Aims: The aim of this study was to analyze and understand the characteristics of images of pressure injury (PI) (formerly called pressure ulcer) stored in electronic health records (EHRs).
Background: To improve the quality of PI documentation, photographing PIs and storing the images in the EHR are accepted practice in many hospitals. Clinical decisions regarding the progress and treatment of PI often rely on the information presented in PI images. As new technologies develop, image processing and computer visualization can make the automated measurement of PI size and wound tissue segmentation possible. However, most research in this area deals with strictly controlled conditions for PI images that are confined to the wound region only. Translating these technologies to PI images taken in clinical environments faces the challenges of controlling the complex photographic conditions and of the lack of standardization of wound photography. In this study, PI images stored in an EHR were reviewed to analyze the characteristics of PI photography in clinical settings. Through better understanding of those characteristics, we expect to help image processing experts shorten the gap between laboratory and clinical environments when translating these new image processing technologies. Moreover, we studied possible approaches to designing a standardized wound photography protocol for clinical environments.
Design: This is an observational descriptive study.
Methods: Copies of a set of 289 PI images were obtained from a Western Pennsylvania hospital. The original images were stored in the wound care documentation component of the EHR. The original PI images were taken by the wound care nurses in clinical settings during daily wound care service. In this study, those images were reviewed one by one by researchers to analyze their characteristics, including the relative position of the PI in the images, the shooting angle of the digital camera, inclusion or exclusion of clinical background materials and their colors, and any materials contaminating the wound surface.
Results: Of all the PI images, only 6% were confined to the wound region only. Clinical background including clothes, bedsheets, and gown existed in 94% of the PI images. In 91% of the images, the PI was presented in the central part of image. In 24%, white powder or lotion covered a part of the PI surface. The digital camera lens was not oriented parallel to the plane of the PI in 46% of the images. There were no PI images in the sample that met the strictly controlled image conditions required for commercially available image processing tools for PI segmentations.
Conclusion: The findings of this study indicate that a digital photograph of PI may increase the accuracy of the assessment and documentation. To extract the wealthy information from PI images through novel image processing technologies, developers must consider the characteristics of PI images in clinical settings. Furthermore, clinicians require a standardized PI photography protocol to ensure the accuracy and objectivity of PI recording. The standardization of PI photography may increase use of new technologies such as computer visualization and telemedicine.