Curbside Consultations and the Viaduct Effect

GROUCHO: Here is a viaduct leading over to the mainland.

CHICO: Why a duck? Why no chicken?

GROUCHO: You try to cross over there a chicken, and you'll find out why a duck. It's deep water, that's viaduct.

From the Marx Brothers, The Cocoanuts , 1929

Among the survival skills of the busy clinician is the "curbside consultation," which can be defined as informally obtaining information or advice regarding a patient from another physician who has not directly examined the patient or reviewed the patient's record, and does not document any recommendations. Given how common this practice is, it is surprising how little medical literature there is on the subject. The 2 articles in this issue of THE JOURNAL, by Keating et al¹ and by Kuo et al,² help to fill in the picture considerably.

These descriptive studies, despite slightly different definitions of the term, confirm the prevalence of informal "curbside" consultations. These studies also add to our understanding of the purposes of informal consultations, the frequency of requests of different subspecialties, their relationship to managed care practice, and, importantly, the attitudes of generalists and subspecialists toward curbside consultations. The broad approach augments previous detailed descriptions in specific subspecialties such as infectious disease,³ endocrinology,⁴ and gastroenterology.⁵

There are many potential advantages to a curbside consultation.⁶ From the perspective of the requesting physician, the consultation may be more efficient than searching the literature, it imposes no costs on patient or physician, it can minimize the number of additional consultants in a complicated case, and it can help physicians keep current with medical information. For the consulting physician, the informal consultation can also be efficient (and in a salaried setting not result in a financial loss), it may be intellectually stimulating, it may facilitate future formal consultations, and it may disseminate knowledge of local experts.⁷ A study of formal consultations showed disagreement in the reason for a consultation between requester and consultant in 14% of the cases⁸ ; direct contact in a curbside consultation might reduce or eliminate this.

Informal consultations have potential disadvantages. There is some concern about lack of compensation^{1 ,6} and about legal liability.⁹ However, the most important issue was voiced in the opinions of the consultants, who perceived that the information communicated was incomplete or inaccurate.^{1 - 2} Is communication a critical problem in curbside consultations? There are a number of reasons to be concerned that this perception may be grounded in reality, all of which stem from the fact that all of the consultant's advice is based both on how well the requester gathered information and on the way that information is conveyed.

The first consideration is that what one party says may not actually be the same as what the other one hears. The imprecision and inaccuracy in commonly used descriptors of probability was studied by Bryant and Norman,¹⁰ who presented a list of 30 expressions to a group of physicians and asked them to attach a specific numerical probability to each. The range of responses was disconcerting. For example, "probable" ranged from 0.3 to 0.95, "consistent with" ranged from 0.2 to 1.0, and even "pathognomonic" ranged from 0.6 to 1.0. Why a duck, indeed?

Second, the potential misperception of the actual likelihood of disease that may be occurring on either end of the telephone may be compounded by an unconscious biasing of the requester's presentation. Tversky and Kahneman¹¹ have shown that how a choice is framed can influence the decision. When given an option between taking a risk and maintaining the status quo, both having the same expected value, if the scenario is worded as a gain ("You have a 70% chance of winning."), most people avoid risk. If the identical scenario is worded as a loss ("You have a 30% chance of losing."), most will take the risk. In addition, consultants may also bring their own risk preferences to this situation without either party being aware. In a series of studies, Nightingale showed that a physician's general attitude of being risk-averse or risk-seeking translated to different behaviors in laboratory use,¹² intubation,¹³ and emergency department admitting rates.¹⁴ Moreover, situations in which multiple options are presented can paradoxically influence physicians to choose an option that would have been declined if fewer options were available.¹⁵ Therefore, the way the requester states the problem could inappropriately influence the consultant's choice.

Although communication concerns were raised more by the subspecialists in the studies by Keating et al¹ and Kuo et al,² the requester also may be at a disadvantage. The paradigm of evidence-based medicine promotes a critical assessment of the quality of the information being brought to bear on a medical decision.¹⁶ However, the informal setting works against this approach. Because of the informal nature of the interaction, there may not be any discussion of the sources for the advice and there may be reluctance to challenge it. The requester also may have already made the decision to trust the authority of the consultant or simply may be looking for reassurance about what had already been concluded.

The locations and situations in which curbside consultations occur² also may not be conducive either to reflective consideration of the problem or to in-depth discussion of its subtleties. A physician who is paged away from another activity, is in a noisy, crowded hallway en route elsewhere, or is buttonholed outside the hospital (for example, on the curb) may be distracted from offering the kind of thoughtful opinion that may come from a formal consultation or a thorough discussion of assumptions behind the informal advice. Moreover, with increasing use of electronic mail for physician-to-physician communication, it seems inevitable that more informal consultations will occur this way. Although "e-mail consults" may allow a more contemplative written response, this approach can distance the consultant even farther from an accurate and informative dialogue about the patient.

With all of these concerns, should curbside consultation be avoided, or is it a safe practice? Unfortunately, there is no good prospective evidence either way. Other "standard" aspects of medical care have been looked at critically, such as in the JAMA series on the rational clinical examination and studies of screening tests such as chest x-ray films and Papanicolaou smears. With this technique being such a prevalent part of medical decision making, it deserves no less. Physicians need to know whether this practice is being used appropriately, how often recommendations would be different between informal and formal consultations, and what impact any difference might have on the quality of health care and outcomes. Although this may not be easy, there are models to work from, including an assessment of outcome in formal pulmonary consultations¹⁷ and in telephone care.¹⁸

In the absence of this knowledge, what is the best approach to using curbside consultations? First, the purposes of an informal consultation need to be distinguished. Using a consultant to answer factual questions independent of a specific patient carries less risk of inaccurate communication, although that information still should be viewed with the same caution given to textbook material or review articles. Actual cases have variable complexity, and the simpler ones may be appropriate for curbsides. If so, the requester should try to be precise in the use of language, attempting to quantify likelihoods as much as is reasonable, avoid a biased presentation, and make certain that all of the important complicating details are raised. The consultant in return should directly convey the assumptions underlying the advice, including the strength of evidence in the medical literature to support it.

Recognizing the limitations of informal consultation and the hazards of the viaduct effect, the requester should then use the principles of evidence-based medicine and consider the advice as imperfect information, like most diagnostic tests. By judging its sensitivity and specificity, the uncertainty can be explicitly incorporated into the final medical decision, increasing the likelihood that it will be optimal for the patient.