Simple methods for the early detection of function loss and disease in
the elderly were evaluated. Such methods might be used by non-profess
ional personnel as case-finding tools for early intervention. The aim
of this study was to assess the reliability of functional and clinical
data collected by non-professional nursing assistants. The study grou
p consisted of sixty elderly short-stay patients in a stable clinical
situation, admitted for rehabilitation and aid in a nursing home. Bart
hel ADL index items and clinical symptoms and signs were scored by nur
sing assistants and nurses (the reference method). Agreement between s
cores was assessed by Cohen's kappa statistics and test of marginal ho
mogeneity. All the Barthel items and the clinical symptoms and signs h
ad fair or better agreement (weighted kappa >0.40). For dressing, chai
r-to-bed transfer, ambulation, bathing, paresis and edema, the agreeme
nt was strong (weighted kappa >0.75). The reliability of the Barthel s
umscores was also acceptable. We concluded that observation of functio
ning and clinical symptoms and signs by non-professional nursing assis
tants is reliable.