Automated Laboratory Reporting of Infectious Diseases in a Climate of Bioterrorism.pdf

While newly available electronic transmission methods
can increase timeliness and completeness of infectious dis-
ease reports, limitations of this technology may unintention-
ally compromise detection of, and response to, bioterrorism
and other outbreaks. We reviewed implementation experi-
ences for five electronic laboratory systems and identified
problems with data transmission, sensitivity, specificity, and
user interpretation. The results suggest a need for backup
transmission methods, validation, standards, preserving
human judgment in the process, and provider and end-user
involvement. As illustrated, challenges encountered in
deployment of existing electronic laboratory reporting sys-
tems could guide further refinement and advances in infec-
tious disease surveillance.
The primary purpose of reporting diseases is to trigger
an appropriate public health response so that further
illness can be prevented and public fears allayed. The
threat of emerging infections and bioterrorist attacks has
heightened the need to make disease surveillance more
sensitive, specific, and timely (1,2). Recent advances in
provider and laboratory information management have
facilitated one step towards the modernization of surveil-
lance: the development of automated reporting systems
(3,4). With recent funding for activities to defend the pub-
lic’s health against terrorism and naturally occurring dis-
eases, development of automated reporting systems has
been accelerated (5).
However, technologically innovative reporting systems
need to be consistent with the purpose of disease reporting.
Wholesale adoption of automated electronic reporting sys-
tems in their current form might instead represent a quick
response to the pressures of the moment rather than a fully
considered decision that acknowledges some of the docu-
mented problems with the new technology. We review here
current limitations of systems that provide automated noti-
fication of reportable conditions identified in clinical labo-
ratories. A more thorough understanding of the pitfalls of
such existing systems can provide insights to improve the
development and implementation of new media in infec-
tious disease surveillance.
With the computerization of patient and clinical labora-
tory data, automated notification of reportable events to
health departments is often assumed to be more effective
than conventional paper-based reporting (6). In recent
years, the Centers for Disease Control and Prevention
(CDC) has been funding several states to develop electron-
ic laboratory reporting (7). With electronic reporting, lab-
oratory findings (e.g., Escherichia coli O157:H7 test
results) are captured from clinical laboratory data and
transmitted directly to the state. In turn, the state routes
messages to local health units, as illustrated in the Figure.
The National Electronic Disease Surveillance System
(NEDSS) and bioterrorism preparedness initiatives are
expected to further enhance disease surveillance by sup-
porting integration of electronic data from various sources
(4,8). Evidence from deployed systems shows promise in
the ability of electronic laboratory reporting to deliver
more timely and complete notifications than paper-based
methods (9–12).
At the same time, experiences in Pennsylvania, New
York, Hawaii, California, and other states indicate that
implementation of automated reporting also poses unantic-
ipated challenges. Five problem areas have been identi-
fied: sensitivity, specificity, completeness, coding stan-
dards, and end-user acceptance.
Sensitivity
To achieve the objective of triggering local public
health response, automated electronic systems should con-
sistently report cases that would have been reported by
conventional methods. Contrary to expectations, automat-
ed reports seldom replicate the traditional paper-based sys-
tem. Errors in data transmission reduce sensitivity in auto-
mated electronic reporting systems. An evaluation of elec-
tronic laboratory reporting in Hawaii documented that
automated reports were not received for almost 30% of the
days on which the paper-based method generated a report,
Emerging Infectious Diseases • Vol. 9, No. 9, September 2003 1053
PERSPECTIVES
Automated Laboratory Reporting
of Infectious Diseases in a Climate
of Bioterrorism
Nkuchia M. M’ikanatha,* Brian Southwell,† and Ebbing Lautenbach‡
*Pennsylvania Department of Health, Harrisburg, Pennsylvania,
USA; †University of Minnesota, Minneapolis, Minnesota, USA; and
‡University of Pennsylvania School of Medicine, Philadelphia,
Pennsylvania, USA

suggesting that automated reporting alone was potentially
suboptimal. Lapses in electronic reporting were due to var-
ious causes including ongoing adjustments to the data
extraction program (11). In California, lapses in a semiau-
tomated electronic laboratory reporting were traced to a
failure in forwarding reports from the county of diagnosis
to the county of residence (12). In Pennsylvania, lapses in
automated notification have resulted from the occasional
failure of data extraction at the clinical laboratory comput-
er, difficulties deciphering reportable diseases from test
results which used local terminology rather than Logical
Observation Identifier Names and Codes (LOINC) codes
(available from: URL: http://www.regenstrief.org/loinc),
and problems in the transmission of data files to and access
by local health jurisdictions. To prevent interruption of
reports while the automated system was being refined,
Pennsylvania opted to continue conventional paper-based
reports for 8 months after initiating electronic reporting.
Specificity
Typically, automated reporting increases not only
reportable events data but also the number of extraneous
reports (e.g., nonreportable conditions, unnecessary nega-
tive reports, or duplicate reports). In addition, false-posi-
tive results are increased by automated abstraction of cul-
ture results entered in free-text. For example, in an evalu-
ation of an electronic laboratory reporting system in
Allegheny County, Pennsylvania, negative results of
Salmonella isolates were automatically transmitted as pos-
itive Salmonella results because the software recognized
the organism name (9). Often, automated reporting trans-
mits preliminary test results followed by results of confir-
matory tests for the same condition. This method is desir-
able because some duplicates may actually provide useful
preliminary test results that might trigger timely responses
(9,10). However, multiple test results increase time for
data processing. In addition, low specificity attributable to
extraneous records of nonreportable culture results is also
problematic. While over time automated programs can be
expected to improve, initially erroneous or missing data
will continue to arise and require manual checking and
recoding.
Programming solutions might offer relief in eliminating
extraneous records. But in a climate of bioterrorism, a
complete replacement of human judgment is probably
unacceptable for many. Therefore, in planning new sys-
tems, accounting for the time and effort of an experienced
epidemiologist to review electronic laboratory data before
routing them to investigators will be essential.
Completeness of Case Records
To be useful, case-reports received through convention-
al or automated methods must contain data in key fields
identifying patient and physician (e.g., name, address, and
1054 Emerging Infectious Diseases • Vol. 9, No. 9, September 2003
PERSPECTIVES
Figure. Steps in automated reporting of infectious disease data. The process begins with abstraction of reportable conditions using a soft-
ware program. Data are stored in a file for future transmission or sent directly to the health department in the case of automated report-
ing systems. Typically, there are multiple clinical laboratories, and reports are transmitted in a variety of methods including file transfer
protocol and dial-up modem at arranged intervals. State health departments review data and forward them to local health departments,
where investigations are done and reportable conditions are determined. Local health departments forward data back to the state, where
further analysis and interpretation are accomplished. The state uploads nationally notifiable diseases data to a secure data network at
the Centers for Disease Control and Prevention. That agency sends data quality feedback to the state immediately. The level of feed-
back among states, laboratories, and local health departments is unknown but suspected to vary widely.

telephone number) and specimen (e.g., collection date,
type, test, and result). Lack of sufficient identifying infor-
mation for follow-up investigations is a serious limitation
in many currently operating automated systems.
In addition, experiences in New York and Pennsylvania
indicate that the lack of a patient’s address is a barrier to
routing electronic laboratory data to local health depart-
ments. Locating a patient’s residence is also useful for rec-
ognizing clusters of diseases attributable to natural causes
or intentional acts of terrorism. Automated means were
intended to improve completeness of case record data by
duplicating required fields, but this has not always been
the case (13). Whether the laboratories fail to report miss-
ing data or whether data elements are not provided in the
initial forms submitted with specimens is unclear.
Widespread dissemination of standardized disease report-
ing forms specifying information required by health
departments to both clinical laboratories and providers
could reduce this problem. Such information could also be
made readily available through the Internet. An example of
what laboratories and providers are required to include in
Minnesota is available (URL: http://www.health.state.mn.
us/divs/dpc/ades/surveillance/card.pdf).
Data Standards
To facilitate use of state-of-the-art electronic surveil-
lance tools as envisioned in the NEDSS initiative, adop-
tion of Systemized Nomenclature of Human and
Veterinary Medicine (SNOMED) (available from: URL:
http:// www.snomed.org/), LOINC, and Health Level 7
standards (a national standard for sharing clinical data,
available from: URL: http://www.hl7.org/) by clinical lab-
oratories is essential. However, in practice clinical labora-
tories often use locally developed coding schemes or a
combination of codes and free text. Data often arrive in
multiple file formats or even with multiple formats with-
in one mapping standard (Figure). In practice, file mes-
sages from multiple laboratories are mapped into a stan-
dardized database with desired variables including
patient, physician contact information, specimen identi-
fiers, test name, and results.
To increase use of uniform data coding and Health
Level 7 as the standard for automated electronic reporting,
further studies are needed to understand barriers encoun-
tered by clinical laboratories and ways to overcome them.
Cost or lack of information technology resources might be
factors contributing to slow adoption of standard coding in
small-size clinical laboratories. In addition, variations in
reporting requirements across states may be an extra cost
to laboratories that serve multiple health jurisdictions. In
addition to understanding and assisting in reducing barri-
ers to use of standards, public health officials could help
promote use of coding standards by demonstrating their
benefits to laboratories and providers. For example, use of
standards such as LOINC facilitates integration of micro-
biologic culture data, minimizes chances for data errors in
translating free text or handwritten test results, and makes
it easier for laboratories to monitor antimicrobial resist-
ance patterns. This could be reinforced by introducing reg-
ular data quality feedback to all the stakeholders, as illus-
trated in the Figure.
User Acceptance
The entire process for detecting diseases relies on
acceptance and appropriate intervention by those working
on the front-line of the public health system. As shown on
the Figure, public health surveillance largely depends on
investigation at the local level, where a determination is
made that reported events meet case definitions for
reportable and notifiable conditions. Local health depart-
ments report data to the state level, where nationally noti-
fiable diseases (available from: URL: http://www.cdc.gov/
epo/dphsi/phs/infdis.htm) are transmitted to CDC. That
agency in turn reports internationally quarantinable dis-
eases to the World Health Organization (available from:
URL: http://www.who.int/emc/IHR/ int_regs.html). The
process begins with receiving, managing, and using sur-
veillance data. Automated reports in the form of electron-
ic-mail attachments could be cumbersome for some local
health departments with limited information technology
support. Also, encryption of data for confidentiality rea-
sons increases complexity of the data retrieval process.
Acceptance of automated electronic reporting systems is
likely when assistance on data analysis and management is
given to disease investigators.
During the 2001 bioterrorism outbreak investigation,
labor-intensive methods (i.e., faxes and emails) were used
for surveillance of cases with clinical syndromes compati-
ble with anthrax among patients in selected counties in
New Jersey, Pennsylvania, and Delaware (14). Because of
personnel time demands, automated electronic systems are
attractive in surveillance of syndromes suggestive of bioter-
rorism agents. While automated electronic surveillance sys-
tems using patient encounter records for syndromic surveil-
lance might offer relatively low costs of adoption for physi-
cians (15), other persons in the system may become undu-
ly burdened. For example, when automated reports of syn-
dromes are forwarded to local public health officials, who
should interpret and act upon the results remains unclear.
The key to the success of such innovative systems outside
investigational settings will be their ability to offer mean-
ingful results at an acceptable marginal cost to both
reporters and local health departments. Integration of syn-
dromic surveillance into local public health surveillance is
less understood and needs attention.
PERSPECTIVES

Discussion
Responding to and anticipating the difficulties encoun-
tered by existing automated reporting systems could be
used to improve current systems and guide development of
future infectious disease surveillance. Addressing limita-
tions of automated reporting systems by continuing con-
ventional notification methods during the adjustment peri-
od, promoting use of coding standards, validating data, and
involving end-users is essential.
As illustrated in this study, lapses in data transmission
occur during initial deployment of automated reporting
systems. The potential risks attributable to lapses or errors
in automated electronic reports are great, as are costs asso-
ciated with misdiagnoses and treatment of healthy persons
(16). Experiences in Hawaii and Pennsylvania indicate the
need for continuing with existing reporting mechanisms
during the first year while new systems are being refined.
Our study calls for evaluations to validate new automat-
ed systems before they are integrated into public health
surveillance. While health departments and CDC have typ-
ically collaborated in such efforts, involvement of
providers and laboratorians is likely to yield additional
insights. Participation of public health officials is indicat-
ed in evaluations of automated methods that are being
developed in research settings to capture nonreportable
syndromes for bioterrorism detection.
Partnerships among state health departments, clinical
laboratories, providers, CDC, and other diagnostics sys-
tems are needed to promote widespread use of uniform
coding standards (LOINC and SNOMED) and Health
Level 7 for messaging. As demonstrated in New York
State, involving all users early in the planning stages
enhances the success of automated electronic reporting
system (13). CDC could facilitate laboratory participation
in use of standards by assisting health departments in iden-
tifying benefits such as use of LOINC-coded data for
antimicrobial resistance monitoring.
Current federal funding for emergency preparedness
surveillance and epidemiology capacity (17) is expected to
stimulate widespread use of automated systems in infec-
tious disease reporting. However, automated systems are a
complement rather than a substitute for human involve-
ment in interpreting laboratory findings and screening for
errors. Furthermore, the requirement that providers and
laboratories report immediately by telephone when they
detect organisms indicating an outbreak or an unusual
occurrence of potential public health importance (18) is
expected to continue even when automated reporting sys-
tems are implemented. Complete replacement of human
judgment in reporting conditions suggestive of CDC cate-
gory A bioterrorism agents (available from: URL: http://
www.bt.cdc.gov/Agent/Agentlist.asp) or other conditions
that require immediate investigation is unrealistic.
Despite the limitations we have described, automated
electronic systems hold promise for modernizing infec-
tious disease surveillance by making reporting more time-
ly and complete. Modern technology can translate into bet-
ter public health preparedness by enhancing and comple-
menting existing reporting systems.
Acknowledgments
We thank Lee Harrison, Kathleen G. Julian, Elliot Churchill,
and David Welliver for their insightful comments.
This study was supported in part by the Centers for
Education and Research on Therapeutics (CERTs) grant (U18-
HS10399) from the Agency for Healthcare Research and Quality.
Dr. M’ikanatha is a surveillance epidemiologist in
Pennsylvania. He is interested in the use of new technology to
promote notification of reportable diseases and other conditions
of public health importance including antimicrobial resistance.
References
1. Henderson DA. The looming threat of bioterrorism. Science
1999;283:1279–82.
2. Fine A, Layton M. Lessons from West Nile encephalitis outbreak in
New York City, 1999: implications for bioterrorism preparedness.
Clin Infect Dis 2001;32:277–82.
3. Centers for Disease Control and Prevention. An overview of NEDSS
initiative. [Cited June 12, 2002.] Available from: URL:
http://www.cdc.gov/nedss/About/overview.html.
4. Doyle TJ, Glynn KM, Groseclose SL. Completeness of notifiable
infectious disease reporting in the United States: an analytical litera-
ture review. Am J Epidemiol 2002;155:866–74.
5. U.S. Department of Health and Human Services. FY2002.
Washington, DC: The Department; 2002. [Cited June 4, 2002.]
Available from: URL: http://www.hhs.gov/ophp/funding
6. Brennan PF. AMIA Recommendations for national health threat sur-
veillance and response. Journal of the American Medical Informatics
Association 2002;9:204–6.
7. Centers for Disease Control and Prevention. Electronic reporting of
laboratory data for public health. 1997. [Cited June 4, 2002.]
Available from: URL: http://www.cdc.gov/od/hissb/docs/elr-
1997.pdf
8. Centers for Disease Control and Prevention. Guidance for fiscal year
2001 supplemental funds for epidemiology and laboratory capacity
for infectious diseases (ELC) cooperative agreement [ELC supple-
ment A- NEDSSS FY2001: New activities]: National electronic dis-
ease surveillance system (NEDSS) activities. [Cited June 12, 2002.]
Available from: URL: http://www.cdc.gov/nedss/About/
NEDSS_RFA_2001.pdfhttp://www.cdc.gov/nedss/About/NEDSS_R
FA_2001.pdf
9. Panackal AA, M’ikanatha NM, Tsui FC, McMahon J, Wagner MM,
Dixon BW, et al. Automatic electronic laboratory-based reporting of
notifiable infectious diseases at a large health system. Emerg Infect
Dis 2002;8:685–91.
10. Centers for Disease Control and Prevention. Supporting public health
surveillance through the National Electronic Disease Surveillance
System. Available from: URL: http://www.cdc.gov/od/hissb/docs/
NEDSS%20Intro.pdf
11. Effler P, Ching-Lee M, Bogard A, Ieong M, Nekomoto T, Jernigan D.
Statewide system of electronic notifiable disease reporting from clin-
ical laboratories. JAMA 1999;282:1845–50.
1056 Emerging Infectious Diseases • Vol. 9, No. 9, September 2003
PERSPECTIVES

12. Backer HD, Bissel SR, Vugia DJ. Disease reporting from automated
laboratory-based reporting system to a state health department via
local health departments. Public Health Rep 2001;116:257–65.
13. Smith PF, Davisson M, Fuhrman J, Chang H, Gotham I, Birkhead G,
et al. Lessons learned from implementing electronic laboratory
reporting, New York State. Proceedings of the International
Conference on Emerging Infectious Diseases, 2002. Available from:
URL: ftp://ftp.cdc.gov/pub/infectious_diseases/iceid/2002/pdf-
smith.pdf
14. Tan CG, Sandhu HS, Crawford DC, Redd SC, Beach MJ, Buehler JW,
et al. Surveillance for anthrax cases associated with contaminated let-
ters, New Jersey, Delaware, and Pennsylvania, 2001. Emerg Infect
Dis 2002;8:1073–7.
15. Lazarus R, Kleinman K, Dashevsky I, Adams C, Kludt P, DeMaria A
Jr, et al. Use of automated ambulatory-care encounter records for
detection of acute illness clusters, including potential bioterrorism
events. Emerg Infect Dis 2002;8:753–60.
16. Kohane IS. Contributions of biomedical informatics to the fight
against bioterrorism. Journal of the American Medical Informatics
Association 2002;9:116–9.
17. Centers for Disease Control and Prevention. Public Health
Emergency Preparedness: Focus Area B: surveillance and epidemiol-
ogy capacity. [Cited April 21,2003.] Available from: URL:
http://www.bt.cdc.gov/planning/CoopAgreementAward
18. State of Michigan. Physician-disease reporting. [Cited July 22, 2002.]
Available from: URL: http://www.mdch.state.mi.us/pha/epi/cded/
physicianlistweb.pdf
Address for correspondence: Nkuchia M. M’ikanatha, Division of
Infectious Disease Epidemiology, Pennsylvania Department of Health,
Health and Welfare Building, P.O. Box 90, Harrisburg, PA 17108, USA;
fax: (717) 772 6975; email: nmikanatha@state.pa.us
PERSPECTIVES
Search p
past i
issues o
of E
EID a
at w
www.cdc.gov/eid

Automated Laboratory Reporting of Infectious Diseases in a Climate of Bioterrorism.pdf

More Related Content

Similar to Automated Laboratory Reporting of Infectious Diseases in a Climate of Bioterrorism.pdf (20)

More from Lisa Muthukumar (20)

Recently uploaded (20)

Automated Laboratory Reporting of Infectious Diseases in a Climate of Bioterrorism.pdf