A recent study by Geifman et al, published in the British Journal of Dermatology, described subgroups or ‘trajectories’ of patients with psoriasis with similar patterns of disease severity (Psoriasis Area and Severity Index, PASI). They used a data‐driven latent class mixed-modeling approach, to study the patients, which can be used to divide a heterogeneous population into a few homogeneous groups or ‘trajectories’. Thus, patient characteristics in particular trajectories can be used to predict health outcomes like treatment success. 

They included the data of patients who were treated with various biologics and identified 4 PASI trajectories, with differences in clinical characteristics like body mass index, baseline PASI, psoriasis subtype and the specific biologics.

Pharmacoepidemiological studies include outcomes, exposures and the presence of patient characteristics in analyses according to a binary approach; which may not accurately reflect the real‐world situation, as these elements can change over time. While modeling methods, like trajectory modeling techniques, can summarize complex individual‐level medication‐utilization trajectories or time‐varying exposures, which leads to several groups of patients within a given population who share similar patterns of characteristics over time.

Geifman et al used an unsupervised data‐driven approach to identify subgroups of patients with similar patterns of PASI scores over time, as this approach has the possibility of handling large datasets and the ability to separate them into groups without bias of pre‐existing knowledge. However, such unsupervised methods may have a risk of over-extraction or identifying groups that are not ‘true’; additionally, different methods for clustering groups might produce different subgroups. Nevertheless, these issues were eliminated when the authors analyzed two independent cohorts, which showed overall similarities.

Geifman et al, grouped different biologics due to data availability. Thus, the trajectories were not biologic-specific. However, a sensitivity analysis on adalimumab alone indicated the identified trajectories to be generic instead of treatment-specific. 

From a clinical perspective, defining biologic‐specific trajectories in the future study would be interesting, as different biologics might provide different treatment responses over time. Additionally, Geifman et al suggested that in future studies additional molecular and pharmacological data could add to definitions of subgroups and ultimately lead to a more accurate prediction of treatment responses. Treatment‐specific trajectories and the inclusion of biomarkers in these models would elevate the accuracy of precision medicine.

Thus, the ambitious data‐driven approach utilized by Geifman et al is a promising first step for using and evaluating big (observational) data for the prediction of psoriasis treatment responses and treatment optimization in the future. 

Source: van der Schoot LS, van den Reck JMPA. Data-driven prediction of biologic treatment responses in psoriasis: steps towards precision medicine. Br J Dermatol. 2021;185(4):698-9.