Medical savings accounts: an idea that just won’t die

By Raisa Deber, Evelyn Forget, Leslie Roos and S.E.D Shortt

Brian Lee Crowley, persistent advocate of medical savings accounts, dismisses our studies, published in the July 23 Canadian Medical Association Journal. Recognizing how disconcerting it can be to find that what he terms “one of the most innovative and promising ideas in the health-care debate” makes little economic, clinical, or administrative sense, he resorted to asserting, in The Globe this week, that yet more study is needed.

The implications of medical savings accounts depend heavily upon the actual distribution of health expenditures. These plans are premised on the assumption that there is a great deal of inappropriate use of health care by consumers, which presumably would be reduced in the face of the monetary penalties and rewards associated with MSAs. As Mr. Crowley asserts, “The reduction in demand for medical services among the vast population of healthy people will free up access to services for the sick who truly need care.”

Unfortunately for MSA advocates, the Manitoba data we analyzed does not support this premise. Our analysis of health spending in Manitoba found that most people do not incur large costs for physician and hospital services; most of the spending results from a small number of very sick individuals. Since the “vast population of healthy people” is already spending very little, there is only minimal scope for reduction.

Instead, medical savings accounts would greatly increase costs by giving money to the healthy. We computed that, if one chose to give each Manitoban the average spent on health care per person ($730), it would result in a 54-per-cent cost increase, all going to healthy people. This occurs because 80 per cent of Manitobans spend less (usually far less) than the average, while the sickest 1 per cent accounted for 26 per cent of costs.

To this, Mr. Crowley suggests that our results would be different if we adjusted for age, sex, income, and/or health status.

According to our analysis, individuals older than 75 did, indeed, incur greater expenditures (average $3,340) than those in the 20-to-24 age bracket (average $331). However, the healthiest seniors incurred about $16 in costs, while the sickest people aged 20 to 24 incurred an average cost of $9,808. The bottom line is this: In every age group, 80 per cent of all people incur costs less than the average for that age, meaning that one is still sending money to a lot of relatively healthy people, without generating offsetting savings from the sickest.

If people are not spending much money for medical care in the first place, one cannot expect “incentives” to economize on their use of care to generate meaningful savings.

It’s also not realistic to subdivide the population into relatively homogeneous groups. Not even taking into account the administrative costs or privacy concerns associated with attempting to so personalize the amount transferred, it should be noted that this cannot be done with current information systems, and that even the most elaborate of such systems in other jurisdictions has been unable to explain more than about 15 per cent of the variance in health spending.

Experience in foreign jurisdictions does not support the assertion that MSAs control costs. Instead, they shift a portion of health-care costs to those individuals with the greatest health-care needs. Nobel economist Amartya Sen once observed that what is economically efficient may be morally repugnant. Most Canadians would find asking the very sick to bear higher costs in order to transfer money to the healthy morally repugnant. When the sizeable cost increases associated with most models are added to the mix, the plan does not even have the virtue of making economic sense.

Our results hold for all age groups, and hold whether you are considering all publicly funded health expenditures or physician services only. Given the extreme skewing of the distribution of health expenditures, our results suggest that, no matter how you formulate MSAs, these accounts are not cost saving, unless coverage is cut to the extent that they no longer constitute insurance.

As a prominent health economist is fond of observing, for every complex policy problem there is a simple solution — and it’s invariably wrong. It is well past time to recognize how well this dictum applies to MSAs.

Unfortunately, if present discussions are an indicator, MSAs appear in danger of joining user fees as the zombies of Canadian health policy: No matter how often they are laid to rest by sound scholarship, they rise again to haunt the fringe of academic and policy circles.

Raisa Deber is a professor of health policy, management and evaluation at the University of Toronto; Evelyn L. Forget is a professor of community health sciences at the University of Manitoba; Leslie Roos is a professor of community health sciences at the University of Manitoba; S. E. D. Shortt is director of the Centre for Health Services and Policy Research at Queen’s University.