NAPHSIS Annual Meeting 2010

State Pilot Project Using the Electronic Death Registration System for H1N1

Surveillance

Linette T Scott, MD, MPHDeputy Director

Health Information and Strategic PlanningCalifornia Department of Public Health

June 9, 2010

NAPHSIS Annual Meeting 2010

Path Travelled

• Background• Data Flow in California• Query Development• Other Issues and Next Steps

NAPHSIS Annual Meeting 2010

California Electronic Death Registration System (EDRS)

• Electronic creation and registration of Death Certificates, Amendments, and Disposition Permits

• Went “live” January 2005

• By February 2010, 99.4% of Certificates created in EDRS

• ~ 240,000 – 250,000 deaths in California annually

NAPHSIS Annual Meeting 2010

Cause ofDeath fields

NAPHSIS Annual Meeting 2010

Value of EDRS Surveillance

• Identifies most serious cases for further epidemiological investigation

• Complements other data sources (e.g. Confidential Morbidity Reports, ED visits, hospitalizations)

• Encompasses entire population, not just a sample

• Minimum added resource cost—uses data from existing system

NAPHSIS Annual Meeting 2010

Data Flow in California

GG note: Would say FH transcribes information

NAPHSIS Annual Meeting 2010

Why wait until local registration?

• Quality Check: Local offices will contact physicians if needed (e.g. Cause of Death is non-specific)

• Prior to local registration, medical information can be freely changed without documentation

• After local registration, certificate becomes a legal document and further changes must be done by filing an amendment

NAPHSIS Annual Meeting 2010

Free Text Surveillance

• Free text search: Cause of Death (COD) entries from physician

• Done before ICD-10 coding for COD assigned to Death Certificates

• Influenza disambiguation example:– Haemophilus influenzae

– Parainfluenza

– Misspellings

NAPHSIS Annual Meeting 2010

Cautions of Preliminary Data

• Late Entries

• Pending Certificates

• Amendments

• Duplicate Entries

NAPHSIS Annual Meeting 2010

Average Processing Times(YTD 2010)

• Date of Death to Local Registration

– EDRS (fully electronic): 6.1 days

– Non-EDRS: 4.5 days

• Date of Death to State Registration

– EDRS (fully electronic): 15.7 days

– Non-EDRS: 71.0 days

NAPHSIS Annual Meeting 2010

Processing Delays

• Biggest delays:– Physician COD entry

– Physician signing of certificate

• Delays can also occur on the personal information half of certificate (e.g. clarifying family information)

• Can pull data prior to local registration but at cost of decreased data quality

NAPHSIS Annual Meeting 2010

How to Improve?

Physician education:

• on importance of accurate COD entry

• on importance of timely COD entry and signature

NAPHSIS Annual Meeting 2010

Steps to Develop Query

1. Obtain listing of confirmed cases2. Look up cases in EDRS for COD3. Select most common/promising COD text

strings and do test run4. Of new results, how many are potential

new cases versus “noise”?5. Refine query

NAPHSIS Annual Meeting 2010

Steps 1 & 2

• Confirmed case listing from May to August 13, 2009

• 55 out of 107 cases had “H1N1” or “Swine” or “Pandemic” in COD

• Others were non-specific (“Influenza”, “Pneumonia”, “ARDS”)

• Spelling variants found: “H1 N1” “H1-N1” “N1H1” etc

NAPHSIS Annual Meeting 2010

Steps 3-5

• Test queries run in same time frame in EDRS

• “H1” “Swine” “Pandemic” brought up 21 additional deaths listed as H1N1

• “Pneumonia” “ARDS” brought up too many non-specific results (e.g. aspiration pneumonia, sequelae of cancer)

NAPHSIS Annual Meeting 2010

Final Query Parameters

“H1” or “Swine” or “Pandemic” anywhere in

COD fields (107A, 107B, 107C, 107D)or Other Significant Conditions field (112)

NAPHSIS Annual Meeting 2010

EDRS Pull Confirmed Cases

• Unreported confirmed case• COD not confirmed (“clinical diagnosis”)

• Search term not in COD• Delayed registration or amendment• Out of state death

Case both in EDRS pull and epidemiology-confirmed

N=21 N = 52N = 55

NAPHSIS Annual Meeting 2010

Other Issues & Next Steps

NAPHSIS Annual Meeting 2010

NAPHSIS Annual Meeting 2010

Media & Public

• EDRS deaths are not confirmed• EDRS and epidemiology-confirmed numbers

will be different• Media confusion regarding availability and

interpretation of source data• Not part of Public Records Act

NAPHSIS Annual Meeting 2010

Other Uses

• Seasonal influenza surveillance• Other infectious disease surveillance• Investigative resource (retrospective look)• Chronic diseases?

NAPHSIS Annual Meeting 2010

Lessons Learned

• EDRS surveillance is feasible and of potential benefit

• Importance of partnering with programs (i.e. Vital Records and Communicable Disease Control)

• What is the program business need that EDRS surveillance can fulfill?

NAPHSIS Annual Meeting 2010

Thank you!

Special thanks to:

• Dr. Scott Fujimoto (HISP)

• Dr. Glenna Gobar (U.C. Davis School of Medicine)

• Entire EDRS Staff

• Meileen Acosta and Division of Communicable Disease Control staff