| by Edward H. Kaplan, PhD My colleagues and I believe that Council on Scientific Affairs Report 2 (I-02), "Smallpox: A Scientific Update," inaccurately portrays our smallpox disease transmission model1 and therefore its potential validity in evaluating the relative merits of traced versus mass vaccination in the event of a smallpox bioterrorist attack in a large U.S city. The Kaplan model assumes that disease transmission derives from free mixing among available susceptibles and asymptomatic infectious individuals in the population, and this approach deliberately addresses the worst case scenario. Ro, the reproductive rate of infection, governs the actual number of infections transmitted per infected individual. In our base case studies, this number is equal to 3, which is at the lower end of published smallpox transmission rates. In sensitivity analyses, we considered values much lower than this (and also higher values). The most important result of our model is that even if all contacts of all smallpox cases are found (ie, perfect contact tracing), the number of deaths under traced vaccination will exceed the number of deaths under mass vaccination for any reproductive number Ro >2. The advantage of mass vaccination increases as the initial number of those infected in a smallpox attack increases. In comparing mass and traced vaccination, the CSA report argues that mass vaccination receives an unfair advantage over traced vaccination because mass vaccination does not require the time necessary to trace and locate persons. However, this is not an artifact, but rather is at the very heart of the tradeoff between mass and traced vaccination! Tracing is all about locating "higher yield" subjects to vaccinate (ie, contacts), while mass vaccination is about rapidly building the level of immunity in the population. The tradeoff is the relatively slow vaccination rate of targeted contacts versus the relatively quick vaccination of untargeted individuals. Our analysis clearly demonstrates that for large attacks, speed is more important than specificity, although for small outbreaks the reverse may well be true. Indeed, there is a fundamental difference between mass and traced vaccination that stems from response logistics. The rate with which a population can be immunized via mass vaccination depends on the resources available. If there are sufficient personnel, the population of a city can be vaccinated in 10 days. If there are fewer, the vaccination takes longer. However, the time required to complete the operation depends on the vaccination resources available. Such is not the case for traced vaccination. With tracing, vaccination proceeds at the pace of the epidemic and requires the identification of new cases to trigger the vaccination of new contacts. The time required to complete vaccination is thus much greater than the time needed for mass vaccination. The Council of Economic Advisors has estimated that the cost of a smallpox attack could approach $177 billion dollars per week.2 Speed is very important and this again favors mass vaccination over tracing. Regarding the time required to find and vaccinate contacts, the CSA report maintains that this could happen more quickly than we assume in the model. While we could find no contact tracing data other than the sexually transmitted disease experience we cited, we did not use the actual rates reported as our assumed rate of tracing; rather we increased this from about 10 to 50 per tracer/vaccinator per day. Nonetheless, we believe that as promulgated in the current version of the Centers for Disease Control and Prevention’s (CDC) smallpox emergency response plan, tracing would remain a resource-intensive activity. Simply completing the necessary contact tracing forms would take a considerable amount of time (available at http://www.bt.cdc.gov/agent/smallpox/response-plan/files/guide-a-form-2-d.pdf; other related forms are available at http://www.bt.cdc.gov/agent/smallpox/response-plan/index.asp). The CSA suggests that a hepatitis A outbreak reported in Colorado provides a better guide for the speed with which traced vaccination could progress, and also suggests that many "worried well" requested vaccination during this outbreak due to media reporting. The implication intended, we believe, is that similar use of the mass media would turn out large numbers of citizens in the face of a smallpox outbreak, increasing the vaccination rate. However, to the extent that the majority of those responding in the hepatitis outbreak were actually not true contacts, a form of mass vaccination occurred by default. In the event of smallpox, the analogy might well be panic, which could easily lead to vaccination facilities being overwhelmed if they had not planned to deliver vaccine on a mass basis. On the other hand, depending on where the attack takes place, individuals may not hasten to seek vaccine. Consider what happened in Tel Aviv when suicide bombers struck the neighborhood surrounding an old bus station, a neighborhood frequented by foreign workers, many of whose visas had expired. Several individuals were badly injured, but they did not present for medical care because they were afraid of being deported. Only after Israel granted amnesty did these injured come forth. A comparable situation in the United States could be a smallpox attack that targets an area frequented by undocumented aliens. Such persons, even if ill, might not come forth to be vaccinated for fear of deportation. Our claim is not that such an event would occur with certainty, or even that this is the most likely scenario–but it is not implausible. The point is simply that it is not necessarily the case, as implied in the CSA report, that tracing would move considerably more quickly than we assumed. The CSA report argues that in the Kaplan model, infected persons progress to symptoms too quickly, and that somehow this serves as a disadvantage to tracing. But the whole idea of tracing is to find people who might have been infected while they can still be saved via post-exposure vaccination. To quote from the CDC's response plan: "Contact identification is the most urgent task when investigating smallpox cases since vaccination of close contacts as soon as possible following exposure but preferably within 3-4 days may prevent or modify disease. This was the successful strategy used for the global eradication of smallpox." What matters is the relative duration of the "vaccine sensitive" window period relative to the duration of infectiousness. The longer the duration of infectiousness relative to the window, the more likely it is that by the time a contact is found, he or she will have already exited the vaccine-sensitive stage. Conversely, the longer the vaccine-sensitive window period, the better the chance of saving an infected contact via post-exposure vaccination. If anything, we were very optimistic regarding traced vaccination in this case, because the evidence documenting the existence and efficacy of the post-exposure window is scant. Finally, the CSA report fails to recognize that the progression from "stage 3" to "stage 4" in our model is not identical to the epidemiological progression from the prodromal to the overtly symptomatic stage. Rather, this transition reflects response operations as well as disease progression. As stated in our paper, "When an infected person recognizes his or her symptoms, which may occur some time after the symptoms first appear, [he or she] seeks medical help and is immediately isolated." It is not enough for an epidemiological change-of-state to occur. The infected person must recognize (or be recognized by others) that this is the case before becoming isolated. Incidentally, it is important to note that in our model, all smallpox cases, once recognized, are immediately isolated, which is assumed to immediately prevent all further transmission. We recognize that there remain differing views on this subject. However, as noted above, we believe that this report inaccurately represents our research. Edward H. Kaplan, Ph.D. References
|
Content provided by: CSAPH