The “thermostat model” of health care utilization
Danil Makarov and colleagues have an interesting paper in the April 2012 issue of Health Affairs entitled “Appropriate And Inappropriate Imaging Rates for Prostate Cancer Go Hand In Hand By Region, As If Set By Thermostat.” Using data from the SEER-Medicare database, the researchers examined regional differences in imaging for prostate cancer patients. The SEER program of the National Cancer Institute collects information about cancer site, stage, and histology for cancer patients from sixteen geographic regions. For cancer patients who are included in the SEER database and are covered by Medicare, information is available on Medicare claims for health care services. The sample consisted of 48,148 prostate cancer patients aged 66-85 who were diagnosed with prostate cancer in 2004 or 2005.
The patients were divided into low- and high-risk groups. According to the 2002 guidelines of the National Comprehensive Cancer Network, which were in effect at the time, high-risk patients should receive imaging such as bone scans, MRIs and CT scans under certain circumstances. In low-risk patients, all imaging was considered inappropriate except CT scans for planning purposes in patients undergoing external beam radiation therapy.
The researchers found that overall rates of imaging varied among the different SEER regions. Imaging appropriateness was not uniformly worse in regions with high rates of overall imaging. Rather, regions with high overall imaging rates had higher rates of inappropriate imaging and higher rates of appropriate imaging. Men with high-risk prostate cancer were more likely to receive appropriate imaging if they lived in areas with higher rates of inappropriate imaging. The authors call this the “thermostat model” of health care utilization. Dividing the regions into quartiles according to rates of inappropriate imaging of low-risk men, in quartile 4 (highest rate of inappropriate imaging), the odds ratio for men with high-risk prostate cancer receiving appropriate imaging was 1.75, with odds ratios in the next two quartiles being 1.48 and 1.04.
The authors state that their analysis suggests that
efforts to lower inappropriate use of imaging may simultaneously lower appropriate use of imaging because the two appear to be coupled. Therefore, policy measures aimed simply at limiting inappropriate imaging in regions with high resource use could have the unintended consequence of decreasing imaging for those patients for whom such care is indicated.
Policies will need to be multifaceted to break down the thermostat-like relationship between inappropriate and appropriate health care use. Accountable care organizations will need clearly defined quality metrics for a broad range of conditions. They will also need well-designed systems to ensure that the right patients are getting the right tests and procedures — and that costs are not contained at the expense of quality. Cost-control policies must selectively educate providers to change their behavior and reduce the use of unnecessary care, while still ensuring appropriate care. Such efforts could take on many forms: profiling physicians who inappropriately overuse resources, linking payment to appropriate utilization criteria, or providing rewards and incentives to physicians and organizations that optimize resource use.
Danil V. Makarov, Rani Desai, James B. Yu, Richa Sharma, Nitya Abraham, Peter C. Albertson, Harlan M. Krumholz, David F. Penson, Cary P. Gross. Appropriate And Inappropriate Imaging Rates For Prostate Cancer Go Hand In Hand By Region, As If Set By Thermostat. Health Affairs 31:4 (2012).
Addendum 4/23/2012: see also Jeff Levin-Scherz’s post on his Managing Healthcare Costs blog.
Posted on April 21, 2012, in health care costs and tagged Cary Gross, Danil Makarov, Harlan Krumholz, health care costs, imaging, prostate cancer, thermostat model, variation. Bookmark the permalink. Leave a comment.