The RAND Health Insurance Experiment study demonstrated that people factored in costs when making decisions about utilizing health care services. And cost-sharing reduced hospitalizations by about 25%, with little to no negative effect on health status.
For 25 years, this conclusion has been the basis for decisions on health care policy, insurance cost-sharing and copays, and the recent push (in some quarters) for HSA-based health plans.
But some are now thinking the experiment was flawed, and that this flaw may call into question what we now “know” about cost-sharing’s influence.
In the experiment, participants were randomly assigned to health plans with different benefit designs, some quite generous and others requiring a deductible or cost-share (members had to pay a percentage of the provider’s charges).
The argument against the study’s conclusion revolves around dis-enrollment. During the five years of the study, less than 0.5% of those enrolled in the free plan dropped out. In contrast, 6.7% of those in plans with cost sharing disenrolled, a sixteenfold difference.
The hypothesis advanced in a recent article by Prof. John Nyman of the University of Minnesota about the study is that the higher rate of disenrollment could result from members facing expensive procedures deciding to opt-out of the expensive plan and return to their original plan (which would have had much lower cost sharing requirements). If valid, Nyman’s interpretation would call into question one of the fundamental conclusions of the RAND study, thereby yanking out one of the pillars upon which much health policy has been built.
Nyman’s hypothesis was recently addressed by one of the original RAND researchers, Joseph Newhouse of Harvard University. Newhouse asserted that the disenrollment did not have a material effect on the conclusions, “either because we collected the relevant data on those who withdrew and incorporated it into our findings…or because attrition was largely unrelated to a person’s propensity to use medical services…”.
Which leads us to today – Nyman has published a >critique of Newhouse’s piece, and as one might assume, he stands by his original statement. This is pretty dense stuff, so I’ll leave it to experts (Jason Shafrin, perhaps?) to weigh in on the validity of Nyman’s arguments. But a couple of Nyman’s findings are pretty compelling.
For example, the actual difference in utilization between those with and without cost-sharing is closer to 19% than 25% (after accounting for disenrollment).
And, the utilization patterns of those who dropped out (returned to their original plan) are unknown – and therefore we just don’t know (for certain) that their health status was consistent with those who stayed in the experiment. This means that a fundamental conclusion of the RAND HIE may be in doubt.
What does this mean for you?
Two things. It is not clear that making people pay for health services reduces unneccessary utilization.
And always question your assumptions.
I have been working with dermatologists for sometime now. Many dermatologists, as I’m sure we all know, offer clinical and cosmetic services in their offices. The clinical side is covered by insurance and the cosmetic side is not. With this in mind I feel I have seen a pretty good cross section of patients. The conclusions that I have drawn from my experiences lead me to believe that making people pay for health services does not reduce unnecessary utilization. It has just become common place and for these unnecessary uses it has become a regular expense for people, almost like buying groceries or filling up the car with gas. Its just factored in as regular expense. It maybe that people who utilize dermatologists tend to be more health oriented but I think more and more people these days are health oriented.
I would agree with the above comment that procedures confined to cosmetic dermatology or plastic surgery would not be affected in the same ways as primary care procedures.
However, the study produced by the RAND does not focus on elective procedures, but on the concerns of primary care. In that vein (no pun intended), I find that comparing it to the cosmetic market is inaccurate.
This is a serious concern with the rise of healthcare costs, and I think we should look at this study very seriously when decided on how to disincentivize patients from abusing medical insurance policies.