ProblemWorried Well Effect
The Question
How to decrease ‘noise’ to find true burden of disease? How to make sense of syndromic data?
ChallengesHealth seeking behavior change
Increase burden of illnessBaseline illness (Seasonal influenza and Novel
Influenza A (H1N1)Geographic differencesMedia effect: Clinton effectFear
Acute Care Visits for ILI: Boston, 2005-9
1/21
/200
6
2/25
/200
6
4/1/
2006
5/6/
2006
6/10
/200
6
7/15
/200
6
8/19
/200
6
9/23
/200
6
10/2
8/20
06
12/2
/200
6
1/6/
2007
2/10
/200
7
3/17
/200
7
4/21
/200
7
5/26
/200
7
6/30
/200
7
8/4/
2007
9/8/
2007
10/1
3/20
07
11/1
7/20
07
12/2
2/20
07
1/26
/200
8
3/1/
2008
4/5/
2008
5/10
/200
8
<2y2-4y5-17y18-44y45-64y65+yWestern city A<2y2-4y5-17y18-44y45-64y65+yMidwestern state A<2y2-4y5-17y18-44y45-64y65+ySoutheastern state C<2y2-4y5-17y18-44y45-64y65+ySoutheastern state B<2y2-4y5-17y18-44y45-64y65+ySoutheastern state A<2y2-4y5-17y18-44y45-64y65+yNortheastern city B<2y2-4y5-17y18-44y45-64y65+yNortheastern city A<2y2-4y5-17y18-44y45-64y65+yCanadian city A
DiSTRIBuTE Visualizations - Week 2008-21 (ending Saturday, May 24, 2008)
Surface plots depict relative increase in ED syndrome visits as observed / baseline by jurisdiction and age.
Relative increase (observed/linear-baseline).
Region and Age (y)
Date
Ideas Readjust baselines using previous “worried well”
periods (i.e. SARS or other major localized health scares) as a way to distinguish between true increase and shift in health seeking behavior.
Severity adjustment – drop in % ED visits being admitted
Objective clinical criteria – EHR surveillance, subjective vs. objective fever
Ratio of ED visits to 911 dispatches as a way to measure degree of worried well effect
Ideas Adjustment of syndrome definitions: ‘Respiratory +
Fever Reason for Visit’ or ‘Respiratory + Measured Temp > 99.9˚F.’ The BPHC separated these two definitions and found a broadening divergence as subjective fevers climb and objective fevers remain steady.
Auxillary real-time surveillance systems, i.e. school/workplace absenteeism
Observing other syndromes simultaneously: Are GI complaints less susceptible to worried well effect?
Contacts• Settings
– Schools– Work places– Cruise ships and Airplanes– Homeless shelters, prisons– Health care settings
• Identify persons with the most risk of progress to disease
• PEP strategies
ChallengesLimited data:
Location – where people satTime – length of time on a flight, hours in the
classroomQuestion of underlying illnessPEP effect: Treat the whole schoolPPE effect: N95 respirator vs masks
What have we learned from pertussis and TB
ModelsClosing schools –
all schoolssome schools with high rates of illnessnone
What is the effect of PEP and school closingstreat high risk family contactstreat high risk school contactsOthers
Building SyndromesHow can we use multiple syndromes to better
understand H1N1 activity?
ChallengesGi vs respiratory
Reports of Gi illness with H1N1 Combine the syndromes Keep them separate Sensitivity, specificity, positive predictive value
What is the impact on the various statistical models
Top Related