Reader Comments

Post a new comment on this article

Warning Sequence Regarding the 2009 H1N1 Influenza Pandemic

Posted by jmwilsonv on 25 Nov 2013 at 22:36 GMT

Here we offer a perspective from the viewpoint of those who recognized the unusual signature pattern in Mexico in April, later recognized as the 2009 H1N1 influenza pandemic:
http://biosurveillance.ty...

Hindsight recollection of data collected by a computer is insufficient for global surveillance operations. One must have an astute analyst who understands event signature processing, is able to triage those signatures for priority notification human-to-human, and promote the issuance of warning. It is the difference between declaration that the "information was there in the Internet... and it was collected with hundreds of other non-prioritized media articles" versus a human being having a direct phone conversation to explain why a particular signature is worthy of all attention by a government agency.

Report of H1N1 cases in California were documented through laboratory surveillance, however *recognition* of the threat it represented was indeed missing. It is only in hindsight these initial positives were claimed to be the first indicators. They were expressly not the initial *recognized* indicators.

Regarding verification of the situation in Mexico, it was clear that miscommunication between Mexican authorities and PAHO hindered the all-critical verification step needed to issue a worldwide advisory. Here the Canadians, based on their experience with SARS, provided us with the best example of warning to their national press. They stated a recognition of a most unusual situation but did not have verification- therefore, check travel histories in returning travelers.

As someone who designed, operationalized, and executed the mission for several of the capabilities mentioned in this paper, I would suggest that while technological improvements have been made in the detection and warning of serious public health threats, it is the sociology that will continue to impair full mission realization of biosurveillance.

James M. Wilson V, M.D.
Director
National Infectious Disease Forecast Center

No competing interests declared.

Identifying the 2009 H1N1 Influenza Pandemic: Response to Wilson

mstoto replied to jmwilsonv on 16 Dec 2013 at 17:06 GMT

It is hard to know what Dr. Wilson means by his comment that "Hindsight recollection of data collected by a computer is insufficient for global surveillance operations." If he is saying that automated analysis of electronic surveillance data, without human review and interpretation, is insufficient for surveillance operations, I agree. Indeed our analysis of the 2001 H1N1 pandemic provides evidence supporting this conclusion.

If, however, Dr. Wilson is referring to our analysis as a “hindsight recollection of data collected by a computer,” he has misread the paper. What we did was to recreate, from a number of practitioner perspectives and based on contemporaneous records as much as possible, how the outbreak developed and how knowledge about it developed and was shared globally. And what we found was sharply inconsistent with Dr. Wilson’s conclusions, based mainly on his involvement with Veratect.

Dr. Wilson states that "Report (sic) of H1N1 cases in California were documented through laboratory surveillance, however *recognition* of the threat it represented was indeed missing. It is only in hindsight these initial positives were claimed to be the first indicators. They were expressly not the initial *recognized* indicators." As we document in our analysis, however, the Mexican authorities were well aware of a number of outbreak signals for weeks before the Centers for Disease Control (CDC) published its results on the two California H1N1 on April 21, 2009, but the identification of a novel influenza virus was critical, allowing them to "connect the dots" and understand that these were all manifestations of the same epidemiological phenomenon. Indeed, this supports Wilson's call for "an astute analyst who understands event signature processing, is able to triage those signatures for priority notification human-to-human, and promote the issuance of warning."

Dr. Wilson’s claim that "miscommunication between Mexican authorities and PAHO hindered the all-critical verification step needed to issue a worldwide advisory" is not “clear” (as he says) but actually untrue. Mexico was open with PAHO, but before CDC published its results, no one knew the importance of the information that the Mexican authorities had. Our analysis shows that CDC's information about the novel H1N1 virus began to emerge about a week before the public report on April 21. Perhaps if Mexico and PAHO had been aware of this, their discussions during this week would have been different.

Dr. Wilson’s statement that the Canadians, based on their experience with SARS, urged preventive measures even though they did not have verification of the virus is misleading.
Canada's statement came one day after CDC's, and was based not only on CDC’s results, but on their communication with the Mexican authorities. Indeed, CDC’s confirmation of the novel viral strain, and global communication, was critical for everything, including Canada’s response.

Dr. Wilson ends with the suggestion that “while technological improvements have been made in the detection and warning of serious public health threats, it is the sociology that will continue to impair full mission realization of biosurveillance." If "sociology" means the social capital represented by the global surveillance systems and communication within it, I agree about the importance of this factor, and indeed our analysis shows how improvements in global communications made a positive difference in 2009.

No competing interests declared.

RE: Identifying the 2009 H1N1 Influenza Pandemic: Response to Wilson

jamesmwilsonvmd replied to mstoto on 26 Aug 2015 at 22:22 GMT

It is key for the readers to know that Dr. Stoto requested information of me personally regarding the Veratect timeline of events. Unfortunately, he failed to acknowledge the fact that timeline was provided to him by email and hence, worthy of acknowledgement in his manuscript.

Further, I am the individual who called CDC's Director's Emergency Operations Center on behalf of the 50 analysts who worked at Veratect and conversed with the senior health intelligence officer in charge of global disease detection the evening of April 20th. I do not recall at any point in time when Dr. Stotos was involved in the warning sequence of this event. Perhaps he can provide clarification on this point.

Regarding Stotos' comment,
"If, however, Dr. Wilson is referring to our analysis as a “hindsight recollection of data collected by a computer,” he has misread the paper. What we did was to recreate, from a number of practitioner perspectives and based on contemporaneous records as much as possible, how the outbreak developed and how knowledge about it developed and was shared globally. And what we found was sharply inconsistent with Dr. Wilson’s conclusions, based mainly on his involvement with Veratect."

This is difficult to reconcile given Veratect's input directly to Dr. Stotos was not acknowledged and further, Dr. Brownstein, who has made claims as the "first to publicly detect the pandemic", was acknowledged in the review of the manuscript. This indicates a potential conflict of interest.

The comment I would direct at Stotos and Brownstein relates to the comment regarding "hindsight recollection of data collected by a computer"- the proof is simply, what analyst at HealthMap recognized that that one signal among thousands of live threads was "the one" to call the CDC DEOC and have them page the senior health intelligence officer immediately, as I did? There is no evidence such action occurred at HealthMap. Further, Stotos resides at Georgetown, the onetime home of Project Argus. These are conflicts of interest that should have been disclosed, given these are both programs designed to provide warning of such public health threats and hence considered "competitors" of Veratect.

"And what we found was sharply inconsistent with Dr. Wilson’s conclusions, based mainly on his involvement with Veratect." *** I strongly encourage Dr. Stotos to be specific about this comment given what is stated above.***


"Dr. Wilson’s claim that "miscommunication between Mexican authorities and PAHO hindered the all-critical verification step needed to issue a worldwide advisory" is not “clear” (as he says) but actually untrue. Mexico was open with PAHO, but before CDC published its results, no one knew the importance of the information that the Mexican authorities had. Our analysis shows that CDC's information about the novel H1N1 virus began to emerge about a week before the public report on April 21. Perhaps if Mexico and PAHO had been aware of this, their discussions during this week would have been different."

Let us examine this paragraph more closely, with the final sentence being relevant:
"Perhaps if Mexico and PAHO had been aware of this, their discussions during this week would have been different."

1. I (nor Veratect writ large) did not have a personal or professional relationship with anyone in the Mexican Ministry of Health at the time. Similarly with both PAHO and Winnipeg. That is known as "gross miscommunication" in operational language. It is only with tremendous regret and hindsight that we recognize, "if Mexico and PAHO had been aware of this, their discussions during this week would have been different."

2. We did brief PAHO and WHO-Geneva and provided them both access to the Veratect system gratis prior to the emergence of H1N1. Similarly with CDC/ASPR. Not the case with Winnipeg, unfortunately. Again, lack of human contact and trusted communication among humans grossly inhibited recognition of a serious threat to international public health.

Regarding this comment, "Dr. Wilson’s statement that the Canadians, based on their experience with SARS, urged preventive measures even though they did not have verification of the virus is misleading." Again we will strongly disagree. The Canadians issued public disclosure of their concern about the events in Mexico ahead of both CDC and HHS. The backbone of community resilience in any community is the frontline healthcare providers. We consider the warning sequence, of which public communication is a part of, to have failed in this example. Writ large, it is comedy of errors and poor public-private communication that led to unnecessary delays in notifying the public and the healthcare providers upon which the public depends. It was apparent to us the Canadians did not delay and issued a press statement indicating concern and acknowledged a lack of ground verification versus waiting on the results of an investigation. To be blunt, we simply do not have time to wait when it comes to some rapidly evolving public health crises.

Regarding this comment, "Dr. Wilson ends with the suggestion that 'while technological improvements have been made in the detection and warning of serious public health threats, it is the sociology that will continue to impair full mission realization of biosurveillance'."

Allow me to clarify: by "sociology" I infer the social constructs that inhibit effective communication among key stakeholders to discuss time-sensitive threat awareness. The fact that Veratect was unable to have a healthy, trusted discussion with ASPR, CDC, Mexico, or Winnipeg impaired a full appreciation by all parties of the scale and magnitude of the involved threat. While CDC and the US Navy were focused on a sparse number of H1N1 cases in Texas and San Diego, we were observing reports of SARS rule-outs in Mexico, ICUs with "atypical pneumonia", and other grossly abnormal reporting patterns.

In the end, it is human behavior is the #1 obstacle that impairs effective operational biosurveillance. Indeed, the miscommunication even in this thread is evidence.

There have been many perspectives written regarding many key public health events of which are but partial representations of what actually happened. It is wise, when attempting to write an unbiased manuscript, to identify the individuals and organizations who were involved in the such operational activities and conduct an appropriate and unbiased investigation of the facts.


James M Wilson V, MD
Director, Ascel Bio World Infectious Disease Forecast Center
Director, Ascel Bio Nevada State Infectious Disease Forecast Station @ the University of Nevada-Reno
jamesmwilson@unr.edu



Competing interests declared: Former Chief Scientist, Veratect Corporation
Current Founder and Director, Ascel Bio World Infectious Disease Forecast Center