Use of digital photography and image analysis techniques to quantify erythema in health care workers

Nosocomial infections affect up to 26% of ICU patients. Compliance with the required hand hygiene procedures remains low, rarely exceeding 50%, with skin irritation as one of the main reasons for lack of conformity. The aim was to quantify the epidermal barrier effects, particularly erythema, from repetitive exposure to hand hygiene procedures among health care workers (HCWs) using an objective, reliable clinical method that could evaluate the entire hand surface.

The hand skin of HCWs was evaluated before the initial scrub and at the end of multiple 2-3 days work cycles during spring (n=54) and winter (n=60) trials. The skin condition was measured with live visual skin evaluation (LSG), digital image analysis (DIA) of high-resolution digital images, and visual perception evaluation (VPS) of image pairs.

The HCWs had significantly higher values of erythema than the non-HCW control group with all methods. Knuckle erythema increased over the cycle in both seasons. It decreased during recovery in spring and continued to increase during recovery in winter. For the DIA area of excess redness, the quantitative measure of erythema, the decrease over the cycle in spring was significantly different than the increase over the cycle in winter. Minimal changes in area of excess redness occurred during recovery in both seasons. With the VPS, both judges found a decrease in erythema during recovery in spring and an increase during recovery in winter, indicating significant differences for spring vs. winter (P<0.05). No differences in VPS erythema were seen for either product set over the work cycle during spring. Correlations were observed for (1) results for the VPS vs. the LSG method and (2) between excess erythema (mu+sigma) from DIA and the VPS erythema scores. Relatively low correlations were found between the DIA and VPS methods, i.e., knuckle mu+sigma and VPS erythema.

Significant work cycle effects for spring vs. winter were observed with DIA, while significant effects were found during the recovery period with VPS. DIA produced an objective quantitative measure of erythema that was not limited or influenced by other aspects of skin irritation (e.g., dryness, scaling) or texture encountered in the visual methods of LSG and VPS. The DIA method minimizes the difficulty in differentiating erythema severity. Standardization of image capture and processing allows assessment of skin condition across clinical locations. The VPS is a more reliable way to compare skin condition at different times, i.e. beginning vs. the end of a treatment cycle, because images are viewed simultaneously and can be carefully examined for differences.