A Comparison of the Clinical Effectiveness and Cost-Effectiveness of Treatments for Moderate to Severe Psoriasis

Psoriasis affects approximately 2% to 3% of the population in the United States^1-3 and is characterized by epidermal hyperproliferation, incomplete epidermal differentiation, vascular changes, and inflammation.⁴ Current evidence implicates the immune system and inflammatory mechanisms, in particular T lymphocytes and inflammatory cytokines, in the pathogenesis of psoriasis.⁵

Psoriasis exerts a substantial clinical and economic toll. The degree of disability experienced by patients with psoriasis is comparable to that of patients with other major chronic medical conditions, such as diabetes, arthritis, and congestive heart failure.⁶ In 2004, there were 36,400 inpatient stays, 158,000 outpatient hospital visits, 18,800 emergency department visits, and about 2.26 million physician outpatient visits attributable to psoriasis-related care in the United States.⁷ Annual direct costs of psoriasis have been estimated at $700 million.^8,9 Of these direct costs, 23% were attributed to prescription medications, 13% to outpatient medical visits, 5% to hospitalization, 4% to phototherapy, and the remainder (55%) to over-the-counter medications.⁸ Estimated indirect costs, due to lost productivity, were $1.2 billion in 2004.⁷

FDA-approved treatments for moderate to severe psoriasis, which affects approximately 33% of psoriasis patients,² have burgeoned since 2003, when only acitretin, cyclosporine, methotrexate, and phototherapy were available. As of June 2008, FDA-approved treatments for moderate to severe psoriasis included 3 oral systemic medications, 5 systemic biologic medications, and phototherapy. In addition, there are more than 20 treatments in clinical trials or under FDA review.¹⁰

Given the array of treatments for moderate to severe psoriasis and the lack of head-to-head comparative trials, dermatologists and health care insurers face a daunting task in making evidence-based decisions about the selection of treatments. To that end, we conducted an analysis of published clinical trials of treatments for moderate to severe psoriasis and applied these findings to compare their relative clinical effectiveness and cost-effectiveness.¹¹ Our goals were to create a transparent and flexible comparative-effectiveness methodology that could support current and future evidence-based clinical, as well as coverage and reimbursement, decisions for treatments of moderate to severe psoriasis^11,12 and to establish a benchmark against which to compare the clinical efficacy and costs of future treatments.

Methods

Selection of clinical trials for analysis. Randomized trials of treatments for moderate to severe psoriasis were identified through a MEDLINE search for clinical trials published in English from January 1966 through June 2008. On the basis of these results, a manual search was also conducted of reference lists for seminal clinical trials and review articles. Key word search terms included “psoriasis” and each of the following: acitretin (or etretin), adalimumab, alefacept, cyclosporine (or ciclosporin or ciclosporine), efalizumab, etanercept, infliximab, methotrexate, phototherapy, PUVA (or PUV-A or psoralen), or UVB (or UV-B or ultraviolet B).

One of the authors examined each article abstract for relevance. Articles that could not be ruled out as meeting the predefined inclusion or exclusion criteria were obtained while those that failed to meet one or more criteria were rejected. Next, each full article was closely examined to ensure that all inclusion or exclusion criteria were met. The following data from each article that was included were abstracted by a research associate under the supervision of 2 of the authors: citation; study design; treatment types, duration, and doses; participant inclusion or exclusion criteria; sample size; analytic approach; study end points; results; author conclusions; and reviewer comments.

All study authors evaluated the abstracted information during an on-site 1-day Working Group meeting. The Working Group comprised 2 pharmacist formulary directors, 3 physician medical directors of health plans, and a health economist. Group consensus was required for all determinations. Reviewers were not blinded to authors, institutions, or journals because such methods do not appear to affect systematic review outcomes.¹³ A board-certified dermatologist reviewed the findings to ensure that results were clinically meaningful and relevant.

Consistent with managed care coverage and reimbursement review procedures, we included information from trials of targeted treatments that had been published in peer-reviewed journals. Study inclusion and exclusion criteria were created to minimize the introduction of bias.¹⁴ English-language studies that enrolled adults 18 years or older who had a diagnosis of predominantly plaque-type moderate to severe psoriasis were included. Efficacy analyses were based on the intent-to-treat sample or on samples with no dropouts. Efficacy was reported as either mean percentage improvement from baseline to end point on the Psoriasis Area and Severity Index (PASI); percentage of patients who achieved a mean PASI improvement of at least 75%; or percentage of patients who achieved a 0 or 1 (clear or almost clear) score on the physician-completed patient global assessment at end point. Efficacy was also reported over a 6- to 14-week acute treatment period (as opposed to a maintenance or follow-up period).

Excluded studies were duplicates of other reviewed studies, did not include any of the treatments or outcomes of interest, or employed inappropriate treatment doses that exceeded or fell below the recommended dose or regimens. These studies did not specify the mean dose, treatment regimen, or duration, or they included fewer than 10 patients receiving active treatment. Also not included were studies with patients who had a limited range of skin types and studies that were potentially biased or of poor quality, such as having unaccounted dropouts or a flawed design. Studies that were not clinical trials, such as clinical case reports or review articles, were also not included. The predetermined inclusion and exclusion criteria were then further refined by the exclusion of all nonrandomized trials and those that did not report mean percentage PASI improvement from baseline to study end point.

Effectiveness end point. In psoriasis clinical trials, the FDA requires that the PASI be used as an efficacy end point¹⁵; the PASI is the most often cited measurement of psoriasis treatment efficacy.¹⁶ This measure provides a composite score of a variety of physical signs of psoriasis, including erythema, infiltration, desquamation, and body surface involvement.¹⁷ Total scores range from 0 to 72, and higher scores denote greater severity of disease. In clinical trials, the PASI is usually assessed as the mean percentage change from baseline to end point. A 75% improvement in PASI (PASI 75) reflects a clinically meaningful outcome.^18,19

Cost-effectiveness analysis. Cost-effectiveness was calculated as the sum of total annualized costs for each treatment divided by treatment efficacy. This yielded the cost to achieve a 1% PASI improvement (PASI 1). Because a PASI 75 reflects a clinically meaningful outcome,^18,19 we multiplied the cost to achieve PASI 1 by 75 to derive the annual cost to achieve PASI 75.

Where appropriate, total annualized costs included drug acquisition at the wholesale acquisition cost (WAC); clinical procedures, such as administration of intravenous infusion or administration of phototherapy at Medicare 2008 reimbursement rates; and screening and monitoring as recommended by the product label for drugs and in consensus statements for phototherapy at Medicare 2008 reimbursement rates. For systemic biologic medications that required clinician-administered procedures, such as alefacept intramuscular injection and infliximab intravenous infusion over 2 hours, the costs at Medicare 2008 reimbursement rates were included in total costs; all other systemic biologic agents were assumed to be patient-administered.

Internal server error