Aims and objectives To verify the ability of infrared thermography in objectively identifying pressure injury and its application value in the early warning of pressure injury. Background There is subjectivity in assessing the risk of pressure injury as well as diagnosis in clinical settings, which makes early detection and prevention difficult. Design Prospective, cohort study. Method Four hundred and fifteen patients admitted to the adult intensive care units were enrolled by a convenience sampling method, and they received a follow‐up monitoring for 10 days. The risk of pressure injury was assessed via Braden scale, and thermal images of sacral area were obtained by infrared thermal imager once a day. The predictive effects of infrared thermography and Braden scale on pressure injury were compared by the receiver operating characteristic curve from which the optimal cut‐off value of skin temperature for predicting pressure injury was determined. The effect of skin temperature on pressure injury was described and compared, using Kaplan–Meier curve and Cox proportional hazard regression model, respectively. We followed STROBE checklist for reporting the study. Results The relative temperature of sacral area was negatively correlated with the risk of pressure injury. The efficiency of infrared thermography for diagnosing pressure injury was better than that of Braden scale. Based on the relative temperature optimal cut‐off value (−0.1°C), Kaplan–Meier curve and Cox proportional hazard regression model analysis showed the incidence of pressure injury with relative temperature below −0.1°C was higher than the group with relative temperature above −0.1°C. Conclusions Infrared thermography can objectively and accurately identify local hypothermia warnings of pressure injury before visual recognition. The application of infrared thermography into routine pressure injury risk assessment provides a timely and reliable method for nursing practitioners. Relevance to clinical practice Infrared thermography has great value of clinical application in daily pressure injury assessment. It is of great significance to make a faster and more objective clinical judgement for patients at risk of pressure injury.

No abstract

Hospital-acquired pressure injury is difficult to identify in the early stage, accompanied with increased morbidity but considered to be preventable. For helping the nurses to monitor the status of the patients' skin, the infrared thermal imaging and the convolutional neural networks were integrated to identify and prevent pressure injury. In the first stage, infrared thermal images were shoot and labelled with the normal group and the pressure injury group by the clinical nurses. In the second stage, the convolutional neural networks and two machine learning algorithms, the random forest and the support vector machine, were applied to classify these two classes of the collected images. The classification model was trained on 164 images and was tested on the special image dataset consisted of 82 infrared thermal images of 1 day before pressure injury. Gray level co-occurrence matrix was utilized to extract the texture features of the infrared thermal images and we chose the pearson correlation coefficient and the Chi square test as the feature selection methods. The classification accuracy of the proposed convolutional neural networks model was 95.2% and the area under curve was 0.98. Moreover, the classification results from the test dataset were conformed to the experience of the experts. After feature selection, variance and entropy were proved to the best distinguishable features. Finally, we concluded that combining the infrared thermal imaging and convolutional neural networks could contribute to the prevention of pressure injury. This measure should be performed in high-risk populations to reduce the incidence of pressure injury.INDEX TERMS Convolutional neural networks, thermal image, pressure injury.

Background It is challenging to detect pressure injuries at an early stage of their development. Objectives To assess the ability of an infrared thermography (IRT)‐based model, constructed using a convolution neural network, to reliably detect pressure injuries. Methods A prospective cohort study compared validity in patients with pressure injury (n = 58) and without pressure injury (n = 205) using different methods. Each patient was followed up for 10 days. Results The optimal cut‐off values of the IRT‐based model were 0·53 for identifying tissue damage 1 day before visual detection of pressure injury and 0·88 for pressure injury detection on the day visual detection is possible. Kaplan–Meier curves and Cox proportional hazard regression model analysis showed that the risk of pressure injury increased 13‐fold 1 day before visual detection with a cut‐off value higher than 0·53 [hazard ratio (HR) 13·04, 95% confidence interval (CI) 6·32–26·91; P < 0·001]. The ability of the IRT‐based model to detect pressure injuries [area under the receiver operating characteristic curve (AUC)lag 0 days, 0·98, 95% CI 0·95–1·00] was better than that of other methods. Conclusions The IRT‐based model is a useful and reliable method for clinical dermatologists and nurses to detect pressure injuries. It can objectively and accurately detect pressure injuries 1 day before visual detection and is therefore able to guide prevention earlier than would otherwise be possible. What is already known about this topic? Detection of pressure injuries at an early stage is challenging. Infrared thermography can be used for the physiological and anatomical evaluation of subcutaneous tissue abnormalities. A convolutional neural network is increasingly used in medical imaging analysis. What does this study add? The optimal cut‐off values of the IRT‐based model were 0·53 for identifying tissue damage 1 day before visual detection of pressure injury and 0·88 for pressure injury detection on the day visual detection is possible. Infrared thermography‐based models can be used by clinical dermatologists and nurses to detect pressure injuries at an early stage objectively and accurately.

No abstract