In the plainest terms, a risk-adjusted mortality rate means that a mortality rate has been statistically adjusted to account for the predicted risk of death. It’s used to assess the performance of institutions, doctors, procedures, healthcare systems, etc. while taking into account other factors that could influence risk of death, such as characteristics in the patient population that make individuals in that population more or less likely to die.
Deeper dive
Comparing mortality rates across different procedures, populations, healthcare providers and institutions is ripe fodder for investigative reporting. However, it’s important to recognize the major limitations of comparing raw percentages, especially when comparing rates across different hospitals.
Nearly all mortality rates published across different institutions require risk-adjustment—statistical modifications to raw rates that attempt to account for factors that could affect individual mortality risk. In short, “risk adjustment levels the playing field to allow comparisons of the quality and safety of hospitals and healthcare providers,” explains an article in American Nurse Today.
Risk adjustment must take into account a variety of factors which vary according to the procedure and outcomes of interest. At the least, it needs to take into account the expected mortality rate, the predicted number of deaths that would be expected in a particular patient population over a set period of time regardless of other interventions.
Both the expected mortality rate and the overall risk adjustment model will have to consider the population’s demographics and illness/health characteristics, such as age, income, illness severity, comorbidities, geographic location, payer type and similar factors. The overall risk adjustment model will further have to take into account hospital characteristics—large or small, composition of patient demographics, total procedures done annually, urban/rural/suburban, teaching hospital or not, etc.
Though risk-adjusted mortality rates are commonly used in hospital performance/quality measures, they have their own limitations, including differences in documentation across institutions (and even across departments in the same institution).