• Rising disease prevalence not the main driver of health care spending growth

    The evidence keeps mounting that high and rising US health spending is due mostly to how intensively care is provided and the prices we pay. Disease prevalence is not the main driving force.

    In recent paper in Health Affairs, Charles Roehrig and David Rousseau find that between 1996 and 2006,

    increases in treated prevalence account for about one-fourth of overall growth in real per capita spending, with the remainder attributable to growth in cost per case. If cost per case had remained constant, the rate of increase in health spending would have been less than GDP growth by more than one percentage point. […]

    Our evidence also suggests that most, and perhaps all, of the treated-prevalence effect is due to an increase in the share of eligible people being treated rather than an increase in clinical prevalence.

    We conclude that health spending growth in excess of GDP growth is not primarily due to increases in clinical disease prevalence. Consequently, efforts to reduce future growth in disease prevalence—such as through prevention— although certainly providing many societal benefits if successful, are unlikely to reduce overall health care cost growth to levels lower than or equal to GDP growth. [Emphasis  mine.]

    Over the period of study, the growth rate in real per capita health spending was 3.8 percent and GDP per capita grew 2.1 percent. If the cost of care had remained constant, the rate of growth due to disease prevalence alone would have been 1.1 percent. Contrast this with the conventional wisdom that holding health spending below GDP + 1 percentage point growth would be a monumental victory in cost control.

    So, not only does the cost of care drive the rate of increase in health care spending, it is responsible for all of the increase in health spending above GDP growth. In fact, it is responsible for more than that. Here it is in a chart:

    Some additional details:

    • The authors’ analysis is based on National Health Expenditure, Medical Expenditure Panel Survey, and National Health and Nutrition Examination Survey data.
    • I’ve blurred the distinction between treated prevalence and clinical prevalence (which includes those with a disease who are not treated) because the authors’ analysis shows the distinction does not matter a great deal, having no impact on the qualitative conclusions. At most, clinical prevalence contributed 0.3 percentage points to the 3.8 percent increase in real per capita spending.
    • There is considerable variation by disease in the degree of contribution of prevalence to increased spending. For instance, all of the increase in spending of esophageal disorders was due to prevalence and almost none of the increase in nervous system disorders was due to prevalence.

    Bottom line: prevention can’t play a major role in curbing the rate of growth in health care spending. The rationale for prevention is for health and other economic and social benefits (greater ability to work, enjoy leisure, etc.).

    • Austin
      Based on this study, and of course its not the last word, what would be a realistic GDP +/- “x” for health care growth target for next decade? The GDP+1 has taken on a magical significance. Its arbitrary and needs better tuning.

      Additionally, do we assume preventative strategies can impact the 25% contribution (probably)?

      This is important. The benchmark’s impact on the health policy discussion debate has implications. This study needs attention, and random anchors erodes expediency in giving right signals to providers.


    • The evidence keeps mounting that high and rising US health spending is due mostly to how intensively care is provided and the prices we pay. Disease prevalence is not the main driving force.

      So we need a Wal-Mart of healthcare. It seems that making it easier to become a provider (MD, PA, RN etc.) would help with the pricing, of course the argument has been made that if you allow more doctors the intensively care of care would rise. But IMO it is worth a try.

      BTW in a way the NHS in the UK seems Wal-Mart like.

    • I find this summary of the study unsettling. To suggest that disease prevention measures will not significantly effect the cost of health care and relegate prevention to a trivial status is dangerous. It is akin to suggesting that the reason we have high gas prices is because each car is such a gas guzzler and that reducing the number of cars on the road is inconsequential. Hidden within the verbiage is a resignation that diseases will always be. When, in fact, many of the diseases we treat are completely preventable. Studies of this nature are done in a bit of an information vacuum anyway. Real costs associated with diseases prevalence are mostly unmeasurable, for many reasons. One big reason is that we don’t have a good platform to capture behavioral data to tease out true correlations. Epidemiologists may prove these conclusions wrong over time. We just need phenotype inputs. A suggestion that prevention, or dare I say reversal, of disease can be a casual by product of our push for health care reform is less than desirable.

      • You have to understand the context. In terms of outcomes, I think prevention is incredibly important. I think prevention leads to better health. I think prevention is a worthy goal.

        But prevention may not reduce costs. And that’s how it’s often touted. Costs and outcomes are not the same thing. Sometimes good things (better outcomes) cost more. I’m OK with that. But they don’t always cost less.

        • If there were no diseases there would be no cost for treatment. We cannot neglect this under the guise that it won’t do anything to reduce costs when in fact, if we prevent a disease, no cost will be incurred. The contrary suggestion actually seems illogical.

          It’s quite possible we are in an era where we should leverage the rising costs of per case treatment to propel us onto a trajectory of significant investment in the knowledge of how to truly prevent disease. I think lip service is all that has ever been given to prevention to date. Let’s not give fuel for the fire of more neglect.

          • If you prevent disease x and, by doing so, increase length of life, it increases the costs of the later-developing disease y.

            If everyone died at birth, health care costs would be quite low. Though true, it’s a poor rationale for policy. Disease should be prevented not because it necessarily dramatically decreases health care spending. They should be prevented so that life is more enjoyable and productive. The reason to stress this argument is if you do not want those who oppose funding of prevention to point to the lack of reduction in health care spending as a justification for cutting support for prevention. Beware of overselling it. (Note the president proposed to cut it this week.)