On pages 190 thru 192 of the text, the process of “Risk Adjustment” is defined and described. Referenced is the use of the CMS “30 Day Readmission” metric, which is commonly used throughout acute care hospitals. Using the research tools that you have acquired, please find two examples where CMS or other Payers use this metric. Please identify the clinical entity that is being evaluated by either ICD-10 or DRG code and write a brief summary (2-3 sentences maximum) that addresses the rationale of the “30 Day Readmission” metric.
Risk Adjustment
The increased focus on value in healthcare and quality measurement provides excellent internal and external benchmarking opportunities. When comparing performance with other facilities, the patient mix treated by each facility may result in a comparison that is biased based on the mix of patients treated by each facility. Many quality indicators are defined with risk adjustment as a part of the measurement process. Although many of the statistical techniques used to perform the adjustment are beyond the scope of this text, an analyst should have a basic understand of the application of risk adjusted measures.
The most common type of risk adjustment used in healthcare quality measures uses the demographics of the patient and sometimes provider-level variables to determine the expected rate of the outcome. The CMS 30-day readmission measures serve as an excellent example of this type of risk adjustment technique.
The Hospital Readmissions Reduction Program was implemented during the federal fiscal year 2012 inpatient prospective payment rule. The IPPS payment received by a facility for an inpatient admission is adjusted based on an excess readmission ratio, which is equal to the risk-adjusted predicted readmissions divided by the risk-adjusted expected readmissions. This ratio is then calculated for each of the measured admissions types: acute myocardial infarction, heart failure, pneumonia, chronic obstructive pulmonary disease, and total hip or knee replacements. The readmissions adjustment factor is then based on the proportion of the total MS-DRG payments for all discharges that is represented by the excess readmissions. The readmission payment adjustment was limited to 1 percent in 2013 and increased to 3 percent in 2015.
The expected number of readmissions for each facility is risk adjusted to ensure that facilities with a higher level of patient acuity are not penalized by the measure. The risk adjustments for the CMS readmission reduction program are based on hospital characteristics, the comorbidities of the patients served at the hospital, the age of the patients, and the patient gender. Taken together, these factors are used to generate the expected number of readmissions per 1,000 admissions at a hospital. This is then compared to the actual number of readmissions.
Another example of a risk adjusted quality measure is the rate of central line-associated bloodstream infections (CLABSI). The CLABSI rate is part of the CMS hospital value based purchasing system for inpatient payment and is one of the indicators available for comparison on Hospital Compare. Figure 9.6 displays a comparison of the CLABSI performance for three hospitals in the Cleveland, OH, area.
Figure 9.6 Central line-associated bloodstream infections for Cleveland, OH area hospitals
Source: https://www.medicare.gov/care-compare/compare?providerType=Hospital&providerIds=360059,360180,360137&city=Cleveland&state=OH
Notice that the actual rates of CLABSIs are not displayed, but instead the ratio of the observed vs the risk-adjusted expected number of infections. This ratio is adjusted so that the US national benchmark is 1.0. University Hospitals of Cleveland and the Metrohealth System both have rates significantly lower than the US national benchmark. Cleveland Clinic’s rate is not statistically different than the benchmark rate. The ranges depicted in the graph represent the 95 percent confidence interval for the ratio of the observed to expected infection rate. The narrower confidence intervals are primarily due to larger sample sizes as found in the Ohio and Cleveland Clinic rates.
The risk adjustment for the CLABSI rate is based on a standardized infection rate (SIR) that is maintained by the Centers for Disease Control (CDC) via the National Healthcare Safety Network (NHSN) (CDC 2019). The SIR is the ratio of the observed infections divided by the expected infections. The expected number of infections is based on a logistic regression model. The formulation of a negative binomial regression model is not presented here, but examining the variables considered for risk adjustment and their relative importance can be understood by viewing the risk adjustment model. Some of the variables included in the risk model are:
Patient location (critical care, oncology, burn, etc.)
Facility bed size
Medical school affiliation
Facility type
Although the techniques used to actually fit the negative regression model and calculate the expected number of infections was not presented here, a basic understanding of the factors included in the model and their relative importance is still quite possible. Analysts that understand the basic principles of risk adjustment and the interpretation of standardized ratios have the ability to apply risk adjusted measures for external benchmarking.
Analysis in Practice
The three hospitals with the highest volume of cases in the practice dataset are Riverside Methodist Hospital, Miami Valley Hospital, and the Cleveland Clinic. The following figures include the comparison of these three hospitals using Hospital Compare data.
Review Questions
