Catheter Infection
Conditions
Keywords
catheter infection, deep learning, local signs
Brief summary
Deepcath is the first step to the introduction of artificial intelligence in catheter care. A better use of visualisation of catheter exit site should be used not only by the HCWs but also by the patients and their family. A deep learning system able to detect visual abnormalities of the catheter exit site will be an helpful tools to develop a continuous follow-up of intravascular catheters.
Interventions
DIAGNOSTIC_TESTPhotographs collection phase
Three medical experts have been selected to review the photo collected. Each expert medical assesses the presence of local signs of infection on the photographs by annotating them directly via a dedicated software. They will annotate local signs: redness, perfusion extravasation, necrosis, hematoma, edema, non-purulent discharge, and purulent discharge. A convolutional neural network model will determine the probability of local sign presence. Each picture will be annotated to determine the main characteristics of the catheter. A dataset preparation with photo cropping will be performed for modelling.
Sponsors
Assistance Publique - Hôpitaux de Paris
National Network Surveillance and Prevention of Infections Associated with Invasive Devices SPIADI
University Hospital, Clermont-Ferrand
University Hospital, Grenoble
UNICANCER
University Grenoble Alps
Outcome Rea
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
Patients over 18 years of age Patients with one or more implanted central venous, midline, piCCline, arterial, or peripheral catheters. Patient and/or trusted person and/or family who have verbally stated their non-objection to the study Patient affiliated or beneficiary of a social security plan
Exclusion criteria
Patients presenting a peripheral identification sign close to the catheter insertion point cannot be masked when the photograph is taken. Thus, jewelry, clothing, tattoos, scars, and birthmarks are identifying features. Patients whose catheter insertion point is not visible.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| The number of correct predictions for redness divided by the total number of predictions | through study completion, an average of 1 year | Overall classification accuracy of the learning model compared to the assessment of three independent medical experts on the detection of the presence of redness greater than or equal to 5 mm at the catheter insertion site. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| The ratio of true positives and total positives predicted: | through study completion, an average of 1 year | The precision metric focuses on Type-I errors(FP). A Type-I error occurs when we reject a true null Hypothesis. |
| The number of correct predictions for indurated venous cord divided by the total number of predictions | through study completion, an average of 1 year | Evaluate, on the basis of images of peripheral catheter insertion sites, the reliability of the learning model on the assessment of the presence of indurated venous cord. |
| The link between presence of local signs and infection | through study completion, an average of 1 year | Measure the correlation between the appearance of the catheter puncture site and the presence of signs consistent with local and systemic infection. |
Countries
France
Contacts
Primary ContactJean-François TIMSIT, Pr
Outcome results
None listed