Cost-effectiveness analysis of using innovative therapies for the management of moderate-to-severe ulcerative colitis in Spain

ORIGINAL RESEARCH Cos -e ec i eness analysis o using inno a i e he apies o he managemen o mode a e- o-se e e ulce a i e coli is in Spain Ca los Taxone a a , Fe nando de And és-Nogales b , San iago Ga cía-López c , Amelia Sánchez-Gue e o d , Belén Menchén d , Ca men Pe al e , Ana Cábez e , Susana Gómez e , Alejand a López-Ibáñez de Aldecoa e , Miguel Ángel Casado b and Luis Menchén ,g a Depa men o Gas oen e ology. Hospi al Clínico Uni e si a io San Ca los, And Ins i u o De In es igación Del Hospi al Clínico San Ca los [Idissc], Mad id, Spain; b Heal h Economics, Pha macoeconomics & Ou comes Resea ch Ibe ia (PORIB), Mad id, Spain; c Depa men o Gas oen e ology. Hospi al Uni e si a io Miguel Se e , Za agoza, Spain; d Hospi al Pha macy Depa men Hospi al Uni e si a io Pue a De Hie o-Majadahonda, Mad id, Spain; e P ize S.L.U, Alcobendas, Mad id, Spain; Depa men o Gas oen e ology. Hospi al Gene al Uni e si a io G ego io Ma añón, And Ins i u o De In es igación Sani a ia G ego io Ma añón, Mad id, Spain; g Depa men o Medicine, Uni e sidad Complu ense De Mad id, Mad id, Spain ABSTRACT Backg ound: To e alua e he cos -e ec i eness o o aci inib in compa ison o edolizumab o he ea men o mode a e- o-se e e ulce a i e coli is (UC) a e ailu e o in ole ance o con en ional he apy (bio-nai e) o i s -line biologic ea men (bio-expe ienced), om he Spanish Na ional Heal h Sys em (NHS) pe spec i e. Me hods: A li e ime Ma ko model wi h eigh -week cycles was de eloped including i e heal h s a es: emission, esponse, ac i e UC, emission a e su ge y, and dea h. Response and emission p obabili ies ( o induc ion and main enance pe iods) we e ob ained om a mul inomial ne wo k me a-analysis. D ug acquisi ion – biosimila p ices included – (ex- ac o y p ice wi h manda o y deduc ions), adminis- a ion, su ge y, pa ien managemen , and ad e se e en managemen cos s (€, yea 2019) we e conside ed. A 3% discoun a e (cos /ou comes) was applied. P obabilis ic and de e minis ic sensi i i y analyses (PSA) we e conduc ed. Resul s: To aci inib was dominan e sus edolizumab (bo h in bio-nai e and bio-expe ienced pa ien s) en ailing o al cos sa ings o €23,816 (bio-naï e) and €11,438 (bio-expe ienced). Di e ences in quali y- adjus ed li e-yea (QALY) we e smalle han 0.1 o bo h popula ions. PSA esul s showed ha o aci inib has a high p obabili y o being cos -e ec i e (bio-naï e: 82.5%; bio-expe ienced: 90.6%) e sus edolizumab. Conclusions: F om he Spanish NHS pe spec i e, o aci inib could be a dominan ea men (less cos ly and mo e e ec i e) in compa ison o edolizumab, wi h ele an cos sa ings and simila QALY gains. ARTICLE HISTORY Recei ed 6 Augus 2020 Accep ed 20 Janua y 2021 KEYWORDS JAK inhibi o ; biologic ea men ; cos -e ec i eness ; an i-TNF; Spain; o aci inib; ulce a i e coli is; edolizumab 1. In oduc ion Ulce a i e coli is (UC) is a ch onic in lamma o y bowel disease wi h impai men limi ed o he colon and ec um. The usual symp oms a e dia hea, ec al bleeding, u gency, enesmus, and c ampy abdominal pain [1]. In mos pa ien s, UC ollows a cou se wi h al e na ing pe iods o emission and elapses [2]. The global incidence and p e alence o UC ha e been inc eas- ing o e ime. I is mo e common in no he n and wes e n Eu opean coun ies as well as in indus ialized coun ies such as he Uni ed S a es, Canada, and Aus alia [2,3]. In Spain, i s p e alence is a ound 88.7 cases pe 100,000 popula ion, and i s annual incidence is 5.7–8 cases pe 100,000 popula ion [4]. UC is associa ed wi h a high economic bu den, especially in se e e cases whe e hospi aliza ions and su ge ies accoun o he majo i y o di ec cos s [5]. Su ge y is no he cu e, wi h a signi ican p opo ion o pa ien s needing addi ional su - ge ies o hospi aliza ions due o pos -su gical complica ions such as ch onic pouchi is [6,7]. In he las yea s and due o he in oduc ion and widesp ead use o expensi e biologic he a- pies, he e has been a shi away om hospi aliza ions and owa d pha maceu icals as he p edominan d i e o di ec heal h-ca e cos s in IBD pa ien s [8,9]. The impac o UC on pa ien quali y o li e and wo k p oduc i i y gene a es signi i- can indi ec cos s, which may e en exceed di ec heal h-ca e cos s [5]. The main objec i e o managemen in pa ien s wi h UC is o induce emission and main ain i o e a long- e m pe - iod, wi h no need o co icos e oids, in o de o p e en disabili y and su gical p ocedu es, limi he incidence o colo ec al cance , and imp o e quali y o li e [2]. Va ious pha macological ea men s a e a ailable o mode a e- o- se e e UC: con en ional ea men s (co icos e oids, hio- pu ines and calcineu in inhibi o s), con en ional biologic ea men s ( umo nec osis ac o [TNF] inhibi o s), new bio- logic ea men s (an i-in eg ins [ edolizumab], in e leukin 12/23 inhibi o s [us ekinumab]) and Janus kinase inhibi o s ( o aci inib) [10]. Howe e , up o 15% o pa ien s will CONTACT Fe nando de And és-Nogales [email p o ec ed] Pha macoeconomics & Ou comes Resea ch Ibe ia (PORIB) Paseo Joaquín Rod igo, 4 Le a I, 28224 Pozuelo De Ala cón, Mad id, Spain EXPERT REVIEW OF PHARMACOECONOMICS & OUTCOMES RESEARCH h ps://doi.o g/10.1080/14737167.2021.1880324 © 2021 The Au ho (s). Published by In o ma UK Limi ed, ading as Taylo & F ancis G oup. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion-NonComme cial-NoDe i a i es License (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/), which pe mi s non-comme cial e-use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed, and is no al e ed, ans o med, o buil upon in any way. ul ima ely equi e su gical ea men (colec omy) due o ea men ailu e o he de elopmen o colo ec al dysplasia o ca cinoma [3]. Gi en he a ailabili y o a ious ea men op ions wi h di e en mechanisms o ac ion, i is essen ial o choose he bes ea men op ion o a pa ien based on his o he indi- idual needs and cha ac e is ics. To do his, con aindica ions o use, pa ien s’ ea men in ole ances, he a e o p ima y non- esponse, and he po en ial o seconda y loss o esponse o e ime mus be conside ed. Seconda y loss o esponse o biologics can be associa ed wi h immunogenici y, which is especially ele an o TNF an agonis s [11,12]. Consis en wi h se e al compa a i e ne wo k me a- analyses (NMAs) [13,14], o aci inib, edolizumab, and all o he biologic ea men s e alua ed a e conside ed sa e and e ec i e he apies ecommended o he ea men o mode - a e- o-se e e ac i e UC, bo h in induc ion and in main enance and bo h in a bio-nai e popula ion and ollowing biologic ailu e. Speci ically, a ecen s udy by Singh e al. [14] classi ied in liximab as he bes d ug o inducing emission and muco- sal healing in bio-nai e pa ien s. I also ound o aci inib and us ekinumab o be he bes ea men s o inducing emission and mucosal healing in pa ien s p e iously exposed o TNF inhibi o s and possibly e en mo e e ec i e han edolizumab and adalimumab in hese pa ien s [14]. Gi en he b oad ange o ea men s o UC, compa a i e economic e alua ion and posi ioning o inno a i e d ugs such as o aci inib and edolizumab a e cu en ly o special in e es o paye s and o p o ide mo e e idence o suppo decision- making. I is impo an o cla i y ha us ekinumab, ano he d ug wi h an inno a i e mechanism o ac ion ecen ly app o ed by he Eu opean Medicines Agency (EMA) o he indica ion o mode a e- o-se e e ulce a i e coli is, was excluded om he s udy because, a he ime o he analysis, his d ug lacked a p ice and unding o his indica ion in Spain. The objec i e o his s udy was o e alua e, om he pe - spec i e o a Eu opean Public Heal h Sys em (Spain), he e icacy o he use o o aci inib e sus edolizumab o he ea men o pa ien s wi h mode a e- o-se e e ac i e UC ollowing ailu e, loss o esponse, o in ole ance o con en- ional ea men (bio-nai e) o o an an i-TNF biologic ea - men (bio-expe ienced). 2. Pa ien s and me hods 2.1. Popula ion This cos -e ec i eness analysis conside ed pa ien s wi h mod- e a e- o-se e e ac i e UC who did no exhibi a sui able esponse, expe ienced loss o esponse, o had an in ole ance o con en ional ea men , con en ional biologic ea men , o new biologic ea men s. Two subpopula ions we e iden i ied based on p io ea men ecei ed: pa ien s nai e o biologics (bio-naï e) and pa ien s ha ing al eady been ea ed wi h an i- TNF biologics (bio-expe ienced). Thei cha ac e is ics we e based on he pa ien s en olled in he OCTAVE Induc ion 1 and 2 s udies [15] (mean age 41.2 yea s; 59.2% male). A pa ien weigh o 71.93 kg was de i ed om he weigh ed mean weigh o indi iduals o e 40 yea s o age in Spain [16]. 2.2. Economic model A de no o Ma ko model was de eloped in Mic oso Excel o show he cou se o UC o e ime acco ding o he ea men ecei ed (Figu e 1). I was designed in line wi h p e iously published analyses on his disease [17,18] and ollowing he cu en ecommenda ions [19,20]. An expe panel composed o gas oen e ologis s and hospi al pha macis s alida ed he pa ame e s o he model and con ibu ed in o ma ion ega d- ing ou ine clinical p ac ice. T ea men consis ed o wo phases: induc ion and main e- nance. In he induc ion phase, he ea men objec i e is o induce esponse o emission o ac i e UC. In he main enance phase, he goal is o main ain ha esponse o emission. Fou heal h s a es in addi ion o dea h we e conside ed: emission (Mayo sco e = 0–2 and all subsco es ≤1), esponse ( educ ion in he Mayo Clinic sco e o a leas 3 poin s and a dec ease o a leas 30% om he baseline sco e, wi h a dec ease o a leas 1 poin on he ec al bleeding subsco e o an absolu e ec al bleeding sco e o 0 o 1), mode a e- o-se e e ac i e UC (Mayo sco e ≥6), and emission ollowing su ge y. An ini ial coho o 1,000 pa ien s wi h mode a e- o-se e e ac i e UC ansi ed be ween heal h s a es in 8-week cycles. Once hey ecei ed induc ion ea men , hey could exhibi clinical emission o clinical esponse o main ain ac i e UC. The induc ion phase las ed 8 weeks, he same as ha o he OCTAVE s udies [15] and simila o ha seen in o he UC clinical ials (6–10 weeks) [14,21]. The ea men al e na i es conside ed in he analysis we e o aci inib and edolizumab. When pa ien s did no achie e esponse du ing induc ion o expe ienced loss o esponse du ing main enance, hey dis- con inued he ea men . Bio-nai e pa ien s could ecei e a second line o biologic ea men ollowing discon inua ion o he ini ial ea men . The subsequen ea men ollowing esponse ailu e was in liximab (Table 1). Pa ien s wi h ac i e UC in he main enance phase emained in ha heal h s a e un il hey died o unde wen su ge y. Fo he pu poses o he analysis, al hough he e a e o he indica ions o su ge y, only A icle highligh s ●Ulce a i e coli is is a ecu en ch onic disease wi h a g owing inci- dence, especially in indus ialized coun ies. ●Since a signi ican pe cen age o pa ien s do no espond, lose he esponse o a e in ole an o he a ailable ea men s, he e is a need o new ea men s, especially o hose who explo e new mechanisms o ac ion. ●A li e ime Ma ko model was de eloped o e alua e he e iciency o he use o o aci inib e sus edolizumab in pa ien s wi h mode a e- o- se e e ac i e ulce a i e coli is ollowing ailu e, loss o esponse o in ole ance o con en ional ea men (bio-naï e) o o a TNF-inhibi o biologic ea men (bio-expe ienced), om he pe spec i e o he Spanish Na ional Heal h Sys em. ●Signi ican sa ings in cos s pe pa ien wi h o aci inib e sus edoli- zumab we e ound in bo h pa ien popula ions — €23,816 (bio-naï e) and €11,438 (bio-expe ienced) — wi h simila gains in quali y-adjus ed li e yea s (<0.05). ●T ea men wi h o aci inib compa ed o edolizumab is a dominan (less cos ly and mo e e ec i e) al e na i e om he pe spec i e o he Spanish Na ional Heal h Sys em. 2C. TAXONERA ET AL. pa ien s wi h ac i e UC could be candida es o su gical ea - men , wi h he possibili y o de eloping complica ions and achie ing emission ollowing su ge y (Figu e 1). This analysis was pe o med om he pe spec i e o he Spanish Na ional Heal h Sys em (NHS). The ime ho izon con- side ed was he pa ien ’s li e ime, wi h a maximum o 60 yea s. A 3% annual discoun was applied bo h o cos s gene a ed and o heal h ou comes. T ea men e icacy was e alua ed using he inc emen al cos –e ec i eness a io (ICER) and he inc emen al cos –u ili y a io (ICUR) and exp essed in e ms o cos pe li e yea (LY) gained and cos pe quali y-adjus ed li e yea (QALY), espec i ely. These a ios Figu e 1. Diag am o Ma ko model and ea men sequences. EXPERT REVIEW OF PHARMACOECONOMICS & OUTCOMES RESEARCH 3 we e es ima ed using he di e ences in e ms o cos s and heal h ou comes be ween he al e na i es s udied. A willingness- o-pay h eshold was se a €25,000/QALY o e alua e whe he one al e na i e is cos -e ec i e compa ed o ano he . This h eshold is no an o icial one in Spain hough he la es publica ions ha e sugges ed anges o €20,000-€25,000/QALY [22] and, mos ecen ly, €25,000- €60,000/QALY [23]. 2.3. E icacy Wi hou di ec compa isons be ween ea men s, he ela i e e icacy o o aci inib e sus edolizumab was ob ained om a ixed-e ec s mul inomial NMA, using placebo as a compa a o o e e ence [24]. P obabili ies o ansi ion be ween he heal h s a es o esponse and clinical emission we e calcula ed o he induc ion phase ( esponse and emis- sion a 8 ± 2 weeks) and he main enance phase ( esponse and emission a 52 ± 4 weeks) [24] (Table 1). The la e phase had wo independen analyses depending on he design o he s udies included: ea - h ough, whe e he induc ion- phase ea men was main ained, and e- andomized, in which some pa ien s we e andomized ollowing he induc- ion phase (Table 1). Gi en ha , a he ime o he s udy, he e we e no da a o he ea - h ough analysis wi h edolizumab, and he ela i e isk be ween edolizumab and placebo a ailable in he e- andomized analysis was applied o each popula ion (bio-nai e, bio-expe ienced). Fo a be e adjus men o i s -yea and long- e m e icacy ou comes, p obabili ies o all ea men s we e adjus ed o wo pe iods: weeks 0–24 and weeks 24–52. This adjus men was based on a pos -hoc analysis o he OCTAVE Sus ain s udy, in which di e ences be ween he wo pe iods in e ms o emission and esponse we e obse ed. In addi ion, he second pe iod seemed o show a cons an end o e ime and ha e ec was be e sui ed o long- e m ex apola- ion han ha ob ained o e he en i e pe iod (weeks 0–52) (Table 1). 2.4. Ad e se e en s, isk o su ge y, and mo ali y Ra es o ea men - ela ed se ious ad e se e en s we e included based on induc ion and main enance s udies o he medicines conside ed [15,21,24–27], wi h no dis inc ion be ween bio-nai e and bio-expe ienced pa ien s (Table 2). An annual a e o su ge y in pa ien s wi h ac i e UC o 1.44% was used in line wi h an ex ensi e epidemiology s udy conduc ed in Spain [28]. Pa ien mo ali y was de e - mined based on mo ali y a es in he gene al Spanish popu- la ion by age and gende [16]; in addi ion, a pe iope a i e mo ali y a e o 1.18% was used in pa ien s ha ing unde - gone colec omy [29]. Table 1. P obabili ies o ansi ion o esponse and emission (induc ion and main enance phases). E icacy o ea men al e na i es P obabili ies o ansi ion (ne wo k me a-analysis – weeks 0–52) Bio-nai e Bio-expe ienced T ea men % SD (%) % SD (%) Induc ion phase F om ‘No Response’ o ‘Response’ To aci inib 57.18 4.53 59.57 4.06 Vedolizumab 66.54 6.34 42.67 7.06 In liximab 72.36 3.02 50.18 8.35 F om ‘No Response’ o ‘Clinical Remission’ To aci inib 22.04 3.49 17.90 2.92 Vedolizumab 29.99 6.10 8.90 3.13 In liximab 36.03 3.39 12.63 4.30 Main enance phase F om ‘Response’ o ‘Response’ To aci inib 58.53 10.21 30.51 10.10 Vedolizumab 46.79 a 5.09 36.81 a 7.92 In liximab 45.72 6.28 31.87 6.86 F om ‘Response’ o ‘Clinical Remission’ To aci inib 46.67 10.45 17.34 7.88 Vedolizumab 40.25 a 5.40 21.07 a 6.01 In liximab 34.22 5.87 18.07 5.16 P obabili ies o ansi ion o o aci inib 5 mg (OCTAVE Sus ain) o adjus men o weeks 0–24 and weeks 24–52 Bio-nai e Bio-expe ienced Ini ial s a e/ inal s a e Response (%) a Remission (%) Response (%) a Remission (%) Main enance weeks 0–24 Remission 31 70 27 64 Response a 29 9 30 14 No esponse 40 20 43 23 Main enance weeks 24–52 Remission 53 80 25 52 Response a 12 7 20 31 No esponse 35 13 55 17 a Response wi h no emission. SD: s anda d de ia ion. a) Assump ion based on ela i e isk in e ms o e icacy be ween edolizumab and placebo in he e- andomized mul inomial analysis. 4C. TAXONERA ET AL. Table 2. Uni cos s (adminis a ion, pa ien managemen , su ge y, and ad e se e en managemen ). Uni cos s Pha macological cos s (p ice pe pack) Ac i e subs ance S eng h and pack size Induc ion phase Main enance phase P ice (EFP-deduc ion RDL 8/2010 a ) To aci inib (o al) 5 mg 56 coa ed able s 10 mg wice daily o 8 weeks 5 mg wice daily €762.20 Vedolizumab (IV) 300 mg 1 ial o powde o concen a e o in usion 300 mg [weeks 0, 2, 6] 300 mg e e y 8 weeks €3,206.05 In liximab (IV) 100 mg 1 ial o powde 5 mg/kg [weeks 0, 2, 6] 5 mg/kg e e y 8 weeks €402.21 Adminis a ion cos s (cos pe adminis a ion) Resou ce I em Uni cos (€, 2019) In a enous adminis a ion Cos o adminis a ion pe nu sing s a membe , pe pa ien In usion ime >2 h €262.62 In usion ime 0.5–2 h €160.49 Pa ien managemen cos s by heal h s a e (cos pe cycle) Resou ce I em Uni cos (€, 2019) Remission Mayo sco e = 0–2 (pa ien s being ea ed wi h o aci inib o biologics) €199.53 Response Dec ease om baseline Mayo sco e by ≥3 poin s and a leas 30%, wi h ec al bleeding subsco e dec ease ≥1 poin o a alue o 0–1 €426.08 Ac i e UC Mayo sco e ≥6€1,149.84 Su ge y cos s (cos pe cycle) Resou ce I em Uni cos (€, 2019) Su ge y (cos o p ocess) Cos s o colec omy- ela ed su ge y and hospi aliza ion (comple e p ocess) €26,918.56 O he heal h-ca e esou ces (yea s 0–2) Cos o o he heal h-ca e esou ces gene a ed by pa ien managemen : hospi aliza ion wi h no su ge y, diagnos ic and labo a o y es s, isi s, and d ugs (pe iod 0–2 yea s om colec omy) €426.90 O he heal h-ca e esou ces (>2 yea s) Cos o o he heal h-ca e esou ces gene a ed by pa ien managemen : hospi aliza ion wi h no su ge y, diagnos ic and labo a o y es s, isi s, and d ugs (pe iod >2 yea s om colec omy) €194.38 Ad e se e en managemen cos s E en To aci inib [15] (% o pa ien s) In liximab [25–27] (% o pa ien s) Vedolizumab [21] (% o pa ien s) Uni cos (€, 2019) Se ious in ec ions a 0.80% 2.49% 1.34% €5,293.57 Uppe espi a o y ac in ec ions b 3.80% 16.32% 8.39% €3,737.70 Malignancies c 0.20% 0.00% b 0.12% €9,842.51 Acu e in usion ad e se e en d 0.00% b 8.29% 3.51% €3,462.45 He pes zos e e 1.30% 0.00% b 0.00% b €4,450.39 a To aci inib and edolizumab ha e a manda o y deduc ion o 7.5%. b Da a no epo ed on clinical ials. IV: in a enous; EFP: ex- ac o y p ice; RDL: Spanish Royal Dec ee-Law; SC: subcu aneous; UC: ulce a i e coli is. ICD-9-CM codes: (a) 041.9 Bac e ial in ec ion, unspeci ied, in condi ions classi ied elsewhe e and o unspeci ied si e; (b) 465.9. Acu e uppe espi a o y in ec ions o unspeci ied si e; (c) 202.8. 202.80. O he malignan lymphomas, unspeci ied si e, ex anodal and solid o gan si es; (d) 995.3 Alle gy, unspeci ied, no elsewhe e classi ied, 999.88 O he in usion eac ion, 995.27 O he d ug alle gy; (e) 053. He pes zos e . EXPERT REVIEW OF PHARMACOECONOMICS & OUTCOMES RESEARCH 5 2.5. Quali y o li e U ili y alues d awn om he li e a u e we e inco po a ed o e lec quali y o li e o pa ien s wi h UC acco ding o heal h s a e [30,31]: emission (0.87), esponse (0.76), ac i e UC (0.41), and pos -su ge y emission (0.68). 2.6. Cos s In acco dance wi h he analysis’s pe spec i e ( ha o he NHS), di ec heal h-ca e cos s we e included (Table 2): pha macolo- gical cos s, adminis a ion cos s, disease managemen cos s by heal h s a e, su ge y cos s, and se ious ad e se e en manage- men cos s. Uni cos s we e ob ained om Spanish na ional heal hca e cos da abases [32,33]. Pha macological cos s we e calcula ed based on he dosage egimens o he ea men al e na i es desc ibed in he summa ies o p oduc cha ac e - is ics [34] and p ices we e exp essed in e ms o ex- ac o y p ice (EFP) [32], wi h applica ion o he manda o y deduc ions acco ding o Royal Dec ee Law 8/2010 [35], co esponding o a 7.5% discoun on he p ice o o aci inib and edolizumab (in liximab does no ha e a manda o y deduc ion as i is included in he e e ence p ice sys em). The p ice o he biosimila medicine was used, i a ailable. In d ugs adminis- e ed in a enously, he p ice pe millig am was conside ed, wi h no ial was age allowed (Table 2). In a enous adminis- a ion cos s we e calcula ed based on he cos o nu sing s a and he in usion ime o he d ugs (30–90 minu es o edo- lizumab and 150 minu es o in liximab, acco ding o he expe panel). O al adminis a ion was assigned no cos . Pa ien managemen cos by heal h s a e was de e mined based on heal h-ca e esou ces consumed, which o hei pa we e de e mined by he expe panel acco ding o clinical p ac ice. Cos s o su ge y and i s complica ions, including pa ien managemen ollowing colec omy, we e included in agg ega e acco ding o da a om a eal clinical p ac ice s udy in Spain [36]. Ad e se e en cos s we e ob ained om he cos pe p ocess in he eco d o discha ges om he da abase o he Spanish Minis y o Heal h acco ding o he ICD 9-CM code bes ma ching each e en [33,37]. 2.7. Sensi i i y analyses De e minis ic sensi i i y analyses (DSAs) and p obabilis ic sen- si i i y analyses (PSAs) we e pe o med o e alua e he unce - ain y o he alues o he pa ame e s and he obus ness o he model. Some scena io analyses we e pe o med modi ying alues o he ime ho izon (5, 10, and 20 yea s), discoun a e (0% and 5%), annual su ge y isk (al e na i e alue: 4.2% [38]), and pe iope a i e mo ali y ( ange 0–2.9% [29]). Uni a ia e and mul i a ia e DSAs modi ied he ollowing pa ame e s by ±20%: uni p ice o o aci inib, uni p ice o edolizumab, cos s o adminis a ion, managemen by heal h s a e, su ge y, and managemen o ad e se e en s as well as u ili y alues o heal h s a es. In he PSAs, he pa ame e s (u ili y alues, an- si ion p obabili ies, isk o su ge y, pe iope a i e mo ali y, pha macological cos s, adminis a ion cos s, pa ien manage- men cos s, and ad e se e en cos s) we e simul aneously a ied by pe o ming 1,000 Mon e Ca lo simula ions on he 1,000-pa ien coho e alua ed. The pa ame ic unc ions used we e be a o u ili ies, su ge y isk and mo ali y isk due o colec omy, gamma o cos s, and Di ichle o dis ibu ion p obabili ies. 3. Resul s The o al pa ien managemen cos s o pa ien s wi h mode - a e- o-se e e ac i e UC a e shown in Table 3. T ea men wi h o aci inib in pa ien s wi h mode a e- o-se e e ac i e UC yielded a educ ion in o al cos s compa ed o edolizumab. In bio-nai e pa ien s, in he ea men sequence ha s a ed wi h o aci inib, his educ ion was €23,815.58 e sus edoli- zumab, whe eas in he bio-expe ienced pa ien popula ion, he educ ion in o al cos s wi h o aci inib was €11,437.56. In bo h subpopula ions, he di e ences in cos s we e mainly due o he ac ha he pha macological and adminis a ion cos s o edolizumab a e subs an ially highe han o o aci- inib, whe eas o all o he dimensions s udied, he esul s o he wo d ugs we e simila hough hey we e sligh ly lowe wi h o aci inib in all cases. Rega ding heal h ou comes, simila gains in e ms o LYs gained we e obse ed due o he disease’s limi ed impac on pa ien su i al, wi h di e ences be ween he ea men op ions o less han 0.0005 LYs gained. In e ms o QALYs, he di e ences be ween o aci inib and edolizumab we e somewha highe in he bio-expe ienced popula ion (0.042 QALYs) (Table 3). In bo h he bio-nai e and bio-expe ienced pa ien popula ions, ea men wi h o aci inib was a dominan ea men al e na i e (lowe cos s and highe e ec i eness) compa ed o edolizumab (Table 3). The esul s o he DSAs and scena io analyses did no show signi ican a ia ions ela i e o he base case, especially in he bio-expe ienced popula ion, in which o aci inib always emained a dominan al e na i e (Figu e 2, Table 3). In bio- nai e pa ien s, o aci inib emained a dominan al e na i e, excep when conside ing a sho e ime ho izon, a highe discoun a e, o a highe su ge y isk; hese cases showed sligh di e ences in QALYs while main aining he cos sa ings associa ed wi h o aci inib (Table 3). O he a iables wi h g ea e impac on he esul s we e he p ice o edolizumab and o aci inib and u ili y alues. Howe e , in hese cases, o aci inib was s ill a dominan al e na i e (Figu e 2). The esul s o he PSA showed ha he p obabili y o o aci inib being a cos -e ec i e al e na i e o edolizumab o a willingness- o-pay h eshold o €25,000/QALY was 82.5% in a bio-nai e popula ion and 90.6% in a bio-expe ienced popula ion (Figu e 3). The a e age cos s and QALYs ob ained in he PSA we e -€10,997.79 and 0.062 QALYs o he bio-nai e popula ion and -€8,687.92 and 0.038 QALYs o he bio- expe ienced popula ion, being o aci inib a dominan al e na- i e in bo h scena ios (Figu e 3). 4. Discussion This analysis showed ha o aci inib is a dominan – i.e. less cos ly and mo e e ec i e – ea men al e na i e compa ed o edolizumab in ea ing pa ien s wi h mod- e a e- o-se e e ac i e UC, ega dless o whe he i is used 6C. TAXONERA ET AL. Table 3. Resul s o cos -e ec i eness analysis (base case and scena io analyses). Resul s o cos –e ec i eness/cos –u ili y analysis Bio-nai e Bio-expe ienced Base case analysis To aci inib (+ in liximab) Vedolizumab (+ in liximab) Inc emen al To aci inib Vedolizumab Inc emen al To al cos s €181,584.46 €205,400.03 -€23,815.58 €160,907.67 €172,345.23 -€11,437.56 Pha macological cos €35,635.81 €57,250.42 -€21,614.61 €8,351.09 €18,123.27 -€9,772.18 Adminis a ion cos €3,374.21 €5,312.54 -€1,938.33 €0.00 €907.22 -€907.22 Disease managemen cos €129,323.13 €129,326.92 -€3.79 €138,701.66 €139,105.33 -€403.67 Su ge y cos s €12,000.05 €12,001.40 -€1.35 €13,593.00 €13,691.53 -€98.53 AE managemen cos €1,251.26 €1,508.76 -€257.49 €261.92 €517.87 -€255.95 To al LYs gained 23.41 23.41 −0.0000046 23.41 23.41 0.00048 To al QALYs 11.94 11.94 0.00014 11.06 11.02 0.042 ICER (€/LY gained) Lowe cos and lowe e ec i eness Dominan ICUR (€/QALY) Dominan Dominan Scena io analysis Bio-nai e Bio-expe ienced Modi ied pa ame e Inc emen al cos s Inc emen al QALYs ICUR (€/QALY) Inc emen al cos s Inc emen al QALYs ICUR (€/QALY) Time ho izon (20 yea s) -€23,829.81 −0.000207 Lowe cos and lowe e ec i eness -€11,519.76 0.0452 Dominan Time ho izon (10 yea s) -€23,472.38 −0.007868 Lowe cos and lowe e ec i eness -€11,596.81 0.0483 Dominan Time ho izon (5 yea s) -€21,226.95 −0.016366 Lowe cos and lowe e ec i eness -€11,610.99 0.0500 Dominan Discoun a e (0%) -€24,946.59 0.005504 Dominan -€11,341.74 0.0359 Dominan Discoun a e (5%) -€23,162.25 −0.002531 Lowe cos and lowe e ec i eness -€11,449.39 0.0439 Dominan Annual su ge y isk (4.21%) -€23,822.65 −0.000201 Lowe cos and lowe e ec i eness -€11,211.67 0.0308 Dominan Pe iope a i e mo ali y (0%) -€23,815.57 0.000143 Dominan -€11,438.21 0.0417 Dominan Pe iope a i e mo ali y (2.9%) -€23,815.58 0.000135 Dominan -€11,436.61 0.0425 Dominan AE: ad e se e en ; ICER: inc emen al cos –e ec i eness a io; ICUR: inc emen al cos –u ili y a io; LY: li e yea ‘ QALY: quali y-adjus ed li e yea . EXPERT REVIEW OF PHARMACOECONOMICS & OUTCOMES RESEARCH 7 ollowing ailu e wi h con en ional ea men o ollowing an an i-TNF biologic ea men . To aci inib yields a educ ion in o al cos s pe pa ien compa ed o edo- lizumab in bo h pa ien popula ions (bio-nai e and bio- expe ienced) la gely de i ing om he pha macological cos s o he ea men s. Among he mos ele an indings we e hose ela ed o adminis a ion cos s: o ally adminis- e ed ea men s accoun o subs an ial sa ings in ea - men cos s compa ed o in a enously adminis e ed al e na i es, which equi e g ea e heal hca e esou ce consump ion. The esul s e ealed li le di e ence in all o he cos s analyzed, in which he cos o o aci inib is shown o be sligh ly lowe . Speci ically, disease manage- men wi h edolizumab esul s in a highe cos due o a highe numbe o pa ien s who emain in a s a e o ac i e UC and, he e o e, wi h a highe p obabili y o unde going su ge y compa ed o ea men wi h o aci i- nib. Conce ning he di e ence in cos s due o ad e se e en s, he highes cos wi h edolizumab de i ed om a highe incidence o se ious in ec ions, uppe espi a o y ac in ec ions, and acu e in usion eac ions, acco ding o he s udies included. On he o he hand, di e ences in QALYs a e sligh , wi h o aci inib p esen ing quali y- adjus ed su i al gains in bo h popula ions s udied. A ecen cos -e ec i eness analysis was published compa - ing o aci inib and biologic ea men s, such as in liximab, adalimumab, golimumab, and edolizumab, in an i-TNF nai e pa ien s wi h mode a e- o-se e e UC in Spain. In his 10-yea analysis, edolizumab and o aci inib we e he ea men s wi h highe e ec i eness, wi h an ICUR o €45,253/QALY o edoli- zumab compa ed o o aci inib. To aci inib had di e ences in cos s be ween -€13,596 and €68,043 and di e ences in e ec- i eness be ween −0.30 and 0,98 QALYs, when compa ed wi h he biologic ea men s [39]. This au ho ecen ly published a budge impac analysis conduc ed in Spain acco ding o which he use o o aci inib esul s in sa ings o up o €3,600,177 a e 5 yea s [40]. In e na ionally, cos –e ec i eness analyses including o aci i- nib ha e indeed been published [41,42]. The esul s o hese publica ions a e consis en wi h he indings o his analysis. A s udy by Lohan e al. in he Uni ed Kingdom also analyzed wo popula ions acco ding o p io exposu e o a TNF inhibi o . This s udy ound cos educ ions associa ed wi h o aci inib e - sus in liximab and edolizumab in bo h popula ions; o aci inib Figu e 2. De e minis ic sensi i i y analysis: o nado diag am (A: bio-nai e, B: bio-expe ienced). Figu e 3. P obabilis ic sensi i i y analysis: cos -e ec i eness plane (A: bio-nai e, B: bio-expe ienced). 8C. TAXONERA ET AL. was dominan in hese wo compa isons and cos -e ec i e (< ₤18,000/QALY) e sus adalimumab and golimumab [41]. Ano he s udy e alua ed he e iciency o o aci inib in wo egions ( he Uni ed Kingdom and China) using biologic and con en ional ea men sequences. The sequences wi h o aci i- nib and edolizumab we e he mos cos -e ec i e ea men s in he Uni ed Kingdom, and he sequences wi h o aci inib we e he mos cos -e ec i e ea men s in China. When sequences wi h a biologic we e compa ed (simila o he bio-expe ienced analysis), o aci inib was dominan compa ed o in liximab, edo- lizumab, and golimumab, yielding a ios o ₤11,819-₤22,515/ QALY e sus adalimumab [42]. In line wi h hese esul s, in an economic e alua ion by Mile e al. [43], o aci inib ollowed by in liximab was associa ed wi h a lowe cos pe pa ien pe mon h compa ed o ea men wi h edolizumab ollowed by in liximab, in a bio-expe ienced popula ion. Ano he s udy was ecen ly published in G eece, in which o aci inib was dominan o edolizumab in bo h bio-nai e and bio-expe ience popula- ions wi h a 97% p obabili y o being cos -e ec i e [44]. Finally, a cos -e ec i eness analysis in Poland compa ed o aci inib and o he biologics wi h con en ional ea men , being o aci inib he mos e ec i e ea men in bio-expe ienced pa ien s and in liximab in bio-nai e pa ien s [45]. This analysis did ha e some limi a ions. Gi en he absence o e icacy in di ec compa ison be ween he wo al e na i es, i was necessa y o pe o m indi ec compa isons using an NMA. As s udies wi h a ea - h ough design wi h edolizumab we e no a ailable a he ime o he analysis, es ima es had o be made based on he da a a ailable om s udies wi h a e- andomized design. This model did no conside adminis a ion o con en ional ea men ollowing ailu e wi h a second line o biologic ea men o o aci inib, which is included in some published analyses [17,41,42]. Howe e , o he exis ing models also do no inco po a e con en ional ea men [39]. The expe panel el ha he adminis a ion o con en ional ea men ollowing biologic ea men s did no e lec ou ine clinical p ac ice; wha is ypical is o swi ch o a new biologic ea men a he clinician’s disc e ion. In line wi h his app oach, a s udy o pa ien s wi h UC being ea ed wi h biologics was iden i ied in which app oxima ely hal he pa ien s who s opped ea men did no subsequen ly es a he same ea men o swi ch o ano he ea men [46]. The bio-nai e pa ien popula ion could only ecei e a second line o biologic ea men ollowing ail- u e wi h he i s s udy ea men . To minimize he e ec s o he subsequen ea men , he same second-line ea men (in lix- imab) was used in he wo compa a o s, whose selec ion was in acco dance wi h he opinion o he expe panel based on ou ine clinical p ac ice. Hence, no u he subsequen lines o ea men we e used as hey could obscu e he e ec achie ed wi h he ea men s o be e alua ed and in iew o he limi ed e idence o de e mining he e icacy o ad anced lines o ea men . Finally, possible dose escala ions included in he summa y o p oduc cha ac e is ics, which could lead o inc eased pha macological cos s o bo h ea men s, we e no conside ed. Ano he ma e o ake in o accoun is ha ea men s we e conside ed in mono he apy, whe eas i has been epo ed ha a high pe cen age o pa ien s being ea ed wi h edolizumab ake i in combina ion wi h aza hiop ine. This would inc ease he pha macological cos o his al e na- i e and could be linked o a highe isk o ad e se e en s such as lymphoma, oppo unis ic in ec ion, and non- melanoma skin cance [47]. Fu he mo e, i should be no ed ha he OCTAVE ials on o aci inib ea u ed a s ic e de ini- ion o emission (wi h he addi ional equi emen o a ec al bleeding subsco e o 0) han ha used in he s udies on edolizumab, and also equi ed a cen alized and he e o e mo e demanding endoscopy eading [15]. To e lec su ge y isk in pa ien s wi h mode a e- o-se e e ac i e UC, a ixed annual colec omy a e was used h oughou he analysis ime ho izon al hough in clinical p ac ice su ge y isk may ha e a ied o e ime in hese pa ien s wi h ac i e UC. Howe e , in his case, he use o mo e ecen da a published in Spain was p e e ed. Finally, as a limi a ion o his s udy, i should be no ed ha while he s udy was unde way, us ekinu- mab, ano he biologic adminis e ed in a enously o he ea - men o mode a e- o-se e e ac i e UC, was app o ed by he EMA. This d ug was no included in he analysis as i lacked a p ice in Spain and eimbu semen a he ime o he s udy. O e all, i is impo an o s ess ha he ul ima e choice o a ea men o UC should be he esul o a comple e assess- men ha includes pa ien cha ac e is ics, d ug e icacy and associa ed isks, as well as managemen -de i ed cos s [48]. Economic e alua ions a e o g ea u ili y in assessing a medicine’s he apeu ic posi ioning. 5. Conclusions F om he pe spec i e o he NHS, ea men wi h o aci inib could be a cos -e ec i e and dominan (less cos ly and mo e e ec i e) al e na i e han edolizumab o he managemen o pa ien s wi h mode a e- o-se e e ac i e UC ollowing ail- u e, in ole ance, o loss o esponse o con en ional o an i- TNF biologic ea men . Da a a ailabili y s a emen The e is no a ailable da a se associa ed wi h he pape . Au ho con ibu ions CP, AC, CT, FdA and MAC we e esponsible o he concep ual de elop- men o he analysis. FdA de eloped he economic model. CP, AC, CT, LM, SG, ASG, BM, ALIA, SG, MAC and FdA alida ed he s udy’s s uc u e and p emise and p o ided in o ma ion on clinical p ac ice in Spain. FdA and ALIA pe o med he analyses. All he au ho s ook pa in he in e p e a- ion o he esul s. FdA and ALIA d a ed he manusc ip . All he au ho s e iewed and app o ed he inal e sion o he manusc ip . CP is he gua an o o he gene al con en o his manusc ip . Decla a ion o in e es CT has ecei ed speake , ad iso y boa d o esea ch ees om MSD, Abb ie, P ize , Takeda, Janssen, Fe ing, Faes Fa ma, D . Falk Pha ma, Amgen, and Tillo s. FdA and MAC a e employees o Pha macoeconomics & Ou comes Resea ch Ibe ia (PORIB), a consul ing i m specialising in economic e alua- ion o heal hca e in e en ions, which ecei ed unding suppo om EXPERT REVIEW OF PHARMACOECONOMICS & OUTCOMES RESEARCH 9