In a modern healthcare system, it is still a serious challenge for inpatients to deal with pressure injuries. In hospitals, nurses are responsible for the assessment of pressure injuries. However, a comprehensive wound assessment is time-consumption and depends on the experi-ence of nurses. However, its assessment is not validated to initiate a targeted therapeutic strategy for each patient. The use of image-based artificial intelligence (AI) in wound assessment with machine learning techniques may be a promising strategy to provide automated clinical infor-mation and decision-making aids. To develop an image-based artificial intelligence in the assessment of pressure injuries of inpatients, which image data were collected from electronic health records. The methodology of this study can be divided into three stages. The first stage is data preprocessing to generate training data sets; the second stage is model training, putting images into the InceptionV3 model for training; the third stage is model verification. The dataset has a total of 561 images of pressure injury wounds, which images are grouped into six categories such as stage 1,2,3,4, deep tissue injury (DTI), and Unstageable. In this study, an ideal AI wound assessment model was established. It could reduce the nursing burden and increase the accuracy of wound assessment.

Download