The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention

1. Evaluation of Media Reports as a DataSource for
Death During Hurricane Sandy — United States, 2012
Olaniyi Olayinka, MD, MPH
EIS Officer, Health Studies Branch
Disaster Epidemiology Community of Practice
April 17, 2014
National Center for Environmental Health
Division of Environmental Hazards and Health Effects

2. Background
 October 29,2012
 Hurricane Sandy
 Category 2 on landfall
 Storm surge
 Power outage

3. Public Health Impact of Hurricane Sandy
 Public health
o Deaths
o Severe morbidity (injuries,
disabilities)
o Mental health
 Economic loss
o $68 billion

4. Current Problems with Reporting Disaster-Related Mortality
 There is no national-level active surveillance system for disaster-
related fatalities
 There is a time-lag in reporting by state vital statistics
 Death certificates do not always indicate disaster-relatedness

5. Description of CDC/HSB Media Report Pilot Surveillance
to Track Hurricane Sandy-Related Deaths
 From October 29–November 5, 2012 CDC/HSB tracked Hurricane Sandy-
related deaths using Google search engine to search the Internet for
demographics and circumstance of death
 Keywords
 Removed duplicates for deaths reported from multiple sources
 Reported actively tracked Sandy-related deaths to CDC EOC every 24 hrs.
o Death
o Disaster
o Drowning
o Hurricane Sandy
o Memorial
o Sandy or
o Storm

6. Objectives of Evaluation
 Assess CDC/HSB media report tracking for active mortality
surveillance during Hurricane Sandy
 Establish
o Accuracy
o Usefulness

7. Stakeholders
 One city and eight state health departments
o New York City (NYC) and Connecticut (CT), Maryland (MD), New Hampshire
(NH), North Carolina (NC), New Jersey (NJ), Pennsylvania (PA),Virginia (VA),
and West Virginia (WV)
 Media
 CDC/HSB
 Local disaster management officers

8. Evaluation Design
In 2013, we requested vital statistics records or line lists of all
Hurricane Sandy-related deaths from NYC and eight states
1. Requested vital statistics offices to identify Sandy-related deaths using
the following:
 ICD-10-CM (X37) code
 Text string searches using keywords (e.g.,“Hurricane Sandy”or“storm
related” to identify deaths)
2. If a state could not provide a comprehensive list of Sandy-related
deaths, we requested vital statistics records based on the media list we
compiled

9. Evaluation Design (Continued)
 Compared data based on media reports with information
based on vital statistics records
• Determined percent agreement for all deaths that were able to
be match
• Considered a match between media-reported death and vital
statistics if first and last name matched or sex and one of the
following key attributes matched with vital statistics data:
o Age (+/- 1 year)
o Date of death
o Place of death (i.e., state,city,county,borough, or specific place of
death)
 Calculated sensitivity and positive predictive value (PPV)

10. Flow Chart of Process Used in Selecting Deaths Reported in Media
for Comparison with Vital Statistics
(N=75)
115 Media-reported deaths
from all states
99 Media-reported deaths
from jurisdictions participating in
evaluation
(CT,NC,NJ,NYC,PA,VA)
75 deaths reported
from media matched with vital statistics
records
16 media-related deaths from
non-participating states (WV,MD,
NH) excluded
24 deaths excluded:
• Did not match a vital statistics
record (n=18)
• Were outside study area (n=5)
• Were outside study time period (n=1)

11. Number of Hurricane Sandy-related deaths reported by the media
and vital statistics, NYC and five Northeastern states
(CT, NC, NJ, PA, VA)
(N=75)
CT
n (%)
NC
n (%)
NJ
n (%)
NYC
n (%)
PA
n (%)
VA
n (%)
TOTAL
Media
Reports
5 (5) 4 (4) 24 (24) 47 (47) 17 (17) 2 (2) 99
Vital
Statistics
5 (6) 2(2) 23 (26) 41 (46) 17 (19) 2 (2) 90
Matched
n (%) 5 (100) 2 (50) 19 (79) 34 (72) 13 (76) 2 (100) 75 (76)

12. 0 10 20 30 40 50 60 70 80 90 100
Cause of Death
Date of Death
Circumstance of Death
Place of Death
Age
Sex
Name
Percent Agreement of Hurricane Sandy-Related Deaths
reported by media and vital statistics death records in NYC
and five Northeastern states (CT, NC, NJ, PA, VA)
(N=75)
Percent Agreement
KeyAttributes

13. Sensitivity and Positive Predictive Value
New York City Vital Statistics Record
Yes No
Yes
No ___
34 7
7
41
 Sensitivity = 34 / 41 = 83%
 Positive Predictive Value = 34 / 41 = 83%
41
ReportedintheMedia

14. Limitations
 Keywords used in conducting text searches of vital statistics death
records were not exhaustive
 PPV calculated only for NYC,not able to calculate for other five
states
 Media-reported deaths did not have unique identifiers other than
the key attributes (e.g.,name,age,sex) thus accuracy of some key
attributes might be overestimated

15. Conclusions
 Media reported deaths provided timely information on Hurricane
Sandy-related deaths
 Information provided by media was moderately sensitive and for
many of the key attributes there was high agreement
 If deaths are not actively tracked during a disaster it can be
difficult to identify deaths that were related to the event

16. Recommendations
 Use media death reports as supplemental source of information on
disaster-related deaths
o Situational awareness
o Immediate public health decision-making during the initial stages
following a disaster
 Use more traditional sources of information when more accurate
information is needed, such as specific details of cause of deaths

17. National Center for Environmental Health
Division of Environmental Hazards and Health Effects
The findings and conclusions in this report are those of the authors and do not necessarily represent the
official position of the Centers for Disease Control and Prevention.
Acknowledgment
 CDC Operation Dragon Fire Workgroup
• Brant Goode
 Participating City and States
 Connecticut State Vital Records Office
 New Jersey Office of Vital Statistics
 New York City Office of Vital Records
 North Carolina Vital Records
 Pennsylvania Department of Health
 Virginia Office of Vital Records
 CDC/NCEH
 Tesfaye Bayleyegn
 Amy Wolkin
 Sherry Burrer
 Lauren Lewis
 Nicole Nakata