The letter below was written by Dr. Patrik Bavoil, the Director of the University of Maryland-Baltimore’s Infectious Disease & Immunology Track (background on Dr. Bavoil: http://bms.dental.umaryland.edu/faculty.asp?ID=7). Using three examples of flawed CDC investigations—investigations of the 1984 Rajneesh cult bioterrorism case, the 2000 West Nile virus outbreak, and the 2003 laboratory infection of a USDA researcher (the investigation occurred in 2004)—Bavoil argues that the CDC’s investigations suffer from poor training which leads them to sometimes jump to incorrect conclusions with implications for public health and safety. It is a follow-up to a POGO July 2005 letter to several Congressional committees.
Patrik Bavoil, PhD
August 12, 2005
House Committee on Energy and Commerce
The Honorable Joe Barton, Chairman
The Honorable John D. Dingell, Ranking Member
The Honorable Nathan Deal, Chairman, Subcommittee on Health
The Honorable Ed Whitfield, Chairman, Subcommittee on Oversight and Investigations
2125 Rayburn House Office Building
Washington , DC 20515
Dear Representatives Barton, Dingell, Deal and Whitfield:
I write this letter as a concerned citizen and in reference to the letter dated July 6, 2005 , titled “POGO letter to Congress raising concerns with Centers for Disease Control and Prevention investigative ability"1, signed by Danielle Brian, Executive Director, Project on Government Oversight (POGO). In this letter, POGO investigators alerted you to the shortcomings of CDC investigations using the investigation of the recent laboratory-acquired E. coli O157:H7 infection of Dr. Ru-ching Hsia at an ARS-USDA research facility in Maryland as an illustration. I am an infectious disease microbiologist with over 25 years experience and I currently run an active research program at the University of Maryland , Baltimore , studying bacterial pathogens of the genus Chlamydia, including the dangerous Chlamydia psittaci, trachomatis LGV, pneumoniae and abortus. I also have specific knowledge of the ARS-USDA accident and of the investigation by officers of the Epidemic Intelligence Service (EIS), CDC, alluded to in POGO’s letter as a primary witness in both. I wish in this letter to detail some of the EIS investigative failures, draw parallels with earlier events and draw your attention to implications on the national response to the bioterrorist threat.
The investigation of Dr. Hsia’s laboratory-acquired infection has confirmed the expertise of CDC epidemiologists in tracing infectious agents from the point of contamination to the infected victim using sophisticated molecular methodologies. Paradoxically, this investigation has also revealed their comparative incompetence at investigating people, particularly people who have an incentive to cover their role in a contamination. In the case of Dr. Hsia, the investigators unequivocally traced E. coli O157:H7 recovered from the victim to the experiment using a DNA-based method called pulse-field gel electrophoresis. However, they were ultimately unable to identify the individuals who designed, performed and supervised the experiment and therefore caused the contamination. Instead, the CDC report from the investigation placed the blame on Dr. Hsia who, based on the scientific record, had not worked on E. coli O157:H7. While she laid in a coma, Dr. Hsia was blamed for supervising the faulty experiment and the technician performing it by the technician himself and his formal supervisor, who by doing so escaped further investigation. The administrative record however indicated that Dr. Hsia had no supervisory role. Soon after she woke up, Dr. Hsia testified that she was neither involved in the project nor responsible for the technician, yet her testimony was entirely ignored by the CDC investigators. Because of these investigative shortcomings, the true cause of the initial contamination was not identified and as a result at least two other individuals may have been infected later.
There exists a startling parallel between this flawed investigation in 2004 and a previous case also involving a CDC investigative team. In 1984, several hundred people were infected with Salmonella enterica serovar Typhimurium in a rural community in Oregon , in what is widely recognized as the first bioterrorist attack in the United States : the case of the Rajneesh religious sect. In this case again, EIS investigators were able to trace the infectious organism to its source using state-of-the-art methods and established that it had first appeared in various salad bars and restaurants in the area. However, as in Dr. Hsia’s case, they were unable to identify the real cause of the outbreak. They were unable to recognize the obvious, i.e. that the simultaneous appearance at different sites of a unique strain that had not been seen in outbreaks for decades --a strain that was in fact from a commercially available diagnostic strip-- was most likely caused by willful, man-made contamination. Since the outbreak stopped and miraculously no one died, CDC investigators went back to Atlanta recording this episode as a natural Salmonella outbreak. As in Dr. Hsia’s case, EIS epidemiologists ignored the testimony of the victims, many of whom asserted that the sect was responsible. The correct conclusion was never reached and the bioterrorist nature of this ‘outbreak’ was revealed to the FBI only a year later by a member of the sect.
An egregious parallel in the failures of EIS investigators to even get close to the cause of the contamination in 1984 and 2004 is their striking inability/unwillingness to weigh in circumstantial evidence. Science, by its very nature, tends to regard circumstantial evidence, even overwhelming, as suspect. In contrast, circumstantial evidence is critical to police investigations and may be used in court proceedings. For instance, there were strong indicators that the Oregon outbreak was man-made: bacteria from a diagnostic strip do not normally occur simultaneously at different places separated by miles without the means to travel from one place to the other. Some of the victims became ill after visiting the Rajneesh community where they were offered a drink that tasted strange. In the Maryland case, two other individuals became ill with identical symptoms months after Dr. Hsia had left for the emergency room. This alone should have raised suspicion that the cause of the first contamination had not been identified. However, one infection was not deemed to be laboratory-acquired because a different bacterial strain was recovered, not E. coli O157:H7. The other case was not investigated further because E. coli O157:H7 was not cultured from the patient. Since environmental contamination of these two individuals could not be ruled out, neither was officially linked to the first accident. Closer examination however indicates that the chance of environmental contamination was infinitely small while the probability that these two new cases were lab-acquired was indeed very high. One infection involved a strain of Campylobacter that had been rarely isolated from environmental sources and only years ago. The facility however maintained dozens such isolates for its research. The other infection was demonstrated to be linked to a recent E. coli O157:H7 infection by serology. Moreover, the victim of this infection was none other than the formal supervisor of the technician and of the experiment that led to the first contamination, i.e. the same individual who had initially placed the blame on Dr. Hsia and was caught distorting facts by the reporter from The Scientist.2
Although the scale of the outbreaks in Oregon and Maryland , the times --1984 and 2004-- and the organisms involved –Salmonella and E. coli-- are different, the constants are the relevant training of the investigators and investigative “tunnel vision” characteristic of both incidents. In both cases, testimony from the victims and multiple circumstantial evidence were ignored. In both cases, the individuals responsible for the contamination were exonerated by the EIS. In both cases, investigative shortcomings of the EIS investigators highlighted their lack of training at investigating people, particularly people who have clear incentives to distort facts. EIS investigators are indeed trained as epidemiologists, not police inspectors, and in as much as one would not expect a police inspector to be competent at investigating microbial behavior, one should not expect an epidemiologist to be competent at investigating human behavior.
Investigative “tunnel vision” by CDC was also a feature of the investigation of the West Nile virus outbreak in the summer and fall of 1999, during which West Nile virus was initially misidentified as St. Louis encephalitis virus3. In this case, CDC had its vision overly colored by initial tests that pointed to St. Louis encephalitis virus and ignored mounting clinical and epidemiological evidence that were inconsistent with the St. Louis virus.
A troubling corollary of the above comes from transposing these investigative failures to an actual small scale bioterrorist attack, for example a “test” attack or “dry run” by a terrorist cell. In such a scenario, the local State Dept of Health would receive the first call as most such attack would first be indistinguishable from a natural outbreak. State Health may then select to call in the CDC and the CDC may send EIS officers to investigate, if the outbreak is judged severe enough. Based on the investigation of Dr. Hsia’s accident, EIS epidemiologists would likely trace the culprit agent to its source. Unfortunately, they would also likely be ineffective at recognizing the roles played by individuals in the outbreak and would not recognize distortions of facts. It is likely that a small scale bioterrorist attack would be as likely to escape early detection today as the Rajneesh bioterrorist attack did in 1984. The recent, much publicized anthrax attack of 2001 was readily recognizable as an act of bioterrorism because it was targeted to well-known individuals. However, it was relatively small in scale and a valid question to ask is whether it would have been detected as an attack, had it not been targeted.
Current training in epidemiology of the CDC EIS officers is glaringly insufficient to investigate cases involving individuals who may lie to hide negligence or criminal intent. EIS officers should be trained in police work, particularly at recognizing and valuing circumstantial evidence that, in the absence of reliable testimony, may be the only evidence available to link a contamination to someone’s ill intent. Alternatively, an oversight commission should be created to ascertain that EIS investigators have not overlooked human distortions of epidemiologic findings and responsibilities. Indeed if EIS officers are unable to recognize fallacious testimony about an accidental contamination caused by someone’s negligence, how can one expect them to recognize lies a bioterrorist would make about a biological attack? An essential element of an efficacious response to a bioterror attack will be our ability to quickly and decisively distinguish such an event from a natural outbreak. I have little confidence that this would happen in view of the recent failed investigation of Dr. Hsia’s lab-acquired infection and the history of similar “tunnel vision” by CDC EIS officers.
I will be grateful for your renewed attention to this problem.
Patrik Bavoil, PhD
PS: In the event I cannot be reached, please contact POGO investigator Nick Schwellenbach at (202) 347 1122.
CC: Senate Committee on Health, Education, Labor and Pensions
The Honorable Mike Enzi, Chairman
The Honorable Edward Kennedy, Ranking Member
The Honorable Richard Burr, Chairman, Subcommittee on Bioterrorism and Public Health Preparedness
House Committee on Homeland Security
The Honorable Christopher Cox, Chairman
The Honorable Bennie Thompson, Ranking Member
The Honorable John Linder, Chairman, Subcommittee on Prevention of Nuclear and Biological Attack
Senate Committee on Homeland Security and Governmental Affairs
The Honorable Susan Collins, Chair
The Honorable Joe Lieberman, Ranking Member
The Honorable Norm Coleman, Chair, Permanent Subcommittee on Investigations