Cost-effectiveness of a primary care-based exercise intervention in perimenopausal women. The FLAMENCO Project

(1)

Original

Cost-effectiveness

of

a

primary

care-based

exercise

intervention

in

perimenopausal

women.

The

FLAMENCO

Project

Zuzana ˇSpacírová

a,∗

,

David

Epstein

a

,

Leticia

García-Mochón

b

,

Virginia

A. Aparicio

c,d

,

Milkana

Borges-Cosic

e

_,

_M.

_Puerto

_López

_del

_Amo

a

_,

_José

_J.

_{Martín-Martín}

a

a_Department_of_Applied_Economics,_Faculty_of_Economics,_University_of_Granada,_Granada,_Spain b_Andalusian_School_of_Public_Health,_Granada,_Spain

c_Department_of_Physiology,_Faculty_of_Pharmacy,_Faculty_of_Sport_Sciences,_and_Institute_of_Nutrition_and_Food_Technology,_University_of_Granada,_Granada,_Spain d_VU_University_and_EMGO+_Institute_for_Health_and_Care_Research,_Amsterdam,_The_Netherlands

e_Department_of_Physical_Education_and_Sport,_Faculty_of_Sport_Sciences,_University_of_Granada,_Granada,_Spain

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received15December2017 Accepted23May2018 Availableonline16October2018

Keywords:

Cost-benefitanalysis Quality-adjustedlifeyears Menopause

Exercise Women

a

b

s

t

r

a

c

t

Objective: Adequatephysical activity levels anda healthylifestyle may preventall kindsof non-communicable diseases, promote well-being and reduce health-care costs among perimenopausal women.Thisstudyassessedanexerciseprogrammeforperimenopausalwomen.

Method:Atotalof150women(aged45-64years)notengagedinregularphysicalactivitywererandomly assignedtoeithera16weekexerciseinterventionortothecontrolgroup.Thestudywasconductedfrom theperspectiveoftheNationalHealthSystem.Healthoutcomeswerequality-adjustedlifeyears(QALYs), measuredbytheEuroQol-5D-5Lquestionnaire.Thetotaldirectcostsoftheprogrammewerethecostsof visitstoprimarycare,specialtycare,emergency,medicines,instructorcostandinfrastructurecost.The resultswereexpressedastheincrementalcost-effectivenessratio.Sensitivityanalysiswasundertaken totesttherobustnessoftheanalysis.

Results: MeanQALYsover16weekswere.228inthecontrolgroupand.230intheinterventiongroup (meandifference:.002;95%confidenceinterval[95%CI]:−0.005to0.009).Improvementsfrombaseline weregreaterintheinterventiongroupinalldimensionsoftheEuroQol-5D-5Lbutnotstatistically sig-nificant.Thetotalcostsattheendoftheinterventionwere160.38Dinthecontrolgroupand167.80D

intheinterventiongroup(meandifference:7.42D;95%CI:−47to62).Theexerciseprogrammehadan incrementalcost-effectivenessratioof4,686D/QALY.

Conclusions:Theprogrammecouldbeconsideredcost-effective,althoughtheoveralldifferenceinhealth benefitsandcostswasverymodest.Longertermfollow-upisneeded.

Análisis

de

coste-efectividad

de

un

programa

de

ejercicio

físico

en

mujeres

perimenopáusicas.

Proyecto

FLAMENCO

Palabrasclave:

Análisisdecoste-utilidad A˜nosdevidaajustadosporcalidad Menopausia

Ejercicio Mujeres

r

e

s

u

m

e

n

Objetivo:Elejerciciofísicopuedepromoverelbienestaryreducirloscostesdeatenciónmédicaenlas mujeresperimenopáusicas.Esteestudioevalúaunprogramadeejerciciofísicoenmujeres perimenopáu-sicas.

Método:Untotalde150mujeres(deedadescomprendidasentre45y64a˜nos)fueronasignadas aleatoria-mentealgrupodeintervenciónoalgrupodecontrol.Elestudiohatenidounaduraciónde16semanas.Los resultadosensaludsehanmedidoena˜nosdevidaajustadosporcalidad(AVAC)medianteelcuestionario EuroQol-5D-5L.Sehaconsideradoeltotaldecostesdirectosdelprograma,integradoporloscostesde lasvisitasenatenciónprimaria,atenciónespecializadayurgencias,medicamentos,costedelmonitory costedelasinstalaciones.Losresultadossehanexpresadocomoratiocoste-efectividadincremental.La robustezdelmodelosehacontrastadoconunanálisisdesensibilidad.

Resultados:Alfinaldelaintervención,losAVACfueron0,228enelgrupodecontroly0,230enelgrupo deintervención(diferenciamedia:0,002;intervalodeconfianzadel95%[IC95%]:−0,005a0,009).La mejoríafuemayorenelgrupodeintervenciónentodaslasdimensionesdelEuroQol-5D-5L,perosin significaciónestadística.Loscostestotalesalfinalizarlaintervenciónhansidode160,38D enelgrupo decontroly167,80Deneldeintervención(diferenciamedia:7,42D;IC95%:−47a62).Elprogramade ejerciciofísicohatenidounaratiocoste-efectividadincrementalde4686D/AVAC.

∗ Correspondingauthor.

E-mailaddress:[email protected](Z. ˇSpacírová).

https://doi.org/10.1016/j.gaceta.2018.05.012

(2)

Z. ˇSpacírováetal./GacSanit.2019;33(6):529–535

Conclusión:Elprogramadebeconsiderarsecoste-efectivo,aunqueladiferenciaenresultadosdesaludy costeshasidomuymoderada.Senecesitaunseguimientoamáslargoplazo.

Introduction

MenopauseisdefinedbytheWorldHealthOrganizationasthe lastdayofmenstruationwhichisduetothelossofovarianfollicular activity.Itoccursonaveragearoundtheageof51.1_{Perimenopause} is an imprecise period that begins with thefirst alterations of theovariancycleandendsoneyearafterthelastmenstruation.1 Symptomsassociated withperimenopausecanbe quitevaried: vasomotorsymptomssuchashotflushes,2 _bone_loss_which_can leadtoosteoporosis,3_bodily_changes_such_as_increased_waist cir-cumference,increasedadiposetissue,decreasedmuscletissue4_or evenincreasedriskofheartdisease.5,6_Many_of_these_changes expe-riencedduringperimenopausecanbemademorebearableoreven preventedbyhealthylifestylehabits.7,8

Numerousstudiesdemonstratethepositiveimpactofphysical activityorexerciseondiminishingriskfactorsassociatedwith car-diovasculardisease,9,10 _promoting_weight_loss11_and_preventing bonelossorosteoporosis.12_In_addition,_women_suffering symp-tomsofmenopausetendtousemoremedicationandotherhealth care.13,14 _Although _the _health _benefits _of _physical _activity _are stronglyestablished,theprevalenceofphysicalactivityinmidlife womenusetobeinadequate,andevidenceremainsinconclusive ontheroleofphysicalactivityonmenopausalsymptoms.15,16_Our hypothesisisthataphysicalexerciseprogramwillimprovequality oflife atanacceptable cost.Eventhough menopauseis a phe-nomenonthatultimatelyconcernsallwomen,onlythreeprevious studieshaveexaminedthecost-effectivenessofaphysicalexercise programinwomenaroundtheageofmenopause.17–19_All_studies concludedthatthephysicalexerciseprogramfortargetedwomen wascost-effective.

TherandomizedcontroltrialFitnessLeagueAgainstMENopause COst (FLAMENCO)20 _investigated _symptoms, _health _related quality of life (HRQoL) and costs of a physical exercise program (Trial Number NCT02358109, https://clinicaltrials.gov/ ct2/show/NCT02358109,dateofregistration:September23,2014). Theobjectiveofthisarticleistostudythecost-effectivenessofthe physicalexerciseprogramforperimenopausalwomen,measured intermsofcostperquality-adjustedlifeyears(QALYs),whichwas theprimaryoutcomeofthestudy.

Methods

Thestudywasdesignedasarandomizedcontrolledtrialand wascarried outover a period of 16 weeks fromthebeggining ofMarchtotheendofJune 2015ata primarycarecenter.The studypopulationwerewomenfromGranada(Spain)whowere notengagedinregular physicalexercise,butotherwise healthy andabletoexercise,agedbetweenabout45to60years coincid-ingwiththeperimenopausalperiod.Theywererandomlyassigned toeither anexerciseintervention group (N=74) or tocontrol group(N=76).Bothgroups receivedfourconferences inwhich generaladvice aboutthe positive effects of a physical exercise programandoftheMediterraneandietweregiven.Theexercise interventionwasperformedinfourgroups.Thegroupstrained3 days/week(60minutes/session)fora16-weekperiodatthe pri-marycarecentre. Eachexerciseincludeda10-minutewarm-up periodwithwalksandmobilityexercises,followedby40-minute ofamainpartwhich variedacrossweekdays.Sessionsfinished

witha10-minutecool-downperiodofstretchingandrelaxation exercises.Theweeklyprogramofexercisesconsistedofresistance strengthonMonday,balanceorientedactivitiesonWednesdayand combinationofaerobic,resistancestrengthandcoordination exer-cisesonFriday.Outcomeassessorsanddataanalystswereblinded totheallocation.FulldetailsoftheFlamencoprojectdesignand methodologyaredescribedelsewhere.20

ThestudyfollowedtheConsolidatedHealthEconomic Evalu-ationReportingStandards(CHEERS)21_and_{recommendations}_for economicevaluationappliedtohealthtechnologiesinSpain.22_The studywasapprovedbytheEthicsCommitteeforResearchInvolving HumanSubjectsattheUniversityofGranadaandwasconducted fromtheperspectiveoftheNationalHealthSystem.The partici-pantsprovidedwritteninformedconsenttoparticipate.Aliterature reviewwasconductedtoidentifyothereconomicevaluationsin thisarea(seeAdditionalmaterialonlineforsearchterms).

Healthcareresourceuse(visitstoprimarycare,specialitycare andemergencyrooms)andpharmaceuticalconsumptionofeach patientbeforeandduringthestudywasobtainedthroughmedical historyfromDirayasystem23_used_by_the_Public_Health_System_of Andalusia.

Costs werecalculated at 2015prices. The salary cost ofthe instructor in charge of carrying out the exercise program was 8.74D/hour.24_Assuming₁₂_hours/week_of_{instructor‘s}_work_(four groups, 3hours/week/group)and 16 weeks of intervention,the personnelcostoftheexerciseprogram/womanwas(8.74D/h×

12h/week×16weeks)/74women=22.68D/woman.Pricesper

visitinprimarycare,specialitycareandemergencyserviceswere estimatedfromstandardhealthservicecostsof200525_and_updated forinflationto2015prices.26_The_consumption_cost_of_prescribed pharmaceuticals wascalculatedbased ontheprices, prescribed doseandscheduleofadministrationinDiraya.

Intheclinicalstudy,theexerciseprogramfacilitieswere pro-vided by the health service at nofinancial cost. In practice in othersettingstheremaybeafinancialcostoranopportunitycost (anotheractivitythatisdisplacedbytheexerciseprogram).We assumedthecostofhiringasuitablefacilityinGranadaforcarrying outtheexerciseprogramwouldbe500D/monthbasedonmarket pricesin2015inGranada,andthecostofutilities(cleaning, light-ing,power,etc.)wouldbe165D/month.27_Assuming_a_maximum utilizationof55hours/week(238hours/month),andagroupsize of18women,theinfrastructurecostperwomanperhouris0.16

D,calculatedas(500+165)/(238×18).

TheHRQoLwasmeasuredbytheSpanishversionof EuroQol-5D-5L(EQ-5D-5L).Thequestionnairewascompletedatthebeginning, atthemiddleandattheendofthestudy.Utilitywasestimated usingthe publishedtariff.28 _QALYs_were_calculated _as_the _area underthecurve.Arandomeffectsorderedlogisticmodel(xtlogit commandin Stata)wasusedtomeasurethedifferencein each dimensionofEQ-5D-5Lbetweengroups.

Missingdatacanleadtobiasedestimatesandreduced preci-sion.Biasmaybeespeciallylikelywhenthereisabigdifference inmissingdatabetweengroups.Becauseresourceusedatawere collectedfromprimarycarerecords,therewerenomissingcost data.However,thereweremissingEQ-5D-5Ldataatbaseline,8 weeksand16weeks.Baselinedatawereimputedwiththemeanof thegroup.29_Missing_intermediate_and_final_EQ-5D-5L_index_scores wereimputedusingmultipleimputationwithchainedequations.

(3)

Themultipleimputationmissingdatamodelincludedaspredictive variablesEQ-5D-5Lindicesatbaselineandfollow-up,costsandage. First,themissingdatawereimputedunderthemissingatrandom (MAR)assumption.Thiswasthemainanalysis(base-case).Second, completecaseanalysiswasperformedwhichwouldcorrespondto ascenariowherewomenwhocompletedallfollow-upcanbe con-sideredfullyrepresentativeofallthewomenwhoinitiallyagreed toparticipate.Thisassumesdataaremissingcompletelyatrandom (MCAR).Thethirdscenarioconsideredwhatmightoccurifdata weremissingnotatrandom(MNARorinformativemissingness). Inthispaper,asimplepatternmixturemodelwasimplemented, followingtheapproachrecommendedinFariaetal.29_For_example, theMNARmodelallowsforthepossibilitythattheprobabilityof attendingfollow-upwasrelated(eitherpositivelyornegatively)to thehealthofthewomenatthattime.TheimputedEQ-5D-5Lof thewomenwhodidnotreturnaquestionnaireat4monthswere modifiedin 1%variationsabove(below)thevaluepredictedby MAR.Thiscorrespondedtoascenariowherewomenthatfailto attendtheexerciseprogramweremore(less)healthythan aver-age.TheaimwastosearchforthethresholdincrementinEQ-5D-5L abovethevaluepredictedbyMARwhichchangedthedecisionat thecommonly-usedwillingness-to-pay(22,000D/QALY).30

Incremental cost and QALYswere calculated using bivariate regression (sureg command in Stata). QALYswere adjusted for baselineEQ-5D-5L31_to_account_for_differences_between_groups_at baseline.Coefficientswerecombinedacrossthemultipleimputed datasetusingRubin’srules.29 _The_probability_that_the interven-tion wascost-effective was calculated assumingthe datawere bivariatenormaldistributed.32_The_analyses_were_performed_using STATA14.

To assess the robustness of the results additional sensitiv-ity analysiswas performed,alongsidethe missing datamodels described above. The firstmodel, the main analysis,uses mul-tiple imputation assuming missing data are MAR and includes

infrastructurecosts.Inthesecondmodelweremovedthe infras-tructurecost.Inthethirdmodel(completecase)weremovedall thewomenwhodidnotreturnaHRQoLquestionnaire(assuming missing dataareMCAR). The fourthmodelexcludes the infras-tructurecostsandalsoallthosewomenwhodidnotreturnthe questionnaire.Fifth,theassumptionthatmissingdataareMNAR wasusedtofindthethresholdofimprovementinnon-attending women‘shealththatmakesthemodelnotcost-effective.

Results

Ofthe150womenwhoparticipatedinthestudy,76(51%)were inthecontrolgroupand74(49%)wereintheinterventiongroup. Nosignificantdifferencesinbaselinevariableswerefoundbetween groups(Table1).Figure1showstheConsortflowdiagram.

Theaveragetotal costperwomanwasslightlyhigherin the interventiongroupthaninthecontrolgroupbutthedifferencewas notsignificant(167.80D and160.38D,respectively;difference: 7.42;p=0.8,95%confidenceinterval[95%CI]:−47to62)(Table2).

Theinterventioncostperpersonwas30.36D (22.68D instructor and7.68D infrastructure),representing18%oftotalcostsinthe interventiongroup,butthiswaspartlycompensatedforbylessuse ofhealthcareservices.Excludingtheinterventioncost,totaldirect costswere16.7%lowerintheinterventiongroupthaninthecontrol group(137.45Dand160.38D,respectively;difference:−22.94;p=

0.38;95%CI:−75to29).However,differenceswerenotstatistically

significant(Table2).

SupplementaryTableIonlinecontainstheresponsestothe var-iousitemsofEQ-5D-5Lforeachgroup.Theunadjustedutilitywas higherintheinterventiongroupthaninthecontrolgroupatthe endofthestudy(SupplementaryTableIIonline).However,there weresmalldifferencesinEQ-5D-5Lscorebetweengroupsbaseline. Althoughthesewerenotstatisticallysignificant,thesedifferences can affect the results of a cost-effectiveness analysis because

Table1

Baselinecharacteristicsofinterventionandcontrolgroups.

Interventiongroup (N=74)

Controlgroup (N=76)

p

Age,years(mean,SE) 54.0(0.52) 53.22(0.88) 0.45a

Education 0.99c

Noeducation(frequency,%) 2(2.74) 2(2.63) Primary(frequency,%) 18(24.66) 21(27.63) Secondary(frequency,%) 16(21.92) 15(19.74) Professionalexperience(frequency,%) 12(16.44) 14(18.42) Bachelor(frequency,%) 15(20.55) 13(17.11) Master(frequency,%) 10(13.70) 11(14.47)

Regularoroccasionalsmoker 0.16c

Dailysmoker(frequency,%) 12(16.67) 16(21.62) Occasionalsmoker(frequency,%) 7(9.72) 2(2.7) Formersmoker(frequency,%) 38(52.78) 33(44.59) Neverhavesmoked(frequency,%) 15(20.83) 23(31.08)

Civilstatus 0.33c

Married(frequency,%) 50(68.49) 56(73.68) Single(frequency,%) 10(13.70) 7(9.21) Separated(frequency,%) 7(9.59) 3(3.95) Divorced(frequency,%) 4(5.48) 9(11.84) Widow(frequency,%) 2(2.74) 1(1.32)

Employment(frequency,%) 36(49.32) 44(59.46) 0.19c

Children(mean,SE) 1.95(0.12) 1.99(0.11) 0.80a

Useofhealth-careservicesintheprevious8weeks(mean,SE)

Visitstoaprimarycare 0.93(0.13) 0.70(0.12) 0.17a

Visitstoaspecialist 0.25(0.07) 0.19(0.05) 0.50a

Visitstoanemergency 0.07(0.03) 0.04(0.02) 0.54a

Medicationcostinprevious8weeks(median,IQR)(D) 8.70(22.86) 7.58(21.73) 0.92b

EQ-5D-5Lindexscore(between0and1)(mean,SE) 0.839(0.02) 0.854(0.01) 0.53b

EQ-5D-5L:EuroQol-5D-5L;IQR:interquartilerange;SE:standarderror.

a_t-test. b_U-test.

(4)

fj Assessed for

elegibility (n=214)

Randomized (n=150)

Excluded (n=6) -Did not meet inclusion criteria (n=39)

-Declined participation (n=14) -Other reasons (n=5)

Intervention group (n=74)

Usual care group (n=76)

Lost follow-up (did not return

EQ-5D questionnaire) (n=20) Lost follow-up

(did not return EQ-5D questionnaire)

(n=9)

Completed the program (n=56) Analyzed (n=76) Completed the

program (n=65) Analyzed (n=74)

Enrol

lm

ent

Al

loc

ati

on

Ana

ly

sis

Fo

llo

w-up

(a

t

16

w

ee

ks)

Figure1.Flowchartofpatientrecruitment.

Table2

Costsofinterventiongroupandcontrolgroup. Unit

cost (D)

Resourceuse Meancost/person Totalcosts/person 8-0

weeks 0-8 weeks

8-16 weeks

8-0 weeks

Diff.mean costper person(D)

0-8 weeks

8-16 weeks

0-16 weeks

C I C I C I C I C I C I C I

Primarycare(visits) (numberof patients>0)

21.62 71 52 60 59 51 39 20.20 15.19 −5.01 17.07 17.24 0.17 14.51 11.39 −3.11 31.58 28.63 −2.94

44 34 35 30 33 27

Specialist(visits) (numberof patients>0)

62.74 19 14 33 22 33 24 15.69 11.87 −3.82 27.24 18.65 −8.59 27.24 20.35 −6.89 54.48 39.00 −15.48

13 12 22 15 25 17

Emergencyrooms (visits)(number ofpatients>0)

59.58 5 3 7 7 14 6 3.92 2.42 −1.50 5.49 5.64 0.15 10.98 4.83 −6.14 16.47 10.47 −6.00

4 3 7 6 12 5

Medicine 18.59 27.50 8.90 26.52 29.02 2.49 31.34 30.33 −1.01 57.86 59.35 1.49

Totalcosts 58.40 56.98 −1.42 76.32 70.55 −5.77 84.07 66.90 −17.17 160.38 137.45 −22.94e Costsofthe

intervention

Instructor 8.74 0 0 0 96h 0 96h 0.00 0.00 0.00 11.34 0.00 11.34 0.00 22.68 Infrastructure 0.16 0 0 0 24h 0 24h 0.00 0.00 0.00 3.84 0.00 3.84 0.00 7.68

Totalcosts 58.40 56.98 -1.42a _76.32 _85.73 _9.41b _84.07 _82.08 _-1.99c _160.38 _167.80 _7.42d

C:controlgroup(N=76);Diff.:difference;I:interventiongroup(N=74).

Theremaybedifferencesindirectcostsandintotalcostsasaresultofroundingout.

Numberofpatients>0referstopatientsattendingprimarycare,specialitycareoremergencyrooms. 8-0weeksreferstoeightweeksbeforethestartofthestudy.

Allp-valueswerecalculatedbybootstrapmethod.

a_p₌_0.90. b _p₌_0.57. c _p₌_0.90. d _p₌_0.80. e_p₌_0.38.

baselineEQ-5D-5LisanelementofthecalculationofQALYs.Once adjustedforthedifferencethatexistedbetweenthetwogroupsat baseline,31_and_imputing_for_missing_data,_the_difference_in_QALYs was0.002(p =0.66;95%CI: −0.005to0.009). The incremental

cost-effectivenessratio(ICER)was4,686D/QALY(Table3).

ThesensitivityanalysisareshowninTable3.Inmodel2,thetotal meancostsintheinterventiongroupwereloweranddeliveredan improvementinhealth,sothatthephysicalexerciseprogramissaid to“dominate”usualcare.Theresultsofthefirsttwomodelswere different,duetothesmalldifferenceinQALYsbetweenthetwo

(5)

Table3

Sensitivityanalysis.

Model Costs Difference QALYs Difference ICERa

Intervention Control Intervention Control

1.Multipleimputationmodel(base-case) 167.80 160.38 7.42 0.2295 0.2279 0.0016e _4,686

2.Multipleimputationwithoutinfrastructurecostsb _160.12 _160.38 _-0.26 _0.2295 _0.2279 _0.0016 _Intervention_dominates

3.Completecaseanalysisc _153.15 _168.53 _-15.37 _0.2304 _0.2272 _0.0032 _Intervention_dominates

4.Completecaseanalysiswithoutinfrastructurecosts 145.47 168.53 -23.05 0.2304 0.2272 0.0032 Interventiondominates 5.IncreasedQALYsofallindividualswithimputed

utilitiesby2%d

167.80 160.38 7.42 0.2300 0.2293 0.0007e _10,748

ICER:incrementalcost-effectivenessratio;QALYs:qualityadjustedlifeyears.

a_ICER_is_the_result_of_dividing_the_difference_between_costs_and_QALYs,_both_without_rounding_out._In_model₁_that_is_{7.416473/0.0015827,}_and_in_model₅_that_is

7.416473/0.00069.

b_{Infrastucture}_costs_refers_to_the_cost_of_hiring_a_suitable_facility_and_to_the_cost_of_utilities_as_cleaning,_power,_etc. c _Analysis_of_only_those_women_who_returned_{EuroQol-5D-5L}_{questionnaire.}

d _This_analysis_was_used_to_find_the_threshold_of_improvement_in_{non-attending}_women’s_health_that_makes_the_model_not_{cost-effective.} e_p_>0.05_corresponds_to_ordinary_linear_square_regression_model.

Table4

Associationofquality-adjustedlifeyearsandincremental-costeffectivenessratiowithnon-attendingwomen’shealth. Non-attendingwomen’shealthat

16weeksoffollow-up

96% 97% 98% 99% 100% 101% 102% 103% 104%

Differenceincost(mean,SD) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) 7.42(27.71) DifferenceinQALYs(mean,SD) 0.0034 0.0029 0.0025 0.0020 0.0016 0.0011 0.0007 0.0002 −0.0002a

(0.0036) (0.0036) (0.0036) (0.0035) (0.0035) (0.0036) (0.0036) (0.0036) (0.0037) ICER(D/QALY) 2,202 2,538 2,996 3,655 4,686 6,526 10,748 30,438 −36,590

ICER:incrementalcost-effectivenessratio;QALYs:quality-adjustedlifeyears;SD:standarddeviation. AlldifferencesinQALYsareroundedout.

a_Dominated._QALYs_in_the_intervention_group_are_lower_than_QALYs_in_the_control_group.

groups.Thismeansthatasmallchangeincostshadalargeimpact ontheICER.Inmodels3and4theresultwassimilartomodel2,i.e. theinterventionwasdominant(Table3).Model5assumesa sce-nariowherewomenmightnotattendfollow-upforreasonsrelated totheirhealthonthatday,thatis,missingdatawereMNAR.Itwas foundthattheexerciseprogramstartstobenotcost-effectiveif theimputedEQ-5D-5Lofthewomenwhodidnotreturna ques-tionnaireat4monthswas3%greaterthanthevaluepredictedby MAR,thatis,ifitisassumedthatwomenwhodidnotattendwere 3%morehealthythanwouldbeexpectedbytheirage,useofhealth servicesandothercharacteristics.Table4showstheresultsofthe MNARanalysis.

The probability that the intervention was cost-effective in differentscenariosandatdifferentlevelsofwillingnesstopayis showninFigure2.Theprobabilityofbeingcost-effectivewhen thethresholdis25,000D/QALYwas63%inmodel1,66%inmodel 2,81%inmodel3,83%inmodel4and54%inmodel5.

Discussion

Thebasecase forthisstudyfoundthatthecostperQALYof aphysicalexerciseprogramwas4,686D/QALY.ThemeanQALYs over 16 weeks were0.230 vs. 0.228(p =0.66) and costswere 167.80D vs.160.38D (p=0.8)in theintervention andcontrol group,respectively.ThereisnoofficialcostperQALYthresholdin Spainbutsomeauthorshaverecommendedthatinterventionswith anICER22,000-25,000D/QALYshouldbeaccepted.30

Theintervention cost perperson was30.36D, but this was partlycompensatedfor bylessuseof otherhealthcareservices in the intervention group, especially in specialist visits. This supportsresultsfromotherstudiesthatphysicallyactivepeople tendtouseless health-careservices.33 _{Unexpectedly,}_specialist visits increased from baseline in both groups, but considering thewaiting list,theappointmentscouldhavebeenagreedeven beforethebeginning ofthestudy. Thatis,we cannot conclude that women’s health worsened and for that reason they went moreoftentothespecialist.Alsounexpectedly,useofmedication

increasedincontrolgroupfrombaseline.Healthcarecanbevery variable,soitisdifficulttodrawinferences.

Otherstudieshaveshownphysicalactivityimprovesqualityof life.34,35_The_trial_protocol_considered_for_the_power_calculation_that theclinicallymeaningfulchangeinEQ-5D-5Lindexover16weeks shouldbe0.07units.20_The_actual_change_in_EQ-5D-5L_in_the inter-ventiongroupbetweenbaselineand 16weekswas0.039units, whichisstillsubstantial,buttherewasalsoasimilarimprovement inthecontrolgroup.Theseimprovementsinhealthinthecontrol groupmaybebecausethecontrolgroupreceivedmorethanusual care,asbothgroupsunderwentfitnesstestingatbaselineand fol-lowup,andreceivedrecommendationsonexerciseanditsbenefits forthelongevity,preventionandtreatmentofdiseases,aswellas thebenefitsoftheMediterraneandiet.Thesecostswerenottaken intoaccountforthecalculationoftotalcosts,becausetheywere identicalinbothgroups.

Thecurrentstudyisoneofthefewcostutilityanalysesofan exerciseinterventionprogramwithperimenopausalwomen.What makesourstudyuniqueisthattheexerciseprogramwas espe-ciallydesignedforperimenopausalwomen.Oneofthelimitations ofthisstudywasthat20%ofwomendidnotcomplete question-nairesattheendofthestudy.However,thisrateofwithdrawal wasallowedforinthesamplesizecalculation.20_Resource_use_and costswereavailableforallparticipantsfromadministrativedata. Furthermore,wehavetakenaccountofmissingqualityoflifedata intheanalysesusingapublishedmethodology.29_Results_are gen-erallyrobusttoassumptionsaboutmissingdata.Theincremental cost-effectivenessratiowaslessthan22,000D/QALYinall scenar-iostestedinsensitivityanalysis.Infrastructurecostsareuncertain, butarenotlikelytoinfluencetheoverallresult.However,the deci-sionisverysensitivetoassumptionsaboutthetruevaluesofthe missingdata.Itappearsnottobecost-effectiveunderthescenario thatwomenwhodidnotattendfollow-upwere3%morehealthy thanwouldbeexpectedgiventheirageandothercharacteristics. Thisresultoccursbecausemorewomenfailedtoattend follow-upinthecontrolgroupthantheexercisegroup.Weshouldalso considerthatthesewomenmightnotbefullyrepresentativeof

(6)

0 10 20 30 40 50 60 70 80 90

0 2.500 5.000 7.500 10.000 12.500 15.000 17.500 20.000 22.500 25.000

Threshold for cost-effectiveness (€/QALY)

Model 1. Multiple Imputation Model. Missing at random (basecase)

Model 2. Multiple Imputation Model without Infrastructure costs

Model 3. Complete case analysis (missing completely at random)

Model 4. Complete case analysis without Infrastructure costs

Model 5. Increased QALYs of all individuals with imputed utilities by 2%

Figure2. Probabilitythattheinterventioniscost-effectivefordifferentscenarios.

theperimenopausalpopulation,sincewomenwhopracticed phys-icalactivityregularlywereexcludedfromthestudy.Althoughthe exerciseprogrammightbeconsideredcost-effectiveonaverage, thedifferenceinimprovementinhealthbetweeninterventionand controlgroupisverysmallandstatisticallynotsignificant.

Threeotherstudieshaveinvestigatedthecost-effectivenessof anexerciseprograminperimenopausalwomen.Koluetal.17 evalu-atedaprograminFinlandwhere151womenaged40-63yearswere dividedintoacontrolgroupandinterventiongroup.The interven-tiongroupunderwenta6-monthexerciseprogram4times/week for50minutes.Themeandifferenceincostswas53D.However, thestudydidnotreportthedifferenceinQALYsover6months. Instead,theyreportedacrudeprojection,assumingthatthe differ-enceinHRQoLattheendofthestudywouldbemaintainedforthe restofthepatient’sexpectedlife.Usingthisprojection,thereported meandifferenceinQALYsoverthepatient’slifetimewas1.16,with anICERof46D/QALY.However,thisextrapolationseemshighly optimisticandisthereforelikelytobebiased.Thefailuretoreport actualoutcomesat6monthsmeanswecannotcomparethisresult withourstudy.

AnotherstudyconductedinCáceres,Spain,assessedthe cost-effectivenessofanexerciseprogramwhere106womenaged60 yearsandolderparticipated.18_The_intervention_group_underwent a6-monthwalking-basedsupervisedexerciseprogramwiththree 50-minutesessions/week.Thecontrolgroupreceiveda recommen-dationofphysicalactivity.Thedifferencein costwas41D (no p-valuegiven)andthedifferenceinQALYsatsixmonthswas0.132 (95%CI:0.104to0.286)thereforetheICERwas311D/QALYgained. Thisisamuchgreaterhealthbenefitthanfoundinourstudy. How-ever,thestudypopulationisolderandmoreunfitthanourstudy andsoresultsmaynotbegeneralizable.

Goranitisetal.19_evaluated_the_cost-utility_of_an_individual_and asocial versionofanexerciseinterventionrelative toa control groupinWestMidlands,UnitedKingdom,with261womenaged between48and57years.Bothinterventiongroupsfolloweda 6-monthcourseofmoderateintensityaerobicexercisefor30minutes

onatleast3days/week.ThedifferencefoundinQALYat12months of0.013(95%CI:₋0.010to0.036)inexercisesupportversuscontrol, andaverysmalldifferenceincost

£

18(95%CI:−68to105)might

beconsidereda slightlygreaterhealth benefitthanobservedin ourstudy,thoughstillnotstatisticallyinsignificant.Nobenefitwas seenintheindividualinterventiongroup(withoutsocialsupport). Accordingtoourfindings,theprogramiscost-effectiveon aver-age.However,thedifferenceinhealthbenefitoftheintervention groupcomparedtothecontrolgroupatfourmonthsissmalland statisticallyinsignificant.Furtherstudiesinthis areamight con-siderwhetheralongerexerciseprogram,oraprogramtargeted atspecificriskgroups,oraprogramthatreinforcessocialbonds withinthegroup,mighthavemoreimpact.Theoptionofusing anotherquestionnaire tomeasureQALYsshouldbeconsidered. Longertermfollowupisalsorequired.

Whatisknownaboutthetopic?

Inperimenopausalwomen,theirregularmenstrualperiods canbeaccompaniedbyhotflashes,vaginaldryness,trouble sleeping,bodilychanges,bonelossorevenbyahigherriskof heartdisease.Despitethepositiveimpactofphysicalactivity inhealth andqualityof life,thenumber ofperimenopausal womenexercisingregularlyisinadequate.

Whatdoesthisstudyaddtotheliterature?

After16 weeksofspecializedphysical exerciseprogram, theinterventiongroupimprovedmorethanthecontrolgroup, butthedifferenceswerenotstatisticallysignificant.Thecostof theprogramisrelativelysmall.Accordingtothisstudy,policy makersshouldconsiderfinancingthisexerciseprogram. Fur-therresearchshouldfocusonlongerfollow-upandifamore targetedapproachwouldofferbettervalue-for-money.

(7)

Editorincharge

MiguelÁngelNegrínHernández.

Transparencydeclaration

Thecorrespondingauthoronbehalfoftheotherauthors guar-antee the accuracy, transparency and honestyof the data and informationcontainedinthestudy,thatnorelevantinformation hasbeenomittedandthatalldiscrepanciesbetweenauthorshave beenadequatelyresolvedanddescribed.

Authorshipcontributions

Z. ˇSpacírovácontributed toacquisition ofdata,analyzedand interpreteddataand wrotethearticle. D.Epsteinanalyzed and interpreteddata,contributedtodesignofthestudyandtorevising themanuscriptcritically.L.GarcíaMochóncontributedtodesignof thestudy.V.Apariciocontributedtoacquisitionofdata.M.Borges Cosiccontributedtoacquisitionof data.M.P.López contributed todesignofthestudyand torevisingthemanuscriptcritically. J.J.Martíncontributedtodesignofthestudyandtorevisingthe manuscriptcritically.Allauthorsgavefinalapprovaloftheversion tobepublished.

Funding

TheprojectreceivedfundsfromConsejeríadeEconomía, Inno-vación,CienciayEmpleo,JuntadeAndalucía(PI-0667-2013).

Conflictsofinterest

None.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,atdoi:10.1016/j.gaceta.2018.05.012.

References

1.UtianWH.TheInternationalMenopauseSociety.Menopause-related terminol-ogydefinitions.Climacteric.1999;2:284–6.

2.Elavsky S, Gonzales J. Effects of physical activity on vasomotor symp-toms: examination using objective and subjective measures. Menopause. 2012;19:1095–103.

3.BilezikianJP,HolickMF,NievesJW,etal.Theroleofcalciuminperi-and postmenopausalwomen: 2006position statement ofthe NorthAmerican MenopauseSociety.Menopause.2006;13:862–80.

4.Lovejoy JC, Champagne CM, deJonge L, etal. Increasedvisceral fat and decreasedenergyexpenditureduringthemenopausaltransition.IntJObes (Lond).2008;32:949–58.

5.GarcíaSotoZM,MontoroGarcíaS,LealHernándezM,etal.Valoracióndelcontrol delosfactoresderiesgocardiovascularenmujeresmenopáusicasobesastras elseguimientodeunprogramaestructuradodeeducacióndietéticayejercicio físico(ProgramaSÍSIFO).HipertensyRiesgoVasc.2016;33:103–10.

6.Soares-MirandaL,SiscovickDS,PsatyBM,etal.Physicalactivityandriskof coronaryheartdiseaseandstrokeinolderadults.Circulation.2016;133:147–55.

7.LuotoR,MoilanenJ,HeinonenR,etal.Effectofaerobictrainingonhotflushes andqualityoflife—arandomizedcontrolledtrial.AnnMed.2012;44:616–26.

8.MoilanenJM,AaltoAM,RaitanenJ,etal.Physicalactivityandchangeinqualityof lifeduringmenopause—an8-yearfollow-upstudy.HealthQualLifeOutcomes. 2012;10:8.

9.DrenowatzC,SuiX,FritzS,etal.Theassociationbetweenresistanceexercise andcardiovasculardiseaseriskinwomen.JSciMedSport.2015;18:632–6.

10.WilleyJZ,MoonYP,PaikMC,etal.Lowerprevalenceofsilentbraininfarctsinthe physicallyactive:TheNorthernManhattanStudy.Neurology.2011;76:2112–8.

11.DiBlasioA,RipariP,BucciI,etal.Walkingtraininginpostmenopause:effects onbothspontaneousphysicalactivityandtraining-inducedbodyadaptations. Menopause.2012;19:23–32.

12.SiegristM.Roleofphysicalactivityinthepreventionofosteoporosis.Med MonatsschrPharm.2008;31:259–64.

13.KleinmanNL,RohrbackerNJ,BushmakinAG,etal.Directandindirectcosts of women diagnosed with menopause symptoms. J Occup Environ Med. 2013;55:465–70.

14.KjerulffKH,FrickKD,RhoadesJA,etal.Thecostofbeingawoman.ANational StudyofHealthCareUtilizationandExpendituresforFemale-Specific Condi-tions.Women’sHealthIssues.2007;17:13–21.

15.PetteeGabrielK,MasonJM,SternfeldB.Recentevidenceexploringthe associa-tionsbetweenphysicalactivityandmenopausalsymptomsinmidlifewomen: perceivedrisksandpossiblehealthbenefits.Women’sMidlifeHealth.2015;1:1.

16.DaleyAJ,ThomasA,RoalfeAK,etal.Theeffectivenessofexerciseastreatment forvasomotormenopausalsymptoms:randomisedcontrolledtrial.BJOGAnInt JObstetGynaecol.2015;122:565–75.

17.KoluP,RaitanenJ,NygårdCH,etal.Cost-effectivenessofphysicalactivityamong womenwithmenopausesymptoms:findingsfromarandomisedcontrolled trial.PLoSOne.2015;10:e0135099.

18.GusiN,ReyesMC,González-GuerreroJL,etal.Cost-utilityofawalking pro-grammeformoderatelydepressed,obese,oroverweightelderlywomenin primarycare:arandomisedcontrolledtrial.BMCPublicHealth.2008;8:231.

19.GoranitisI,BellancaL,DaleyAJ,etal.Aerobicexerciseforvasomotormenopausal symptoms:acost-utilityanalysisbasedontheActiveWomentrial.PLoSOne. 2017;12:1–15.

20.Carbonell-BaezaA,Soriano-MaldonadoA,GalloFJ,etal.Cost-effectivenessofan exerciseinterventionprograminperimenopausalwomen:theFitnessLeague AgainstMENopauseCOst(FLAMENCO)randomizedcontrolledtrial.BMCPublic Health.2015;15:555.

21.Husereau D, Drummond M, Petrou S, et al. Consolidated Health Eco-nomicEvaluationReportingStandards(CHEERS)statement.EurJHealEcon. 2013;14:367–72.

22.LópezBastidaJ,OlivaJ,Anto˜nanzasF,etal.Propuestadeguíaparalaevaluación económicaaplicadaalastecnologíassanitarias.GacSanit.2010;24:154–70.

23.Servicio Andaluz de Salud. La historia clínica electrónica de Andalucía. (Accessed 2017/05/5.) Available at: http://www.juntadeandalucia.es/ servicioandaluzdesalud/principal/documentosacc.asp?pagina=prdiraya

24.Boletín Oficial del Estado. 2014. (Accessed 2017/05/5.) Available at:

http://www.minetur.gob.es/

25.BoletínOficialdelaJuntadeAndalucían.◦210.Ordende14deoctubrede

2005,porlaquesefijanlospreciospúblicosdelosserviciossanitarios presta-dosporCentrosdependientesdelSistemaSanitarioPúblicodeAndalucía.2005. (Accessed 2017/07/7.) Available at: http://www.juntadeandalucia.es/boja/ 2005/210/boletin.210.pdf

26.Instituto Nacional de Estadística. (Accessed 2017/01/10.) Available at:

http://www.ine.es/varipc/

27.CarmonaLópezG,PérezRomeroC,FornielesY.Costessanitariosdeprocesos asistencialesintegrados.Granada:EscuelaAndaluzadeSaludPública.2006. 28.HerdmanM,BadiaX,BerraS.ElEuroQol-5D:unaalternativasencillaparala

medicióndelacalidaddevidarelacionadaconlasaludenatenciónprimaria. AtenPrimaria.2001;28:425–9.

29.Faria R, GomesM,EpsteinD, et al.Aguideto handlingmissing datain cost-effectivenessanalysisconductedwithinrandomisedcontrolledtrials. Phar-macoeconomics.2014;32:1157–70.

30.Vallejo-Torres L, García-Lorenzo B, Serrano-Aguilar P. Estimating a cost-effectivenessthresholdfortheSpanishNHS.HealthEcon.2018;27:746–61.

31.MancaA,HawkinsN,SculpherMJ.EstimatingmeanQALYsintrial-based cost-effectivenessanalysis:theimportanceofcontrollingforbaselineutility.Health Econ.2005;14:487–96.

32.O’BrienBJ,BriggsAH.Analysisofuncertaintyinhealthcarecost-effectiveness studies:anintroductiontostatisticalissuesandmethods.StatMethodsMed Res.2002;11:455–68.

33.SariN.Physicalinactivityanditsimpactonhealthcareutilization.HealthEcon. 2009;18:885–901.

34.MansikkamakiK,NygardCH.Hotflushesamongagingwomen:a4-year follow-upstudytoarandomisedcontrolledexercisetrial.Maturitas.2016;88:84–9.

35.Kim MJ,Yim G, AhnY, etal. Association between physicalactivity and menopausalsymptomsamongperimenopausalwomen.BMCWomensHealth. 2014;14:122.