Appendix C
Summary of Research into the Costs of Enhanced Public Health Surveillance Systems
Henry Roman
Industrial Economics, Incorporated
In 2003, the U.S. Department of Homeland Security (DHS) initiated an environmental surveillance system intended to detect the intentional aerosolized release of biological pathogens. The program, called “BioWatch,” deploys air samplers, primarily in outdoor locations, in more than 30 jurisdictions throughout the United States. Over the next 4 years, DHS intends to upgrade the capabilities of the current system through improved technology and expanded coverage, both of existing BioWatch jurisdictions and expansion to additional jurisdictions and indoor facilities.
At the request of the Congress, the Institute of Medicine (IOM) and the National Research Council (NRC) have been asked by DHS’s Office of Health Affairs (OHA) to evaluate the effectiveness of the current BioWatch program (BioWatch) relative to planned improvements to the existing program and to current surveillance performed through the U.S. public health and hospital systems. This effort will be conducted by the Committee on Effectiveness of National Biosurveillance Systems: BioWatch and the Public Health System (the committee). The committee consists of experts in areas that include biological threat assessments, evaluation of biological detection systems, environmental monitoring technologies, biological assays, microbiology, virology, epidemiology, syndromic surveillance, health information technology, the U.S. public health sector, hospital systems, emergency medicine, laboratory operations, statistical methods, systems engineering, operations research, and economic analysis.
IOM/NRC engaged Industrial Economics, Incorporated (IEc), to provide analytical support to the committee to address cost-related aspects of its charge. This support included (1) an assessment of the costs of imple-
menting current and future realizations of the BioWatch monitoring system; and (2) research into the costs of “the current and a potential ‘enhanced national surveillance system’ to provide a basis for a rapid response to bioterrorist attacks or other biothreats, including initiation of pre-infection prophylaxis and expedited response and recovery.”1 This memo focuses on IEc’s efforts to acquire and analyze cost data on the public health system both as it currently exists and for cutting-edge programs aimed at improving surveillance and response capabilities. Note that this memorandum is not intended to provide information about the merits of the current or “enhanced” public health system or provide a snapshot of current public health surveillance capabilities nationwide. Furthermore, for the reasons detailed below, IEc and the committee have concluded that the available data do not support a comprehensive cost analysis of either current or enhanced public health activities related to biosurveillance and outbreak response. Thus, this memo does not present such an analysis; instead, it describes the major challenges to performing a cost analysis of public health programs, describes IEc’s effort to obtain cost data for example programs operating at the state or local level that match key biosurveillance and response activities identified by the committee, and discusses IEc’s observations regarding the limited data set we were able to obtain.
We include example costs for specific state and local programs in Table C-2.
OBSTACLES TO COST ANALYSIS
Significant obstacles preclude our ability to generate a comprehensive cost estimate for the current or an enhanced U.S. public health system’s biosurveillance and response efforts. The primary problem is a general lack of financial transparency and accountability across the U.S. public health system. A number of papers and reports have documented this issue (e.g., Hebert et al., 2007; Honoré et al., 2007; TFAH, 2008). In part, this lack of transparency reflects differences in the organization of the public health system from state to state, where some have a more centralized system of public health responsibilities and others spread those responsibilities across multiple agencies. More importantly, there is no uniformly recognized classification system for expenditures on public health, leading to confusion about what constitutes public health spending. The report from the Trust for America’s Health (TFAH, 2008) presents examples of inconsistencies in cost accounting for specific public health initiatives both across states and
also year-to-year within the same state.2 This lack of transparency makes it difficult to do a top-down analysis of the portion of federal bioterrorism preparedness funds that are spent on biosurveillance and response activities by state and local health departments. It also precludes a “bottom up” approach to estimating funds spent on bioterrorism surveillance and response, given the difficulties both identifying and isolating spending in specific public health categories.
Another challenge for the committee is defining the entity—the “enhanced national surveillance system”—for which costs are to be assessed. No current definition of such a system exists, though efforts to define such a system are ongoing. Given the tight time frame of the committee’s analysis, and not wishing to preempt current efforts in this area, the committee opted to instead highlight opportunities for enhancement of surveillance through the public health and health care systems in broad categories of legally mandated reporting, automation of health care information systems and public health linkages, laboratory and diagnostic testing capacity, and information integration. We have used these categories to guide our cost research.
APPROACH AND DATA SOURCES
Faced with these challenges and the timeframe of this analysis, IEc adapted its cost analysis to focus on providing example cost data for activities or programs at the state or local level that match the surveillance- and response-related categories described above. The results were intended to provide some insight into the potential order of magnitude or range of magnitudes of the costs of developing and maintaining systems that improve reporting, surveillance, and response. IEc focused its analysis on state or local public health programs and initiatives, though it did review one source reporting cost estimates for implementing electronic laboratory reporting (ELR) nationwide.
Data Sources
The cost information we obtained came from two categories of sources:
-
Published reports and journal articles. IEc conducted a thorough review of all cost-related documents on the IOM/NRC portal website submitted by committee members, IOM/NRC staff, and the U.S. Centers for Disease Control and Prevention (CDC) for relevant cost estimates for the categories provided by the committee. IEc also searched for additional pub-
-
lic health cost analyses using the National Library of Medicine’s PubMed database.3 A table listing all the documents reviewed by IEc has been previously submitted to the committee.4
-
Telephone interviews. Most of our data came from e-mail exchanges and telephone interviews with state and local health officials and representatives of public health organizations such as the Association of State and Territorial Health Officials (ASTHO), the National Association of County and City Health Officials (NACCHO), the Council of State and Territorial Epidemiologists (CSTE), and the Association of Public Health Laboratories (APHL). The individuals contacted in December 2008 and January 2009 were identified through discussions with the aforementioned organizations, the committee, and IOM/NRC staff. IEc sent inquiries to individuals from state public health departments in Florida, Massachusetts, Michigan, Minnesota, Nebraska, New York, North Carolina, Pennsylvania, and Washington. IEc obtained cost data or conducted follow-up interviews with representatives of all of these states except Michigan, New York, and Pennsylvania. IEc also spoke with committee member James Buehler, who provided cost information about health departments in Georgia. At the local level, IEc interviewed a representative of the Tarrant County, Texas, health department, and we obtained data for the New York City health department from committee member Marci Layton. IEc also spoke with a University of Pittsburgh researcher currently developing a biosurveillance grid covering areas of Pennsylvania and Ohio. Individuals contacted by IEc are shown in Table C-1.
RESULTS AND OBSERVATIONS
The results of IEc’s research efforts are included in Table C-2. The data are quite limited; however, we can make a few observations.
-
The level of detail of these costs was generally rather broad, as expected, with responses often given in terms of development and operations and maintenance (O&M) costs.
-
To the extent we can tell from limited data, the relative scale of the magnitude of costs across states and localities appears reasonable.
-
We obtained one national-level cost estimate—the cost of implementing a national ELR infrastructure, as estimated during a 2007 meeting of
3 |
IEc also reviewed CDC budget documents online (http://www.cdc.gov/fmo/topic/Budget); however, due to the difficulties in tracking CDC funding to the state and local level, we did not derive any cost estimates from these data. |
4 |
See e-mail messages from Henry Roman, IEc, to Lois Joellenbeck, NAS/IOM, dated December 20, 2008, and January 24, 2009. |
-
local, state, federal, and private-sector experts sponsored by CDC through its cooperative agreement with APHL. The initial estimates included $50 million for planning and building basic infrastructure where needed, $600 million for development, and $100 million per year for O&M costs.
-
IEc was most successful in identifying cost information for the larger information technology (IT)-related efforts—ELR, syndromic surveillance systems, health information exchanges (HIEs), and electronic disease surveillance and outbreak management software. However, these data were often incomplete—for example, missing O&M costs in some cases because the systems are too new. IEc was least successful obtaining cost estimates for marginal expansion of laboratory testing and diagnostic capabilities, clinical decision support tools and point-of-care diagnostic tools, health alert networks, and fusion centers.
-
The interconnectedness of some of these systems makes it difficult to separate out costs of individual systems. For example, to implement ELR in North Carolina, the health department needed to develop a National Electronic Disease Surveillance System (NEDSS) function in order to receive ELR information. To do this, North Carolina funded the development of the Maven software system at a cost of over $3 million, with $650,000 annually in O&M costs. To estimate costs across the current categories, one would need to develop an approach to dividing the development (or purchase) and O&M costs of Maven across categories when such overlap exists, to avoid double-counting.
-
Another complicating factor in assessing the costs of these initiatives relates to accounting for costs to all parties involved. For example, the true costs to hospitals that participate in ELR or syndromic surveillance systems may not be included in these estimates. Also, in the case of Washington State, one of their major syndromic surveillance systems (Electronic Surveillance System for the Early Notification of Community-based Epidemics, or ESSENCE) was provided to them by Johns Hopkins University at no cost, including an estimated $50,000 in hardware (Wayne Turnberg, WADOH, personal communication).
-
Customization can be a driving factor in the costs of syndromic surveillance systems, and the difference in cost can be dramatic. In addition to the ESSENCE example above, the Real-time Outbreak and Disease Surveillance (RODS) system, developed at the RODS laboratory at the University of Pittsburgh, is an open-source system that is available at no cost to health departments. If a department wishes to customize RODS, it either needs to have the IT resources available in-house to modify the code, or it needs to contract out the modifications to an IT firm (Rich Tsui, University of Pittsburgh, personal communication). The Outbreak Detection Information Network (ODIN) syndromic surveillance system, developed by Washington State to complement ESSENCE, features greater customization
-
options and incorporates additional data sources; the cost of developing ODIN was approximately $5 million over 4 years.
-
In some cases, notably the HIEs, we observe a significant range in development costs—from a Tarrant County system with development costs under $200,000 to the PA-OH surveillance grid system ($1 million) to the Washington State HIE, which is being developed under a $5 million CDC grant. The Tarrant County effort, described as a “rudimentary” HIE that covers 50 to 60 hospitals, illustrates a less highly tailored approach, using open-source syndromic surveillance systems (RODS and ESSENCE) “off the shelf” to implement surveillance (William Stephens, Tarrant County (TX) Health Department, personal communication). While some of the cost differences with the other systems may be due to scope or other factors, we expect much of the difference reflects a focus on front-end development, providing greater customization and flexibility and features.
REFERENCES
Hebert, K., N. Henderson, and E. Gursky. 2007. Building preparedness by improving fiscal accountability. Journal of Public Health Management and Practice 13(2):200–201.
Honoré, P.A., R.L. Clarke, D.M. Mead, and S.M. Menditto. 2007. Creating financial transparency in public health: Examining best practices of system partners. Journal of Public Health Management and Practice 13(2):121–129.
Neumann, P.J., P.D. Jacobsen, and J.A. Palmer. 2008. Measuring the value of public health systems: The disconnect between health economists and public health practitioners. American Journal of Public Health 98(12):2173–2180.
TFAH (Trust for America’s Health). 2008. Shortchanging America’s health 2008: A state-by-state look at how federal public health dollars are spent. Washington, DC: Trust for America’s Health. http://healthyamericans.org/reports/shortchanging08/ (accessed March 27, 2009).
TABLE C-1 Public Health Cost Data Contacts
Health Department/Organization |
Contacts |
APHL |
Scott Becker Chris Mangal Michelle Meigs Patina Zarcone (January 23, 2009) |
ASTHO |
James Blumenstock |
CSTE |
Lisa Dwyer |
Emory University |
James Buehler* (December 17, 2008) |
Florida |
Richard Hopkins (January 14, 2009) |
Massachusetts |
Dina Caloggero |
Michigan |
James Collins |
Minnesota |
Richard Danila |
NACCHO |
Jack Herrmann, William Stephens (Tarrant County (TX) Public Health, December 18, 2008) |
Nebraska |
Thomas Safranek (January 14, 2009) |
New York City |
Marci Layton* |
New York State |
Robert Burhans |
North Carolina |
Jeffrey Engel (January 16, 2009) |
Pennsylvania |
Veronica Urdaneta |
University of Pittsburgh |
Fu-Chiang (Rich) Tsui (December 18, 2008) |
Washington State |
Michael Davisson John Erickson Margaret Hansen Judy May Wayne Turnberg (all January 20, 2009) |
NOTE: Interview dates in parentheses. APHL, Association of Public Health Laboratories; ASTHO, Association of State and Territorial Health Officials; CSTE, Council of State and Territorial Epidemiologists; NACCHO, National Association of County and City Health Officials. *IOM/NRC committee member. |
TABLE C-2 Public Health Cost Matrix: Examples of Surveillance Program Costs
Category |
Example Programs |
Environmental Monitoring Programs |
|
Enhance animal, wildlife, vector testing |
Canada’s National Wildlife Disease Strategy |
Legally Mandated Reporting and Surveillance Systems: Notifiable Diseases, Outbreak Reporting, Vital Event Registration |
|
Enhance electronic laboratory reporting systems |
1. APHL/ANSER/CDC estimate for developing a national ELR system |
|
2. NC (online for 2 labs) |
|
3. WA: PHRED (online for 5 of larger hospital labs, about 50% of test results) |
|
4. NE (developing) |
|
5. MA |
Cost Information |
Cost Information Reference |
Comments |
No data yet on Canadian program. Paper identified by committee member Seth Foldy suggests $4–5 million over 3 years to design and implement global animal surveillance system for zoonotic pathogens |
Kuiken et al. (2005) |
|
1. Development: $650 million to plan ($50 million) and implement ($200 million/yr over 3 years). Planning includes building state public health LIMS where needed ($25 million). O&M: $100 million/yr estimated O&M costs |
1. APHL et al. (2007) |
MA values assume ~$53,000 labor, $140,000 in software, and ~$3,000 in hardware support for 6 servers |
2. Development: $400K for STARLIMS product for labs; cost per hospital can vary, can be as much as $40,000. O&M: no data yet. Costs to Receive ELR: needed to develop NEDDS systems; Development: $3.25 million to develop and implement Maven system; O&M: $650,000 ($500,000/yr for Maven; $150,000/yr for ELR coordinator) |
2. Jeffrey Engel, NC, personal communication |
|
3. Development: do not have yet; O&M: about $200,000/yr (half technology, half operations) |
3. Michael Davisson, Wayne Turnberg, Judy May, WA, personal communication |
|
4. Development: estimates $600,000 ($20,000 per lab); O&M: estimates $50,000–75,000/yr personnel costs |
4. Thomas Safranek, NE, personal communication |
|
5. Development: pending; O&M: $196,000/yr |
5. Dina Caloggero, MA, personal communication |
|
Category |
Example Programs |
Enhance notifiable disease reporting by clinicians (outreach, electronic reporting procedures, 24/7 call line) |
|
Enhance electronic death reporting systems |
1. Boston surveillance system receives death certificate data issued in City of Boston |
|
2. WADOH (5 counties) |
Automation of Healthcare Information Systems and Public Health Linkages |
|
Enhance use clinical decision support tools for diagnosis, reporting, & management (e.g., triage, isolation, treatment) |
eTriage VisualDx |
Cost Information |
Cost Information Reference |
Comments |
Outreach programs generally involve 1 staff member working full-time as liaison (~$54,000/yr) |
Assume 1 junior epidemiologist FTE for Outreach “liaison” in public health department in GA (James Buehler, personal communication) “though staff varies with size of department, e.g., percentage of several MDs’ time employed in outreach in NYC” (Marci Layton, personal communication). Outreach materials: Marci Layton, personal communication |
|
Outreach materials can cost up to several hundreds of thousands of dollars per year (upper bound, NYC) |
|
|
For Internet-based systems, there is a startup cost plus 1 IT FTE to maintain it (~$48,000/yr). |
|
|
|
Cost of full-time staff from the State of Hawaii Department of Health Bioterrorism Preparedness Response Program–FY 2005 Salary Survey |
|
1. $74,389 (development; also includes costs for 911 and EMS data) |
1. Kirkwood et al. (2007) |
|
2. Development: awaiting costs on latest version; O&M: unknown |
2. Michael Davisson, WADOH, personal communication |
|
VisualDx: $20,000 per hospital |
|
|
Category |
Example Programs |
Enhance use of information from EMRs and other electronic health information sources to detect unusual cases of disease or trends (public health collects and collates information from multiple institutions) |
Syndromic Surveillance systems such as RODS, ESSENCE |
Systems in many states, including MA, NY, NC (NCDetect), WA |
|
Also local HIE efforts to increase integration with EHRs, beyond typical syndromic surveillance features—more public health user control over data reporting, reachback features (e.g., WA state, Indiana, NY?) |
|
|
Electronic Support for Public Health (ESP): an automated, secure, portable public health detection and messaging system for cases of notifiable diseases. Uses electronic medical data and supports an optional case management workflow system for case notification control. All relevant clinical, laboratory, and demographic details are transferred to the local health authority |
Cost Information |
Cost Information Reference |
Comments |
Boston, MA: An electronic, emergency department-based syndromic surveillance system for the Boston Public Health Commission cost $422,899 between December 2003 and July 2005. This included: |
Kirkwood et al. (2007) |
|
|
|
|
NYC: O&M for syndromic surveillance system estimated to be $130,000–$150,000/yr |
Heffernan et al. (2004) |
|
NC: NCDetect, statewide. Development: $1.9 million over 3 years (does not include some costs to hospitals); O&M: $600,000/yr. Also, 60 hospitals have reachback enhancement for an extra $3 million ($50,000 per hospital) |
Jeffrey Engel, NC, personal communication |
|
WA: uses ESSENCE and ODIN; ESSENCE provided at no cost to WADOH; minimal staff maintenance/review time. ODIN (Kitsap and Pierce counties) Development: ~$5 million over 4 years; O&M: unclear, some minimal staff maintenance and review time. HIE with EMR extraction for syndromic surveillance Development: $5 million in CDC grant over 3 years |
Michael Davisson, Wayne Turnberg, Judy May, WA, personal communication |
40 hospital partners for HIE in WA; CDC grant for HIE went to University of Washington, INHS (a RHIO), SAIC |
No cost data yet on ESP |
|
ESP developed through a collaboration between CDC, the CDC-funded Center of Excellence in Public Health Informatics based at Harvard, Harvard Vanguard Medical Associates, and the Massachusetts Department of Public Health |
Category |
Example Programs |
|
Massachusetts eHealth Collaborative, which uses health information technology through community-based implementation of EHRs and health information exchange |
Enhance use of regional health information exchanges to detect unusual cases of disease or trends (public health interfaces with area-wide exchange that involves multiple institutions) |
1. Tarrant County, TX system |
|
2. PA-OH BiG (under development) |
Enhance public health capacity to electronically alert healthcare sector of important public health events (e.g., inform evaluation, triage, isolation, treatment) |
|
Laboratory and Diagnostic Testing Capacity |
|
Expand access to rapid diagnostic tests, “point of care” tests |
Luminex |
Extend capacities to characterize pathogens and coalesce reports to identify emergence or spread of related cases of infectious disease (e.g., PulseNet model) |
|
Cost Information |
Cost Information Reference |
Comments |
eHealth Collaborative: $50 million was pledged by Blue Cross Blue Shield of Massachusetts to establish an all-stakeholder organization that would implement three full funded community-wide EHR demonstration projects across Massachusetts over 3 years |
Goroll et al. (2009) |
|
1. Development costs: $150,000; O&M: $70,000–$75,000/yr; Future enhancements (includes adding ELR): $150,000–$200,000 |
1. William Stephens, manager, Southwest Center for Advanced Public Health Practice; Tarrant County Public Health, personal communication |
|
2. Development costs: $1 million ($500,000 over 2 years); O&M: similar to maintenance costs in current public health system |
2. Rich Tsui, University of Pittsburgh, RODS laboratory, personal communication |
|
NYC: Several million in startup costs plus 3–4 FTEs to operate/maintain (~$200,000/yr) |
Marci Layton, personal communication |
|
Cost of full-time staff from the State of Hawaii Department of Health Bioterrorism Preparedness Response Program, FY 2005 Salary Survey |
|
|
Development costs: $4–6 million over 2 years; unit cost: <$50,000 |
|
|
Cost Information |
Cost Information Reference |
Comments |
1. Maven in NC: Devevelopment: $3 million; O&M: $500,000/yr (overlap with ELR?); Maven in MN: $800,000 to purchase; no O&M given |
1. Jeffrey Engel, NC, personal communication; Rich Danila, MN, personal communication |
|
2. WATrac Development: $160,000; O&M: $45,000 |
2. Margaret Hansen, WADOH, personal communication |
WATrac: bed tracking and ED status in 82% of WA hospitals; resource and pharmaceutical tracking anticipated to be implemented in 2009 statewide |
exchange; IT, information technology; LIMS, laboratory information management system; NEDSS, National Electronic Disease Surveillance System; NBIC, National Biosurveillance Information Center; NBIS, National Biosurveillance Information System; O&M, operation and maintenance; PHRED, Public Health Reporting of Electronic Data; RHIO, regional health information organization; RODS, Real-time Outbreak and Disease Surveillance; WADOH, Washington State Department of Health. |
REFERENCES
APHL (Association of Public Health Laboratories), ANSER, and CDC (Centers for Disease Prevention and Control). 2007. Framework for electronic laboratory reporting: Recommendations to policymakers. A report of the Electronic Laboratory Reporting Consensus-Building Meeting, October 30, 2007. Arlington, VA: Analytic Services Inc. http://www.aphl.org/policy/Documents/Electronic_Lab_Reporting_2007.pdf (accessed March 26, 2009).
Goroll, A.H., S.R. Simon, M. Tripathi, C. Ascenzo, and D.W. Bates. 2009. Community-wide implementation of health information technology: The Massachusetts eHealth Collaborative experience. Journal of the American Medical Informatics Association 16:132–139. http://www.jamia.org/cgi/content/full/16/1/132 (accessed August 10, 2009).
Heffernan R., F. Mostashari, D. Das, M. Besculides, C. Rodriguez, J. Greenko, L. Steiner-Sichel, S. Balter, A. Karpati, P. Thomas, M. Phillips, J. Ackelsberg, E. Lee, J. Leng, J. Hartman, K. Metzger, R. Rosselli, and D. Weiss. 2004. New York City syndromic surveillance systems. Morbidity and Mortality Weekly Report 53(Suppl.):23–27.
Kirkwood, A., E. Guenther, A.T. Fleischauer, J. Gunn, L. Hutwagner, and M.A. Barry. 2007. Direct cost associated with the development and implementation of a local syndromic surveillance system. Journal of Public Health Management Practice 13(2):194–199.
Kuiken, T., F.A. Leighton, R.A.M. Fouchier, J.W. LeDuc, J.S.M. Peiris, A. Schudel, K. Stohr, and A.D.M.E. Osterhaus. 2005. Public health: Pathogen surveillance in animals. Science 309(5741):1680–1681.