Frequent Flares Signal Worse Disease Course in Atopic Dermatitis, Machine Learning Study Finds
Denmark: A recent study published in JAMA Dermatology has revealed that the frequency, severity, and duration of flares in patients with atopic dermatitis (AD) are strong predictors of future disease severity and quality of life. The research, led by Mia-Louise Nielsen from the Department of Dermatology at Copenhagen University Hospital-Bispebjerg in Denmark, utilized advanced machine learning techniques to identify which factors most significantly impact the progression of the chronic skin condition.
Atopic dermatitis is known for its unpredictable course, marked by recurring flare-ups and symptom fluctuations. However, despite the clear burden flares place on patients, they are often underrepresented in clinical evaluations and severity classifications. This study aimed to bridge that gap by analyzing how flares influence long-term outcomes and determining which variables could best predict disease trajectory.
Drawing on data from the Danish Skin Cohort, the researchers analyzed information from 878 individuals diagnosed with AD. Participants’ flare activity from 2022 was compared to their reported disease severity in 2023 using quantile regression models. Additionally, machine learning models—specifically boosted random forests—were applied to identify key predictors of flare frequency and AD severity.
The findings were noteworthy and are outlined as follows:
- Patients with more frequent flares in 2022 reported greater disease severity in 2023 and a lower quality of life.
- Among the 878 participants, 26 reported no flares, 405 reported 1 to 5 flares, 169 had 6 to 10 flares, and 278 experienced more than 10 flares in 2022.
- Flare-related factors, such as number, duration, and severity, were identified as key predictors of future disease severity.
- Baseline disease severity was also a strong predictor of flare frequency, indicating a two-way relationship between flares and overall disease burden.
- After adjusting for baseline PO-SCORAD scores, flare frequency showed a significant association with future scores on the Patient-Oriented Eczema Measure and the Dermatology Life Quality Index.
These insights have important implications for clinical practice. The researchers emphasized that information on flare activity should be more thoroughly incorporated into treatment planning and severity assessments. Recognizing flare patterns as a core component of disease management could help clinicians make more informed decisions, aiming not just to reduce symptoms but to prevent disease worsening and improve patients' day-to-day experiences.
The study advocates for defining a clear threshold for what constitutes an "acceptable" number of flares, which could serve as a treatment target in managing atopic dermatitis more effectively. Ultimately, these findings support a more personalized, proactive approach to AD care, emphasizing flare prevention as a key strategy for improving patient outcomes.
Reference:
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
