A new assessment for elders admitted to acute care: reliability of the MDS-AC.

Assessment of older people rarely includes functional domains critical for ensuring optimum outcome of treatment in acute hospital care. We report the development of a new assessment instrument, and illustrate how differences between pre-hospital and hospital admission status can be systematically evaluated using the Minimum Data Set for Acute Care (MDS-AC). Content was developed by literature review and consultation with professionals working in acute areas. Dual independent assessments were conducted on hospital in-patients in 4 countries. Inter-assessor reliability coefficients were calculated for each item. Kappa was calculated for all binary and multi-level nominal variables. Quadratically weighted Kappa was estimated for all ordinal multi-level variables. Where one level of the variable contained 90% or more of the subjects, total observed agreement is reported. Separate reliability estimates were calculated for pre-hospitalization and inpatient items. Subjects had a mean age of 78. Completion of pre-hospitalization and hospital period assessment (combined) required 20 and 30 minutes. Excluding items for which 90% or more of subjects were classified into a single scoring level, average inter-assessor reliability coefficient for the pre-hospital period items was 0.57 and for in hospital 0.58. Overall exact agreement was 83% for pre-hospitalization assessment items, and 79% for the in-hospital items. The reliability achieved in the highly unstable situation of the acute admission phase is sufficient for use in clinical care and research. Differences in pre-hospital and admission status necessary for case-mix adjusted comparison of outcomes were illustrated. Development of a means for systematically comparing changes in older people during the course of illness is of increasing importance when addressing questions of the appropriate and inappropriate use of medical technology.