For these arguments to be correct, two things must be true:
- Health care providers will try to use all available facilities, regardless of their utility for the patient (creating the problem in the first place)
- Patients will not shop around, or be induced to shop around, for the lowest cost additional service that the health care provider is recommending (preventing downward price competition from happening in any case)

Doctors and nurses may certainly have an emotional bias toward doing something rather than nothing, but these arguments presume that they will be so overcome by emotion (or, I suppose, greed) and out of sight of reason that they will always use up all available supply of any facility. That is hard to believe. Since most doctors and nurses want to give patients the right treatment, it seems likely that they would take studies showing overtreatment seriously.

On the second point, there are certainly situations in health care where search costs are very high, such as emergencies. But there are many situations where search costs are not very high, particularly when assisted by an sophisticated insurer.

But you don’t have to believe either of these counterarguments, because induced demand is testable. If providers really do have an insatiable urge to make use of all available facilities, then deviations such show up according to the level of search costs in a given situation. For example, comparing jurisdictions with stricter and looser certificate of need panels, we should see looser jurisdictions have longer hospital stays (because it is nearly impossible to choose where to spend an additional day in the hospital) but similar numbers of admittances (because it is much easier to choose where to be admitted in the first place). Granted, restrictive practices in hospital privileges have made choosing where to be initially admitted much harder, but still not nearly as hard as switching hospitals mid-stay.

It would be nice to see some research trying to tease out these principles by applying economic theory, not by gathering up an agglomeration of anecdotes.

As far as care being “supply sensitive,” it generally should be. Facilities like hospital beds have very high fixed costs, but very low variable costs. If a bed is just sitting there, and a night’s stay could have any benefit for the patient at all, then it is efficient to use it. (Of course, a certificate of need panel could always decide to impose a reimbursement ceiling on additional facilities, rather than barring them from being built, to ensure they reduce costs rather than induce demand…)