Background
Psoriasis is an immune-mediated disease that affects 3.2% of adults in the United States (US) [
1,
2]. The estimated percentages of patients in the US with mild, moderate, or severe psoriasis are 83.3, 11.4, and 5.3%, respectively [
3]. Psoriasis is associated with significant morbidity and clinically significant comorbidities including diabetes, cardiovascular disease, metabolic syndrome, autoimmune diseases, and psychiatric impairment [
4‐
10]. The disease impacts overall quality of life and productivity due to its physical and psychological components [
11]. Currently, retrospective research in psoriasis is limited by knowledge gaps that exist in commonly used data sources, but electronic health records (EHR) may help address this problem [
12].
Population studies in psoriasis have typically relied on claims databases or publicly available national databases [
12]. However, these databases often lack point-of-care data collected at actual clinic visits. Furthermore, these databases are often not specialty-specific and therefore do not ask clinically relevant questions to a particular specialty [
12]. Disease registries or post-marketing drug registries collect useful and structured information [
13‐
16], but these registries may have circumscribed focus, assess a characteristically or geographically limited population, and/or require substantial human resources and financial support [
12]. Specialty-specific EHRs provide an opportunity to conduct research from a large, diverse population using clinically relevant data that are collected at point-of-care as part of a typical provider consultation. Because specialty-specific EHRs are completed by specialist providers, the specialty-specific data are much less likely to be subject to misclassification errors compared to data collected by non-specialist providers. EHR systems that collect large amounts of structured data from diverse dermatology practices can fill substantial knowledge gaps not filled with claims databases, publicly available national databases, or data from single institutions [
12].
In this study, we used point-of-care, real-world clinical data from a widely used dermatology-specific EHR in the US to examine patient-perceived treatment effectiveness, patterns of medication use (duration, switching, and/or discontinuation), healthcare-resource utilization, and medication costs among psoriasis patients.
Methods
Data source
Data were collected from Electronic Medical Assistant (EMA) Dermatology, a HIPAA-compliant dermatology-specific cloud-based EHR (Modernizing Medicine, Inc., Boca Raton, Florida, US). EMA is a widely implemented dermatology-specific EHR platform, used by over 4500 dermatology providers (30% of the market share) across the US. EMA houses data for over 500,000 psoriasis patients from 49 US states and 2 territories. Dermatology providers input data directly into this EHR during clinical visits at point-of-care. Data were de-identified to ensure patient privacy. Research using de-identified records was approved by the New England Independent Review Board.
Study design
This multicentre, longitudinal, observational cohort study retrospectively examined adults (≥ 18 years) with psoriasis who visited participating dermatology practices in the US during the study period (September 1, 2014-September 1, 2015).
Study population
The study population included adults who were diagnosed with psoriasis by a dermatologist, were classified as having moderate-to-severe psoriasis, and visited a dermatology provider during the study period. Patients were considered to have “moderate-to-severe” disease if they were scored ≥3 on the static physician’s global assessment (sPGA [0 = very clear, 5 = very severe]), ≥ 3% body surface area (BSA), received phototherapy, oral systemics (methotrexate, acitretin, cyclosporine, or apremilast), or biologic therapies (etanercept, adalimumab, infliximab, ustekinumab, or secukinumab) during the study period or 6 months prior to study initiation. Patients were categorized into the following treatment groups for analyses: topical treatments, phototherapy, oral systemic treatments (methotrexate, acitretin, cyclosporine, and apremilast), biologic treatments (i.e., etanercept, adalimumab, infliximab, ustekinumab, and secukinumab), combination treatments, and other interventions. Other interventions included patients who were not on topical, phototherapy, oral systemic, biologic, combination therapies commonly used for psoriasis, or patients not receiving psoriasis treatment during the study period. At the time of the study, limited data were available for secukinumab, and data were not available for ixekizumab or brodalumab.
Patient-perceived treatment effectiveness
Patient-perceived treatment effectiveness was assessed by patient response to the following question at follow-up: “I believe this treatment is effective in clearing my skin of psoriasis.” Responses were graded on a 5-point Likert scale (1 = “strongly agree,” 5 = “strongly disagree”). Adherence was determined by a “yes” or “no” patient response in EMA to the following question: “The treatment was followed as directed.”
Patterns of medication usage: duration, switching, and discontinuation
Healthcare resource utilization
Discussion
Health outcomes of psoriasis patients in the real world are critical to inform clinical practice. However, these data are scarce in the US. This is due at least in part to difficulties in synthesizing data across disparate EHR systems. Studies using structured, point-of-care clinical data that are supplied directly by dermatology providers can provide in-depth understanding of clinical interactions occurring at the visit-level in the real world.
This study uses real-world, point-of-care data obtained from dermatology providers on a widely used dermatology specific EHR-platform in the US in order to address clinically relevant questions in patients with moderate-to-severe psoriasis. Many data elements are unique to the present study, such as patient-perceived treatment effectiveness and collection of validated psoriasis outcome measures.
Patient-perceived treatment effectiveness affects clinical decision-making in substantial ways because patient input during clinical encounters often influence management plans. In this study, the majority of patients under the care of dermatology providers agreed that their treatments were effective. Specifically, those receiving biologic medications reported the highest rate of strong agreement that the treatment “is effective in clearing my skin of psoriasis.” The results derived from EMA for patient-perceived treatment effectiveness on biologics are in agreement with published studies [
18]. Of note, greater patient-perceived treatment effectiveness was associated with treatment adherence. Because perceived treatment effectiveness and treatment adherence are interdependent, addressing treatment adherence directly with patients during visits is paramount.
The finding from this study that the majority of psoriasis patients thought their treatment was effective differs somewhat from a previous study [
19]. The previous study consisted of randomly selected psoriasis patients regardless of provider type; more than half of these patients reported dissatisfaction with their treatment [
19]. Two factors may contribute to the differences in findings. First, it is likely that patients under the care of dermatology providers are more likely to be satisfied with their treatments than patients cared for by generalists. Second, temporal differences in the study periods can affect the findings because the availability of advanced therapies for psoriasis has improved over time.
We investigated changes in treatments including switching and discontinuation. The data showed that both treatment switching and discontinuation were common in the study population. Switching was lowest in patients treated with topical therapies even though the highest proportion of those on topical therapies reported that they did not perceive their topical treatment to be effective. One potential explanation is that, even though topical therapies may only be modestly effective, patients continue on them due to perception of these therapies being safer. This explanation is supported by results from Armstrong and colleagues [
19] indicating that a high proportion of psoriasis patients participating in National Psoriasis Foundation (NPF) Surveys with all spectrum of disease severity chose to receive topical treatment alone due to “fewer adverse events than other treatments”. Another potential explanation is that certain providers may have a higher threshold for what may trigger a therapeutic escalation from topical medications alone to initiating systemic therapies.
Across treatment categories, the most frequent reasons reported for discontinuation of treatment were loss of efficacy and side effects. Those on biologics or oral systemics reported the highest frequencies for treatment discontinuation. Among patients who discontinued biologics and oral systemics, the most common reason for discontinuing biologics was loss of efficacy, and the most common reason for discontinuing oral medications was side effects. Similarly, Levin and colleagues [
20] also observed that discontinuations were most frequently due to lack of efficacy for biologics and adverse events for traditional systemic therapies. These results may help providers better understand why patients continue, discontinue, or switch treatments. In addition, these results highlight the need for newer treatments (biologics and oral) that are more capable of providing long-term disease control.
With regards to healthcare resource utilization, patients with more severe disease had greater visit frequency and visit complexity compared to those with milder psoriasis. Therefore, concerted efforts at controlling psoriasis severity are important not only to reduce physical and psychosocial burden, but also to reduce overall healthcare resource utilization. The association between severity and visit complexity (measured as cost) is consistent with findings by Evans [
21]. Evans reported that patients with moderate-to-severe psoriasis have 5-fold higher total healthcare costs versus patients with mild psoriasis [
21].
In addition, healthcare resource utilization also varied across treatment groups. Notably, being on oral therapies required highest visit complexity, possibly due to the need for more intense monitoring and management, greater frequency of laboratory evaluation, and greater incidence of adverse events [
22]. While treatment with biologics was associated with higher costs, biologic medication duration was longer than oral systemics. Phototherapy was associated with the highest visit frequency, which was expected due to the regularity of visits necessary for in-office light treatments. These results highlight the differences in healthcare resource utilization and costs associated with differing levels of disease severity and treatments.
This pilot study highlights the potential for using an EHR such as EMA to conduct real-world health outcomes research. A unique advantage of a dermatology-specific EHR is the minimization of the risk of misclassification because the data are entered directly by dermatology providers. This is supported by a prior study which reported psoriasis diagnoses (as ≥3 ICD-9 codes) provided by dermatologists had a 97.7% positive prediction value [
23]. Both combination therapy data and patient adherence data were limited. In addition, documented medications reflect prescription behaviours and may not accurately represent ultimate utilization by patients. Our grouping of apremilast with other oral systemics may affect cost and healthcare utilization data.
Conclusions
This study used point-of-care, real-world data from a widely implemented US EHR platform to examine several clinically relevant and important questions including patient-perceived treatment effectiveness, medication duration, reasons for switching/discontinuation of treatments, healthcare resource utilization, and costs. With continued development and improvement, EHRs with structured and validated data can serve as a powerful tool for real-world research. Real-world research using EHRs provides valuable insights and help clinicians and payers address questions on treatment patterns, costs of care, real-world effectiveness of treatments, and patient satisfaction in clinical practice.