Atopic dermatitis (AD) is a common chronic skin
disease characterised by recurrent skin inflammation
and weak skin barrier, and is known to be a
precursor to other allergic diseases such as asthma.
AD affects up to 25% of children worldwide
and the incidence continues to rise. There is still
uncertainty about the optimal treatment strategy
in terms of choice of treatment, potency, duration
and frequency. This study aims to develop a
computational method to design optimal treatment
strategies for the clinically recommended "proactive
therapy" for AD. Proactive therapy aims to prevent
recurrent flares once the disease has been brought
under initial control. Typically this is done by using
an anti-inflammatory treatment such as a potent
topical corticosteroid intensively for a few weeks
to "get control", followed by intermittent weekly
treatment to suppress subclinical inflammation to
"keep control". Using a hybrid mathematical model
of AD pathogenesis that we recently proposed,
we computationally derived the optimal treatment
strategies for individual virtual patient cohorts, by
recursively solving optimal control problems using
a differential evolution algorithm. Our simulation
results suggest that such an approach can inform the
design of optimal individualised treatment schedules
that include application of topical corticosteroids
and emollients, based on the disease status of
patients observed on their weekly hospital visits.
We demonstrate the potential and the gaps of our
approach to be applied to clinical settings.