seniormetrix: data - quality - savings

reduce practice variation in post-acute care

Predicting Patient Scores

Williams, B.C., Li, Yi, Fries, B.E. and Warren, R. (1997) Predicting patient scores between the Functional Independence Measure and the Minimum Data Set: development and performance of a FIM-MDS “crosswalk”. Archives of Physical Medicine and Rehabilitation; 78: 48-54

Objective: to determine whether MDS items measure the same patient behavior as comparable FIM items.

Design: One hundred seventy three patients from the SNF setting were tested with both MDS and FIM. While both scales have similar items (e.g., transfer), FIM measures “typical help needed” while MDS measures “frequency of accident or help needed”. In addition, some of the scales of ADL function measure different aspects of patient behavior. Accordingly, a group of “Psuedo – FIM” items were created as a cross-walk between the two scales by an expert panel. These items were then used to test the relationship between Psuedo – FIM and FIM.

Results: For the 18 FIM items, 12 Psuedo – FIM items were identified (8 motor & 4 cognitive). Mean FIM and Psuedo – FIM items were not significantly different for 5 of 12 items. Remaining items were similar, but not statistically.

Conclusion: Not all FIM items can be predicated with seemingly comparable MDS items. & MDS items, especially at the patient level. For population comparisons, FIM motor and cognitive domain scores can be predicted using MDS.

Relevance: When predicting individual patient outcome using FIM, the applied database must also provide FIM item scores.