Introduction

Mr. Chairman and Members of the Subcommittee, my name is John R. Lumpkin, MD, MPH. I am a board-certified emergency physician and the Director of the Illinois Department of Public Health. I also serve as Chairman of the National Committee on Vital and Health Statistics (NCVHS), the advisory committee to the Secretary of the Department of Health and Human Services on health information policy. I appreciate the opportunity to speak at today’s hearing on bioterrorism preparedness, an issue that is so vital to our public health system and our national security.

Last Labor Day, few would have imagined that our nation would be rocked on September 11 by terrorist acts through the air, or that we later would be attacked by biologic agents through the mail. Now we must face and address this threat as the unthinkable – bioterrorism – has invaded our nation, including the very buildings from which this body governs.

Under the leadership of Governor George Ryan, the State of Illinois has been actively preparing to meet the challenges posed by bioterrorism since well before September 11. Our State Task Force on Bioterrorism has established plans and identified infrastructure needs to enhance our readiness. Centers for Disease Control and Prevention (CDC) funding appropriated by Congress has allowed us to upgrade our laboratories, enhance our communications, and train more than 1000 public and private health care workers to prepare them to meet the bioterrorism threat.

However, since September 11, we have reevaluated our preparedness and have identified areas where further enhancements are needed. I would like to focus my testimony today on one such area; specifically, the need for and importance of electronic disease reporting within the context of the health care data standard development process and the National Health Information Infrastructure (NHII).

History of Disease Reporting

Developments in England in the mid 1800s demonstrated the importance of collecting and analyzing reports of illness to identify outbreaks of disease. In 1854, Dr. John Snow identified an outbreak of cholera in a London neighborhood and linked it to a contaminated water source. He then went to the water pump on Broad Street and removed the handle. In so doing, Dr. Snow stopped the outbreak and pioneered the field of modern epidemiology.

Since then, improvements in statistical science have increased the level of confidence that epidemiologists can give to their analysis. However, the effectiveness and timeliness of interventions have depended upon the rapidity and quality of reporting from health care providers. The most important advance in timeliness occurred with the incorporation of the telephone in the last century. Despite the use of computers to analyze disease reports, however, health care providers’ reporting of diseases has changed little since the 1920s.

Challenges of Disease Reporting

Each State has laws requiring physicians to report communicable diseases as well as some chronic illnesses. When a physician diagnoses a reportable illness, he must complete a paper report form and mail it to the appropriate local health department. Highly contagious and dangerous diseases must be reported immediately by phone to local health departments, which then forward these reports to the state health department. State infectious disease staff members selectively record case data into an information system to allow analysis and report preparation.

However, physicians do not always report mandatory diseases, for a variety of reasons. For example, physicians sometimes fail to correctly diagnose the disease. Recent events with anthrax have shown how difficult it can be to diagnose an uncommon and unexpected disease. Obviously if the disease is not diagnosed, it cannot be reported. Another reason why physicians do not report mandatory diseases is because they are unfamiliar with reporting requirements, or simply are too busy to fill out the paper documents. Nevertheless, experience has shown that, once diagnosed, cases of very rare diseases like anthrax, plague, malaria, or smallpox will be reported quickly.

Clinical laboratories are also required to report infectious diseases. Although this requirement may lead to duplicative reports, clinical labs generally are more reliable and consistent in reporting than are physicians. Most clinical labs link laboratory machines with a computerized laboratory information system to manage large volumes and to automate reporting to physicians. However, since most State infectious disease information systems are unable to receive electronic data from clinical laboratories, clinical labs must print out hard copy lists to comply with reporting requirements. Laboratory staff then complete public health infectious disease report forms by hand for submission. Public health staff then extract data by hand from these reports for data entry into the State information system.

The Need for an Electronic Disease Reporting System

As we in public health struggle with this antiquated system, we know that an electronic disease reporting system could be developed using currently available technology. Such a system would permit a much quicker, more sensitive, and more efficient monitoring of unusual patterns of disease, which could result from the use of weapons of mass destruction or a naturally occurring emergent disease. To be fully functional, the electronic disease reporting system should be constructed following national data standards adopted pursuant to the Health Insurance Portability and Accountability Act of 1996 (HIPAA)(1), standards for patient medical information as proposed for adoption by the NCVHS(2), and the National Public Health Data Model (PHDM)(3) developed by the CDC for the National Electronic Disease Surveillance System (NEDSS)(4).

Efficient detection of abnormal disease patterns requires the collection, aggregation, and analysis of data at both the State and national level. Occasionally, however, outbreaks are detected by luck. Two years ago, a nationwide outbreak of Salmonella Agona was first identified in Illinois. Three cases of this rare strain were reported to local health departments. A clerical employee performing routine data entry noted an unusual number of cases in a single month. The outbreak would not have been identified if this connection between these three seemingly unrelated cases had not been noticed. Without this accidental discovery, hundreds more would have been infected before the offending product was pulled from the shelves. We will never know how many other outbreaks have gone unreported or undetected.

Unlike our experience in Illinois, a medium- to large-scale bioterrorism event might produce sentinel cases appearing in a number of communities with reports to many local health departments. We cannot rely on luck to detect such outbreaks in the future. The public health surveillance system must be sensitive enough to identify these incidents as early as possible to minimize the spread of disease and maximize the number of lives saved.

Existing surveillance systems are too slow and insensitive to meet this challenge. Clinical symptoms may not appear to be significant when taken one patient at a time. However, when viewed from a community, state, or nationwide perspective, patterns become apparent indicating the need for public health response.

A major obstacle for better disease surveillance is the difficulty in getting timely and accurate clinical data into public health data systems. Current systems are too labor-intensive, creating barriers for participation by physicians, who increasingly must see more and more patients in less time. Two years ago, an outbreak of invasive Group A Streptococcus associated with eleven deaths was not reported until three months after the first death occurred.

An electronic reporting system based upon national data standards would help overcome this obstacle. National standards would allow vendors of medical information systems to build reporting modules into current and future products. Use of HL7 standards for laboratory reports would facilitate the electronic transmission of positive laboratory tests. Internet technology with public key infrastructure (PKI) and data encryption would allow disease reports to securely travel from caregiver to the public health agency.

Data mining technology then would enable public health epidemiologists to identify patterns which may represent disease outbreaks. The funding of the public sector component of the system should include adequate resources to allow for sophisticated analysis of the data collected. Of course, strong privacy protections must be implemented to restrict access to individual personal health information.

The diffusion of electronic medical record keeping has been slow in health care practice. The Institute of Medicine(5) and the Leapfrog Group(6) documented the adverse impacts on the quality and efficiency of health care due to a lack of automation of health care business processes. This presents a barrier to the development of a truly comprehensive electronic disease surveillance system.

In the interim, while health care information systems mature, sentinel surveillance systems could supplement electronic disease reporting systems. For example, the current influenza reporting system is such a sentinel system. Influenza currently is not reportable due to the immense burden such a requirement would place on the health care system in a non-automated environment. Instead, a sample of emergency departments and primary care physicians have been identified to report cases of influenza and submit samples for culture and identification.

A significant percentage of emergency departments use electronic systems or dictation to generate medical records, but they are then transcribed onto paper. Electronic forms of these records, stripped of identifiers, could be accessed through data mining technology to monitor abnormal patterns of known diseases or the emergence of common symptom complexes indicative of a disease outbreak. However, the operation of this system would require the use of national standards for data format and contact. Stand-alone systems with proprietary data definitions and format would impede the rapid movement of information.

Scenarios to Illustrate an Effective Disease Reporting System

I would like to describe two scenarios that illustrate possible applications of currently available technology to improve public health surveillance and health care delivery. These scenarios demonstrate the need to envision the development of electronic disease surveillance as part of an overall strategy in the development of a National Health Information Infrastructure.

Scenario 1:

Fred S. moves across the country to start a new job. He has already chosen a medical practice in his new town because it offers the same online health support service as his previous doctor, even though he is covered by a different health plan. Fred can set up appointments, get prescription drug refills and lab test results, e-mail doctors or nurses and manage his personal health history. A week after he arrives at his new home, he develops fever and a runny nose. He e-mails his new doctor a list of his symptoms, fearing that he may have anthrax or smallpox.

The doctor's computer system automatically does several things. First, it asks Fred to give his full itinerary over the previous fourteen days and compares it with the public health database of anthrax and smallpox occurrences -- which is updated constantly by the CDC -- to see if Fred has been in places where he could have been exposed. Next, the system runs his symptoms against his own personal health record, including his list of medications, which Fred had updated before moving. Finally, it sends an urgent alert to the doctor to review the reports the system has generated.

Upon receiving the report, the doctor sees no likely source of exposure for Fred but notices a potential drug interaction among his medications. The doctor calls Fred and tells him that the new drug he just started for his urinary condition could have caused an adverse reaction. She feels confident that Fred does not need to come in for tests or take unnecessary antibiotics. Rather, she simply prescribes a new drug (the new prescription is automatically sent to his closest pharmacy) and asks Fred to e-mail her in 24 hours with an update of his symptoms. The next day, his e-mail message confirms that his fever and other symptoms are gone.

This system affords enormous savings for providers, payers and patients by avoiding unnecessary lab tests (reimbursement for the eHealth consultation is about the same as for a clinical visit), hours of investigation by public health authorities, hours of anxiety for Fred (which could have become a full scale panic shared with friends and family), and the risk of contributing to bacterial resistance by prescribing an unneeded antibiotic. This "non-event" is the happiest of all endings for Fred, his doctor and the health of the public.

Scenario 2:

The FBI notifies the Illinois Department of Public Health of a credible threat of bioterrorism. It is believed that a foreign-funded group has acquired a cache of Y. Pestis (plague) bacteria and is planning an attack on the anniversary of the collapse of the al Qaeda terrorist network. The Illinois Department of Public Health sends out an alert through the Health Alert Network (HAN) to all local health departments. In addition, a similar alert is sent to all hospitals and emergency departments.

The signs and symptoms of all forms of plague are incorporated into software that is then downloaded to the clinical information systems of clinicians throughout the state. Dr. Smith’s system identifies two patients with a matching clinical profile in his practice. After approval by Dr. Smith, the system notifies the two patients and their local health information system by phone. They agree to come in to his office later that day.

That morning, Dr. Smith sees another patient with what appears to be a pneumonia and hemoptysis (i.e., he is coughing up blood). Dr. Smith prepares to send the patient to the hospital for x-rays and cultures when his office information system warns him that this patient’s symptoms fit the recently updated public health surveillance profile. He forwards a notice to the public health department and sends the patient into the hospital for further evaluation. For this patient, the public health laboratory assists in making the diagnosis of a common pneumonia. In this manner, patterns of reports by Dr. Smith and other physicians are shared with the FBI as they continue to be alert to possible terrorist acts.

Proposed Action: Develop a National Health Information Infrastructure

The NCVHS has proposed the development of a National Health Information Infrastructure (NHII) to enhance the movement of data between individuals, caregivers and community-focused health organizations such as public health agencies. The ‘NHII’ is a set of technologies, standards, and applications that support communication and information to improve clinical care, monitor public health, and educate consumers and patients. It is not a unitary database.

The broad goal of the NHII is health knowledge management and delivery, so that the full array of information needed to improve the public's health and health care is optimally available for professionals, policy makers, researchers, patients, care givers and consumers. The NHII as a system should seek to improve and enhance privacy and confidentiality of personal health information.(7) The NHII would enhance preparedness by allowing the early identification of abnormal patterns of disease, providing decisional support for clinicians to improve quality of care, and identifying significant public health symptom complexes. Further, the NHII would facilitate the dissemination of health information to the public to guide individual health decision making and for alerts related to bioterrorism. National leadership and resource commitment is necessary to make this possible.

Conclusion

Over the last few weeks, we have been forced to accept the reality that biologic agents can be used as tools of terrorism. Prior to this, our image of terrorism involved a dramatic event with explosions, fires, or even chemical attacks. As we have learned, bioterrorism can be silent and insidious.

In fact, the recognition that such an attack has occurred is dependent upon a national system of disease reporting. To assure rapid and accurate identification of an attack, which is necessary to allow appropriate responses, our disease reporting system needs to be upgraded, updated and automated. The National Health Information Infrastructure as proposed by the NCVHS describes the conceptual model of the environment in which this system would be most effective. However, in the short term, the automation of current disease reporting systems, coupled with electronic emergency department sentinel systems, would provide a heightened level of surveillance.

Mr. Chairman, this concludes my prepared statement. I would be pleased to answer any questions you or other Members may have.

(1) http://www.ncvhs.hhs.gov/stdslet1.htm
(2) http://www.ncvhs.hhs.gov/hipaa000706.pdf
(3) http://www.cdc.gov/od/hissb/docs/phcdm.htm
(4) http://www.cdc.gov/od/hissb/docs/NEDSS%20Intro.pdf
(5) http://books.nap.edu/html/to_err_is_human/
(6) http://www.leapfroggroup.org/safety1.htm#CPOE
(7) http://www.ncvhs.hhs.gov/NHII2kReport.htm