ORIGINAL RE5EARCH ARTICLE PharmacoEconomics 1996 Sep; 10 (3): 262-2681170-7690/96/0C1J9-Q262/S03.50/0

© Adis International Limited. Ali rights reseNed.

A Comparison of the Cost Effectivenessof Alternative Prophylactic Therapiesin Childhood AsthmaPhilip c. Booth, Nicholas E.I. Wells and Alan K. Morrison

The Health Economics Group, Glaxo Wellcome UK Ltd, Uxbridge, Middlesex, England

Summary An economic analysis was conducted comparing the cost effectiveness offluticasone propionate with that of sodium cromoglycate (cromolyn sodium) ina group of children aged 4 to 12 years old with asthma, who required inhaledprophylactic therapy.

Over an 8-week study period, 115 patients received sodium cromoglycate20mg 4 times daily, via the spin operated dry powder inhaler, and 110 patientsreceived fluticasone propionate 50llg twice daily, via the Diskhaler™(trademarkheld by the Glaxo Wellcome Group of Companies). Patient healthcare resourceuse was examined in terms of study medication, the use of rescue medication[salbutamol (albuterol) 200llg] and the number ofhospitalisations. The effective­ness of both treatments was examined over a range of success and failure criteriaembracing peak expiratory flow rate (PEFR) improvement, symptom control andthe level of adverse events related to the study medication.

Results indicate that, for each UK pound (f) spent, fluticasone propionate wasassociated with twice as many successfully treated patients as sodium cromo­glycate, using a range of outcomes based on the goals of treatment defined in theBritish Thoracic Society's asthma guidelines.

It is concluded that fluticasone propionate was more cost effective than sodiumcromoglycate in improving PEFR and symptom control in this group of childrenwith asthma who had a clinical requirement for prophylactic therapy.

Prevalence studies conducted in recent years in­dicate that 12 to 15% of children experience epi­sodes of wheezing over a 12-month periodJl] Ofthis group, approximately one-third to one-halfhave actually been diagnosed as having asthma.Applying these figures to UK population esti­mates[2] indicates that there are probably between1.3 and 1.6 million children less than 14 years oldin the UK with wheeze, of whom between 0.5 and0.75 million will formally have been diagnosed ashaving asthma.

The morbidity associated with asthma variesmarkedly between patients and within the same pa­tient at different points in time. Therefore, it is dif­ficult to generalise about the frequency, severityand duration of symptoms as well as about the im­pact that the disease has on normal daily activities.Nevertheless, available evidence indicates that, forsome groups at least, symptoms occur frequentlyand give rise to significant social costs such as ex­tensive absence from schoolJ3,4] Furthermore, it isclear that asthma/wheezing in childhood places an

Asthma Prophylaxis in Childhood

important burden on health services. For example,about 50% of hospital admissions for asthma in­volve children aged between O and 14 years oldand, in total, childhood asthma is estimated to costthe UK National Health Service (NHS) in excessof 100 million UK pounds (f) each year.[3]

Against this background, it is not surprising thatthere has been a growing desire to promote greatereffectiveness in the treatment of asthma and itssymptoms. In particular, the British Thoracic Society(BTS) has published guidelines[5] for the manage­ment of asthma in adults and children, which em­phasise the importance of preventing the develop­ment of symptoms by controlling the inflammatoryprocesses in the airways. Of the prophylactic treat­ments available, sodium cromoglycate (cromolynsodium) is one of the most frequently employedmedicines in childhood asthma. Inhaled cortico­steroids constitute an effective alternative thera­peutic approach but, historicalIy, there has tendedto be some reluctance on the part of clinicians toemploy these agents.

However,. the introduction of a new inhaledcorticosteroid, fluticasone propionate, could po­tentialIy reduce the apparent reluctance to employthis class of compound among children because itcombines high topical anti-inflammatory activity[6]with negligible arai systemic bioavailability.[7]One of the questions raised by this possible thera­peutic development is whether it would entail sig­nificant resource implications for the NHS. In anincreasingly cost-conscious healthcare environ­ment, ever more attention is being given to the eco­nomic costs and benefits associated with medi­cines and other means of intervention.

This study examines the cost effectiveness offluticasone propionate relative to that of sodiumcromoglycate in a group of children aged between4 and 12 years with asthma, who had a clinicalrequirement for prophylactic therapy. The resultsare expressed in terms of cost-effectiveness ratios(defined in 'Methods' section) for fluticasone pro­pionate and sodium cromoglycate to assess whichof the 2 treatments offers better value for moneywithin this patient group.

© Adis International Limited. Ali rights reserved.

263

Methods

Study Design

The economic analysis was based on a clinicaltrial, which was a multicentre, open, paralIel­group, randomised study of 8 weeks duration pre­ceded by a 2-week baseline period, the results ofwhich are reported elsewhere.[8]

Prior to the baseline period, patients (aged 4 to12 years) had to have had a clinical history ofasthma, already be receiving bronchodilator ther­apy and could, in the investigator's opinion, beconsidered for inhaled prophylactic therapy. Theconsent of a parent/guardian must also have beengiven. Children were excluded at this stage if theyhad: (i) previously received inhaled prophylactictherapy; (ii) received a course of arai corticoste­roids in the previous 6 weeks or more than 3 shortcourses of corticosteroids in the previous 6 months;or (iii) experienced a respiratory tract infection inthe 2 weeks prior to the baseline periodo

On entry into the 2-week baseline period, alIpatients stopped their current bronchodilator treat­ment and were given salbutamol (albuterol) 200f..lgvia a Rotahaler™ (trademark heldby the GlaxoWelIcome Group of Companies) as and when re­quired. From then on, a patient was entered into thestudy if he or she was able to display satisfactoryuse of the Mini-Wright™ (trademark held byClement Clarke International) peak flow meter,Rotahaler™ and Diskhaler™/spin operated drypowder inhaler. For at least 7 of the baseline dayspatients had to have recorded a daytime or night­time symptom score of at least 1 (range O to 3)[table I] and used at least 1 salbutamol 200f..lgRotacapTM (trademark held by the Glaxo WelI­come Group of Companies) during the same 24-hourperiodo Alternatively, they had to have morningpeak flow readings, for at least 7 of the baselinedays, that were less than 80% of their maximum[each patient's maximum achievable peak expir­atory flow rate (PEFR) was ascertained at the startofthe baseline period by measuring PEFR 15 min­utes after inhalation of 400f..lg salbutamol via theRotahaler™].

PharmacoEconomics 1996 Sep; 10 (3)

264

Table I. Symptom scoring schedule and definitions of successfultreatment used in the study

Symptom scoringO- no cough or wheeze, good day (slept well)

1 - cough and/or wheeze on 1 occasion (woke early or oncewith cough and/or wheeze)

2 - cough and/or wheeze on several occasions but activities notrestricted (woke 2 or 3 times with cough and/or wheeze)

3 - a bad day, cough and/or wheeze restricting activities (badnight, awake most of the night with cough and/or wheeze)

Definitions of successful treatmenta

1. No serious adverse effects and freedom from daytimesymptoms in the last 3 weeks of the studyb

2. No serious adverse effects and freedom from night-timesymptoms in the last 3 weeks of the studyb

3. No serious adverse effects and a mean morning PEFR of atleast 90% of predicted in the last 3 weeks of the studyC

4. No serious adverse effects and a mean evening PEFR of atleast 90% of predicted in the last 3 weeks of the studyC

a Serious adverse effects unrelated to the study were notdesignated a treatment failure.

b 'Freedom from symptoms' is defined as a median symptomscore of zero in the last 3 weeks of the study.

c When a patient's PEFR was already >90% of predicted at thestart of study, any subsequent improvement in predictedPEFR was designated a treatment successo

Abbreviation: PEFR =peak expiratory flow rate.

Children meeting these criteria were then random­ised to receive either sodium cromoglycate 20mg4 times daily via the spin operated dry powder in­haler or fluticasone propionate 50Jlg twice dailyvia the Diskhaler™. Relief medication in the formof salbutamol 200Jlg, via the Rotahaler™, wasavailable and could be used as and when required.

Cast Data

In addition to the use of study drugs, resourcedata collected in the clinical trial included the useof rescue medication (salbutamoI200Jlg) by eithergroup throughout the study period and the inci­dence of hospitalisations. A mean cost per patientof treatment for either therapy arm was calculatedbased on these 3 resource items.

Data on other aspects of resource utilisation, suchas unscheduled GP contacts, were not collected inthe trial and are not therefore included in the analysis.

© Adis International Limited. Ali rights reseNed.

Booth et al.

The implications of this omission are addressed inthe 'Discussion' section.

Efficacy Data

The effectiveness measures used in the study arebased upon the originaI BTS adult recommenda­tions of 1990,[9] the only BTS guidelines availableat the time the clinical trial was undertaken. TheseoriginaI guidelines suggested that the goals ofasthma management in adults should be to recog­nise asthma, abolish symptoms, restore normal orbest possible long term function of the airways, andreduce the risk of a severe attack. This paper fo­cuses on 2 of these aims, the abolition of symptomsand the restoration of normal or best possible func­tion of the airways. These objectives are consistentwith the recently revised BTS guidelines[5] (1993)which address the long term management of asthmain children.

Efficacy and safety parameters assessed in thetrial included morning and evening PEFR measure­ment, the incidence of symptoms during the day andnight, and the occurrence of adverse events.

For the purposes of the analysis, 4 differentmeasures of effectiveness are used to define a treat­ment success (table I). Each measure combines asafety and efficacy component reflecting 2 of theobjectives specified by the BTS in 1990[9] with re­spect to asthma management: the abolition ofsymptoms, and the restoration of normal or bestpossible long term lung function. The safety com­ponent constitutes the avoidance of drug-relatedserious adverse events during the study periodoSuch an event was defined as:

• death• an event that was life-threatening, disabling or

incapacitating

• hospitalisation• congenital anomalies

• cancers• drug overdoses• clinical or laboratory events that led to drug

withdrawal.

PharmacoEconomics 1996 Sep; 10 (3)

Asthma Prophylaxis in Childhood 265

Cost-Effectiveness Ratias

Cast Data

The proportion of successfuHy treated patientsin either treatment group was then calculated foreach measure of effectiveness.

The proportion of successfuHy treated patientsin either treatment group was then divided by themean treatment cost per patient to produce cost­effectiveness ratios. These ratios define the cost ofachieving treatment success and were calculated asfoHows:

Efficacy Data

the average sodium cromoglycate recipient took 58rescue doses of salbutamol, incurring a slightlyhigher cost of f3.07. The lower acquisition cost offluticasone propionate relative to sodium cromo­glycate thus produces a total drug cost per patientin the fluticasone propionate group of f18.06,which is nearly half that of the sodium cromo­glycate patients at f32.40. Drug costs did not in­clude pharmacy expenses.

In terms of hospitalisations, only 3 were re­corded across both treatment groups during thestudy period - 1 for a patient receiving fluticasonepropionate and 2 for patients receiving sodiumcromoglycate. However, since in no case was thehospitalisation deemed by the clinician to be re­lated to the patient's therapy, none was included inthe assessment of mean treatment cost per patient.

Table III outlines the trial findings with respectto the various measures of effectiveness. Across aHof the measures employed, the proportion of pa­tients achieving successful treatment is signifi­cantly higher for fluticasone propionate than forsodium cromoglycate; fluticasone propionate had,on average, a success rate 20% higher than sodiumcromoglycate. The actual difference between theproportion of successfuHy treated patients, acrossaH effectiveness criteria, ranges from a minimumof 15% to a maximum of 27% in the favour offluticasone propionate.

mean cost per patient of treatmentprobability of being successfully treated

Results

cost effectiveness

305 children were recruited into the study. How­ever, 3 had unverifiable data and were sub­sequently excluded from the analysis. Another 77patients withdrew and received no study treatment,7 patients because of an adverse event, 9 becauseof failure to comply with the treatment protocoland 62 because of failure to meet the entry criteria(Note: this totals 78 because 1patient was recordedin 2 separate exclusion criteria columns of the phy­sician's record form - failure to comply with treat­ment protocol and failing the entry criteria). Of theremaining 225 patients, 110 were randomised toreceive fluticasone propionate and 115 to receivesodium cromoglycate.

Table II shows the clinical trial results for the 2treatment groups with regard to the cost of studymedication and mean cost of rescue medication re­quired during the study periodo

The study drug cost for a patient receivingfluticasone propionate was f15.89 (1995 values)compared with f29.33 for a patient receiving so­dium cromoglycate.[10] Patients receiving flutica­sone propionate also required, on average, less res­cue medication throughout the study period thanthose receiving sodium cromoglycate. While theaverage fluticasone propionate recipient took 41rescue doses of rescue medication at a cost of f2.17,

Table Il. Mean treatment cost in UK pounds (E; 1995 values) anduse of salbutamol (albuterol) rescue medication during the 8-weekstudy period

Fluticasone Sodiumpropionate cromoglycate(n =110) (n =115)

Cost of study drug (E)a,b 15.89 29.33

Mean number of rescue doses 41 58

Cost of rescue medication (E)b 2.17 3.07

Total cost per patient (E) 18.06 32.40

a Based on the assumption that patients received the full 56days of treatment.

b Acquisition cost obtained from the Monthly Index of MedicaISpecialities.[10]

© Adis International Limited. Ali rights reserved. PharmacoEconomics 1996 Sep; la (3)

266

Table III. Proportion of successfully treated patients based onvarious measures of successo Ali patients were free from seriousadverse effects in addition to the definitions shown

Definition of successful Proportion of successfully p-Valuetreatment treated patients (%)

fluticasone sodiumpropionate cromoglycate(n =110) (n =115)

Freedom from daytime 71 55 0.018symptoms in last3wk of study

Freedom from night-time 80 65 0.016symptoms in last3wk of study

Mean morning PEFR 70 43 <0.001>90% of predicted inlast 3wk of study

Mean evening PEFR 70 46 <0.001>90% of predicted inlast 3wk of study

Abbreviation: PEFR =peak expiratory flow rate.

Tolerability

Only 9 patients experienced adverse events dur­ing the study period that were deemed by the clini­cian to be related to, and led to withdrawal of, a studydrug. In the fluticasone propionate group, 1 patientdeveloped a cough. In the sodium cromoglycategroup, 8 patients experienced adverse events: 2 de­veloped episodes of heaving and vomiting, 2 de­veloped a cough and 2 reported an unpleasant tasteafter use of the study drug, 1experienced a burningsensation in the chest and the other developed asore throat.

Cost-Effectiveness Ratios

Table IV outlines the cost-effectiveness ratiosderived for each treatment group by combining the

Booth et al.

mean treatment cost per patient with the proportionof successfully treated patients, for each measureof successo The average cost per success for fluti­casone propionate was approximately f25.00 com­pared with f63.00 for sodium cromoglycate. Thedifference between the 2 groups' cost-effectivenessratios is consistently above f2? .00, resulting influticasone propionate being at least twice as costeffective as sodium cromoglycate, irrespective ofthe measure of effectiveness used in the analysis.

Sensitivity Analysis

The influence of the choice of effectivenessmeasure on the final results needs to be considered.Table V outlines the resuIts of sensitivity analysesusing 8 alternative measures of success: 4 in whichthe target value of percentage of predicted PEFR isvaried to 80% and 100% (for morning and eveningPEFR measurement) and 4 in which the conditionof 'no serious adverse events' is removed from themeasure of effectiveness.

The results indicate that, even if the definitionof success is altered across a range of variables,fluticasone propionate remains consistently morecost effective than sodium cromoglycate. When thetarget value of percentage predicted PEFR is low­ered to 80%, higher success rates are achieved byboth treatment groups, but the cost-effectivenessadvantages of fluticasone propionate over sodiumcromoglycate remain greater than 2 to 1. Raisingthe target to 100% of predicted PEFR yields suc­cess rates for fluticasone propionate that are approx­imately twice that of the sodium cromoglycate groupand increases the cost-effectiveness advantages toover 3 to 1 in favour of fluticasone propionate, for

Table IV. Cost-effectiveness ratios for fluticasone propionate (FP) and sodium cromoglycate (SC) based on various measures of successoAli patients were free from serious adverse effects in addition to the definitions shown

Definition of successful treatment Cost-effectiveness ratios (f:/success)

FP (n =110) SC (n =115)

SC: FP ratio

Freedom from daytime symptoms in last 3wk of study

Freedom from night-time symptoms in last 3wk of study

Mean morning PEFR >90% of predicted in last 3wk of study

Mean evening PEFR >900/0 of predicted in last 3wk of study

Abbreviation: PEFR =peak expiratory flow rate.

© Adis International Limited. Ali rights reserved.

25.44 58.91

2~58 4~85

2~80 7~35

25.80 70.43

2.3: 1

2.2: 1

2.9: 1

2.7: 1

Pharmac,oEconomics 1996 Sep; 10 (3)

Asthma Prophylaxis in Childhood 267

Table V. Sensitivity analysis results using alternative measures of effectiveness for patients treated with fluticasone propionate (FP) or sodiumcromoglycate (SC)

Definition of successful treatment Successfully treated patients Cost-effectiveness ratios SC: FP ratio(%) (E/success)

FP (n = 110) SC (n = 115) FP (n = 110) SC (n = 115)

No serious adverse effects and mean morning 82 55 22.02 58.91 2.7: 1PEFR >80% of predicted in last 3wk of study

No serious adverse effects and mean evening 78 61 23.15 53.11 2.3: 1PEFR >80% of predicted in last 3wk of study

No serious adverse effects and mean morning 58 29 31.14 111.72 3.6: 1PEFR >100% of predicted in last 3wk of study

No serious adverse effects and mean evening 60 33 30.10 98.18 3.3: 1PEFR >100% of predicted in last 3wk of study

Mean morning PEFR >90% of predicted in last 71 46 25.44 70.43 2.8: 13wk of study

Mean evening PEFR >90% of predicted in last 70 49 25.80 66.12 2.6: 13wk of study

Mean morning PEFR >80% of predicted in last 83 60 21.76 54.00 2.5: 13wkof study

Mean evening PEFR >80% of predicted in last 78 66 23.15 49.09 2.1 : 13wk of study

Abbreviation: PEFR = peak expiratory flow rate.

both moming and evening measurements. The re­moval of serious adverse events from the choseneffectiveness criteria does little to alter the finalresults, with fluticasone propionate remaining atleast twice as cost effective as sodium cromoglycate.

Examining the costs, there is only one compo­nent that can be varied realistically within the con­fines of this study - the amount ofrelief medicationrequired in the 2 patient groupings. Assuming pa­tients in the sodium cromoglycate group used norelief medication at all, this agent is stilliess costeffective than fluticasone propionate. This alterationcauses a small reduction in the average cost perpatient in the sodium cromoglycate group, but failsto compensate for the difference in success ratesbetween the 2 groups.

Discussion

This study examined the cost effectiveness offluticasone propionate and sodium cromoglycatein a group of children with asthma, aged 4 to 12years, who were candidates for prophylactic ther­apy. In terms of basic drug cost, an 8-week supplyof sodium cromoglycate costs over ;(13.00 morethan fluticasone propionate. The latter's cost advan-

© Adis International Limited. Ali rights reseNed.

tage is further extended when resource utilisationcosts - the use of rescue medication - are included.

In this study, fluticasone propionate also achievedbetter outcomes than sodium cromoglycate, irre­spective of the symptomatic or clinical measuresof success that were used. Therefore, when out­comes and costs are considered together, the ad­vantage of fluticasone propionate extends evenfurther, with the cost per successfully treated pa­tient associated with fluticasone propionate beinghalf that found for sodium cromoglycate.

The greater part of the difference in effective­ness between the 2 treatments was thus derivedfrom the superior efficacy of fluticasone propio­nate over sodium cromoglycate, rather than fromshort term tolerability differences.

As highlighted earlier, nondrug costs have notbeen included in deriving a mean treatment costper patient, since relevant data, such as GP surgeryvisits or emergency home visits, were not collectedin the clinical trial. If such data had been collectedit is unlikely that they would have significantlydiffered between the 2 treatment groups. Althoughthis omission may have implications in absoluteterms by understating the mean cost of achieving

PharmacoEconomics 1996 Sep; 10 (3)

268

success in either treatment group, it is unlikely toaffect the relative cost effectiveness result in favourof fluticasone propionate.

The length of study was predetermined by theclinical study data and, although a longer durationwould have been preferable, the prescribing notesfor sodium cromoglycate use as presented in theMonthly Index of MedicaI Specialities state that,after a 4- to 6-week trial, children who fail to re­spond may be given an inhaled corticosteroid.[10]There have been discussions with regard to studyduration and readers are directed to the paper byRutten-van M6lken et al.[11]

There exists no single, unequivocal, measure ofclinical outcome for asthma treatment. The effec­tiveness measures selected in this study were basedon the originaI BTS[9] guidelines, which concen­trated on asthma in adults. The measures of effec­tiveness that were used in this study were symptomcontrol, an assessment oflung function and the pre­sence of adverse events. These criteria have sub­sequently been recognised as appropriate effective­ness measures for asthmatic children in the revisedBTS guidelines. [5] While the revised guidelinesindicate that restoring best possible airway func­tioning, minimising adverse event incìdence, andcontrolling symptoms should be some of the aimsin paediatric asthma management, they also statethat there are additional aims for the managementof asthma, namely: (i) to enable normal growth tooccur in children and (ii) to minimise absence fromschool.

The clinical study used as the basis for this anal­ysis did not collect data on either of these factors.Both aims could be considered as possible outcomemeasures in the future, but would clearly requireclinical trials of a longer duration than this study.

© Adis International Limited. Ali rights reserved.

Booth et al.

The indirect costs of absenteeism from school arediscussed by Rutten-van M6lken et al. [11]

This study has shown that fluticasone propio­nate has cost and effectiveness advantages over so­dium cromoglycate and that, ofthe 2 treatments, fiuti­casone propionate was the most cost effective wayof improving PEFR and symptom control in agroup of children with asthma who had a clinicalrequirement for prophylactic therapy.

Acknowledgements

This study was supported by Glaxo Wellcome resources.

References1. Anderson HR. Is the prevalence of asthma changing? Arch Dis

Child 1989; 64: 172-52. CentraI Statistical Office. Annual Abstract of Statistics 1993.

London: HMSO, 1993: 83. Lenney W, Wells NEJ, O'Neill BA. The burden of paediatric

asthma. Eur Respir Rev 1994; 4 (18): 49-624. Anderson HR, Bailey PA, Cooper JS, et al. Morbidity and school

absence caused by asthma and wheezing illness. Arch DisChild 1983; 58: 777-84

5. British Thoracic Society. Guidelines for the management ofasthma. Thorax 1993; 48 Suppl.: Sl-24

6. Phillips GH. Structure - activity relationships oftopically activesteroids: the selection of fluticasone propionate. Respir Med1990; 84 Suppl. A: 19-23

7. Harding SM. Human pharmacology of fluticasone propionate.Respir Med 1990; 84A: 25-9

8. Price JF, Weller PH. Comparison of fluticasone propionate andsodium cromoglycate for the treatment of childhood asthma(an open parallel group study). Respir Med 1995; 89: 363-8

9. British Thoracic Society. Guidelines for the management ofasthma in adults. 1- Chronic persistent asthma. BMJ 1990;301: 651-4

lO. Prescribing notes for anti-inflammatories, non-steroid. MonthlyIndex of MedicaI Specialities 1995 Feb: 203

Il. Rutten-van Molken MPMH, Van Doorslaer EKA, Jansen MCC,et al. Cost effectiveness if inhaled corticosteroid plus bron­chodilator therapy versus bronchodilator monotherapy inchildren with asthma. PharmacoEconomics 1993: 4 (4): 235-10

Correspondence and reprints: Mr A.K. Morrison, The HealthEconomics Group, Glaxo Wellcome UK Limited, StockleyPark West, Uxbridge, Middlesex VB11 1BT, England.

PharmacoEconomics 1996 Sep; 10 (3)