Embed Size (px)
Citation preview
Influenza Mortality Surveillance… Making Real-Time National Mortality
Surveillance a Reality
National Center for Health Statistics
Division of Vital Statistics
Mortality Surveillance Team Lead
NAPHSIS Annual Meeting 2014Seattle, Washington
June 8-11, 2014
Paul D. Sutton, Ph.D.
Pneumonia and Influenza Mortality Surveillance
Partnership with CDC/NCIRD/Influenza Division
Objective: Use the National Vital Statistics System (NVSS) for real-time surveillance of Pneumonia and Influenza (P&I) mortality.
122 Cities Mortality Reporting System
122 Cities Mortality Reporting System – cont.
Incomplete reporting About 25% of all deaths Increasingly difficult to get reports from some cities
Inconsistent reporting Not all jurisdictions follow the case definitions (e.g. some only
report underlying cause of death) Certificate review automated or semi-automated in some
jurisdictions, manual review in other jurisdictions. Deaths reported by week the death certificates
were reviewed (not by actual date of death) Lag between date of death and report is unknown and varies
from place to place Deaths reported only by place the death occurred
Place of death not necessarily the same as place of residence
Future NVSS based Influenza Mortality Surveillance
Improvements Reported by date of death Automated and consistent process for certificate review
based on coded cause of death information Deaths available by place of occurrence and residence
Potential new capabilities More focused regional, state, or local surveillance may
be possible Ability to switch from weekly to daily reporting in a
pandemic
What is it Mortality Surveillance Data?
Final Data Surveillance Data
Based on final/best version of each death record
Based on most recent version of each death record
Based on all deaths Based on some percentage of all deaths
Data files are static Data are continually changing as new records and updates are received
Data released as annual files Data are shared on an ongoing basis with federal public health surveillance partners
Making it a Reality
Reporting lag evaluation Difference between date of death and the date NCHS
receives the cause of death Develop an efficient interface to the data Parallel surveillance for 2013-14 influenza
season Real-time comparison of vital statistics and 122 city
based influenza mortality surveillance Compute new baselines
• Seasonal baseline• Epidemic threshold -- 1.645 standard deviations above the
seasonal baseline
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90 93 960
10
20
30
40
50
60
70
80
90
Completeness of Death Record for U.S. NCHS P&I Multiple-Cause Death Data, by Year, 2011-2013 (thru 7/13/13)
2011 2012 2013
Weeks Between Death Occurrence and Record Completion
% C
ompl
etion
of P
&I D
ata
1/1/2
011
1/29/2
011
2/26/2
011
3/26/2
011
4/23/2
011
5/21/2
011
6/18/2
011
7/16/2
011
8/13/2
011
9/10/2
011
10/8/2
011
11/5/2
011
12/3/2
011
12/31/2
011
1/28/2
012
2/25/2
012
3/24/2
012
4/21/2
012
5/19/2
012
6/16/2
012
7/14/2
012
8/11/2
012
9/8/2
012
10/6/2
012
11/3/2
012
12/1/2
012
12/29/2
0120
2
4
6
8
10
12
Comparison of NCHS Multiple-Cause Cumulative P&I Death Ratio and Final P&I Ratio 1 Week After Week of Death
(r = 0.5704)
Final P&I Ratio NCHS P&I Ratio - Lag Week 1
Weekending Date
% o
f Dea
ths
Due
to P
&I
1/1/2011
1/29/2011
2/26/2011
3/26/2011
4/23/2011
5/21/2011
6/18/2011
7/16/2011
8/13/2011
9/10/2011
10/8/2011
11/5/2011
12/3/2011
12/31/2011
1/28/2012
2/25/2012
3/24/2012
4/21/2012
5/19/2012
6/16/2012
7/14/2012
8/11/2012
9/8/2012
10/6/2012
11/3/2012
12/1/2012
12/29/20120
2
4
6
8
10
12
Comparison of NCHS Multiple-Cause Cumulative P&I Death Ratio and Final P&I Ratio 2 Weeks After Week of Death
(r = 0.9015)
Final P&I Ratio NCHS P&I Ratio - Lag Week 2
Weekending Date
% o
f Dea
ths
Due
to P
&I
1/1/2011
1/29/2011
2/26/2011
3/26/2011
4/23/2011
5/21/2011
6/18/2011
7/16/2011
8/13/2011
9/10/2011
10/8/2011
11/5/2011
12/3/2011
12/31/2011
1/28/2012
2/25/2012
3/24/2012
4/21/2012
5/19/2012
6/16/2012
7/14/2012
8/11/2012
9/8/2012
10/6/2012
11/3/2012
12/1/2012
12/29/20120
2
4
6
8
10
12
Comparison of NCHS Multiple-Cause Cumulative P&I Death Ratio and Final P&I Ratio 3 Weeks After Week of Death
(r = 0.9146)
Final P&I Ratio NCHS P&I Ratio - Lag Week 3
Weekending Date
% o
f Dea
ths
Due
to P
&I
Mortality Data NCHS had Received and Coded as of 3:00 AM Last Sunday
(6/8/2014)
Date of Death(Week
Ending)Record Count
% of Expected*
6/7/2014 1,707 4%5/31/2014 10,161 22%5/24/2014 15,519 33%5/17/2014 17,600 37%5/10/2014 20,881 44%5/3/2014 23,365 49%
4/26/2014 25,329 52%4/19/2014 26,440 54%4/12/2014 27,228 54%4/5/2014 27,910 56%
3/29/2014 30,105 60%3/22/2014 34,090 67%
* Average number of deaths for comparable weeks in 2010, 2011, and 2012.
Mortality Data Warehouse and Data Cubes
Updated nightly with most recent data and updates
Mortality Review Cube Available to DVS staff only Large number of variables including literal text Using Excel PivotTables users can slice and dice data
very easily Users can also drill through to individual records
Mortality Surveillance Cube Available to DVS staff and CDC Surveillance Partners
(within the CDC firewall) Customizable views… users can only access what they
have been given permission see Users can not drill through to individual records
Excel PivotTable Interface to Surveillance Cube
Comparison of 122 Cities and NVSS Based
P&I Ratios
Comparison of 122 Cities and NVSS Based
P&I Ratios with 2 week adjustment
NVSS – Tuesday January 7
122 Cities – Released Friday January 3
122 Cities – Released Friday January 10+3 days
122 Cities – Released Friday January 17+1.5 weeks
122 Cities – Released Friday January 24+2.5 weeks
Mortality Data NCHS had Received and Coded as of 3:00 AM Last Sunday
(6/8/2014)
* Average number of deaths for comparable weeks in 2010, 2011, and 2012.
Jurisdictions 172733374042434445474649
Range in Jurisdictio
ns % of Expected*
<1 - 23%2 - 75%
<1 - 83%<1 - 84%11 - 93%3 - 91%8 - 95%2 - 94%
<1 - 100%<1 - 100%<1 - 100%<1 - 100%
Date of Death(Week
Ending)Record Count
% of Expected*
6/7/2014 1,707 4%5/31/2014 10,161 22%5/24/2014 15,519 33%5/17/2014 17,600 37%5/10/2014 20,881 44%5/3/2014 23,365 49%
4/26/2014 25,329 52%4/19/2014 26,440 54%4/12/2014 27,228 54%4/5/2014 27,910 56%
3/29/2014 30,105 60%3/22/2014 34,090 67%
The Example of Region X
122 CitiesP&I Ratio =
15%
NVSSP&I Ratio =
7%
6 States
Challenges and Opportunities for National Mortality Surveillance
Timely data from all Jurisdictions Culture Change Electronic Death Registration Systems (EDRS)
• Goal is to have 80% of all deaths reported to NCHS within 10 days of the date of death
• Opportunity to improve data quality
Robust National Mortality Surveillance IT Infrastructure Automation Robust tools for analysis, visualization, and dissemination
Building and Maintaining Surveillance Partnerships Improved and/or more efficient surveillance Investment in the larger National Vital Statistics System
Other Surveillance Projects and Partnerships
Rare Vaccine Preventable Disease Mortality Surveillance and Validation National Center for Immunization and Respiratory
Diseases (NCIRD), CDC Respiratory Syncytial Virus (RSV) Mortality
Surveillance National Center for Immunization and Respiratory
Diseases (NCIRD), CDC Suicide Surveillance
National Institute of Mental Health (NIMH), National Institutes of Health (NIH)
Mortality Surveillance Indicators Project: Objectives
Identify key causes of deaths that require real-time public health surveillance.
Evaluate the reliability and validity of the selected mortality estimates derived from the near real-time mortality surveillance data.
Determine optimal frequency, format, and mechanisms to timely disseminate the selected mortality estimates.
Pilot the process to identify and address barriers and challenges to build a feasible, reliable, and timely system for rapid mortality surveillance.
Mortality Surveillance Indicators Project:
Proposed Indicators
1. Influenza2. Suicide3. Firearm-related deaths4. Drug poisoning deaths5. Poisoning deaths6. Homicide7. Asthma among people <35 years8. HIV/AIDs9. Deaths for infants <1 year10. Unintentional fall age 65+ 11. Heart disease12. Stroke13. Alzheimer's14. Diabetes
Influenza Mortality Surveillance… Making Real-Time National Mortality
Surveillance a Reality
For more information please contact
Paul D. Sutton3311 Toledo Road, Hyattsville, MD 20782Telephone: (301) 458-4433E-mail: [email protected]
The findings and conclusions in this presentation are those of the author and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
National Center for Health Statistics
Division of Vital Statistics
