Studying more than 800,000 families from across the United States, RAND researchers found that when people shifted into health insurance plans with high deductibles, their health spending dropped an average of 14 percent when compared to families in health plans with lower deductibles:

“We discovered that costs go down dramatically during the first year people are enrolled in high-deductible health plans, as long as the deductible is at least $1,000 per person,” said Amelia M. Haviland, a study co-author.

But as families reduced their medical spending, they eliminated some care that is clearly beneficial. Among families in traditional health plans, childhood vaccination increased increase slightly by 2.7 percent. Among families in high-deductible health plans, childhood immunizations fell by 3.6 percent. Mammography, cervical cancer and colorectal cancer screening rates were also slightly lower — just over two percentage points— among those with high-deductible health plans compared to families in traditional health plans.

RAND concedes that it doesn’t know why these results occurred:

The drop in preventive care happened even though the high-deductible plans in the study waived the need to pay a deductible when receiving such care. This suggests that enrollees in high-deductible plans either did not understand this part of their policy or some other factor discouraged them from getting preventive care.

Comments (6)

Trackback URL | Comments RSS Feed

  1. Jeff says:

    Good study. Lousy interpretation of results.

  2. Jeff says:

    What’s wrong with choosing between valuable health care and other uses of money? Isn’t that what we want people to do?

  3. Devon Herrick says:

    I suspect the moderately lower preventive care received by people enrolled in high-deductible plans has to do with consumer preference. People who like to run to the doctor numerous times each year naturally prefer comprehensive health plans with low deductibles. People who don’t want to run to the doctor numerous times a year probably don’t want to pay for a comprehensive plan. Just because the preventive care is free does not mean that all people will have equal demand for it.

  4. Ken says:

    I second Jeff’s second comment. For that matter, I also second his first comment.

  5. Linda Gorman says:

    They looked at two years of claims. Their measure of preventive care included some child immunization construct that wasn’t specified and that was supposed to indicate if a kid was “on track” to get all of his immunizations. They had specific criteria for sample membership. Whether that affected results was not discussed.

    A two year window to look at colon cancer, pap smears, and mammograms when samples are quite different in age (CDHC set was younger, more likely to be single male) might miss people stretching mammograms out for just slightly longer than a year, say 18 months, something that saves money but hasn’t been shown to be harmful. Plus, not everyone with high deductibles files claims for routine care even if an insurer pays. There is no way to know if that differs between conventional and CDHC insurance.

    There was a lot of adjustment done to make the very different CDHC/traditional groups statistically similar. There is a paper out showing that the propensity score adjustment used can, when models aren’t properly specified, bias the logistic regression coefficients used to determine the costs and amount of preventive care.

    Finally, diabetics in CDHC got more recommended care when CDHC types were skimping on the other measures.

  6. Ken Houghton says:

    Re: jeff’s second comment

    Nothing, however:

    “The drop in preventive care happened even though the high-deductible plans in the study waived the need to pay a deductible when receiving such care.”

    So they could have accepted “valuable health care” and made other uses of the money. That requires a Utility Curve where I would rather spend money on other things than choose “valuable health care.”

    To call that rational, you have to then conclude that I would rather spend dollars on later-stage illness than on “other uses of money.”

    I can construct the Agent who would do that, but considering such to be Representative would doom cost controls to failure.