Skip To Content

Voices On Value RSS Feed

What’s all this about Risk Adjustment and Burden of Illness?

(The first installment in a series dedicated to helping Coloradans better understand health care claims data on!)

If you’ve spent time looking at health care cost, utilization and quality data (and who hasn’t!), you have no doubt come across the terms risk adjustment and burden of illness. Unless you are a down in the weeds health policy data wonk (that term is used here in the most affectionate way!), you may be utterly baffled as to what all of this means. In the following paragraphs I’ll explain these terms in (more or less) plain English and shed some light on the purpose and implications of risk adjustment and burden of illness measures.

Attempts to explain risk adjustment and burden of illness measures generally start with the need to be able to make apples to apples comparisons. In the health policy world, one key goal is to make meaningful comparisons between the cost and amount of health care used by different groups that are in some sense fair. For example, imagine two hypothetical physicians who both treat diabetic patients. The first physician appears to provide more health care (or treatment) per diabetic patient and at higher cost than the second. Taken at face value, one might be tempted to conclude that the first physician is not efficient in treating diabetic patients because they receive more health care services and cost more.

But what if the diabetic patients treated by the first physician are less healthy, on average, than diabetic patients treated by the second physician? In that case, the amount of health care and higher cost associated with the first physician’s diabetic patients may be appropriate and entirely justified. Risk adjustment and burden of illness measures provide one objective way to help determine if this is the case.

In claims data, risk adjustment is a process that looks at all medical conditions or diagnoses (e.g., diabetes, asthma, high blood pressure, etc.) and procedures (e.g., office visits, tests, surgeries, etc.) at the level of de-identified individual patients. Although the computer software used to perform risk adjustment can see all of the medical claims information for a particular patient, de-identified means that we have no idea who that patient actually is. Based on all diagnoses and procedures for a particular year, each de-identified patient is assigned to a unique risk category based on their most significant chronic disease or medical condition. The risk adjustment tool then further categorizes patients based on their overall severity of illness. Severity of illness is determined by the seriousness of the primary medical condition, the presence of additional and less severe (or minor) diagnoses, numbers of hospitalizations and ER visits, etc. and also takes into account various patient characteristics including age and gender.

In practical terms, risk adjustment can help to level the playing field and enhance physician willingness to participate in alternatives to traditional fee for service payment. Under a capitated payment system, physicians are paid a fixed dollar amount to treat a defined group of patients for an entire year. Without risk adjustment, both physicians in our example might be paid the same amount per patient even though the first physician’s diabetic patients require more health care. Risk adjustment specifically takes into consideration health status differences between the two groups of patients and would result in physician one receiving additional dollars to provide care to his/her relatively less healthy patients.

Burden of illness is a measure of the relative health of a defined group of patients based on risk adjustment and also reflects the cost of health care services. Burden of illness measures can be calculated for an entire population (e.g., all patients in the Colorado All Payer Claims Database) or for specific sub-groups (e.g., at the county level or for patients having knee replacement surgery at a specific hospital). At the state level, average burden of illness for all patients in the Colorado APCD is set (or normalized) to a value of one. This normalization allows comparison of the burden of illness score for patients in each county to the state as a whole. Residents of counties with a burden of illness score of less than one are “healthier” and/or use less health care services than the statewide average, whereas residents of counties with scores greater than one are less healthy based on the amount of health care they receive and the cost of that care. Similarly, individuals with higher burden of illness scores generally have more diagnoses, undergo more medical procedures and have higher costs than patients with lower scores. In our example, if the first physician’s diabetic patients have higher burden of illness scores, they would be considered less healthy based on this measure and this would provide one potential explanation for their higher cost and greater use of health care.

Burden of illness scores appear in the map and report views of information available on the Colorado APCD website On the maps tab, under the population sub-category, an illness burden score map allows comparisons of relative population health at the county and three digit Zip Code levels. On the reports tab, all tables include the illness burden score to help users better understand one potential reason for observed variation in health care utilization and spending. In general, counties with higher illness burden scores have relatively less healthy populations and would be expected to use more health care and have higher total costs than counties with lower scores.

The next installment in this series will expand on information presented above and explore in greater detail how risk adjustment and burden of illness scores are reflected in reports currently available on In future blog posts, I will (attempt to) explain how this information:

  • Can be used to better understand variation in health care utilization and spending,
  • Is used to generate expected values and how to interpret compared to expected (C2E) utilization and spending reports.

Stay tuned!

If you have questions about risk adjustment and burden of illness measures for the APCD, please contact Jonathan Mathieu, Director of Data and Research, at

Share on: DiggFacebookRedditDeliciousTwitter


Blog post currently doesn't have any comments.

Leave comment


 Security code