Transcript Vital statistics system approaches for public health surveillance

Infectious Disease Surveillance
& National/Health Security
Michael A. Stoto
CNSTAT Workshop on Vital Data for National Needs
April 30, 2008, Washington DC
Outline
• “Biosurveillance” for national/health
security
– Detection of bioterrorist attacks vs.
– Situational awareness for public health
emergencies
• Relationship to the U.S. vital statistics
system
– Using mortality data to monitor disease
outbreaks
– Applying vital statistics system ideas
for public health surveillance
Stoto CNSTAT 4/08
Syndromic surveillance
• Biosurveillance / Syndromic surveillance
– Near real-time acquisition and use of prediagnostic health data for monitoring
population health
• Builds on existing data systems
• Health care, ED chief complaints, medication
sales, absenteeism, …NOT vital data
• Usually electronic/networked
• Statistical analyses used to detect changes
•  disease outbreak, especially a covert
bioterrorist attack
Stoto CNSTAT 4/08
20
10
0
Count
30
40
Gastrointestinal, Winter 2003
1 Jan 03
1 Feb 03
A
1 Mar 03
date1
B
C
D
1 Apr 03
E
H
1 May 03
I
Stoto CNSTAT 4/08
20
10
5
1
0
Daily Cases
30
50
75
Gastro: Flu Season 2003
15Jan03
1Feb03
15Feb03
1Mar03
15Mar03
1Apr03
15Apr03
30apr03
Red = Hospital A, Orange = Hospital B, Green = Hospital C, Blue = Hospital D, Pink = Hospital E, Purple = Hospital H, Brow n = Hospital I
Shewhart
Mild
Medium
Ultra
CUSUM
Expo
Stoto CNSTAT 4/08
20
10
0
Count
30
40
Gastrointestinal, Winter 2003
1 Jan 03
1 Feb 03
A
1 Mar 03
date1
B
C
D
1 Apr 03
E
H
1 May 03
I
Stoto CNSTAT 4/08
Statistical issues
• Obtaining and integrating accurate data
quickly from a variety of sources
• Determining when something is “unusual”
– In the presence of highly variable and
possibly unstable background variation
– When there may be other reasons for
changes in the data
• Tradeoffs among
– Sensitivity
– False positive rate (specificity)
– Timeliness
Stoto CNSTAT 4/08
Practical issues
• Patient privacy concerns
– HIPAA regulations, state law, etc.
• Proprietary concerns and turf issues
• Concern about secondary uses of data by others
• Operational costs of sharing data
– including personnel and IT costs
• Information overload
– excess flagging at health department end
• Ability to interpret and respond at higher levels
• Inability to detect small outbreaks
• Sense that utility of SS is not yet proven
Stoto CNSTAT 4/08
Goals and purpose shifting
• From “statistical” approaches
– increase in the number of people with
common symptoms to detect BT events
– cannot detect small numbers of cases, even
if very unusual
• To “situational awareness”
– facilitate physician reporting
– active surveillance / case finding
– aid in outbreak investigations
– real-time monitoring of disease outbreaks
• including, when to cease interventions
– information sharing within public health
Stoto CNSTAT 4/08
Vital statistics for 21st Century
health surveillance?
• Given emphasis on timeliness, can vital
statistics be used for biosurveillance?
– Can Lemuel Shattuck’s 18th Century
approach to monitoring health with mortality
rates be updated to the 21st Century?
Stoto CNSTAT 4/08
Mills et al, Nature. 2004; 432:904-906
Stoto CNSTAT 4/08
Stoto CNSTAT 4/08
Downloaded from CDC 4/26/08
Stoto CNSTAT 4/08
Vital statistics for health
surveillance?
• Given emphasis on timeliness, can vital statistics
be used for biosurveillance?
– Excess P&I mortality
• key to what we know now (perhaps then) about the
1918 flu pandemic
• major current flu surveillance system
– How can modern IT (e.g. E-Vital, EDR) make mortality
data more useful for real-time monitoring?
• increased timeliness
• more geographically representative
– improve accuracy of P&I coding
• include causes of death other than P&I
Stoto CNSTAT 4/08
Vital statistics system approaches
for public health surveillance
• Notifiable disease reporting system
– state & local function with limited national coordination
• mainly what to report
– 19th century concept of operations
• post card  fax, phone, Internet
• CDC Biosense
• Notifiable Disease Cooperative Program
– state & local ownership of case reports for local action
– federal ownership (and standardization and funding) of
statistical data for disease surveillance as a national
security resource
Stoto CNSTAT 4/08