The New Standard for Wound Care
Clinical variability can critically undermine care quality, often contributing to improper treatment decisions, adverse events, and poor patient outcomes. With a shortage of wound specialists (1 specialist for every 500 patients), minimal formal wound care education for most providers (less than 9 hours of formal education), an abundance of wound types (over 80 wounds types), and predominantly manual, error-prone methods of assessing wounds (i.e. paper-rulers), it is no surprise that clinical variability is a massive challenge in wound care - especially at scale. Critical parameters like wound size and margins, tissue types, and 3-Dimensional topography have previously been impossible to collect objectively, or required large, expensive, single-purpose devices to capture. These and the hundreds of other parameters governing wound healing, make the scale of the problem of making data-driven decisions optimized for the patient intractable for individual providers. Further, the process of caring for wounds and making these clinical decisions has forever been a lengthy, painful, and isolating experience for both patient and clinician.
To solve this challenge we’ve developed Swift Skin and Wound, incorporating several novel AI and computer vision approaches, aimed at solving each element of this challenge under a unified umbrella of patient and clinician empathy. We drive computation to the bedside with deep learning and ML being used to inform critical care decisions in the moment, and getting out of the way so that the clinician can spend more time listening to the patient. We autonomously identify the wound boundaries, measure the wound in three dimensions, and identify tissue types within, using deep learning and vision architectures developed specifically to run in milliseconds on mobile devices, speeding clinicians through a once-arduous workflow to let them focus on what matters. Benefiting from the combined experience of tens of thousands of clinicians who have treated hundreds of thousands of wounds before this one, we then generate detailed prognostic risk scores custom to the wound and the patient, allowing clinicians an instant view of the future while still at the bedside. And then, to support organizations managing populations of hundreds or thousands of patients, in hospitals, skilled nursing facilities, or in their own home, we provide intelligent aggregation tools using AI to stratify and identify patients at highest risk, so that limited care resources can be focussed where it’s most needed. We’ve taken an industry with a single patient, single day view, and turned it into a platform, an AI powered operating system for improving quality of wound care, and healing patients faster.
Other tools on the market aim to solve single elements of the problem, or exist all within a single segment of care. Swift's breadth and partnerships provide these tools across the continuum, allowing our technology to ‘walk with the patient’ throughout their entire care journey, from hospital, to wound clinic, to skilled nursing facility, to ‘hospital-in-the-home’. Our AI models see much more dense data, connecting the dots into a true healing picture, and giving caregivers data-driven insights and superpowers they need to operate at the top of their license. Fundamentally, we differ in our focus on “understanding” healing. Not just correlating in-situ data, but using it to obtain a deeper understanding of the underlying physics of healing. For example, instead of inferring tissue composition from colour, we are directly inferring tissue types and wound severity from the images with our unique deep learning architectures that were trained using vast amounts of chronic wound imaging data . Our FDA registered HealX marker enables our system to calibrate every image, regardless of the differences visit to visit in device, environmental lighting, and user, into a common colorspace reference. This edge gives our dataset, numbering millions of wound images and tens of millions of patient and wound assessments, longitudinal predictive powers unmatched by any other wound dataset in the world. Our push to create custom, mobile capable, deep learning and vision pipelines, rather than using off-the-shelf tools, allows us to get deeper into the clinical process and actually assist with empathy at the bedside.
Swift has conclusively demonstrated the accuracy, consistency, and objectivity improvements afforded by our solution in a paper published in PLOS One https://pubmed.ncbi.nlm.nih.gov/28817649/, and in award winning publications at industry conferences like SAWC: (PI-09) AI-powered Wound Tissue Classification and Segmentation https://swiftmedical.com/poster-clinical-validation-of-digital-wound-management/, https://swiftmedical.com/poster-fluorescence-based-quantification-of-bacterial-presence/ https://www.hmpgloballearningnetwork.com/site/woundcare/poster/can-big-data-provide-predictive-analytics-wound-trajectories). Our tools have improved accuracy in the industry, eliminating 88% of measurement inaccuracy vs. paper methods known to overestimate wound size by more than 40%. Measurements taken using Swift provide greater accuracy and improved inter-rater reliability, even when measurements were taken by non-wound experts. Swift helped reduce the prevalence of pressure ulcers by 10% at a skilled nursing facility as detailed in this peer-reviewed paper published in the International Wound Journal https://pubmed.ncbi.nlm.nih.gov/30864302/. Pressure Ulcers in skilled nursing facilities can range in cost from $8,000-$20,000+ depending on the severity. In a benefits evaluation with Swift home health client, Reach Healthcare, we demonstrated a 20% reduction in wound visits and a savings of 85$ per patient per month post adoption. (https://www.reachhealthcareservices.com/uploads/userfiles/files/documents/Swift%20Medical%20Case%20Study%20-%20Reach%20Healthcare%20-%20April%202020.pdf) During COVID, Swift leaned in, creating a Telewound Coalition (https://www.prnewswire.com/news-releases/groundbreaking-coalition-of-industry-leaders-bring-telehealth-to-wound-care-at-no-cost-during-coronavirus-pandemic-301039773.html) and winning a $2+ million grant project to deploy our solutions toward reducing in-person clinical wound care visits, reducing risk to patients and caregivers, while ensuring the standards of care could be maintained. In one example of the success of that program, a homeless patient who was refusing in-patient care due to COVID risk, was willing to be treated and assessed in the field by emergency care personnel, with wound data and measurements returned to a remote specialist.
Swift is a true innovator in application of AI to bedside care. We are constantly striving to improve our product, and the use of AI allows our solution to provide clinical superpowers, to help clinicians on the frontline better monitor, measure, and manage wounds across the continuum of care.