Evaluation of Minnesota’s Long-Stay Nursing Home Quality Indicators

The Minnesota Nursing Home Report Card provides 19 clinical quality indictor (QI) ratings. Currently, face validity and expert opinions are employed to group the 19 long-stay QIs into 10 different domains. However, we do not know whether these domains are supported by the data. Under the current scoring program, some QIs may not discriminate very well between facilities. The objective was to evaluate the dimensionality of the QIs and the current scoring approach used to assign points to the domain and total QI scores. Risk-adjusted facility-level rates for the 19 QIs over the 2012-2019 period were used. Our findings indicate it is reasonable to categorize these QIs into 4 domains. Moreover, the current scoring approach is best suited for a facility QI distribution that is approximately normal. However, 11 QIs display a skewed distribution with facilities tightly grouped at the very bottom (floor) or top (ceiling) of the QI distribution. Our findings suggest that the current scoring approach may distort or exaggerate the differences in the QI rates with skewed distributions, assigning widely varying points to facilities that vary little in their QI rates. We recommend a zero-error approach for highly skewed QIs where the QI outcome is achievable and it reflects a serious quality problem. Our study of the QI scoring system is part of a package of recommendations to improve the Minnesota Nursing Home Report Card and value-based reimbursement system. Lessons learned from the study are readily applicable to Medicare’s Nursing Home Compare report.