Real-world US healthcare costs of psoriasis for biologic-naive patients initiating apremilast or biologics

This retrospective, observational cohort study used administrative claims data of patients identified from the IBM MarketScan^® Commercial and Medicare Supplemental Databases (Truven Health Analytics, now IBM Watson Health, MA, USA). The study period extended from 1 January 2013 to 30 June 2016. The MarketScan databases contain the pooled healthcare experience of >100 million commercial and Medicare supplemental enrollees during the study period. The databases include medical claims for healthcare services performed in the inpatient and outpatient settings as well as enrollment data, including member demographic information, eligibility and benefits data. The medical claims files include service dates, provider reimbursement amounts, patient copayments and deductible amounts. All database records are deidentified and fully compliant with US patient confidentiality requirements, including the Health Insurance Portability and Accountability Act (HIPAA) of 1996. The databases have been evaluated and certified by an independent third party to be in compliance with the HIPAA statistical deidentification standard. The databases were certified to satisfy the conditions set forth in Sections 164.514 (a)–(b)1ii of the HIPAA privacy rule regarding the determination and documentation of statistically deidentified data. Because the study used only deidentified patient records and did not involve the collection, use or transmittal of individually identifiable data, institutional review board approval to conduct this study was not required.

Patients were identified if their index date for treatment initiation with apremilast or a biologic agent for psoriasis and/or psoriatic arthritis (PsA) was between 1 January 2014 and 30 June 2016. Biologic agents included adalimumab, certolizumab, etanercept, golimumab, infliximab, secukinumab or ustekinumab; medications that were approved after 30 June 2015 were not evaluated because the patient population initiating these medications would not have satisfied the study inclusion and exclusion criteria. Inclusion criteria were as follows: at least two medical claims with an International Classification of Diseases (ninth and tenth edition) Clinical Modification diagnosis of psoriasis and/or PsA during the 12 months prior to the index date; aged ≥18 years of age on the index date; continuous enrollment with medical and pharmacy benefits for the 12 months prior to the index date (preindex/baseline period); continuous enrollment with medical and pharmacy benefits during the 12 months following the index date (postindex/follow-up period); no evidence of other biologic-indicated autoimmune conditions in the 12-month pre-index period or 12-month post-index period (ulcerative colitis, Crohn's disease, rheumatoid arthritis and other inflammatory polyarthropathies, including ankylosing spondylitis and juvenile idiopathic arthritis); no evidence of a cancer diagnosis at any time during the study period (pre- or post-index period); and no evidence of multiple index medications on the index date. Analysis was limited to psoriasis patients who were naive to both apremilast and biologic agents; patients with PsA were excluded from this analysis.

The two primary outcomes analyzed in this study were treatment patterns (treatment persistence, persistence-based medication possession ratio [MPR], adherence) and all-cause healthcare costs. Treatment persistence was defined as the proportion of patients without discontinuation of the index medication based on a >60-day gap in therapy following the exhaustion of the previous days’ supply or a switch to apremilast or a biologic prior to the end of the 12-month postindex period. The postindex persistence period was defined as the total number of days from the index date to the earliest discontinuation of the index medication based on a >60-day gap in therapy, or a switch to apremilast or a biologic if the patient was nonpersistent on therapy, or the end of the 12-month postindex period if the patient was persistent on therapy. Persistence-based MPR was the sum of days’ supply for all prescription claims during the postindex persistence period divided by the total number of days in the postindex persistence period. Adherence was defined as a persistence-based MPR of ≥80%.

All-cause healthcare expenditures were reported by type of service (outpatient medical, outpatient pharmacy, total) and measured while patients were persistent on their index treatment per-patient per-month (PPPM) and during the 12-month postindex period. The total costs of 12 months of persistence were calculated among persistent apremilast initiators and among persistent biologic initiators. The total costs for nonpersistent switchers and nonpersistent nonswitchers were calculated among nonpersistent apremilast or biologic patients who switched treatments and among nonpersistent apremilast or biologic patients who did not switch treatments, respectively. Healthcare costs were based on paid amounts of adjudicated claims, including insurer and health plan payments as well as patient cost-sharing in the form of copayment, deductible and coinsurance. Cost for services provided under capitated arrangements were estimated using payment proxies that were computed based on paid claims at the procedure level using the MarketScan Commercial and Medicare Supplemental Databases. All dollar estimates were inflated to 2016 dollars using the medical care component of the Consumer Price Index.

Covariates included age (18–24, 25–34, 35–44, 45–54, 55–64, 65–74, ≥75 years of age), sex, insurance plan type (comprehensive, exclusive provider organization, health maintenance organization, point of service, preferred provider organization, point of service with capitation, consumer driven health plan, other), US Census Bureau geographic region (Northeast, North, Central, South, West, or unknown), urban/rural residence and index year. Comorbid conditions identified during the baseline period included anxiety, cardiovascular disease, depression, headache, hepatotoxicity, hypertension, serious infections, moderate or severe renal disease, osteoarthritis, osteoporosis, postmenopausal symptoms and Type II diabetes mellitus. Concomitant medications were antidiabetics, antidepressants, atypical antipsychotics, benzodiazepines, bisphosphonates/selective estrogen receptor modulators and opioids. The Charlson Comorbidity Index, baseline total healthcare expenditures, psoriatic-related medications and procedures were also noted during the baseline period and while patients were persistent on the index therapy (phototherapy, nonsteroidal anti-inflammatory drugs/cyclooxygenase-2 inhibitors, corticosteroids, systemic agents [acitretin, cyclosporine, methotrexate, sulfasalazine, leflunomide]).

All variables were summarized descriptively as means (standard deviation) or counts (percentages). The t-tests/Wilcoxon rank-sum tests were used to test for significant differences between continuous variables and χ² tests were used for categorical variables.

Apremilast and biologic users were 1:2 propensity score matched using the nearest neighbor matching with calipers (0.25 width) of the pooled standard deviation. The propensity score was calculated from logistic regression model predicting the probability of treatment assignment conditional on observed baseline patient demographics and clinical characteristics. The matching characteristics included in the model were age group, gender, health plan, payer, region, index year and urbanicity.

Generalized linear models (GLMs) with log link and γ-error distribution were used on the matched cohorts to estimate treatment effect on healthcare costs, adjusting for patient demographics, baseline clinical characteristics and baseline total costs. Models were fit separately by type of service (outpatient medical, outpatient pharmacy, total) while the patient was persistent on their index treatment and during the 12-month postindex period.

From the MarketScan Commercial and Medicare Supplemental Databases, 128,927 patients were identified as having initiated a new treatment episode with apremilast or a biologic agent for psoriasis and/or PsA between 1 January 2014 and 30 June 2016. After applying the inclusion and exclusion criteria, 8519 eligible patients were identified (Figure 1). The apremilast and biologic patients were then propensity score matched; of those successfully matched, the biologic-naive patients were retained to provide a final matched, apremilast- and biologic-naive sample composed of 703 psoriasis patients initiating apremilast and 1378 psoriasis patients initiating a biologic.

After propensity score matching, the two biologic-naive study cohorts had similar patient demographics and clinical characteristics. The mean age was 49 years, 50% were female and the mean Charlson Comorbidity Index score was 0.4 (Table 1). Concomitant medication use, comorbid conditions and baseline healthcare costs were also similar; however, baseline medication history differed somewhat in that fewer patients initiating apremilast had prior methotrexate use in the 12-month pre-index period compared with those initiating a biologic (11.0 vs 18.7%; p < 0.001).

Treatment persistence in the 12-month postindex period was similar for psoriasis patients initiating apremilast compared with patients initiating a biologic (44.4 vs 44.6%; p = 0.914). Persistence-based MPRs were similar for patients initiating apremilast compared with patients initiating a biologic (0.86 vs 0.86; p = 0.654), as were adherence rates (73 vs 74%; p = 0.516). The patients initiating apremilast were less likely to be nonpersistent because of switching compared with patients initiating biologics (Table 2).

Unadjusted PPPM total healthcare costs while persistent on therapy were lower among apremilast patients compared with biologic patients (US$3426 vs US$5649; p < 0.001). Compared with biologic users, apremilast users had lower PPPM outpatient pharmacy costs (US$2478 vs US$4922; p < 0.001) and lower PPPM outpatient medical costs (US$407 vs US$640; p < 0.05). Differences in outpatient medical costs were due to lower PPPM costs for other outpatient services, which included services such as radiology, infusion administrations, rehabilitation and physical therapy (US$272 vs US$488; p < 0.05).

Unadjusted total healthcare costs during the 12-month follow-up period were lower among apremilast patients compared with biologics patients (US$32,304 vs US$49,875; p < 0.001); differences were driven by lower outpatient pharmacy costs (US$23,376 vs US$41,803; p < 0.001).

Adjustment for differences in baseline clinical characteristics with GLMs did not change the magnitude or the direction of the results: Apremilast users experienced lower costs while persistent on therapy and during the 12-month follow-up period compared with biologic users. PPPM total costs were US$2458 lower among apremilast patients while persistent on therapy (US$2484 vs US$4942; p < 0.001), which were driven by lower PPPM outpatient pharmacy costs and PPPM outpatient medical costs (Figure 2). During the 12-month postindex period, total healthcare costs were US$19,327 lower for apremilast patients compared with biologic patients (US$31,337 vs US$50,664; p < 0.001) because of lower total outpatient pharmacy costs (US$23,171 vs US$42090; p < 0.001; Figure 3).

The total healthcare costs for 12 months of persistence were US$19,287 lower among persistent apremilast patients compared with persistent biologic patients (US$37,159 vs US$56,446; p < 0.001; Figure 4). The difference in total cost over 12 months decreased to US$17,556 among nonpersistent nonswitchers initiating apremilast compared with nonpersistent nonswitchers initiating a biologic (US$30,278 vs US$47,834; p < 0.001). Compared with persistent patients and nonpersistent nonswitchers, the total costs over 12 months were highest among the nonpersistent patients who switched treatments (apremilast: US$51,542, biologics: US$61,825; p < 0.001). The cost advantage among apremilast patients who switched treatments decreased to US$10,283 less than biologic initiators who switched treatments because of higher pharmacy costs.