Dear Editor,


We read with great interest the article “Long-term psoriasis control with guselkumab, adalimumab, secukinumab, or ixekizumab in the USA” recently published by Fitzgerald et al. in this journal. The authors compared treatment persistence, reinitiation, and rates of discontinuation in patients with psoriasis initiated on guselkumab, adalimumab, ixekizumab, or secukinumab using administrative claims data. The results demonstrated better persistence and higher remission rates with guselkumab compared to the other biologics studied [1]. With a growing body of literature related to real-world treatment patterns of biologics in psoriasis, we feel there is a particular need for better understanding the methodologies and interpretations of these types of studies to inform clinical context [2].

Methodology

When considering treatment pattern data, it is important to keep in mind that variability in definitions and methodologies employed among studies can lead to drastically different results, even when using the same data source. This makes it difficult to compare treatment outcomes between studies. When using claims data sets, persistence on a medication is based on prescription refill history. For many reasons (e.g., change of insurance or drug holidays), patients may experience a gap between refills, which is defined as a treatment gap. This gap may be based on a fixed time interval or, alternatively, based on varying gap lengths that correspond to different drug dosing frequencies [3,4,5]. The published study by Fitzgerald et al. utilized a 2× dosing interval gap definition, which introduces some limitations. For example, in comparing persistence of the available biologics for the treatment of psoriasis and psoriatic arthritis, one biologic is considered non-persistent after 14 days of missed doses, while another biologic is considered persistent even after close to half a year of missed doses. The methodology used by the authors could therefore lead to a bias that unfairly favors IL-23 inhibitors, which are dosed infrequently, and against other classes of biologics that are dosed more frequently. Therefore, it would be helpful to include sensitivity analyses using treatment gap definitions independent of dosing schedules in future studies measuring treatment persistence. Table 1 gives an illustration of the different treatment gap definitions used for evaluating biologics in psoriasis or psoriatic arthritis. Additionally, the rationale for not including ustekinumab in the current study, because a similar study comparing treatment persistence with ustekinumab and other biologics was recently published by the same group [6], is not scientifically robust. Moreover, inclusion of other IL-23 inhibitors would have helped conclude if study outcomes are related to a class effect specific to IL-23 inhibitors versus guselkumab alone.

Table 1 Definition of persistence and treatment gaps used in psoriasis studies
Full size table

Second, days of supply in claims databases could vary and may differ from the recommended dosing duration. Imputation on days of supply for only one of the drugs being compared that differed from the recommended dosing duration could introduce bias in persistence analysis [7].

Third, Fitzgerald et al. did not consider adjustment of baseline heterogeneity between treatment groups for potential bias, which is inherent in observational studies. Assessment of heterogeneity in observational studies should consider methods to control confounding, such as matching and re-weighting, for comparative effectiveness [8].

Furthermore, when assessing treatment patterns using administrative claims databases, one should consider evaluating both adherence, which measures the actual days on treatment without counting the gaps, and persistence to provide a more comprehensive assessment of real-world treatment effectiveness. The medication possession ratio (MPR) and proportion of days covered (PDC) are two common methods to estimate adherence or compliance to treatment, which help determine whether a patient is taking their medication as prescribed, regardless of the dosing schedules or treatment gaps. Fitzgerald and colleagues did not report these measures.

Clinical interpretation

We challenge the authors on their use of the terms “disease control” and “remission.” While persistence is accepted as a proxy for effectiveness, it is inappropriate to generalize persistence data as “disease control,” since persistence data are influenced by many factors, including effectiveness, but also by safety, tolerability, healthcare systems, and patient access to therapy. Regarding remission, a patient with no claims for psoriasis-related treatment post-discontinuation with ≥ 6 months of follow-up was defined as in “remission.” Yet, claims for topical therapies were allowed, which is incongruent with the definition of remission. Furthermore, this definition of “remission” does not consider the wide variety of factors that contribute to treatment patterns, including loss of insurance coverage or termination of systemic therapies while a patient is pregnant. Importantly, “disease control” and “remission” are complex to define and are both influenced by multiple factors, all of which would require an evaluation of clinical outcome measures that are not contained within administrative claims databases. For data related to disease control and remission, the best sources will be large registries, such as CorEvitas [9], or prospective observation trials, such as the Psoriasis Study of Health Outcomes [10].

Indeed, administrative claims data are helpful in analyzing adherence and persistence patterns of different therapies. It is important to keep in mind, however, the nuances associated with such analyses and to interpret the results with caution. Finally, we assert there is a need for consistent and standardized definitions, assessments, and analyses across studies, especially for those evaluating biologics used to treat psoriasis, given the differing dosing intervals. Such standardization would improve interpretation of results, which may help healthcare practitioners utilize real-world treatment data to make treatment decisions for patients with psoriasis.