A new simplified minimum data set better predicts outcomes for individuals admitted to intermediate care functional recovery units in catalonia

• Intermediate care is a growing transitional level of care including the provision of functional recovery for older adults or adults with complex care needs, and is well consolidated in Catalonia, Spain. • A new simplified Minimum Data Set is the base of “Intermediate care case-mix indexes (ICMI)”, built on the basis of the main diagnosis and functional status, which predict length of stay and discharge destination in Catalan functional recovery units. • ICMI showed better predictive accuracy compared to the currently used Resources Utilization Groups (RUG)-III system, based on an extensive data collection which diverge from clinical assessment tools and represents a considerable burden for care professionals. • The implementation of these new indexes could optimize resource allocation and decision-making in intermediate care. Background Intermediate care functional recovery (ICFR) units in Catalonia, Spain, provide post acute care and geriatric rehabilitation. Providers report a Minimum Data Set (MDS), initially developed to classify patients into resource utilization groups (RUG), to adjust public reimbursement. The MDS, completed by health professionals, is time consuming and diverges from routine clinical assessment. Objectives We developed new intermediate care case-mix indexes (ICMI), based on admission diagnosis and function, and tested whether ICMIs better predict length of stay (LOS) and discharge destination, compared to RUG. Methods We developed the ICMIs using 122,754 ICFRs admissions (years 2017–18–19–22): we categorized individuals by diagnosis and four levels of activities of daily living (ADL). To obtain ICMIs, each category was assigned a weight based on associations with ICFR LOS and “unsuccessful discharge” (death or other service versus home). Then, in the 2023 cohort, we compared the performance of four models (base model (age, sex, income), base+RUG, base+ADL, base+ICMIs) to predict LOS >57 days and unsuccessful discharge. Results The 2023 cohort included 31,640 patients (median age [IQR]=82.0 [75.0;88.0], 57% women). When predicting dichotomous LOS, the area under the ROC curve (AUC) increased from 0.56 [95% CI=0.544–0.57] to 0.56 [95% CI=0.55–0.56] (base+RUG), 0.59 [95% CI=0.58–0.61] (base+ADL), and 0.63 [95% CI=0.62–0.65] (base+ICMI). AUC to predict discharge destination showed similar improvements (0.55 [95% CI=0.54 0.55], 0.59 [95% CI=0.58–0.59], 0.64 [95% CI=0.63–0.64], 0.66 [95% CI=0.65–0.66], respectively). Conclusions In Catalan ICFRs, new, simplified, case-mix indicators improved prediction of LOS and discharge destination, compared to existing tools. This study can inform new policies for intermediate care, adding a validated instrument to improve evaluation and reimbursing systems.