BACKGROUND
Chronic ulcers, and especially ulcers affecting the lower extremities and their protracted evolution, are a health problem with significant socio-economic repercussions. The patient's quality of life often deteriorates, leading to serious personal problems for the patient and, in turn, major care challenges for healthcare professionals. Our study proposes a new approach for assisting wound assessment and criticality with an integrated framework based on a Mobile App and a Cloud platform, supporting the practitioner and optimising organisational processes. This framework, called Clinicgram, uses a decision-making support method, such as morphological analysis of wounds and artificial intelligence algorithms for feature classification and a system for matching similar cases via an easily accessible and user-friendly mobile app, and assesses the clinician to choose the best treatment.
OBJECTIVE
The main objective of this work is to evaluate the impact of the incorporation of Clinicgram, a mobile App and a Cloud platform with Artificial Intelligence algorithms to help the clinician as a decision support system to assess and evaluate correct treatments. Second objective evaluates how the professional can benefit from this technology into the real clinical practice, how it impacts patient care and how the organisation’s resources can be optimised.
METHODS
Clinicgram application and framework is a non-radiological clinical imaging management tool that is incorporated into clinical practice. The tool will also enable the execution of the different algorithms intended for assessment in this study. With the use of computer vision and supervised learning techniques, different algorithms are implemented to simplify a practitioner's task of assessment and anomaly spotting in clinical cases. Determining the area of interest of the case automatically and using it to assess different wound characteristics such as area calculation and tissue classification, and detecting different signs of infection. An observational and an objective study have been carried out that will allow obtaining clear indicators of the level of usability in clinical practice.
RESULTS
A total of 2,750 wound pictures were taken by 10 nurses for analysis during the study from January 2018 to November 2021. Objective results have been obtained from the use and management of the application, important feedback from professionals with a score of 5.55 out of 7 according to the mHealth App Usability Questionnaire. It has also been possible to collect the most present type of wound according to Resvech 2.0 of between 6 and 16 points of severity, and highlight the collection of images of between 0 and 16 cm2 of area 88%, with involvement of subcutaneous tissue 53.21%, with the presence of granulated tissue 59.16% and necrotic 30.29% and with a wet wound bed 61.54%. The usage of app to upload samples increase from 31 to 110 samples per month from 2018 to 2021.
CONCLUSIONS
Our real-world assessment demonstrates the effectiveness and reliability of the wound assessment system, increasing professional efficiency, reducing data collection time during the visit and optimising costs-effectivity in the healthcare organisation by reducing treatment variability. Also, the comfort of the professional and patient. Incorporating a tool such as Clinicgram into the chronic wound assessment and monitoring process adds value, reduction of errors and improves both the clinical practice process time, while also improving decision-making by the professional and consequently having a positive impact on the patient's wound healing process.
Development of a Method for Clinical Evaluation of Artificial Intelligence–Based Digital Wound Assessment Tools
