths each year
• Children 1 of VPD Deaths
• 100 to 300 deaths each year

Source CDC, IOM
28
ACIP Recommendations 2005-06

• Highest Priority Groups
• High risk for serious complications
• Age 65
• Chronic medical conditions
• Conditions that compromise respiratory function
• Residents of LTCFs
• Pregnant women
• Children/adolescents on chronic ASA Rx
• Children 6 to 23 months of age
• Likely to be high risk (ages 5064)
• Persons who can transmit to high risk groups
• Special emphasis on HCWs
• Others ie, anyone wishing to avoid influenza
• Tiering recommendations to be published
  separately for use with vaccine shortage

CDC. MMWR 2005.
29
Influenza Vaccination Rates for US Adults are at
a Plateau
2010 goal
Source www.nccd.cdc.gov/brfss
30
Distribution of Federal Funding for Vaccination
Costs of Vaccination Operations Infrastructure
31
Distribution of VPD Deaths and Federal Funding
for Immunizations
32
Critical Factors for Successful Control of VPDs
thru Vaccination
Patient
Provider
Successful Vaccination
Vaccine
Policy
33
Lessons from Influenza Vaccine Case Study

• Understanding what we are preventing
• Interpreting the evidence appropriately
• Addressing barriers for adult vaccination as we
  have done for children