Study Highlights Predictors of Atopic Dermatitis Flares Using Machine Learning

The duration, frequency, and severity of atopic dermatitis flares are predictive of the future severity of atopic dermatitis, recent machine learning findings suggest, and these elements are also linked to diminished quality of life.¹

These data resulted from a study conducted to validate the predictability of flares as well as atopic dermatitis severity, in addition to quantifying the value of predictive variables. This analysis was led by such investigators as Mia-Louise Nielsen, MSc, PhD, from the department of dermatology at Copenhagen University Hospital-Bispebjerg in Denmark.

While systemic drugs have been shown to be efficacious in their treatment of atopic dermatitis, Nielsen et al highlighted that many of those impacted remain under-treated with inadequate medication results and poor satisfaction rates.²

“Especially patients with mild to moderate disease but persistent flaring and an unstable disease course may be at risk,” Nielsen and coauthors wrote.^1,3 “This study used statistical methods and predictive machine learning models to validate the predictability and identify predictors of flares and disease severity.”

Trial Design and Findings

The investigative team used a comprehensive dataset drawn from the Danish Skin Cohort, the contents of which highlight records on adults in Denmark with atopic dermatitis. These data also include information on disease severity as well as patterns related to flares patterns. The team looked at these findings to examine links between the frequency of flares among such patients in 2022 and these patients' self-reported severity of their disease in 2023.

Nielsen and colleagues used quantile regression models for the purposes of evaluating this relationship. They implemented boosted random forest algorithms with the aim of identifying key predictive variables of both yearly flare frequency and severity of atopic dermatitis. Individuals who had been provided with a dermatologist-confirmed diagnosis during their adulthood were considered eligible for inclusion in the investigators' analysis.

At 2 time points, survey responses were gathered by Nielsen et al: from January 14 - February 6, 2022, and from January 3 - January 31, 2023. Their surveys were designed to capture general data among patients such as age, sex, and comorbid conditions, and to capture specific information related to their atopic dermatitis. The investigative team concluded their research having evaluated 878 participants, all of which had a median age of 49.0 years (interquartile range [IQR], 39.0–59.0 years).

Among the 878 trial subjects, 26 were reported by the team to have no atopic dermatitis flares in 2022.¹ Additionally, 405 reported between 1 - 5 disease flares, 169 reported 6 - 10 flares, and 278 reported more than 10 flares during the same year. In the investigators' quantile regression results, the data demonstrated a significant link between the number of flares in 2022 and several patient-reported severity outcomes reported in 2023.

Even after adjusting for baseline scores on the Patient-Oriented Scoring of Atopic Dermatitis (PO-SCORAD), flare frequency remained significantly linked to scores on the Patient-Oriented Eczema Measure (POEM) and the Dermatology Life Quality Index (DLQI).¹

Additionally, Nielsen and colleagues' machine learning analysis later indicated that flare characteristics—namely duration, frequency, and severity—had been deemed among the strongest predictors of patients' overall severity of disease. They also concluded that atopic dermatitis severity itself had emerged as a key factor in predicting how frequently study subjects reported flares annually.

Overall, this study demonstrated that self-reported numbers of flares in the previous year, as well as duration and severity of disease, are predictive of future atopic dermatitis severity and frequency of flares in the future.

“Although a consensus on how many flares are too many remains to be established, the current findings suggest that flares might serve as an early indicator of disease progression or inadequate disease control, highlighting that flares could be relevant in clinical decision-making to support optimal treatment strategies,” the investigators concluded.¹

References

1. Nielsen M, Nymand LK, Pena AD, et al. Predictors of Flares and Disease Severity in Patients With Atopic Dermatitis Using Machine Learning. JAMA Dermatol. Published online July 16, 2025. doi:10.1001/jamadermatol.2025.2073.

2. Charman CR, Venn AJ, Williams HC. The patient-oriented eczema measure: development and initial validation of a new tool for measuring atopic eczema severity from the patients’ perspective. Arch Dermatol. 2004;140(12):1513-1519. doi:10.1001/archderm.140.12.1513.

3. Wei W, Anderson P, Gadkari A, et al. Extent and consequences of inadequate disease control among adults with a history of moderate to severe atopic dermatitis. J Dermatol. 2018;45(2):150-157. doi:10.1111/1346-8138.14116.