Is the “split benefit” a feasible way to reduce health care costs?

The excellent Kevin Lewis has pointed my attention to this paper by Robertson, Yokum, Sheth, and Joiner:.  The idea will sound like common sense to an economist, namely give people some cash if they turn down special treatments of uncertain value.  The funnier thing is, there is now some evidence it might actually work:

Traditional cost sharing for health care is stymied by limited patient wealth. The “split benefit” is a new way to reduce consumption of high-cost, low-value treatments for which the risk/benefit ratio is uncertain. When a physician prescribes a costly unproven procedure, the insurer could pay a portion of the benefit directly to the patient, creating a decision opportunity for the patient. The insurer saves the remainder, unless the patient consumes. In this paper, a vignette-based randomized controlled experiment with 1,800 respondents sought to test the potential efficacy of the split benefit. The intervention reduced the odds of consumption by about half. It did so regardless of scenario (cancer or cardiac stent), type of split (rebate, prepay, or health savings account), or amount of split (US$5,000 or US$15,000). Respondents viewed the insurer that paid a split as behaving fairly, as it preserved access and choice. Three-quarters of respondents supported such use in Medicare, which did not depend on political party affiliation. The reform is promising for further testing since it has the potential to decrease spending on low-value interventions, and thereby increase the value of the health care dollar.

My concern of course is that on a larger scale eventually this would be gamed, and faux treatment offers will be generated for the purpose of transferring wealth to patients, with doctors and hospitals, one way or the other, in on the act.