Topical therapies for the treatment of localized plaque psoriasis in primary care: a cost-effectiveness analysis

Authors


  • Funding sources National Institute for Health and Clinical Excellence and National Institute for Health Research comprehensive Biomedical Research Centre.
  • Conflicts of interest All authors were members of the National Institute for Health and Clinical Excellence Psoriasis Guideline Development Group (C.H.S. chaired the group, L.S. and D.W. were the health economists and E.J.S. was the research fellow). After completing the guideline analysis but before its publication, L.S. joined Symmetron Limited.

Correspondence

Laura Sawyer.

E-mail: lsawyer@symmetron.net

Summary

Background

Topical therapies are a mainstay of psoriasis treatment, but they vary substantially in terms of cost.

Objectives

To determine the cost-effectiveness and optimal treatment sequence for psoriasis of the trunk, limbs and scalp.

Methods

Probabilities of response from a network meta-analysis were used to determine the short-term efficacy of topical therapies. Longer-term outcomes, including relapse, were informed by published evidence and clinical opinion. Benefits of treatment were measured as quality-adjusted life years (QALYs). Direct costs included topical agents, primary and secondary care visits and second-line therapies for treatment failures.

Results

For the trunk and limbs, initial treatment with a two-compound formulation (TCF) product containing vitamin D and potent corticosteroid provided the most QALYs, followed by separate morning and evening application of vitamin D and potent corticosteroid [two-compound application, TCA (am/pm)], and then twice-daily potent corticosteroids. The use of twice-daily potent corticosteroids was the most cost-effective first-line strategy (incremental cost-effectiveness ratio £20 000 per QALY), followed by TCA (am/pm) (£22 658 per QALY) and TCF product (£179 439 per QALY). For scalp psoriasis, initial treatment with very potent corticosteroids generated the most QALYs, followed by TCF product and then potent corticosteroids. Very potent corticosteroids were the most cost-effective treatment but, if too aggressive, potent corticosteroids were optimal followed by TCF product (£219 846 per QALY). The cost-effectiveness of second- and third-line topical agents varied with the assumptions made.

Conclusions

Potent corticosteroids, used alone or in combination with vitamin D, are the most cost-effective treatment for patients with psoriasis of the trunk and limbs. Potent or very potent corticosteroids are the most cost-effective treatment for patients with scalp psoriasis.

Psoriasis is an inflammatory skin disease that typically follows a relapsing and remitting course and affects an estimated 1·3–2·2% of people in the U.K. Although rarely fatal, its impact encompasses functional, psychological and social dimensions,[1] significantly impairing quality of life. Most people affected have localized disease, and here topical therapy is the principal approach to treatment. A multitude of different topical interventions are available for the treatment of psoriasis, some of which are fairly inexpensive (< £10 per month) while others cost nearly 10 times that. While the unit cost of each topical agent is low in the context of high-cost biological agents, multiplied across the population for whom this is the mainstay of treatment, the total health spend on topical agents is significant, with more than an estimated £44 million spent on topical therapies for psoriasis in England alone in 2011.[2]

The National Institute for Health and Clinical Excellence (NICE) commissioned a clinical guideline to provide recommendations to health professionals in England and Wales on the assessment and treatment of patients with psoriasis.[3] Unlike most clinical guidelines, the NICE guidance explicitly considers cost-effectiveness alongside clinical efficacy when formulating recommendations.[4] Here, we describe an original cost-effectiveness model developed to inform recommendations around topical therapy.

Materials and methods

Model overview

The analysis compares the cost-effectiveness of different topical therapies licensed for patients with psoriasis for whom topical therapy is expected to be practical, effective and safe in the long term. The analysis separately considers treatment of psoriasis of the trunk and limbs and psoriasis of the scalp. Patients with extensive disease or where topical treatment is not effective (e.g. nail disease) are not considered in the analysis.

A schematic of the Markov model care pathway is presented in Figure 1. The model was considered by a panel of general practitioners (GPs), dermatologists, specialist nurses, pharmacists and patient representatives, to reasonably reflect clinical practice. Consequences of treatment are reflected as a set of possible transitions between health states over a series of 4-week cycles. The first part of the structure is common to both populations (trunk/limbs and scalp), whereas the second part of the model, capturing the consequences following a failure of topical therapies, differs. It was considered important to model sequences of first-, second- and third-line topical therapies (Table 1), given that most patients will try several therapies in primary care before moving to more intensive treatments such as specialist applied topical agents, phototherapy and/or systemic therapy.[5]

Table 1. List of possible first-, second- and third-line treatments that may be combined as a sequence for the treatment of psoriasis of the trunk, limbs and scalp
First lineSecond lineThird line
  1. OD, once daily; BD, twice daily; am/pm, separate morning and evening application. aTCA am/pm vitamin D and potent corticosteroid. bScalp model only. cTrunk/limb model only.

Vitamin D OD (Vit D OD)Vit D ODVit D OD
Vitamin D BD (Vit D BD)Vit D BDVit D BD
Potent corticosteroid OD (PS OD)PS ODPS OD
Potent corticosteroid BD (PS BD)PS BDPS BD
Two-compound formulation product (Dovobet®; Leo Pharma, Princes Risborough, U.K.) OD (TCF OD)TCF ODTCF OD
Two-compound application (TCA; am/pm)aTCA (am/pm)TCA (am/pm)
Very potent corticosteroid OD (VPS OD)bVPS ODbDithranol OD
Very potent corticosteroid BD (VPS BD)bVPS BDbCoal tar BD
Coal-tar shampoob
VPS ODb
VPS BDb
Referral to specialist
Restrictions on sequences to ensure their safe and logical use
Any topical OD would not come after a failure of the same topical BD
PS OD or Vit D OD would not come after TCA (am/pm) treatment or TCF OD
TCA (am/pm) would not come after a failure of TCF OD
PS would not come after VPSb
No strategy could include corticosteroids among all three lines of treatment [including as part of a TCA (am/pm) or TCF product]c
Figure 1.

Markov model, a simplified transition diagram representing patient flow, with differentiated pathways for trunk and limb nonresponders and scalp nonresponders. Hypothetical patients either respond or do not respond. Responders stop treatment and either maintain response or relapse. Patients who relapse resume treatment and again respond or do not respond. Patients who do not respond after up to 8 weeks are assumed to return to their general practitioner (GP) for an alternative topical therapy. Patients try up to three topical therapies before being referred for specialist review where second-line treatment options are considered.

The analysis took a U.K. National Health Service (NHS) perspective, with costs expressed in 2011 pound sterling. A 1-year time horizon was considered clinically relevant and sufficiently long to capture important costs and consequences of treatment in primary care. The performance of alternative treatment sequences was estimated using incremental cost-effectiveness ratios (ICERs), defined as the added cost of a given strategy divided by its added benefit compared with the next most expensive strategy.

Benefits

Clinical efficacy

The benefits of each treatment were determined by the proportion of patients responding during a given 4-week cycle. Probabilities of response for each treatment were derived from a systematic review and network meta-analysis (NMA) of randomized controlled trial (RCT) data.[6] Based on trial evidence and clinical opinion, patients were assumed to undergo a maximum of 8 weeks of continuous therapy with a given topical agent, except in the case of very potent corticosteroids, which for reasons of safety we assumed would be trialled for only 4 weeks. Response could occur in the first 4 weeks (early responders) or the second 4 weeks (late responders). Table 2 and Figure 2 present the overall response rates for each treatment from the NMA, and a breakdown of these rates into early and late response.

Table 2. Clinical efficacy and quality-of-life inputs in the models
VariableValueDescription/source
  1. OD, once daily; BD, twice daily; TCF, two-compound formulation product containing potent corticosteroid and vitamin D; TCA, two-compound application of vitamin D and potent corticosteroid; am/pm, separate morning and evening application; NMA, network meta-analysis. awww.mrc-bsu.cam.ac.uk/bugs.

Clear/nearly clear response, %Early/total 
Trunk/limbs
Placebo BD13/13Total responses generated from NMA.[6] Parameter uncertainty is obtained directly from posterior distribution from winBUGS.a Early and late responders based on Fleming[28] for placebo, vitamin D, potent corticosteroid, TCF product and TCA (am/pm), and on Hutchinson[29] for dithranol and coal tar
Vitamin D OD15/44
Vitamin D BD18/54
Potent corticosteroid OD27/48
Potent corticosteroid BD35/62
TCF OD42/71
TCA (am/pm)37/65
Coal tar BD27/55
Dithranol OD21/43
Scalp
Placebo BD7/11Total responses generated from NMA.[6] Parameter uncertainty is obtained directly from posterior distribution from winBUGS. Early and late responders based on Jemec,[30] Kragballe[13] and van de Kerkhof.[20] Maximum 4 weeks' continuous treatment with very potent corticosteroids
Vitamin D OD21/35
Vitamin D BD21/35
Potent corticosteroid OD48/57
Potent corticosteroid BD42/49
Very potent corticosteroid OD69/69
Very potent corticosteroid BD78/78
TCF OD56/64
Coal-tar shampoo OD10/19
Second-line treatments
Phototherapy (trunk and limbs only)87Dawe,[31] Hallaji,[32] Cameron[33]
Intensive scalp treatment75Assumption
Relapse, %
All topical therapies36Camarasa,[7] Christensen,[8] Langley[34]
Referral to specialist, %
Trunk/limbs60Schofield[35]
Of referred, receive phototherapy30Assumption
Scalp100Assumption
Of referred, treated by specialist nurse25Assumption
Of referred, receive intensive scalp treatment25Assumption
Utilities, mean (standard error)
Trunk/limbs
Baseline0·8Bottomley[11]
Nonresponder utility gain0·05 (0·01)Assumption
Responder utility gain0·09 (0·018)Bottomley[12]
Scalp
Baseline0·767Affleck[10]
Nonresponder utility gain0·0111 (0·009)Affleck[10]
Responder utility gain0·0292 (0·010)Affleck[10]
Figure 2.

Probabilities of response from evidence synthesis, broken down into early and late response. OD, once daily; BD, twice daily; PS, potent corticosteroid; TCF, two-compound formulation product of potent corticosteroid and vitamin D analogue; TCA (am/pm), two-compound application of vitamin D and potent corticosteroid (separate morning and evening application); VPS, very potent corticosteroid; CT, coal tar.

Psoriasis is a relapsing and remitting chronic condition and there is no evidence that response to treatment has an effect on its natural history. Evidence to support treatment-dependent relapse rates is limited;[6] therefore, we assumed the same rate for all treatments. Based on trials reporting relapse,[7-9] the average 4-week risk was 35·5%. We assumed relapse could occur at any point following response.

Health-related quality of life

Response and the associated utility gain were used in the model to determine the impact of treatment on overall health. Utility-gain estimates were informed by recent cost–utility analyses[10, 11] and were assumed to depend on response rather than treatment. Bottomley et al.[11] used EQ-5D data (EuroQol, Rotterdam, the Netherlands) gathered by an RCT[12] evaluating topical therapies for trunk/limb psoriasis. Affleck et al.[10] mapped SF-36 data (www.sf-36.org) gathered by an RCT[13] comparing topical therapies for scalp psoriasis onto SF-6D scores using established mapping algorithms.[14]

Patients not achieving clearance or near clearance are still expected to experience some improvement on treatment, and thus a modest utility gain. Patients on treatment awaiting a full response (e.g. late responders), patients having relapsed following response, and nonresponders being managed in the long term by a GP or specialist, are assumed to accrue utility gains associated with nonresponse. Patients who fail to respond and require specialist referral are assumed to return to their baseline utility while they wait to receive – and potentially respond to – more aggressive treatment.

Resource use and costs

Topical therapy

Costs of topical therapies are based on reported mean quantities used in clinical trials[12, 15-21] converted into packs of medication available for clinical practice. Unit costs were taken from the British National Formulary.[22] Interventions were modelled assuming a class effect; therefore, the cost for the product with the lowest unit cost per gram per millilitre was used. Only scalp formulations were considered for scalp psoriasis. Resource use and costs of topical therapies are presented in Table 3.

Table 3. Resource use and unit costs
 4-week quantity (SE; grams)/ no of visitsUnit cost (£)4-week cost (£)Descriptiona/source of quantity
  1. OD, once daily; BD, twice daily; TCF, two-compound formulation product containing potent corticosteroid and vitamin D analogue; TCA, two-compound application of vitamin D and potent corticosteroid; am/pm, separate morning and evening application; BNF, British National Formulary; PSSRU, Personal Social Services Research Unit; NHS, National Health Service. aDiprobase®, Schering-Plough, Welwyn Garden City, U.K.; Silkis®, Galderma, Watford, U.K.; Betnovate®, Dermovate®, GSK, Uxbridge, U.K.; Dovobet®, Leo Pharma, Princes Risborough, U.K.; Psoriderm®, Dithrocream®, Doublebase®, Betacap®, Capasal®, Dermal Laboratories, Hitchin, U.K.

Resource use
Trunk/limbs
Placebo BD152·8 (30·6)6·321·93500 g Diprobase, Guenther[12]
Vitamin D OD142·0 (28·4)13·8727·74100 g Silkis, Kaufman[17]
Vitamin D BD164·9 (33·0)27·74Douglas,[15] Guenther[12]
Potent corticosteroid OD140·0 (28·0)4·05, 1·436·91100 g and 30 g Betnovate cream or ointment, Kaufman[17]
Potent corticosteroid BD144·5 (28·9)6·91Douglas[15]
TCF OD134·0 (26·8)61·27, 32·9994·26120 g and 60 g Dovobet ointment, Kaufman[17]
TCA (am/pm)80·45 each (32·2)17·92Bottomley[11]
Coal tar BD339·2 (67·8)9·4218·84225 g Psoriderm cream, assumed same as dithranol
Dithranol OD339·2 (67·8)3·77, 4·04, 4·66, 5·4235·78 (first 4 weeks), 37·94 (later 4 weeks)50 g 0·1%, 0·25%, 0·5%, 1% Dithrocream, van de Kerkhof[20]
Scalp
Placebo BD77·6 (15·5)5·830·90500 g Doublebase gel, Jemec,[30] Jemec,[16] Tyring[19]
Vitamin D OD89·2 (5·8)26·07, 12·7025·40120 g and 60 g Calcipotriol scalp solution, Jemec,[30] Jemec,[16] van de Kerkhof[20]
Vitamin D BD85·6 (13·9)25·40Affleck[10]
Potent corticosteroid OD87·35 (6·6)3·753·75100 g Betacap scalp application, Buckley[36] Jemec,[30] Jemec,[16] van de Kerkhof[20]
Potent corticosteroid BD90·16 (6·6)3·75Assumption based on once-daily application
Very potent corticosteroid OD60 (12·0)10·42, 3·076·14100 g and 30 g Dermovate scalp application, max suitable quantity BNF[22]
Very potent corticosteroid BD60 (12·0)6·14Max suitable quantity BNF[22]
TCF OD71·4 (6·3)67·79, 36·5067·79120 g and 60 g Dovobet gel, Buckley,[36] Jemec,[30] Jemec,[16] Tyring,[19] van de Kerkhof[20]
Coal-tar shampoo OD250 mL (50·0)4·694·69250 mL Capasal shampoo, assumption
Healthcare consultations    
GP consultations  PSSRU[37]
For treatment changeOne visit£28/visit  
For long-term managementOne visit per 3 months  
Specialist referral (consultant)One following failure of all topicals in a sequence£112/visit NHS reference costs 2009–10[38]
Trunk/limbs
Phototherapy session24 sessions per course£82/visit NHS reference costs 2009–10[38]
Scalp
Specialist follow-up and support (specialist nurse)Three outpatient visits£64/visit NHS reference costs 2009–10[38]
Intensive scalp treatment in outpatient day careThree visits£351/visit NHS reference costs 2009–10[38]

Other healthcare resource use

We assumed that patients would consult their GP following a nonresponse and receive a different topical agent. Following relapse, we assumed that patients would receive a repeat prescription for the same topical agent without another GP visit. Patients failing to respond to all therapies in a sequence were referred for specialist review. Between referral and review, we assumed that patients maintain topical treatment, for which the average cost across all topical treatments was used (£29·78 for trunk/limbs, £17·88 for scalp).

Given the population considered in the model, we assumed that the risks and/or inconvenience of phototherapy (or systemic therapy) would not be justified in most patients. Therefore, once referred, many patients would be discharged back to their GP, with some followed up by a specialist nurse and a minority undergoing more aggressive second-line therapies including phototherapy or intensive scalp treatment (Tables 2 and 3).

Sensitivity analyses

All analyses were conducted in a probabilistic framework, thus capturing the imprecision and uncertainty around parameter point estimates. Probability distributions were based on error estimates from data sources, such as confidence intervals. Uncertainty in the probabilities of response were obtained directly from the joint posterior distributions of the NMA.[6] Beta distributions were used for other probabilities, gamma distributions were used for resource use and disutility values, and lognormal distributions were used for NHS reference costs. The results of 5000 simulations are summarized as mean costs and mean quality-adjusted life years (QALYs).

In addition, scenario analyses and a series of one-way sensitivity analyses (Table 4) were performed to assess how changes in one or more parameters might change the conclusions of the analysis.

Table 4. Description of one-way sensitivity analyses performed
Sensitivity analysisDescription
  1. PAGI, patient assessment of global improvement; NMA, network meta-analysis; TCA, two-compound application of vitamin D and potent corticosteroid; am/pm, separate morning and evening application; TCF, two-compound formulation product containing potent corticosteroid and vitamin D analogue. aGP Pharma, Hitchin, U.K.; bAlmirall, Uxbridge, U.K.

Clinical effectivenessUse of patient-reported response data (PAGI) from NMA
Reduced adherenceAssume 40% of patients do not adhere to twice-daily treatment[39]
60% adhere and receive benefits of twice-daily application
40% do not adhere and receive benefits of once-daily application
Scenario 1. Patients receiving TCA (am/pm) who do not adhere to a twice-daily application are assumed to use only the steroid component of the regimen once daily
Scenario 2. Patients receiving TCA (am/pm) who do not adhere to a twice-daily application are assumed to use only the vitamin D component of the regimen once daily
Early vs. late responsePatients receive a maximum of 4 weeks of treatment to assess response or nonresponse
UtilitiesPatients will receive no benefit from nonresponse
Topical useAmount of TCF product used in 4 weeks reduced. Trunk/limbs: 92·6 g,[11] scalp: 60 g[10]
Acquisition costMost costly per unit brand/formulation for potent corticosteroids and vitamin D assumed
Fluocinolone acetonide (Synalar®) gela
Tacalcitol (Curatoderm®) ointmentb
30% discount applied to cost of TCF product
ReferralHigher (80%) and lower (40%) referral rates for specialist review
Time horizonExtended to 3 years and 10 years, 3·5% discounting rate applied to costs and benefits

Results

Trunk and limbs

Base case

This analysis found that among 118 sequences evaluated for the treatment of trunk/limb psoriasis, the difference between the most and least effective strategies was only 0·0144 QALYs (5·26 quality-adjusted days), reflecting the similar efficacy between various topical therapies. Differences in cost are greater, with per-patient costs ranging from £227 to £664. Results indicate that starting with twice-daily potent corticosteroids and moving to combined vitamin D and potent corticosteroid, applied separately, morning and evening [two-compound application, TCA (am/pm)] and then twice-daily coal tar, is least costly (see Table 5). Switching the order to start with TCA (am/pm) would be most cost-effective if decision-makers are willing to pay (WTP) up to £22 658 per additional QALY, but starting with a two-compound formulation (TCF) product containing vitamin D and potent corticosteroid instead of TCA (am/pm) would be cost-effective only if that threshold were above £179 439. Complete results of the probabilistic analysis are presented in Table S1 and Figure S1 (see Supporting Information).

Table 5. Incremental analysis of base-case results
Treatment sequenceaTotal costs (£)Incremental costs (£)Total QALYsIncremental QALYsICER (£/QALY)
  1. QALY, quality-adjusted life year; ICER, incremental cost-effectiveness ratio; PS, potent corticosteroid; TCA, two-compound application of vitamin D and potent corticosteroid (separate morning and evening application); TCF, two-compound formulation product containing potent corticosteroid and vitamin D analogue; VPS, very potent corticosteroid; OD, once daily; BD twice daily. aAll sequences not presented here were ruled out through dominance (more costly and less effective than a strategy included in the table) or extended dominance (more costly and less effective than a mixture of two other strategies included in the table). Full results for all comparators are presented in Tables S1 and S4 (see Supporting Information).

Trunk/limbs psoriasis only
PS BD – TCA – Coal tar BD227 0·8487  
TCA – PS BD – Coal tar BD23690·84910·000422 658
TCF OD – PS BD – Coal tar BD4281920·85020·0011179 439
Scalp psoriasis only
PS OD – VPS OD – VPS BD163 0·77405  
VPS OD – VPS BD – TCF OD190270·775410·0013619 706

Due to their lower rank in effectiveness and higher acquisition cost, sequences including vitamin D monotherapy were dominated by (i.e. were more costly and less effective than) sequences relying more heavily on corticosteroids. Use of dithranol and referral after just two topical therapies were also dominated by alternatives. Costs for these less effective strategies were driven by referrals and specialist management for nonresponders.

Sensitivity analyses

The use of continuous and high-dose potent corticosteroids may increase the incidence of adverse events (AEs), which are not captured in this analysis. In a scenario analysis, sequences with consecutive use of steroidal products were removed, as were sequences where twice-daily potent corticosteroids appeared as first- or second-line treatment. Results of this restricted-comparator scenario, using base-case parameter values, show that TCA (am/pm) treatment followed by twice-daily vitamin D and then twice-daily potent corticosteroids is likely to be optimal at a WTP threshold of £20 000. Replacing third-line potent corticosteroids or first-line TCA (am/pm) with TCF product results in higher costs (by £67 and £188, respectively), but the associated QALY gains would be justified only at WTP thresholds over £160 000. Figure 3 presents the probability that a given topical therapy is the most cost-effective initial treatment across a range of thresholds up to £100 000, at all of which TCA (am/pm) is optimal.

Figure 3.

Cost-effectiveness acceptability curve for first-line topical therapies for psoriasis of the trunk and limbs under the restricted-comparator scenario. PS, potent corticosteroid; TCF, two-compound formulation product of potent corticosteroid and vitamin D analogue; TCA (am/pm), two-compound application of vitamin D and potent corticosteroid (separate morning and evening application); Vit D, vitamin D or vitamin D analogue; OD, once daily; BD, twice daily; QALY, quality-adjusted life year.

Table 6 presents the results of sensitivity analyses, all within the framework of the restricted-comparator scenario (see also Tables S2 and S3; see Supporting Information). The results suggest that base-case conclusions are insensitive to changes in efficacy, adherence, cost and time horizon. The cost-effectiveness of TCF product compared with TCA (am/pm) improves relative to the base case when adherence to twice-daily strategies is reduced (we assumed in this sensitivity analysis that 40% of patients would apply treatment only once daily), when the cost of TCF product is discounted by 30%, when the most costly corticosteroid and vitamin D products are used and when the time horizon is extended up to 3 years and 10 years; however, the ICER for TCF product remains firmly above the £20 000 WTP threshold. Similarly, third-line treatment with TCF product is unlikely to be cost-effective compared with twice-daily potent corticosteroids, as it is either extendedly dominated by other strategies or the ICERs far exceed conventional WTP thresholds. To be a cost-effective first-line treatment option, threshold analyses showed that the unit cost of TCF product would need to drop by 55–70% given perfect adherence and by 20–50% given reduced adherence.

Table 6. Results of sensitivity analyses within the framework of restricted comparators – trunk and limbs
Sensitivity analysisCost per QALY gainedOptimal strategy (threshold £20 000 per QALY)
TCA – VitD BD – TCF OD vs. TCA – VitD BD – PS BDTCF OD – VitD BD – PS BD vs. TCA – VitD BD – PS BD
  1. QALY, quality-adjusted life year; TCA, two-compound application of vitamin D and potent corticosteroid (separate morning and evening application); VitD, vitamin D or vitamin D analogue; TCF, two-compound formulation product containing potent corticosteroid and vitamin D analogue; PS, potent corticosteroid; OD, once daily; BD, twice daily; PAGI, patient assessment of global improvement. aConcurrent – VitD BD – PS BD ruled out through extended dominance. bRuled out through extended dominance. cDominated by TCF OD – VitD BD – PS BD. Results for all comparators are presented in Tables S2 and S3 (see Supporting Information).

Base case – restricted comparators£160 238£173 028TCA – VitD BD – PS BD
PAGI response data£51 183£154 889PS OD – VitD BD – TCAa
Reduced adherence, Scenario 1£89 740b£69 170TCA – VitD BD – PS BD
Reduced adherence, Scenario 2£89 872b£59 251TCA – VitD BD – PS BD
Four-week treatment£157 627c£91 549TCA – VitD BD – PS BD
Nonresponder utility gain = 0£78 966£133 542TCA – VitD BD – PS BD
Reduced-quantity TCF£118 947b£117 573TCA – VitD BD – PS BD
Unit cost PS and Vit D£122 727b£105 000TCA – VitD BD – PS BD
30% discount on TCF product£98 293£102 170TCA – VitD BD – PS BD
Increased referral (80%)£149 302£171 651TCA – VitD BD – PS BD
Reduced referral (40%)£200 000b£194 068TCA – VitD BD – PS BD
Time horizon (3 years)£110 300b£90 794TCA – VitD BD – PS BD
Time horizon (10 years)£97 668c£73 449TCA – VitD BD – PS BD

Scalp

Base case

Among 169 sequences evaluated for the treatment of scalp psoriasis, the difference between the most and least effective alternatives was modest (0·0102 QALYs; 3·73 quality-adjusted days), but cost differences were substantial (£448). Base-case results showed that potent and very potent corticosteroids were the most cost-effective initial strategies, and third-line TCF product might produce additional health gains at acceptable extra cost (Table 5 and Table S4; see Supporting Information). Results for other topical therapies were consistent with findings in the trunk/limbs analysis, namely that vitamin D monotherapy was dominated by corticosteroids. Coal-tar shampoo alone performed poorly, ranking only slightly above vehicle gel alone. No comparison with TCA (am/pm) could be made due to a lack of data in scalp psoriasis.

On average, the difference between once- and twice-daily vitamin D was very small (£2 more and 0·00003 fewer QALYs). The difference was slightly larger, although still modest, for potent corticosteroids (£23 more and 0·0006 fewer QALYs). Use of once-daily very potent corticosteroids was also more costly and less effective than twice-daily use (£83 more and 0·0015 fewer QALYs), but was less costly and more effective than TCF product (£165 less and 0·0009 more QALYs). A cost-effectiveness acceptability curve for first-line topical therapies for scalp psoriasis is given in Figure S2 (see Supporting Information).

Sensitivity analyses

In addition to concerns about the safety of continuous use of potent corticosteroids, there is concern that very potent corticosteroids are too aggressive. In a scenario analysis, sequences with consecutive use of steroidal products were removed, as were sequences where very potent corticosteroids were used as the first-line treatment or applied twice daily. Additionally, given patient preference for once-daily applications, twice-daily strategies across all agents were removed as possible initial treatments. The results of this scenario under base-case parameter assumptions indicate that a sequence of once-daily potent corticosteroids, twice-daily vitamin D and then once-daily very potent corticosteroids is expected to be optimal at the NICE threshold and across a range of higher WTP thresholds (Fig. 4). TCF product instead of potent corticosteroids in this sequence would be cost-effective only at WTP thresholds over £220 000.

Figure 4.

Cost-effectiveness acceptability curve for first-line topical therapies for scalp psoriasis under the restricted-comparator scenario. PS, potent corticosteroid; TCF, two-compound formulation product of potent corticosteroid and vitamin D analogue; Vit D, vitamin D or vitamin D analogue; CT, coal tar; OD, once daily; QALY, quality-adjusted life year.

Results of one-way sensitivity analyses (Table 7 and Table S5; see Supporting Information) suggest that the conclusions of the restricted-comparator scenario are insensitive to changes in key parameters of efficacy, adherence, cost and time horizon. Even in the more favourable scenarios, TCF product is only likely to be considered cost-effective compared with potent corticosteroids at WTP thresholds in excess of £70 000. A threshold analysis showed that only given a 60–70% discount would TCF product represent a more cost-effective first-line treatment than potent corticosteroids alone.

Table 7. Results of sensitivity analyses within the framework of restricted comparators – scalp
Sensitivity analysisCost per QALY gainedOptimal strategy (threshold £20 000 per QALY)
TCF OD – VitD BD – VPS OD vs. PS OD – VitD BD – VPS OD
  1. QALY, quality-adjusted life year; TCF, two-compound formulation product containing potent corticosteroid and vitamin D analogue; VitD, vitamin D or vitamin D analogue; VPS, very potent corticosteroid; PS, potent corticosteroid; OD, once daily; BD, twice daily. Results for all comparators are presented in Table S5 (see Supporting Information).

Base case – restricted comparators£219 846PS OD – VitD BD – VPS OD
Reduced adherence£221 692PS OD – VitD BD – VPS OD
Four-week treatment£97 597PS OD – VitD BD – VPS OD
Nonresponder utility gain = 0£210 429PS OD – VitD BD – VPS OD
Reduced-quantity TCF£74 545PS OD – VitD BD – VPS OD
Unit cost of PS£143 351PS OD – VitD BD – VPS OD
30% discount on TCF product£150 968PS OD – VitD BD – VPS OD
Time horizon (3 years)£101 203PS OD – VitD BD – VPS OD
Time horizon (10 years)£94 445PS OD – VitD BD – VPS OD

Discussion

We performed a cost-effectiveness analysis of potential treatment sequences for psoriasis of the trunk, limbs and scalp. This model was adapted from published analyses evaluating a more limited range of comparators.[10, 11] Our results indicate that the various topical therapies produce similar health benefits, but they differ substantially in terms of cost.

Based on the costs and benefits of 118 compared sequences for the treatment of trunk and limb psoriasis, the results showed potent corticosteroids to be highly cost-effective, although there is some uncertainty regarding what frequency of application offers the best value for money. Twice-daily application was more cost-effective than once-daily use in the base case, but this was reversed when patient-assessed response outcomes were modelled (Table S2; see Supporting Information). The model also does not capture potential AEs of potent steroids (e.g. skin atrophy, rapid relapse) due to a lack of data. Therefore, even if application twice daily is more effective at inducing a response than once-daily use, the risks of a higher dose are likely to outweigh the potential benefits.

Separate application of vitamin D and potent corticosteroids [TCA (am/pm)] represents the most cost-effective first-line option for trunk and limb psoriasis if twice-daily corticosteroids are deemed too aggressive. Once-daily TCF product was more effective than TCA (am/pm); however, the slight additional benefit (equivalent to 0·40 quality-adjusted days) would come at a cost of nearly £200 more per patient, a difference explained almost entirely by its high cost relative to separate vitamin D and corticosteroid products (two to four times more costly). We explicitly considered whether the practicality of a once-daily combined product might make it cost-effective given the problems that many patients have adhering to twice-daily regimens. A sensitivity analysis wherein 40% of patients prescribed a twice-daily regimen were assumed to apply their topical only once daily showed that the benefits of a once-daily TCF product were still too small to justify the extra cost. This cost, multiplied across the numerous patients using topical therapies to manage their psoriasis, would represent a huge NHS expenditure, with insufficient evidence that it would dramatically improve outcomes. The most efficient placement of TCF product would therefore be after both potent corticosteroids (once and twice daily) and TCA (am/pm) have failed to bring about the desired response.

After comparing 169 sequences of scalp treatments, results suggest that initial treatment with once-daily very potent corticosteroids is likely to offer the best value for NHS resource; however, there is concern that very potent corticosteroids are an aggressive strategy and carry greater risk of steroid-related AEs, which the model did not capture. The second-most cost-effective first-line treatment was once-daily potent corticosteroids. TCF product was dominated by very potent corticosteroids and was not cost-effective compared with potent corticosteroids or vitamin D analogues; the very modest additional benefits compared with these agents (equivalent to 0·24–0·75 quality-adjusted days) might be considered cost-effective only if WTP thresholds were substantially higher than NICE's £20 000. These conclusions were robust to changes in key assumptions including adherence, efficacy, cost and time horizon.

Across both populations, vitamin D is optimal when looking to avoid continuous use of corticosteroids. In scalp psoriasis particularly, the additional benefits of twice-daily compared with once-daily vitamin D are negligible; therefore, other considerations such as convenience and cosmetic acceptability should have greater influence on treatment decisions. Indeed, such considerations are important across all agents, as the added benefits of a second daily application quickly wane if adherence is adversely affected.

Short-contact dithranol, coal tar and coal-tar shampoo performed less well compared with alternatives. For example, the majority of patients offered only coal-tar shampoo failed treatment and were referred onward, making the true costs of coal-tar shampoo much higher than its acquisition cost alone. It should be reserved for less severe scalp psoriasis or combined with other more effective topical agents.

Although these analyses provide the most complete evaluation of licensed topical therapies to date, the results should be considered in light of several limitations related to gaps in the clinical and economic evidence base. Firstly, costs and consequences of treatment-related AEs were not captured. Longer-term, good-quality data on the safety of topical therapies is lacking, but it is generally accepted that long-term use of corticosteroids carries risks of local AEs and, in those with extensive disease, potential systemic AEs. Had these AEs been included, the expected benefits of corticosteroids, particularly in higher doses, may have been outweighed by their risks. Secondly, the model focused on the induction of disease clearance, not on the maintenance of clearance. Maintenance trials were limited[22-26] and, where available, inadequately reported for use in the economic model. The model takes a relatively short time horizon considering that psoriasis is a chronic, long-term condition for which patients may take up treatment intermittently for many years. Longer time horizons of up to 10 years were explored in sensitivity analyses and conclusions were insensitive to these extensions. Frequency and severity of relapse, selection for and speed of onward referral, adherence, methods of self-management and long-term safety are all inadequately addressed by the published literature, and future clinical studies should be designed to address these gaps to improve the value, validity and interpretability of health economic conclusions. Finally, evidence assessing the impact of localized psoriasis on quality of life is limited and quite uncertain. Future studies measuring patients' quality of life using preference-based tools or mapping algorithms from disease-specific measures will help to quantify the disease's impact and the gains associated with effective treatment.

In conclusion, this study demonstrates the likely cost-effectiveness of corticosteroids for patients with chronic plaque psoriasis. Potent corticosteroids, used alone or in combination with vitamin D for patients with trunk or limbs psoriasis, are likely to offer the best value for money, while potent or very potent corticosteroids are likely to be best for patients with scalp psoriasis.

Acknowledgments

This work was undertaken by the National Clinical Guideline Centre, which received funding from the National Institute for Health and Clinical Excellence. The views expressed in this publication are those of the authors and not necessarily those of the Institute. We gratefully acknowledge the advice and input of the following members of the Psoriasis Guideline Development Group and experts who advised the group regarding this evaluation: David Chandler, Paul Hepple, Karina Jackson, Ruth Murphy, Jillian Peters, Natasha Smeaton, Claire Strudwicke, Roderick Tucker, Richard Warren, Jill Cobb, Nancy Pursey, Amelia Ch'ng and Jill Parnham.

Ancillary