The 1990 Florida Dental Investigation: Theory and Fact

In this issue [1], Barr shows that, in some quarters, the question of who transmitted human immunodeficiency virus (HIV) to six dental patients in Florida in the late 1980s is still open. In epidemiologic quarters, however, the answer to this question is known. The patients were infected by David J. Acer, a dentist who was shared by the six patients and who died of complications of the acquired immunodeficiency syndrome (AIDS) in 1990. Investigators at the Centers for Disease Control and Prevention (CDC) figured this out. Why and how Acer infected his patients is unknown and almost certainly will remain so. This hole in the story makes the Acer case, already a unique tragedy, an abiding mystery as well. However, the unresolved issues of motive and means do not in any way diminish the mass of evidence that points to the dentist as the common source of transmission.

The continuing controversy stems from a segment called “Kimberly's Story” that aired on 19 June 1994 on the CBS television program 60 Minutes (the reference is to Kimberly Bergalis, the best known of Acer's patients). The broadcast described, for each of the six infected patients, actual or theoretical exposures, other than contact with Acer, that might have resulted in HIV transmission. It also indirectly questioned the good sense, thoroughness, and competence of the CDC investigators. Late in 1994, some of those investigators responded in this journal [2]. The paper in this issue by Barr [1], one of the journalists who produced the CBS piece, is a response to that response.

Strictly speaking, the journalists didn't say then (and Barr doesn't say now) that their purpose was to refute the dental transmission theory. However, in criticizing the CDC investigation, they implicitly build the argument that the investigation was fatally flawed and that the conclusions of the investigation are probably wrong. What they do not do is present information that affirmatively supports an alternative explanation. They find real, although largely unimportant, flaws in the CDC investigation, and they posit the possible existence of undiscovered information. But they offer no credible contrary explanation of the facts that are known.

This act of critiquing is defensible and, on many levels, desirable. Irksome as it may seem, reporting by nonspecialists keeps specialists on their toes, even when the subject is as abstruse as molecular epidemiology. It remains true, however, that finding fault (or the possibility of fault) in itself does not get one very far. It does not, for instance, necessarily get one closer to the truth.

As it happens, epidemiology is a particularly fertile enterprise with which to find fault. An investigation like that of the Acer case requires that persons reconstruct historical events, albeit recent ones (in that sense, it is much like investigative reporting). It requires judgment and interpolation, and it always leaves unanswered questions. The provisional nature of epidemiology is the reason why, at least among epidemiologists, a case does not need proof “beyond a reasonable doubt” in order to be convincing. For many lay persons, however, this is hard to accept. Journalists in particular are uncomfortable with holes left in narratives. The problem is that this generally healthy desire for “all the facts and just the facts” can lead in an odd and illogical direction.

Early in “Kimberly's Story,” Mike Wallace, the correspondence for the piece, interviews Harold W. Jaffe, a CDC official. Wallace asks, “So you're saying with absolute certainty that the DNA sequencing proves that Dr. Acer infected those six patients?” [3].

Although this passes as a “tough-minded” question in journalism, it is an unfortunate one, at least when posed about a field like epidemiology. Nothing is “absolutely certain” in epidemiology, or in all of medicine, for that matter. If the 60 Minutes journalists didn't know that at the outset, they should have learned it during their investigation, for it is a central truth about the subject. “Absolute certainty” is absolutely irrelevant to reaching an operative version of reconstructed events.

Jaffe's unwillingness (and inability) to answer Wallace's question by saying, “Yes, I am certain,” symbolically sets the stage for what follows: a recitation of oversights, real and imagined, in the CDC investigation. Unanswered questions and unaddressed issues are offered, implicitly, as evidence that Acer may not have infected his patients. But, in my view, the arguments are weak and mostly rhetorical.

Patient A lied about her sexual experience and had a disease that could have been sexually transmitted. This may add a risk factor to her side of the equation, but were any sexual contacts actually found who could have given this patient HIV? No. Patient B, a woman, had an affair in the late 1970s, and that contact was never tested by the CDC. Ignoring the fact that this liaison occurred before the AIDS epidemic began (and thus doesn't truly constitute a risk factor for the patient), did the former lover have AIDS when Barr interviewed him 15 years later? (Fifteen years is much longer than the median time between infection and onset of symptoms, and AIDS almost certainly would have been evident even in the absence of blood testing.) Barr doesn't say. Patient C, a man, had five sexual partners who could not be found and tested (nine others tested negative). That may put Patient C in a high-risk category, but high-risk status should not be confused with actual evidence suggesting that the patient was infected elsewhere or that his connection to Acer is a huge coincidence.

Barr [1] says that the CDC officials “… seem to suggest that the foundations of their epidemiologic field work are firm as long as no one can document some other specific, unassailable source of infection. This is an impossible standard. …”

This is a misstatement of the CDC position. The “firmness” of the field work is based on much more than the standard that Barr posits and rightly dismisses as impossible. The CDC found and tested many sexual contacts, found none to be seropositive for HIV, and discovered high probability matches between the viral strain of the dentist and the strains of the patients [4]. The firmness of the field work is based on this. The notion that the epidemiologists are high-handedly dismissing the critique by 60 Minutes because the journalists lack “unassailable” alternative sources of infection simply does not make sense. Unassailable data are not the standard. The standard is this: A reasonable explanation (such as the CDC's) is more likely to be true than no explanation (as implied by 60 Minutes), keeping in mind, of course, that a reasonable explanation can sometimes be wrong.

Nevertheless, uncertainties remain in the Acer case, even if they aren't big enough to overturn the CDC's conclusions.

The statistical match between the viral strain of the dentist and the strains of his patients is both the clearest and the muddiest of the issues raised by Barr. It is unfortunate that neither Ciesielski and colleagues [2] nor Barr [1] offers a chart quantifying the degree of similarity between the virus samples so that readers can decide for themselves. It may be naive, however, to believe that such a chart could be made, because various methods of statistical analysis, as well as different control groups, would produce different numbers.

Two things, however, seem clear. First, no consensus exists about which analytical method is best. Second, several different ways of looking at the data strongly suggest an infectious link between the dentist and the six patients [4, 5]. One analysis found insufficient evidence to either accept or reject the dentist as the source of the infections [6]. This does not add up to incontrovertible evidence for the CDC's hypothesis, but it does add up to very strong evidence. When the subject involves two fields in which there is either incomplete knowledge (the genomic variability of HIV) or lack of consensus (the use of statistics in molecular epidemiology), this is about as good as it gets.

Ultimately, the point (although probably an unintended one) of “Kimberly's Story” is that every historical narrative is imperfect. Barr and CBS have taught this useful, if unstartling, lesson. Learning the lesson, however, does not require one to reject a painstaking, logical, and compelling—albeit imperfect—reconstruction of events, which is what the CDC epidemiologists created in their investigation of the case of Dr. David Acer.

• Copyright ©2004 by the American College of Physicians