Research on sensitive skin and sensory irritation at the Skincare Products Research and Safety Science Research Laboratories of Kao Corporation (Michitaka Sawada, President) has revealed relationships between various types of sensitive skin and the irritant effects of cosmetic ingredients. Investigators at these laboratories can now predict which ingredients have the potential to trigger sensory irritation in sensitive skin of different types. By adopting this predictive approach, they expect to develop a system to help customers with individual skin sensitivities find the gentlest cosmetics for their skin.
This research was presented at the 30th IFSCC Congress held in Munich from September 18 to 21, 2018.
A growing number of people in the world are coming to feel they have sensitive skin. Many have experienced stinging, itching or burning when using cosmetics in the past. When choosing a cosmetic, a customer with sensitive skin may take pause to consider whether she wants to use the product on her skin. Researchers at Kao have been developing a system that helps customers with sensitive skin select products that pose no risk of sensory irritation.
A wide range of data were collected from Japanese (n=120) and German (n=80) women who felt they had sensitive skin. A cluster analysis of the data showed that the sensitive skin of the subjects could be divided into eight distinct categories, two of which were shared by subjects from both Japan and Germany.
Next, the subjects underwent stinging tests using common cosmetics ingredients such as pH regulators, preservatives, oils, and moisturizers. These tests revealed important differences between the ingredients likely to cause sensory irritation in the eight skin types identified in the cluster analysis. Further analysis showed that the skin properties and physico-chemical properties of individual ingredients were important in determining the risk that sensory irritation would occur.
Identifying which of the more than 10,000 ingredients used in cosmetics might trigger sensory irritation is no easy matter. With the help of artificial intelligence, researchers at Kao created four separate models that use machine learning algorithms to predict sensory irritation. Out of the four, the Random Forest model was found to be the most efficient, with an accuracy rate of 73%.
This research shows that it may be possible to develop a system that helps people with sensitive skin select cosmetics that are less likely to irritate their skin. The prediction models developed at Kao can be applied as soon as convenient methods are available to obtain data on the physico-chemical properties of target cosmetic ingredients and skin measurement data of the customer in question. If the system is introduced, skin measurements performed during in-store counseling will be used to predict ingredients likely to trigger unpleasant skin sensations. New product formulations without those ingredients may also follow.
Kao will continue researching and developing skincare products responsive to the needs of people with sensitive skin.