1. Preservative contact allergy in occupational dermatitis: a machine learning analysis.
- Author
-
Kyritsi, Aikaterini, Tagka, Anna, Stratigos, Alexandros, and Karalis, Vangelis
- Subjects
- *
MACHINE learning , *MEDICAL personnel , *SKIN inflammation , *CLEANING personnel , *ALLERGIES - Abstract
Occupational dermatoses impose a significant socioeconomic burden. Allergic contact dermatitis related to occupation is prevalent among healthcare workers, cleaning service personnel, individuals in the beauty industry and industrial workers. Among risk factors, the exposure to preservatives is frequent, since they are extensively added in products for occupational use. The goal of this study is to investigate the contact allergy patterns in order to understand the linkage among hypersensitivity to preservatives, occupational profiles, patients' clinical and demographic characteristics. Patch test results were collected from monosensitized patients to Formaldehyde 2%, KATHON 0.02%, thimerosal 0.1%, and MDBGN 0.5%; information was also collected for an extended MOAHLFA (Male-Occupational-Atopic-Hand-Leg-Face-Age) index. To assess the relationship between allergen group and occupational-related ACD, the chi-square test for independence was utilized. To uncover underlying relationships in the data, multiple correspondence analysis (MCA) and categorical principal components analysis (CATPCA), which are machine learning approaches, were applied. Significant relationships were found between allergen group and: occupation class, atopy, hand, leg, facial, trunk, neck, head dermatitis, clinical characteristics, ICDRG 48 h and ICDRG 72 h clinical evaluation. MCA and CATPCA findings revealed a link among allergen group, occupation class, patients' demographic and clinical characteristics, the MOAHLFA index, and the ICDRG scores. Significant relationships were identified between the allergen group and various manifestations of dermatitis. The utilization of machine learning techniques facilitated the discernment of meaningful patterns in the data. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF