This study aimed to determine the predictive accuracy and factors deteriorating blanchable erythema. A prospective cohort study was conducted in a 832-bed university hospital in Japan. Skin condition, risk factors, care plans and practices were measured everyday for 4 weeks by direct assessments and chart review by research staff. Blanchable erythema developed in 62 (24·9%) and a pressure ulcer in 8 (3·2%) patients. Six of the 62 blanchable erythema patients worsened to pressure ulcer of either stage I or II. Calculating the accuracy of blanchable erythema for predicting pressure ulcer development, sensitivity was 75%, specificity 77% and positive likelihood ratio 3·26. The factors found to deteriorate blanchable erythema were K-scale ‘pressure’ and inadequate support surface management to a patient whose condition was deteriorating. These results suggested that the value of positive likelihood ratio means small effect to the clinical use. However, if adequate support surface management were implemented to blanchable erythema patients, the incidence of pressure ulcer would probably fall to 0·8%.