抄録
In clinical evaluations of chemotherapy, operational rating scales for clinical states or their changes have usually been employed to integrate and summarize several symptoms or signs.
This report describes an attempt to establish a rating scale for chemotherapy on simple acute cystitis. The data applied here were taken from a double blind study which was carried out lately to evaluate the efficacy of pivmecillinam including 216 cases with simple acute cystitis. For this study, subjective global judgement on the effect of chemotherapy was available together with data of four target symptoms: urinary frequency, pain during urination, WBC in sediment and bacterial count in urine.
Statistical analyses on interrelationship revealed considerable high correlation between global judgement and individual changes in symptoms, strongly suggesting that global judgement could be validly explained by these symptoms. Since changes in symptoms correlated with each other, one of them, pain during urination, could be dropped without loss of statistical efficiency. In adopting the remaining three symptoms, the squared multiple correlation coefficient R2 was 0.835, the correlaton ratio in Hayashi's quantification theory η2 was 0.880, and WBC had the largest weight in the estimation. As the result of these analyses, a simple scale of scores 0-8 was tentatively established, which agreed satisfactorily with global judgement; the agreement ratio was 87.7%.
