An Evaluation of Pharmacy Data for Surveillance of Gastrointestinal and
Respiratory OutbreaksYevgeniy Elbert1, MS
Shilpa Hakre1, MPH, DrPHHoward Burkom2, PhDJulie Pavlin1, MD, MPH
1Walter Reed Army Institute of Research, Silver Spring, MD2Johns Hopkins Applied Physics Laboratory, Laurel, MD
November 15, 2004
Report Documentation Page Form ApprovedOMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering andmaintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, ArlingtonVA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if itdoes not display a currently valid OMB control number.
1. REPORT DATE 15 NOV 2004
2. REPORT TYPE N/A
3. DATES COVERED -
4. TITLE AND SUBTITLE An Evaluation of Pharmacy Data for Surveillance of Gastrointestinal andRespiratory Outbreaks
5a. CONTRACT NUMBER
5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Walter Reed Army Institute of Research, Silver Spring, MD
8. PERFORMING ORGANIZATIONREPORT NUMBER
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)
11. SPONSOR/MONITOR’S REPORT NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for public release, distribution unlimited
13. SUPPLEMENTARY NOTES See also ADM001849, 2004 Scientific Conference on Chemical and Biological Defense Research. Held inHunt Valley, Maryland on 15-17 November 2004 . , The original document contains color images.
14. ABSTRACT
15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT
UU
18. NUMBEROF PAGES
40
19a. NAME OFRESPONSIBLE PERSON
a. REPORT unclassified
b. ABSTRACT unclassified
c. THIS PAGE unclassified
Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std Z39-18
• A U.S. Department of Defense (DoD) system
• Designed to detect infectious disease outbreaks
• Serves military active duty members, their beneficiaries, and retirees
• Uses mainly ICD-9-CM codes from outpatient visits
• Delay of 1- 4 days from patient visit date to data capture date by ESSENCE
INTR
OD
UC
TIO
NESSENCE
Electronic Surveillance System for the Early Notificationof Community-based Epidemics
Pharmacy Data Transaction Service(PDTS)
• Developed by the DoD Pharmacoeconomic Center (PEC) program, managed by WebMD
• Data repository for prescriptions filled by beneficiaries of the military health system
• Approx. 61.1 million prescriptions filled annually1
– 82% of prescriptions from MTFs (n=587) – 18% from:
• Tricare Managed Care Support Contractors Network pharmacies (n=50,000)
• National Mail Order Pharmacy (n=1)
INTR
OD
UC
TIO
N
TricareTricare pharmacy pharmacy transactionstransactions
PDTSPDTS
3.2 seconds
Update every 24 hours
PDTS Fields
• The PDTS formulary and standards maintained through WebMD’ssubscription to the American Hospital Formulary Service (AHFS), the National Council for Prescription Drug Programs (NCPDP), and First Data Bank, Inc.
• Medication id fields in PDTS- Drug name (Label)- Therapeutic class number (GCN)- Therapeutic class code (GC3)- Other
• Previous study by Scott Eader et al. using ESSENCE data for the DC metropolitan area
– Significant positive correlation between outpatient visit and PDTS data for GC3 codes identified for Respiratory (RESP) and gastrointestinal (GI) syndromes
• GC3: developed by First Data Bank, Inc.
3 characters (alpha, numeric, alpha)– organ system, pharmacological class, specific therapeutic
class, respectively
INTR
OD
UC
TIO
N
1.381223731HYPOTENSIVES,ANGIOTENSIN RECEPTOR ANTAGONISTA4F
1.411242950MACROLIDESW1D
1.411247417THIAZIDE AND RELATED DIURETICSR1F
1.431265605GLUCOCORTICOIDSP5A
1.471299466ANTICONVULSANTSH4B
1.481311393ANTI-ANXIETY DRUGSH2F
1.611420677NASAL ANTI-INFLAMMATORY STEROIDSQ7P
1.761557748BETA-ADRENERGIC AGENTSJ5D
1.871654859EXPECTORANTSB3J
1.921700412THYROID HORMONESP3A
2.221966936CALCIUM CHANNEL BLOCKING AGENTSA9A
2.312039720PENICILLINSW1A
2.522230604SELECTIVE SEROTONIN REUPTAKE INHIBITOR (SSRIS)H2S
2.682373543BETA-ADRENERGIC BLOCKING AGENTSJ7C
3.012659102HYPOTENSIVES, ACE INHIBITORSA4D
3.843395463GASTRIC ACID SECRETION REDUCERSD4K
4.313812841ANTIHISTAMINESZ2A
4.584049267LIPOTROPICSM4E
4.814256802ANALGESICS,NARCOTICSH3A
6.345603369NSAIDS, CYCLOOXYGENASE INHIBITOR - TYPES2B
PERCENTPERCENTCOUNTCOUNTDESCRIPTIONDESCRIPTIONGC3GC3
Painkillers
Cholestrolreducers
Allergy relievers
INTR
OD
UC
TIO
NPDTS
Commonly filled medications, 2003
STUDY OBJECTIVES
1) Determine medications commonly prescribed for GI and RESP syndromes during outbreaks
2) Examine trends in daily counts of medications filled for GI, RESP, Asthma visits during outbreaks
3) Conduct retrospective surveillance on drug groups that correlated with GI and RESP outpatient visits during outbreaks
METHODS
• List compiled of known GI and RESP outbreaks
• Commonly prescribed non-refill medications identified for each outbreak– By linking ambulatory and pharmacy data by date of visit/prescription
written and unique encrypted identifiers common to both data sets
• Trends in medications filled and GI/Resp visits investigated during outbreaks– Moving 7-day averages of daily counts
RES
ULT
SG
I Out
brea
k1
SUN MON TUE WED THUR FRI SAT
Field Training Exercise
C 148th Company
35 C company members ill
- Nausea- Diarrhea- Fever- Headache
• Food and water samples from the FTX tested– 5/13 samples tested positive for Campylobacter jejuni
Top 5 ICD-9 codes used during a C. jejuni outbreak
0 10 20 30 40 50 60 70 80 90
Gastroenteritis, noninfctNEC
Diarrhea NOS
Enteritis, viral NOS
Gastritis, acute w/ohemorrhage
Gastritis NOS w/ohemorrhage
558.
978
7.91
008.
853
5.00
535.
50 PERCENT
Top 5 prescriptions written during C. jejuni outbreak
0 5 10 15 20 25 30
Antidiarrheals
Analges ics / antipyretics
Quinolones
Antiem etic/antivertigo agents
Expectorants PERCENTRES
ULT
SG
I Out
brea
k 1
RESULTS
Campylobacter jejuni Outbreak 7-day moving averages
0
5
10
15
20
25
30
35
40
10/1/2002 10/15/2002 10/29/2002 11/12/2002 11/26/2002 12/10/2002 12/24/2002
Date
GI S
yndr
ome
Visi
ts
0
10
20
30
40
50
60
70
80
Med
icat
ions
Fill
ed
GI visits Antiemetics Antidiarrheals Quinolones Antihistamines
OUTBREAK
RES
ULT
SG
I Out
brea
k2
Top 5 ICD-9 codes used during a Salmonella outbreak
0 5 10 15 20 25 30 35
Poisoning, foodNOS
Gastroenteritis,noninfct NEC
Diarrhea NOS
Poisoning, food,bacterial NEC
Gastritis, acute w/ohemorrhage
005.
955
8.9
787.
9100
5.89
535.
00 PERCENT
Top 5 prescriptions during the Salmonella outbreak
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0
Antihistamines
Quinolones
Analgesics/antipyretics
NSAID
Antidiarrheals PERCENT
RESULTS
Salmonella Outbreak 7-day moving averages
05
10152025
11/1/2002 11/15/2002 11/29/2002 12/13/2002 12/27/2002
Date
Med
icat
ions
Fi
lled
0510152025
GI S
yndr
ome
Visi
ts
GI Visits Analgesic/antipyretic AntidiarrhealsQuinolones Antihistamines
OUTBREAK
Analgesics/antipyretics: Acetaminophen, Tylenol, etc.Antidiarrheals: Loperamide, Imodium, Pepto-bismol, etc.Antihistamines: Includes Phenergan, Promethazine (antiemetics)Quinolones: Ciprofloxacin, Levaquin, etc.
RESULTSGI OUTBREAK 3
• At a U.S. Air Force training base in Texas3
– Norovirus outbreak– In all 464 trainees affected– Lasted 7 days– 7/7 stool samples tested positive for
Norovirus
Top 5 prescriptions written during a Norovirus outbreak
0 10 20 30 40 50 60
ANTIDIARRHEALS
ANTIHISTAMINES
PERCENT
Top 5 ICD-9 codes used during a Norovirus outbreak
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00
Gastroenteritis, noninfct NEC
Enteritis presumed infct origin
Diarrhea NOS
Nausea with vomiting
Vomiting alone
558.
900
9.1
787.
9178
7.01
787.
03 PERCENT
RES
ULT
S
GI O
utbr
eak
3
RESULTS
Norovirus Outbreak 7-day moving averages
0102030405060708090
1/1/2004 1/22/2004 2/12/2004 3/4/2004 3/25/2004 4/15/2004
Date
GI S
yndr
ome
Visi
ts
010203040506070
Med
icat
ions
Fill
ed
GI Visits Antidiarrheals Antihistamines
OUTBREAK
RESULTSRESP OUTBREAK 1
• At the Marine Corps Recruit Depot in San Diego3
– Largest Group A Streptococcus outbreak since 1968
– 160 recruits admitted for pneumonia (radiographically confirmed)
– November 1 – December 20– Sputum, blood and throat cultures
• Group A Streptococcus isolated
Top 5 ICD-9 codes used during a Streptococcus A outbreak
0 5 10 15 20 25 30 35
Infct up rsprt mlt sites, acute NOS
Pneumonia, organism NOS
Pharyngitis, acute
Shortness of breath
Cough
465.
948
646
278
6.05
786.
2
PERCENT
Top 5 prescriptions written during a Streptococcus A outbreak
0 5 10 15 20
Analgesics/antipyretics
Expectorants
Sympathomimetic agents
Macrolides
Zinc replacement PERCENTRES
ULT
SR
ESP
Out
brea
k 12
RESULTS
Streptococcus A Outbreak7-day moving averages
0
20
40
60
80
10/15/2002 10/30/2002 11/14/2002 11/29/2002 12/14/2002 12/29/2002
Date
Med
icat
ions
Fill
ed
0102030405060
RES
P Sy
ndro
me
Visi
ts
RespVisits ExpectorantsZincReplacement Analgesics/antipyreticsSympathomimeticAgents
OUTBREAK
Expectorants: Robitussin, Guaifenesin, etc.
Sympathomimetic agents: Sudafed, etc
Zinc replacement: Cepacol Cold Care, etc.
Analgesics/antipyretics: Acetaminophen, Tylenol, etc.
RESULTSRESP OUTBREAK 2
• Influenza in an Advanced Individual Training population in VA in November, 2003– Approx. 187 cases total– Oct. 31st – Nov. 7th
– Rapid flu tests positive; 5 CDC viral cultures Flu A positive, subtyped Korea-like or Fujian-like viruses
Top 5 ICD-9 codes used during an influenza outbreak
0 5 10 15 20 25 30 35
Infct up rsprt mltsites, acute NOS
Bronchitis, acute
Cough
Bronchitis NOS
Pharyngitis, acute
465.
946
6.0
786.
249
046
2 PERCENT
Top 5 prescriptions written during an influenza outbreak
0 5 10 15 20 25 30
Expectorants
Analgesics/antipyretics
Cough and/or coldpreparations
Beta-adrenergic agents
NSAIDS PERCENT
RES
ULT
SR
ESP
Out
brea
k 2
RESULTS
Influenza Outbreak7-day moving averages
0
20
40
60
80
100
120
140
10/1/2003 10/16/2003 10/31/2003 11/15/2003 11/30/2003 12/15/2003 12/30/2003
Date
Med
icat
ions
fille
d
0102030405060708090100
Visi
ts
RespVisits Expectorants Analgesic/antipyretic Antitussives
OUTBREAK
RESULTS
GI groupGI group
1. 1. AntiemeticsAntiemetics2. 2. AntidiarrhealsAntidiarrheals3. Antihistamines3. Antihistamines4. Analgesics/antipyretics4. Analgesics/antipyretics5. Antispasmodics5. Antispasmodics6. Antacids6. Antacids7. 7. QuinolonesQuinolones8. Intestinal adsorbents and 8. Intestinal adsorbents and
protectivesprotectives9. 9. PenicillinsPenicillins
10. Absorbable sulfonamides10. Absorbable sulfonamides
RESP groupRESP group
1. Expectorants1. Expectorants2. Cough and/or cold 2. Cough and/or cold
preparationspreparations3. Analgesics/antipyretics3. Analgesics/antipyretics4. 4. AntitussivesAntitussives, non, non--narcoticnarcotic5. 5. SympathomimeticSympathomimetic agentsagents6. NSAIDS6. NSAIDS7. 7. MacrolidesMacrolides8. 8. QuinolonesQuinolones9. Zinc replacement9. Zinc replacement
10. 10. PenicillinsPenicillins
STUDY OBJECTIVES
1) Determine medications commonly prescribed for GI and Respiratory (RESP) syndromes
2) Examine trends in daily counts of medications for GI, RESP, Asthma visits during outbreaks
3) Conduct retrospective surveillance on GI and RESP syndrome drug groups and GI and RESP outpatient visits at installations where outbreaks occurred
METHODS
• Retrospective surveillance conducted on pharmacy and outpatient visit data for approx. 1 year prior to and including the outbreak timeframe– Regression/Exponential Weighted Moving Average (EWMA) detection
algorithms run on daily total GI/RESP visits, daily total PDTS syndrome groups and each GC3 group → output of p-values
– P-values from individual larger count GC3 groups (>1 prescription in 3 days) were combined using Fisher and Edgington combination p-value methods
• More information on combination p-value methods → Dr. Howard Burkom’s talk at 2:00 p.m.
– Detections with approximately 1 per 6 weeks false alert rate were identified graphically
Salmonella Outbreak(12/9/02 to 12/18/02)
1
11
21
31
41
51
61
71
81
91
10/3
1/20
01
11/1
4/20
01
11/2
8/20
01
12/1
2/20
01
12/2
6/20
01
1/9/
2002
1/23
/200
2
2/6/
2002
2/20
/200
2
3/6/
2002
3/20
/200
2
4/3/
2002
4/17
/200
2
5/1/
2002
5/15
/200
2
5/29
/200
2
6/12
/200
2
6/26
/200
2
7/10
/200
2
7/24
/200
2
8/7/
2002
8/21
/200
2
9/4/
2002
9/18
/200
2
10/2
/200
2
10/1
6/20
02
10/3
0/20
02
11/1
3/20
02
11/2
7/20
02
12/1
1/20
02
12/2
5/20
02
Outpatient Visits ESSENCE Visits EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
1
11
21
31
41
51
61
71
81
91
101
3/3/
2002
3/10
/200
2
3/17
/200
2
3/24
/200
2
3/31
/200
2
4/7/
2002
4/14
/200
2
4/21
/200
2
4/28
/200
2
5/5/
2002
5/12
/200
2
5/19
/200
2
5/26
/200
2
6/2/
2002
6/9/
2002
6/16
/200
2
6/23
/200
2
6/30
/200
2
7/7/
2002
7/14
/200
2
7/21
/200
2
7/28
/200
2
8/4/
2002
8/11
/200
2
8/18
/200
2
8/25
/200
2
9/1/
2002
9/8/
2002
9/15
/200
2
9/22
/200
2
9/29
/200
2
10/6
/200
2
10/1
3/20
02
10/2
0/20
02
10/2
7/20
02
11/3
/200
2
11/1
0/20
02
11/1
7/20
02
11/2
4/20
02
12/1
/200
2
12/8
/200
2
12/1
5/20
02
12/2
2/20
02
12/2
9/20
02
Outpatient Visits ESSENCE VISITS EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
C. jejuni Outbreak(11/16/02 – 11/18/02)
1
51
101
151
201
251
301
3/3/
2003
3/17
/200
3
3/31
/200
3
4/14
/200
3
4/28
/200
3
5/12
/200
3
5/26
/200
3
6/9/
2003
6/23
/200
3
7/7/
2003
7/21
/200
3
8/4/
2003
8/18
/200
3
9/1/
2003
9/15
/200
3
9/29
/200
3
10/1
3/20
03
10/2
7/20
03
11/1
0/20
03
11/2
4/20
03
12/8
/200
3
12/2
2/20
03
1/5/
2004
1/19
/200
4
2/2/
2004
2/16
/200
4
3/1/
2004
3/15
/200
4
3/29
/200
4
Outpatient Visits ESSENCE VISITS EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
Norovirus Outbreak3
(02/17/04 to 02/24/04)
Norovirus Outbreak(11/25/02 to 12/15/02)
1
21
41
61
81
101
121
3/3/20
023/1
7/2002
3/31/2
0024/1
4/2002
4/28/2
0025/1
2/2002
5/26/2
0026/9
/2002
6/23/2
0027/7
/2002
7/21/2
0028/4
/2002
8/18/2
0029/1
/2002
9/15/2
0029/2
9/2002
10/13
/2002
10/27
/2002
11/10
/2002
11/24
/2002
12/8/
2002
12/22
/2002
1/5/20
031/1
9/2003
Outpatient Visits ESSENCE Visits EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
Alerted earlier
1
51
101
151
201
251
301
351
10/31
/2002
11/14
/2002
11/28
/2002
12/12
/2002
12/26
/2002
1/9/20
031/2
3/2003
2/6/20
032/2
0/2003
3/6/20
033/2
0/2003
4/3/20
034/1
7/2003
5/1/20
035/1
5/2003
5/29/2
0036/1
2/2003
6/26/2
0037/1
0/2003
7/24/2
0038/7
/2003
8/21/2
0039/4
/2003
9/18/2
00310
/2/20
0310
/16/200
310
/30/200
311
/13/200
311
/27/200
312
/11/200
312
/25/200
3
Outpatient Visits ESSENCE VISITS EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
Influenza Outbreak(11/1/03 to 11/18/03)
Group A Streptococcus Outbreak12/08/02 to 12/12/02
1
21
41
61
81
101
121
141
161
181
201
12/10
/2001
12/20
/2001
12/30
/2001
1/9/20
021/1
9/2002
1/29/2
0022/8
/2002
2/18/2
0022/2
8/2002
3/10/2
0023/2
0/2002
3/30/2
0024/9
/2002
4/19/2
0024/2
9/2002
5/9/20
025/1
9/2002
5/29/2
0026/8
/2002
6/18/2
0026/2
8/2002
7/8/20
027/1
8/2002
7/28/2
0028/7
/2002
8/17/2
0028/2
7/2002
9/6/20
029/1
6/2002
9/26/2
00210
/6/20
0210
/16/200
210
/26/200
211
/5/20
0211
/15/200
211
/25/200
212
/5/20
0212
/15/200
212
/25/200
2
Outpatient Visits ESSENCE VISITS EDGINGTON Rx FISHER Rx ESSENCE Rx Outbreak
Alerted earliest
CONCLUSIONS• Use of the GC3 classification system in military pharmacy data
can increase sensitivity without significantly affecting the false alert rate.– A limitation: Certain medications were found in more than one
GC3 group. If these don’t contribute to an outbreak, the signal is weakened
• Outbreaks were characterized by increases in certain GC3 groups and combinations of GC3 groups but these varied by outbreak type.
• Syndromic drug groups were formed using medications more commonly used during outbreaks, increasing the overall sensitivity of the system.
CONCLUSIONS
• The ESSENCE outpatient visit detector, the gold standard, alerted for all 6 outbreaks.
• The ESSENCE pharmacy detector performed well and alerted for 5/6 outbreaks.
• The two p-value combination methods have distinct advantages. The Fisher method alerted for 5/6 outbreaks and was more sensitive to individual GC3 increases. Whereas, the Edgington method alerted for 3/6 outbreaks but was more responsive to consensus and in 2 outbreaks alerted earlier.
• Combination methods are more sensitive and specific if low count gc3 streams are accumulated or dropped.
FUTURE RESEARCH
• Investigate performance of algorithms for other outbreaks
• Investigate performance of detectors for a GC3 group switch, i.e., ignore product groups that don’t meet a certain minimum count
• Continue to investigate product groups that give more signal and less noise
• Prospective surveillance using ESSENCE pharmacy and combination method detectors
REFERENCES
1. http://www.pec.ha.osd.mil/ICD/ICD_MTF_040502.doc
2. Centers for Disease Control and Prevention. Outbreak of group A streptococcal pneumonia among Marine Corps recruits – California, November 1 –December 20, 2002. MMWR. 2003;52(6):106-109.
3. Yamane, GK, Shibukawa-Kent, RL. 2004. A large norovirus GI outbreak among trainees at a USAF training base. Seventh Annual Force Health Protection Conference, Aug. 8-12, 2004.
ACKNOWLEDGEMENTS
CPT Jeffrey K. Brown, USAMRID, Fort Detrick, MD
LCDR Scott Thornton, NEPMU-6, Pearl Harbor, HI
MAJ Samuel Jang, CHPPM, Aberdeen Proving Grounds, MD
Asha Riegodedios, MPH, NEHC, Bethesda, MD
Top Related