This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

Phototherapy, oral systemic medication and biological outpatient treatment of moderate to severe psoriasis were examined. Phototherapy comprised ultraviolet B (UV-B) or psoralen with ultraviolet-A (PUV-A)). The oral systemic medications were acitretin, cyclosporine and methotrexate. The biological treatments were alefacept, efalizumab, etanercept and infliximab. Full details of the treatment (specified active and assumed maintenance) regimens were given in the paper. Summary details for each are provided in the 'Cost Results' section of this abstract.

Treatment.

Cost-effectiveness analysis.

The study population comprised adult (18 years or older) patients suffering from moderate to severe psoriasis, or those with a baseline Psoriasis Area Severity Index (PASI) score of at least 8 (signifying at least moderate severity).

The setting was unclear, but it would appear to be secondary care. The economic study was carried out in the USA.

The effectiveness data were derived from studies published between 1989 and 2004. The cost data related to 2004.

The effectiveness data were derived from a review of studies.

Although no modelling was specifically described, the authors did calculate the cost-effectiveness of the different regimens.

The outcomes assessed were percentage improvements in PASI (PASI%; from baseline to end point) for UV-B, PUV-A, acitretin, cyclosporine, methotrexate, alefacept, efalizumab, etanercept and infliximab, and for combined regimens of acitretin with PUV-A or UV-B. Risk factors for serious adverse events associated with each treatment were also reviewed in the papers.

The included studies were peer-reviewed, published studies that reported the mean PASI% improvement from baseline and specified the time to the end point assessment. Priority was given to randomised, double-blind, placebo-controlled trials. Other inclusion requirements for the studies were:

a sample size of 10 or more adult (18 years or older) patients who were suffering from moderate to severe psoriasis;

patient recruitment not based on prior response to the targeted treatment;

a treatment duration of between 6 and 14 weeks;

the mean dose of treatment was specified;

treatment excluded ancillary concomitant systemic therapies or phototherapies;

treatment was initiated as de novo rather than as maintenance following stabilisation; and

medication was administered in dose ranges within the current product recommendation.

Clinical case studies or retrospective case reviews, review articles, news articles and letters were also excluded.

MEDLINE was searched for primary studies.

Excluded studies were:

those in which the targeted treatment was omitted;

those in which the participants were not randomised to treatment conditions (for comparative trials);

duplicates of the studies already reviewed;

those that did not apply intent to treat analysis;

those that did not include drop-outs in the analysis; and

those that did not clarify the disposition of drop-outs.

One of the authors selected 84 out of 3,885 articles. The other authors then evaluated extracted summaries of these 84 articles for final selection. Consensus was achieved through discussion. The reviewers were not blinded to the authors of the articles, their institutions, or the journal in which the article was published.

Sixteen primary studies were included in the review.

The studies were not combined.

One large difference in PASI% improvement using UV-B was investigated. The authors attributed this to a lack of robustness in one particular study, which was based on a small sample and open-label trial in comparison with two other larger, double-blind, placebo-controlled trials.

PASI% improvements were:

for broadband UV-B, 47.0 to 83.7%;

for narrowband UV-B, 73.0%;

for PUV-A, 61.1 to 92.9%;

for acitretin, 60.4%;

for cyclosporine, 33.4 to 52.0%;

for methotrexate, 58.4 to 71.6%;

for alefacept, 45.0%;

for efalizumab, 51.0 to 52.0%;

for etanercept, 64.2%;

for infliximab, 82.8%;

for UV-B plus acitretin, 80.7%; and

for PUV-A plus acitretin, 97.3%.

The risks associated with treatments were as follows:

UV-B-associated nonmelanoma skin cancer (NMSC), 2% annual incidence;

PUV-A-associated NMSC, 5-year risk 1.2%;

acitretin-related hyperlipidaemia, 8.25% of patients will require lipid-lowering agents;

cyclosporine-related hypertension, 18% of patients will require calcium-channel antagonists;

UV-B plus acitretin-related hyperlipidaemia, 4.13% of patients will require lipid-lowing agents and NMSC 5-year risk of 1%;

PUV-A plus acitretin-related hyperlipidaemia, 4.13% of patients will require lipid-lowing agents and NMSC 5-year risk of 2%.

The measures of benefits used were PASI1 (1% improvement), PASI50 (50% improvement) and PASI75 (75% improvement). These were derived directly from the review.

The direct costs included were those for medication or phototherapy, monitoring for potential treatment related to adverse events, and treatment costs for adverse events. Discounting was not relevant given the short period of analysis (one year). The costs and the quantities were reported separately. Average wholesale prices at January 2004 rates were used in calculations for oral systemic medications and biologic drugs. The costs for phototherapy were based on 2004 Medicare reimbursement rates. The price year, therefore, was 2004.

No statistical analysis of the costs was carried out.

The indirect costs were not included.

US dollars ($).

No sensitivity analysis was carried out.

See the 'Effectiveness Results' section.

The total annual direct costs were as follows:

methotrexate 7.5 mg, $1,783;

methotrexate 15 mg, $2,378;

PUV-A, $3,854;

broadband UV-B plus acitretin, 25 mg, $4,464;

broadband UV-B, $4,682 to $4,931;

narrowband UV-B, $4,682;

PUV-A plus acitretin 20 mg, $7,440;

cyclosporine 3 mg/kg, $6,951;

cyclosporine 1.5 mg/kg, $4,721;

acitretin 50 mg, $13,060;

infliximab 5 mg/kg, $26,436;

etanercept 50 mg, $21,158;

efalizumab 1 mg/kg, $17,999; and

alefacept 15 mg intramuscular (IM), $27,098.

The cost-effectiveness was calculated as costs to achieve PASI1, PASI50 and PASI75. Annualised costs to achieve PASI1, PASI50, PASI70 were as follows:

methotrexate 7.5 mg: PASI1 = $31, PASI50 = $1,526 and PASI70 = $2,290;

methotrexate 15 mg: PASI1 = $33, PASI50 = $1,660 and PASI70 = $2,491;

PUV-A: PASI1 = $41 to $63, PASI50 = $2,074 to $3,154 and PASI70 = $3,111 to $4,731;

broadband UV-B plus acitretin 25 mg: PASI1 = $55, PASI50 = $2,766 and PASI70 = $4,149;

broadband UV-B: PASI1 = $56 to $100, PASI50 = $2,822 to $4,981 and PASI70 = $4,233 to $7,472;

narrowband UV-B: PASI1 = $64, PASI50 = $3,207 and PASI70 = $4,811;

PUV-A plus acitretin 20 mg: PASI1 = $76, PASI50 = $3,823 and PASI70 = $5,735;

cyclosporine 3 mg/kg: PASI1 = $134, PASI50 = $6,683 and PASI70 = $10,025;

cyclosporine 1.5 mg/kg: PASI1 = $141, PASI50 = $7,067 and PASI70 = $10,600;

acitretin 50 mg: PASI1 = $216, PASI50 = $10,811 and PASI70 = $16,217;

infliximab 5 mg/kg: PASI1 = $319, PASI50 = $15,964 and PASI70 = $23,946;

etanercept 50 mg: PASI1 = $330, PASI50 = $16,478 and PASI70 = $24,717;

efalizumab 1 mg/kg: PASI1 = $346, PASI50 = $17,307 and PASI70 = $25,960; and

alefacept 15 mg IM: PASI1 = $602, PASI50 = $30,109 and PASI70 = $45,163.

The most cost-effective treatment was methotrexate oral systemic medication but this has serious potential long-term risks which were not calculated in the analysis. Psoralen plus ultraviolet A (PUV-A) phototherapy and ultraviolet B (UV-B) phototherapy, as monotherapy and in combination with acitretin, were the next most cost-effective treatments. They also have excellent short- and long-term safety profiles. Based on these results, the authors suggested that although biologics may present an important option in the "psoriasis treatment armamentarium", the results do not justify positioning biologics as first-line therapy for moderate to severe psoriasis.

The rationale for the choice of the comparators was clear and well covered in the paper. You should determine if these technologies are relevant to your own setting.

The authors undertook what appears to have been a systematic review of the literature to identify potential studies. They clearly described their inclusion and exclusion criteria and the methods used to check the validity of data extraction, using more than one researcher and resolving differences through discussion. The studies were not used to produce summary estimates, but the results were reported in full and utilised in the economic analysis. The approach used suggests that the validity of the estimate of effectiveness should be high. However, the hypothesis might have benefited from decision analytical modelling, which the authors did not undertake, within which sensitivity analyses could have been performed.

The benefit measure used was the percentage improvement in PASI. The basic analysis involved dividing the annual cost by the percentage improvement. This was then multiplied by either 50 or 75 to determine the cost-effectiveness of the technology. This is a condition-specific benefit measure and it does not have the comparability of a generic benefit measure such as the quality-adjusted life-year. Comparisons with other health care programmes would, therefore, be problematic.

The good features of the cost analysis were that the costs and the quantities were reported separately and the sources and price year were given. This would assist in reflation exercises to other settings and time periods. In terms of limitations, no statistical or sensitivity analyses were performed and reimbursement rates were used for some cost items (i.e. not reflecting opportunity costs). Therefore, the generalisability of the results outside the USA is likely to be low.

The authors compared their results with some other studies of the cost-effectiveness of the technologies analysed, indicating that similar results had been found. The issue of generalisability to other settings was not addressed. The authors pointed out several limitations of their study. First, only peer-reviewed published studies were included. Second, studies in languages other than English were excluded. Third, treatment effectiveness was based solely on PASI. Fourth, the analysis was based on patients with moderate to severe skin psoriasis. Finally, the study period for studies was only one year. The authors indicated that previous analyses of this nature were conducted when treatment doses exceeded current recommendations. The results reflected the scope of the data that were available for analysis. The authors used average cost-effectiveness ratios, whereas an incremental cost-effectiveness analysis would have been more informative to a decision-maker.

The findings suggested that the most cost-effective treatment is methotrexate, but the long-term risks were not calculated. The authors suggested that, although biologics are an important option in this patient domain, the results do not justify their use as first-line therapy for moderate to severe psoriasis. Decisions on which treatments to utilise, and the order of their initiation, should take safety, efficacy and costs into consideration.

Funded by Connetics Corporation.