outcomes of a web-based patient education program for asthmatic children &
TRANSCRIPT
Outcomes of a Web-Based PatientEducation Program for AsthmaticChildren and Adolescents*
Claus Runge, PhD; Josef Lecheler, MD; Michael Horn, MD;Jan-Torsten Tews, MD; and Marion Schaefer, PhD†
Background: Asthma is the most common chronic disease among children in Germany. Ap-proaches to reduce the burden of asthma include patient education to improve self-managementskills.Study objectives: We determined whether a continuous Internet-based education program (IEP)as an add-on to a standardized patient management program (SPMP) improves health outcomesof asthma patients at a favorable benefit-cost ratio.Patients and methods: A total of 438 asthmatic patients aged 8 to 16 years in 36 study centers wereenrolled during a 6-month period. We performed a prospective cost-benefit analysis alongside anonrandomized trial. At baseline and at 6 months and 12 months, health service utilization datawere collected.Interventions: Study participants were assigned to a control group and two intervention groups.Patients in both intervention groups participated in an SPMP. Additionally, patients in oneintervention group received the IEP.Results: Utilization of various health-care services decreased significantly in both interventiongroups. From a payer perspective, the benefit-cost ratio of the traditional education program was0.55. Adding the IEP improved the ratio (0.79). For patients with moderate or severe asthma, thebenefit-cost ratios were 1.07 and 1.42 (with IEP), respectively.Conclusions: The IEP offers the potential to decrease the burden of disease and to realizeincremental morbidity cost savings. Subgroup analysis demonstrated that within 1 year, thesavings exceed the intervention costs in patients with moderate or severe asthma.
(CHEST 2006; 129:581–593)
Key words: asthma; cost-benefit analysis; Internet costs; patient education; quality of life
Abbreviations: CG � control group; GP � general practitioner; IEP � Internet-based education program;ITT � intention to treat; PEF � peak expiratory flow; PPP � population per peer protocol; QoL � quality of life;SPMP � standardized patient management program
A sthma is the most common chronic disease inchildren, affecting approximately 10% of the
German pediatric population.1 There is a distinctsocial burden of asthma, not only for the sick child,but also for the caregiver and the society in general.At the same time, the economic impact due toextensive health-care utilization is high. Health-careexpenditures for asthmatic children are roughlythree times higher compared to children withoutasthma.2 A positive correlation has been describedbetween average total costs and degree of asthmaseverity in several investigations.3,4
Increased health-care expenditure and morbiditycan largely be attributed to poor asthma control,including underuse of antiinflammatory therapy, andpoor compliance.5 Therefore, international guide-lines recommend the increased use of controllertherapy to improve the overall asthma management6
and the participation in comprehensive patient edu-cation programs.7,8 Although there is evidence of thepositive economic impact of patient education,9–11
� 1% of all asthmatic children in Germany areinstructed in a systematic manner per annum. Mainreasons for the poor participation rates are insuffi-
Original ResearchASTHMA
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 581
cient incentives for providers, a lack of patientinformation about ways to be recruited, and inade-quate design of the programs when needs andinterests of children and adolescents are consid-ered.12
To address the observed shortcomings of theexisting training programs, an Internet-based educa-tion program (IEP) had been developed for asth-matic children aged 8 to 16 years. The IEP serves asan add-on to traditional approaches in patient edu-cation. The current study assesses the impact of theIEP on health outcomes and economic implications.It is a cost benefit analysis reporting all costs at year2001 level in Euros without further discounting.
Materials and Methods
Study Design
The current investigation was a nonrandomized, nonstratified,multicenter, real-life study comprising two intervention groupsand one control group (Fig 1). The investigation was conductedbetween July 2001 and December 2002 in 36 study centersnationwide, including general practitioner (GP) and specialistoffices as well as asthma outpatient facilities of hospitals.
Intervention Groups
All eligible patients were invited to join a standardized patientmanagement program (SPMP). They had the option to self select
additional educational activities with the additional IEP, providedthat individual access to the Internet was possible at home, at thestudy centers, or at school. Data collection in the interventiongroups was completed at two subsequent visits scheduled at 6months (visit 2) and 12 months (visit 3) after enrollment.
Control Group
Due to limited capacities in several study centers, numerouspatients were in waiting loops for educational courses. Patients onwaiting lists for a future (prescheduled) asthma education pro-gram were asked to build the control group (CG). Havingconsented to take part in the study, these patients were recruited6 months prior to their scheduled educational intervention.During the 6 months prior to the educational program and duringthe 6 subsequent months, patient outcomes under usual carewere documented. Follow-up of CG patients ended with theinitiation of the patient education program after visit 2 (end ofwaiting period; Fig 1).
At the baseline evaluation (visit 1), information was securedabout the child, including recent asthma activity, medicationadherence, use of rescue medication, lung function (peak expi-ratory flow [PEF], FEV1), health-care utilization (eg, number ofhospitalizations, number of outpatient visits at GP or specialist,number of emergency visits, medication costs), and quality of life(QoL). Baseline assessment also included demographic charac-teristics, allergy status, and classification of degree of asthmaseverity according to national recommendations.13 At all visits,the identical protocol was used in all study groups.
Clinical outcomes and health-care resource use data weredocumented by treating physicians based on (electronic) patientrecords. Documentation covered the 6-month periods prior tothe baseline and subsequent follow-up visits. An additionalcaregiver questionnaire provided information about patients’school absenteeism, caregivers’ loss of workdays due to theirchild’s asthma, and the number of asthma-related emergencies.
Finally, a patient questionnaire was distributed at all scheduledstudy visits. It contained a QoL questionnaire (KINDL; Berlin,Germany).14 The KINDL questionnaire is a generic QoL instru-ment applicable for self administration. It consists of 24 Likert-scale items in the dimensions of physical well-being, psycholog-ical well-being, self esteem, family, friends, and socialfunctioning. Furthermore, it is supplemented with a disease-specific asthma module consisting of six further items. Follow-upquestionnaires were mailed to GPs, patients, and caregivers incase patient education had taken place in a remote investigatorsite (four study centers are located in the Alps; country wide, GPsrefer their patients to these centers as the centers are specializedin treating asthma and delivering patient education programs inremote areas with low allergen exposure, eg, dust mite).
Intervention
Patients in both intervention groups took part in an SPMP. Ina series of educational sessions (5 � 2 h), asthma knowledge andasthma self-management skills were strengthened. Patients wereinstructed in several disease-related fields (eg, proper use ofdifferent inhaler devices, classification of reliever and controllermedication, avoidance of asthma triggers, limitation of allergenexposure, peak flow measuring, and its interpretation). Duringrole-playing exercises, critical everyday life situations of asthmapatients were simulated in small groups, and possible solutions todaily disease-specific problems were discussed. Content, scope,and results of the SPMP are reported in detail elsewhere.15–17
*From GlaxoSmithKline GmbH & Co. (Drs. Runge and Tews),Munich; CJD Asthmazentrum Berchtesgaden (Dr. Lecheler),Berchtesgaden; Forum Telemedizin GmbH (Dr. Horn), Munich;and Institut of Clinical Pharmacology (Dr. Schaefer), Humboldt-University Berlin, Berlin, Germany.†A list of participants is given in the Appendix.GlaxoSmithKline Germany has cofinanced the development ofthe IEP. GlaxoSmithKline has also financed the entire studyincluding an investigator bonus of 25 Euros for every completecontrol group patient documentation. Furthermore, GlaxoSmith-Kline has arranged three investigators meetings between 2001and 2003. Travel expenses were refunded at the initial meeting.The IEP was designed by the enterprise Forum TelemedizinGmbH, which is owned and headed by Dr. Horn, who providedprivate funds in addition to the financial support granted byGlaxoSmithKline.Dr. Lecheler has not received any honoraria from GlaxoSmith-Kline in this investigation, but in some cases travel expenses wererefunded by GlaxoSmithKline.Representing an institution that owns the intellectual propertyrights of a traditional patient education program, Dr. Lechelerhas advised Dr. Horn in designing the Internet education tool. Inthe past, Dr. Lecheler was a paid consultant/speaker in severalprojects with GlaxoSmithKline involvement. Professor Schaeferwas the doctoral thesis advisor of the first author, Dr. Runge,whose thesis deals with the IEP. In the past, Professor Schaeferhas raised third-party funds from different pharmaceutical com-panies.Drs. Runge and Tews are employees of GlaxoSmithKline.Manuscript received March 2, 2004; revision accepted October 5,2005.Reproduction of this article is prohibited without written permissionfrom the American College of Chest Physicians (www.chestjournal.org/misc/reprints.shtml).Correspondence to: Claus Runge, MD, Kunstwerkerstrasse 179,45136 Essen, Germany, e-mail: [email protected]
582 Original Research
Additional Education in the IEP Group
Patients in the intervention groups were allowed to self selectthe IEP as an additional module. The IEP (www.forum-telemedizin.de, now www.asthmax.de) includes an additionaleducational module with an asthma-related quiz and an interac-tive adventure game incorporating numerous virtual asthma-related situations that have to be managed adequately. Further-more, a repetition section displaying the educational material ofthe SPMP is included. In addition, the IEP provides a medicalmodule with individual medication plans, scheduled chats withasthma experts, and an on-line peak flow protocol that is to bemaintained by the patient. Communication of registered users,both patients and health-care providers, is feasible in chat roomsor by e-mail. Content for the IEP had been developed by amultidisciplinary team. Over a 6-month period, the IEP wastested successfully in a focus group including 50 children in aninpatient setting.18
Study Subjects
Eligibility criteria for study participation were age between 8and 16 years and the presence of a confirmed asthma diagnosisfor at least 1 year. No other asthma program participation prior tostudy entry was allowed. Sufficient knowledge of the Germanlanguage was assumed. Appropriate information about the con-cept of the study was given to all patients and their caregivers bythe investigators, and informed consent was obtained from allparticipants (intervention and control groups) before study entry.
Methods
Physician- and caregiver-reported resource-use data for asthmacare were collected. The primary study outcome measure was thecalculation of a benefit-cost ratio for each intervention strategyfrom both the perspective of health-care insurance and a societalperspective. Prices were adjusted to year 2001 prices. Since theobservational period did not exceed 1 year, neither prices noroutcomes were discounted.19 Secondary outcome measures in-
cluded QoL, lung function, use of rescue medication, andnumber of days absent from school due to asthma.
Direct Medical Costs
The following asthma-related morbidity measures were calcu-lated for each of the three study groups: average rate ofscheduled and unscheduled medical care visits, rate of emer-gency department visits, and frequency and length of inpatientstays. The monetary values of these measures were determinedaccording to recent estimates and national statistics.20,21 Dailyasthma medication costs were calculated based on physicians’prescription records.22 Costs for traditional patient educationwere derived from existing reimbursement contracts betweenpaymasters and providers and are on a comparable level nation-wide. A sickness fund with a market share of approximately 11%reimburses the additional costs of the IEP depending on fre-quency of use and performance shown by the patient. For thecalculation of incremental costs in the IEP group, it was assumedthat reimbursement was secured for all participating patients.
Direct Nonmedical Costs
Transportation costs were calculated on the basis of publicinner-city transport tariffs, assuming that for every medical carevisit a round-trip ticket was needed. Patients using the IEP hadto bear Internet access costs.
Indirect Costs
Indirect costs included caregivers’ loss of workdays due to theirchild’s asthma. Based on national statistics, the average dailygross earning was determined using the human-capital ap-proach.23 All unit costs and sources are summarized in Table 1.
Cost-Benefit Analysis
In all study groups, the average costs per patient during the 6months prior to baseline visit were determined (costs at visit 1)
Figure 1. Course of the trial in the three study groups. The SPMP started within 4 weeks after thebaseline visit. Patients in the IEP group went through the SPMP in combination with IEP during thesubsequent months.
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 583
and extrapolated to annual costs (costs at visit 1 � 2). In theintervention groups, the average annual costs during the obser-vational period were subtracted from these baseline costs inorder to arrive at the morbidity cost savings (savings � [costs atvisit 1 � 2] � [costs at visit 2 � costs at visit 3]).
We arrived at the adjusted benefit in the intervention groupsby accounting for potential savings or excess expenditures in theCG (adjusted benefit � savings � [costs CG at visit 1 � 2 – costsCG at visit 2 � 2]). The net benefit in the first year afterintervention was calculated by subtracting the intervention costsfrom the adjusted benefit. Finally, the benefit-cost ratio wasdetermined as the ratio of adjusted benefits and interventioncosts. The same method was used to calculate the incrementalsavings and benefits of the educational activities including theIEP when compared to standard patient education.
Analysis
Baseline demographic (sex, age) and baseline clinical variables(lung function) were compared between the groups using �2 testsfor categorical variables and t tests for continuous variables. Forcomparison of study data with normal distribution, we used theKolmogorov-Smirnov test. In case of normal distribution ofcontinuous variables, we used the t test for group-specificbefore/after comparisons. Bonferroni tests were applied for therespective group comparisons.
If no normal distribution was assumed following the Kolmog-orov-Smirnov test, patient data for the baseline and follow-upperiods were compared intraindividually using nonparametricanalysis methods. We used the Wilcoxon signed-rank test forcontinuous variables. Possible differences between study groups
were analyzed using the Mann-Whitney U test. A two-tailed pvalue � 0.05 was regarded as statistically significant. In order toensure comparability between groups, all health-care resourcedata were adjusted to the distribution of asthma severity degreesin the largest of the three study groups.
The weighted average of asthma-severity specific rates (eg,costs), with weights taken from the SPMP group (which wasdefined as the “standard population,” as it is the largest one inpatient numbers), provides for each other population (IEP andcontrol group) a summary rate (eg, costs) that reflects the coststhat would have been expected if the populations compared hadhad identical distributions of asthma severity degrees. We havechosen asthma severity as weighting factor as it has a strongimpact on asthma-related costs.3,4 Similar approaches are appliedin other disease areas, eg, in comparisons of cancer mortalityrates over different decades with changing age structures ofunderlying populations.24
Results are presented as severity-adjusted mean scores for thepopulation per peer protocol (PPP), for subgroups, and for theintention-to-treat (ITT) population. As a standard in healtheconomic evaluations, one-way sensitivity analyses were per-formed in order to test for the influence of variations in drugtherapy adherence, outpatient visit costs, and intervention costson the benefit-cost ratio. The data were analyzed using statisticalsoftware (Version 10.0; SPSS; Chicago, IL).
Results
The original study sample consisted of 358 pa-tients. A further 80 patients were interested in
Table 1—Cost Items, Data Sources, and Unit Costs in 2001 Euros
Cost Items Data Source Unit Cost, €
Direct costsGP consultation fee (first/subsequent visit in a
quarter)*National catalogue of fees for services; definition of
routine services by experts according to Woehl2034.20/23.60
Specialist consultation fee (first/subsequent visit in aquarter)†
National catalogue of fees for services; definition ofroutine services by experts according to Woehl20
43.40/34.00
Hospital day (including allowances granted by officialauthorities)
Deutsche Krankenhausgesellschaft33 346.70
Emergency department visit National catalogue of fees for services; definition ofroutine services by experts according to Woehl20
26.00
Ambulance transport in case of emergency Bavarian Red Cross34 461.78Emergency physician answering an emergency call Bavarian Red Cross34 512.78GP answering an emergency call National catalogue of fees for services; definition of
routine services by experts according to Woehl2032.80
Traditional patient education Contract between health-care insurance companies andthe national working group of asthma patienteducation
541.14
IEP, additional fee Contract between Barmer Ersatzkasse (sickness fund)and IEP provider
43.99
Medication Rote Liste22 Varies by medicationDirect, nonmedical costs
Public transport per scheduled or unscheduled visit atGP or specialist
Ministry of Economics, Energy, and Transport35 3.10
Internet fees (IEP group only) within 6 mo Deutsche Telekom,36 IEP provider 14.40Indirect costs
Productivity loss per day of absence from work21 Statistisches Bundesamt21 94.70
*According to the German fee for service system, this includes consultation fee, therapeutic advice, spirometry, totaling in 855/590 credit points;assumed point value, 0.04€.
†According to the German fee for service system, this includes consultation fee, therapeutic advice, body plethysmography, totaling in 1,085/850credit points; assumed point value, 0.04€.
584 Original Research
participating but could not be included as they didnot meet inclusion criteria. Complete medical re-source use data at all follow-up visits were availablefor 178 patients (49.7%) who were defined as PPP(Table 2). Medical resource use data of 56 studycompleters were insufficient for analysis (15.6%).
Reasons for study discontinuation (n � 124;34.6%) were as follows: relocation of patients(n � 6), withdrawal of caregiver consent (n � 6),nonattendance of scheduled follow-up visits(n � 11), no returning of mail questionnaires(n � 41), and relocation of personnel in two studycenters (n � 19). Reasons for dropouts were notdocumented in 41 cases. Patient characteristics aresummarized in Table 3.
Compared to both intervention groups, signifi-cantly more patients in the CG were male. In theIEP group, significantly more patients presentedwith allergies when compared to the SPMP group.
Although not statistically significant (p � 0.17),the proportion of patients with moderate or severeasthma in the IEP group (65.9%, n � 29) tended tobe higher than in the SPMP group (44.2%; n � 38)and in the CG (39.6%; n � 19). At baseline, a total of116 patients (65.2%) received inhaled corticoste-roids, with no statistically significant differencesbetween groups.
Table 4 shows major characteristics of the PPPand the group of noncompleters at baseline assess-ment. Patients who did not complete the studytended to use more health-care resources. Amongthe patients who were assigned to the interventiongroups, study completers had significantly lowermorbidity costs at baseline when compared to non-completers.
The medical outcomes of the PPP are summarizedin Table 5. Various improvements were reported inall study groups, but patients using the IEP were theonly ones to experience a statistically significant
short-term reduction in asthma-related emergenciesand a long-term decrease in the daily use of rescuemedication. During the follow-up year, morbiditycosts savings in the IEP group were due to areduction in physician consultations (56%), asthma-related emergency treatments (75%), and daily med-ication costs (52%) [Fig 2].
Compared to the year prior to baseline, averagemorbidity cost savings during the first year afterintervention made up 461€ from a health insuranceperspective and 467€ from a societal perspective inthe IEP group. Patients in the IEP group hadsignificant reductions (p � 0.05) in both days offschool (� 75%) and average daily use of rescuemedications (� 77%) when comparing visit 2 datawith baseline.
In the CG, significant improvements (p � 0.05)were observed when analyzing the number of out-patient visits (� 6%) and the number of lost schooldays (� 35%) before and after the trial initiation.Benefits in the usual care group reached 1.55€ froma paymaster perspective and 58€ from a societalperspective. Patients receiving the SPMP alone hadbetter outcomes than control subjects, eg, signifi-cantly (p � 0.05) less physician consultations(� 44%), emergency treatments (� 67%), and daysoff school (� 71%) in the follow-up period.
In contrast to the IEP group, the reduction in theaverage daily use of rescue medication among SPMPpatients did not reach a statistically significant level.Compared to the IEP group, it took longer to realizea statistically significant reduction in the number ofemergencies. Morbidity cost savings in the tradi-tional patient education group (SPMP) were lowerwhen compared to IEP group patients and added upto 301€ from a paymaster perspective and 333€ froma societal perspective. Adjusted benefit, net benefit,and benefit-cost-ratio applicable to the two interven-tion groups are summarized in Table 6.
From a health-care insurance perspective, it cost585€ to deliver education including the IEP; 461€
was saved in health expenses. Adjusting for benefitsin the control group, 0.79€ was saved for every 1.00€
spent in the first year after intervention.Adding the IEP to the SPMP yields incremental
morbidity cost savings of 160€ (direct costs) at anadditional average cost of 44€ when compared to theSPMP alone. This translates into a benefit-cost ratioof 3.65.
Lung function as measured with PEF increasedsignificantly (p � 0.05) in all study groups. FEV1 didnot change significantly over time in any of the threegroups. Changes in lung function are shown inTable 7.
Table 2—Number of Patients Per Study Group andEligibility for Analysis
Variables CG
Intervention Groups
TotalSPMP SPMP Plus IEP
ITT population* 85 127 146 358IEP not used (n � 66) 66† 66Study discontinuation 17 80 27 124Resource-use data
incomplete20 27 9 56
PPP‡ 48 86† 44 178
*All patients with a baseline visit.†Patients who did not start the IEP after completing the SPMP wereevaluated as regular SPMP patients.
‡All patients with complete medical resource-use data.
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 585
Subgroup Analysis
Subgroup analysis of the main study outcomeshows higher benefit-cost ratios for patients withmoderate persistent or severe persistent asthma(1.42 in the IEP group) and for patients with anyemergency 6 months prior to the study entry (1.21 inthe IEP group) [Table 6].
Sensitivity Analysis
In the framework of one-way sensitivity analysis,several cost items and assumptions have been mod-ified in order to test for their impact on the benefit-cost ratio (Table 8). The ratio is most sensitive tochanges in the drug adherence. Assuming a constant
reduction in morbidity cost savings in the yearfollowing the observational period, ie, allocating just50% of the educational costs to the observationalperiod, the benefit-cost ratio becomes � 1 for theIEP group while getting close to 1 for patients withinthe SPMP group. The benefit-cost ratio is morerobust to elevations of GP fees and the IEP reim-bursement level.
QoL
QoL as indicated by the KINDL total QoL scoreincreased significantly (p � 0.05) in both interven-tion groups. QoL decreased to a little but nonsignif-icant extent in the CG. Statistically significant(p � 0.05) improvements were observed in four
Table 3—Demographic and Clinical Variables by Study Group*
Variables CG (n � 48) SPMP (n � 86) SPMP Plus IEP (n � 44)
Age, yr 11.5 � 2.9 11.1 � 2.4 11.0 � 2.2Male gender 38 (79.2)†‡ 47 (54.7) 29 (65.9)Height, cm 152.1 � 19.1 150.1 � 15.8 148.4 � 13.9Weight, kg 51.0 � 29.5 45.6 � 17.4 43.8 � 15.2Degree of asthma severity
Mild intermittent 7 (14.6) 12 (14.0) 3 (6.8)Mild persistent 20 (41.7) 36 (41.9) 12 (27.3)Moderate persistent 19 (39.6) 37 (43.0) 27 (61.4)Severe persistent 1 (1.2) 2 (4.5)Not answered 2 (4.2)
Asthma diagnosed since, yr 5.9 � 4.2 5.5 � 3.5 6.3 � 3.6Any allergy 43 (89.6) 68 (79.1) 43 (97.7)‡Seasonal allergic rhinitis 20 (41.7) 41 (47.6) 24 (54.5)Atopic asthma 13 (27.1) 25 (29.1) 11 (25.0)
*Data are presented as mean � SD or No. (%).†p � 0.05 compared to SPMP plus IEP.‡p � 0.05 compared to SPMP group.
Table 4—Comparison of Study Completers and Dropouts
Variables Completers (n � 178) Noncompleters (n � 180) p Value
Degree of asthma severity at baseline assessment 0.93Mild intermittent 22 (12.4) 28 (15.6)Mild persistent 68 (38.2) 47 (26.1)Moderate persistent 82 (46.1) 62 (34.4)Severe persistent 3 (1.7) 10 (5.6)Missing information 3 (1.7) 33 (18.3)
Female sex 64 (36.0) 69 (38.3) 0.47Six months prior to baseline
Referrals to hospital, No. 0.06 � 0.30 0.13 � 0.50 0.11Inpatient days, No. 0.13 � 0.84 0.67 � 3.01 0.15School absenteeism, d 2.01 � 4.29 3.86 � 11.65 0.33Daily drug therapy costs, € 1.45 � 1.22 1.75 � 1.61 0.10Consultations, No. 3.84 � 4.19 4.32 � 4.86 0.81Average daily use of reliever medication 0.28 � 0.73 0.35 � 1.02 0.60
At baseline assessmentTotal QoL as measured with the KINDL questionnaire 75.5 � 11.78 76.1 � 12.17 0.47
Intervention groups (n � 273)Average total costs at baseline, € 527.78 � 398.19 (n � 130) 669.00 � 648.26 (n � 143) 0.007
*Data are presented as No. (%) or mean � SD.
586 Original Research
Tab
le5—
Med
ical
Res
ourc
e-U
seD
ata,
Use
ofR
escu
eM
edic
atio
n,an
dSc
hool
Ab
sent
eeis
min
the
Ob
serv
atio
nP
erio
d
Prog
ram
CG
(n�
48)
SPM
P(n
�86
)SP
MP
Plus
IEP
(n�
44)
pV
alue
,K
rusk
al-W
allis
pV
alue
,M
ann-
Whi
tney
Med
ian
Mea
n*
pV
alue
(T1
vsT
0,W
ilcox
on)
Med
ian
Mea
n*
pV
alue
(T1/
T2
vsT
0,W
ilcox
on)
Med
ian
Mea
n*
pV
alue
(T1/
T2
vsT
0,W
ilcox
on)
Bas
elin
evi
sit
(6m
opr
ior
tost
udy
entr
y)Ph
ysic
ian
cons
ulta
tions
,No.
23.
53
3.4
45.
20.
11H
ospi
tald
ays,
No.
00.
30
0.1
00.
10.
16E
mer
genc
ies,
No.
00.
40
0.6
00.
40.
99W
orki
ngda
yslo
st(c
areg
iver
s),N
o.0
0.51
00.
650
0.25
0.23
Dai
lyth
erap
yco
sts,
€0.
781.
270.
771,
701.
731.
980.
25D
aily
use
ofre
scue
med
icat
ion,
No.
00.
200.
030.
230
0.26
0.16
Day
sab
sent
from
scho
ol,N
o.0
2.0
03.
41.
54.
00.
78V
isit
1(0
to6
mo
afte
ren
rollm
ent)
Phys
icia
nco
nsul
tatio
ns,N
o.1
3.3
0.04
1.5
2.3
0.01
22.
70.
020.
49H
ospi
tald
ays,
No.
00
0.18
00.
30.
950
0.2
0.93
0.43
Em
erge
ncie
s,N
o.0
0.2
0.53
00.
30.
260
00.
020.
040.
03(I
EP
vsSP
MP)
Wor
king
days
lost
(car
egiv
ers)
,No.
01.
070.
070
0.56
0.67
00.
140.
380.
77T
hera
pyco
sts,
€0.
731.
370.
880.
681.
090.
071.
551.
570.
190.
13D
aily
use
ofre
scue
med
icat
ion,
No.
00.
120.
350
0.21
0.14
00.
100.
180.
10D
ays
abse
ntfr
omsc
hool
,No.
01.
30.
050
1.7
0.01
01.
30.
004
0.01
0.01
(CG
vsSP
MP)
Vis
it2
(6to
12m
oaf
ter
enro
llmen
t)Ph
ysic
ian
cons
ulta
tions
,No.
11.
90.
001
2.3
0.00
20.
63H
ospi
tald
ays,
No.
00
0.07
00
0.56
0.22
Em
erge
ncie
s,N
o.0
0.2
0.01
00.
10.
030.
68W
orki
ngda
yslo
st(c
areg
iver
s),N
o.0
0.17
0.89
00.
250.
310.
93D
aily
ther
apy
cost
s,€
0.41
0.97
0.03
0.49
0.93
0.01
0.15
Dai
lyus
eof
resc
uem
edic
atio
n,N
o.0
0.21
0.13
00.
060.
020.
25D
ays
abse
ntfr
omsc
hool
,No.
01.
00.
001
01.
00.
001
0.14
*Adj
uste
dfo
ras
thm
ase
veri
ty.
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 587
domains of the KINDL questionnaire among pa-tients using the IEP and in three domains amongpatients receiving the SPMP (Fig 3). Statisticallysignificant improvements (p � 0.05) were also ob-served in the asthma-specific module in both inter-vention groups.
IEP Use
On average, patients were logged in to the IEP for2 h/mo, as indicated by the provider who maintainsthe IEP. The peak flow protocol section, the patientchat room, and the asthma adventure game were themost widely used modules.
Discussion
The positive impact of the SPMP is fortified by theadditional IEP as demonstrated in several outcomemeasures, including morbidity costs savings, QoL,absenteeism from school, number of asthma-relatedemergencies, and use of short-acting -agonists. Oursubgroup analysis demonstrates that, within 1 year,morbidity cost savings exceed the intervention costsin patients who belong to risk groups (benefit-costratio � 1). Similar findings that reflect the influenceof target population characteristics on the results canbe found for other asthma education programs in theliterature.8–10
Unsurprisingly, PEF increased significantly in our
growing children in all three groups over time.Presumably, sample size was not big enough todetect any statistically differences in PEF gain be-tween groups if there were any. At study entry, weobserved minor impairment in lung function asmeasured by the FEV1. We did not observe anysignificant improvements in FEV1 values in thecourse of the investigation in any of the groups. Sincelung function (PEF, FEV1) has not been recordedcontinuously throughout the study, missing dataprevent us from drawing conclusions of the effects ofthe patient education program on lung function.However, several aspects remain to be discussed,since particularities in the study design and thepatient population might limit the general transfer-ability of our findings.
Lack of Randomization
As far as patient characteristics are concerned,significantly more patients in the CG were malecompared to the intervention groups. Moreover,lower initial morbidity costs in the CG and thedissimilar distribution of asthma severity degreesreflect a selection bias within the framework of thechosen naturalistic study design. As asthma severityis the main predictor of health-care costs, we haveaddressed this issue by weighting the group-specificcosts according to the underlying distribution ofasthma severity. This approach gave us a valid ap-proximation for the primary study outcome.
Figure 2. Cost development during the observation period. Results are shown as average costs perpatient and 6-month period. T0 � baseline; T1 � visit 2; T2 � visit 3.
588 Original Research
There are various possible explanations for thetrend to include more severely ill patients in theintervention groups. First of all, physicians mighthave a tendency to offer a new technology to more
severely ill patients, a phenomenon that is knownfrom other disease areas as “launch bias.” Addition-ally, asthma-related complications (eg, hospitaliza-tion or frequent asthma attacks) could be drivingfactors. At the same time, negative experiences suchas emergency department visits might increase themotivation of patients and caregivers to take part ineducational activities.
The problem of positive selection could have beenavoided by randomization of patients. This was notdone for two reasons: First of all, potential candi-dates with poor asthma control and an urgent needfor educational activities could have been preventedfrom being randomized to one of the two interven-tion groups. The study investigators had rejected thisapproach categorically at an initial study conference:no one was ready to accept that severely ill patientswould be barred from a scientifically proven inter-vention such as the SPMP in case of being random-ized to a CG.
The second reason for not randomizing patientswas study feasibility. If patients without constantaccess to the Internet had been randomized to the
Table 6—Results of the Cost-Benefit Analysis*
Variables CG (n � 48) SPMP (n � 86) SPMP Plus IEP (n � 44)
Intervention cost 541.14 585.13Cost savings
Paymaster perspective 1.55 300.78 461.45Societal perspective 57.50 333.20 467.05
Adjusted benefitPaymaster perspective 299.23 459.90Societal perspective 275.70 409.55
Net benefitPaymaster perspective � 241.91 � 125,23Societal perspective � 218.20 � 118.08
Benefit-cost ratioPaymaster perspective 0.55 0.79Societal perspective 0.51 0.70
Subgroup analysisAdjusted benefit (paymaster perspective)
Moderate or severe asthma 0 (n � 19) 579.02 (n � 38) 830.88 (n � 29)Any emergency 6 mo prior to study entry 0 (n � 12) 600.67 (n � 31) 708.00 (n � 14)ITT population 0 (n � 85) 183.99 (n � 127) 245.75 (n � 146)
Benefit-cost ratio (paymaster perspective)Moderate or severe asthma 1.07 1.42Any emergency 6 mo prior to study entry 1.11 1.21ITT population 0.34 0.42
Univariate sensitivity analysisBenefit-cost ratio (paymaster perspective)
50% reduction in SPMP program costs 0.98 1.28IEP costs twofold higher
Entire population 0.64Patients with moderate or severe asthma 1.22
25% reduction in drug adherenceEntire population 0.35 0.51Patients with moderate or severe asthma 1.13 0.86
Elevation of GP reimbursement on specialist level 0.52 0.68
*Results are shown as mean per patient in € unless otherwise noted. Calculations are based on severity-adjusted costs.
Table 7—Results of Lung Function Tests
Variables
PEF FEV1
No.Mean � SD,
L/min No.Mean � SD,
L/min
SPMP plus IEPBaseline 42 340 � 86* 35 99.2 � 12.2†Visit 1 33 363 � 89* 16 96.3 � 14.0Visit 2 32 368 � 92* 14 89.6 � 11.4
CGBaseline 46 324 � 98 38 90.8 � 14.5Visit 1 41 341 � 110* 32 93.4 � 11.2
SPMPBaseline 70 347 � 94 69 98.9 � 14.7‡Visit 1 63 366 � 101* 29 98.4 � 18.9Visit 2 49 376 � 97* 25 98.1 � 15.7
*p � 0.05 vs baseline visit (Wilcoxon test).†p � 0.035 vs CG (Bonferroni test).‡p � 0.014 vs CG (Bonferroni test).
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 589
IEP group, patient recruitment numbers would havedropped significantly. Since there is just a limitedpool of providers offering a small number of educa-tional courses per year, the recruitment periodwould have been far too short.
The benefit-cost ratio of 0.55 for the SPMP aloneis comparable to results previously reported by Clarket al9 and Scholtz et al17 for their approaches toeducate preselected patient groups. They realizedbenefit-cost ratios of 0.62 and 0.61, respectively.
As to the efficiency of patient education in se-lected risk groups, our findings are also in line withformer studies: for every US dollar spent in educa-tional activities, Lewis et al,25 Clark et al,9 andO’Brien26 reported savings of $2.43, $11.22, and$1.48, respectively.10 This compares to savings of1.42€ for every Euro spent (SPMP plus IEP) inpatients with moderate persistent or severe persis-tent asthma and savings of 1.21€ in patients with aprior hospitalization, respectively.
When program costs (585€ in the IEP group) arecompared to the average cost-of-illness level amongstudy individuals at baseline (approximately 1,200€
per annum in the IEP group), it is not surprising thatcost savings did not exceed program costs in theentire study population within the first year afterintervention. But given the significant decline inmorbidity costs, it can be assumed that a break-evenfor the entire study population might be arrived at inthe subsequent year. In the light of evidence gener-ated in earlier studies27,28 that demonstrated thatpatient education programs yield further cost savingsin the long term, it is conceivable that this might alsobe true for the SPMP (plus IEP). Nevertheless, thisassumption will need to be established empirically.
As our findings are based on the PPP, which islikely to represent the most compliant and motivatedpatients, we reckon that our results might be toooptimistic. This is underlined by the results calcu-lated for the population ITT. After replacing missingvalues by the group-specific averages, lower benefit-cost ratios were observed for the ITT population.
Noncompleters in the two intervention groups werecharacterized by significantly higher initial morbiditycosts at baseline assessment when compared to studycompleters. Although scope for improvement washigher among these noncompleters, the educationprograms did not turn out to be as effective in thesepatients. Consequently, it will be of importance todevelop strategies that will improve asthma manage-ment and compliance for these target groups.
The majority of patients enrolled in this study(65.2%) received inhaled steroids at baseline assess-ment. In contrast, earlier findings29 showed substan-tial underuse of inhaled steroids in patients withasthma in Germany. According to Lagerlov et al,30
only 31% of all asthmatic patients received inhaledsteroids in a German study sample. An analysis fromMaziak and colleagues31 proved that even amongchildren with diagnosed asthma and � 12 wheezingattacks in the preceding year, only 42% had usedinhaled steroids and only 21% reported regular use.As prescription of inhaled steroids to children withpersistent symptoms is considered an important in-dicator of good adherence to asthma treatmentguidelines, it can be assumed that the investigators inthis study represent a positive selection of physiciansfostering the need for antiinflammatory therapy asrecommended.
In this context, it is remarkable that drug costs inthe intervention groups decreased significantly, al-though past queries have shown the opposite phe-nomenon in the aftermath of educational activities.32
Obviously, there is no backlog demand in the inves-tigated population, and better compliance has givenleeway to dose reductions, at least in the interventiongroups.
Another hint at a relatively good asthma control inthe pre-run to the current survey is the small numberof hospitalizations recorded for study participants. Incontrast to population-based surveys in which hospi-talization costs make up between 53% and 72% of alldirect costs,3 the initial range of 3% (IEP group) to7% (SPMP group) in our study documents the
Table 8—Calculation of Adjusted Costs*
Variables
SPMP Group PlusIEP Group,
No. (%)SPMP Group,
No. (%)Direct CostsGroup A, €
Formula forCalculating
Weighted Costs
ResultWeighted
Cost, €
Intermittent 3 (7) 12 (14) 455 455 � 0.14/0.07/44 21Mild persistent 12 (27) 36 (42) 519 519 � 0.42/0.27/44 18Moderate persistent 27 (61) 37 (43) 582 582 � 0.61/0.43/44 18Severe persistent 2 (5) 1 (1) 1,351 1,351 � 0.01/0.05/44 270Total 44 (100) 86 (100) 327
*In this example, the weighted average cost in group A is 327€, instead of 591€ (unweighted). In other words, if group A had been constitutedas group C in terms of asthma severity distribution, we would have expected costs of 327€. We used this common method to compare groupresults.
590 Original Research
Figure 3. Changes in QoL as measured with the KINDL questionnaire. Results are presented bydomains and as a total score for baseline and follow-up visits. See Figure 2 legend for expansion ofabbreviations.
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 591
potential selection bias of well-controlled asthmaticswhen compared to the average.
Despite the above-average quality level of care atbaseline, study participants benefited from patienteducation. In the second half of the observationperiod, only one patient (2%) in the IEP group andfour patients (5%) in SPMP group had an emergencyevent. This approximates the Global initiativefor Asthma recommendations (available at www.ginasthma.com) that demand the absence of anyasthma-related emergency.
Health-care utilization data and work absence ofcaregivers were recorded retrospectively at each ofthe study visits. This might raise the presumptionthat these data are subject to a recall bias. However,doctors performed out major parts of this documen-tation (eg, number of outpatient visits) based on(electronic) patient records. If necessary, patientsand caregivers provided additional information onnumber and devolution of emergencies (emergencydepartment visits and ambulance transports). Thesedata seem to be easily recollected by patients andcaregivers due to their drastic consequences. Fur-thermore, a precedent study27 has proven a highconsistency between the data provided by patientsand that provided by the health insurance agencies.Therefore we consider the potential overestimationor underestimation of health-care resource use asminimal.
We find our results encouraging enough to antic-ipate further implementation activities in coopera-tion with local and regional paymasters. In eithercase, regional campaigns boosting the use of the IEPwill have to be accompanied by continuous economicevaluations in order to verify its long-term effects onmedical and economic outcomes and to quantify theeffects on the benefit-cost ratio when both patientand provider groups will be expanded.
Appendix
The supporters and active study investigators in the followinginstitutions participated in this study: CJD Berchtesgaden,Klinik Schonsicht, Klinik Santa Maria, Marien-Hospital Wesel,Kinderklinik Gilead, Prinzregent Luitpold Kinderklinik, PraxisDr. Laub, Institut fur Verhaltensmedizin IAV, Praxis Dr.Straub/Dr. Kohler, Praxis Dr. Blum/Dr. Butsch v.d. Heydt,Praxis Dr. Averbeck, Praxis Dr. Bolz, Kinderklinik des Klini-kums Nurnberg-Sud, Praxis Dr. Bulle, Praxis Dr. Maier,Fachkliniken Wangen, Asthmazentrum Rhein-Main, PraxisDr. Weihaar, Jugenddorf Dortmund, Universitatsklinik Er-langen, Praxis Dr. Potthast, Praxis Dr. Weber, St. JohannesHospital, Praxis Dr. Gall, Praxis Dr. Silbermann, Kreisklini-kum Traunstein, Praxis Dr. Lindner, IAV Berlin, Praxis Dr.Bastian, Praxis Dr. Brosi, Praxis Dr. Shetty, Praxis Dr. Wasser,Praxis Dr. Liefring, Praxis Drs. Schleupner/Boudriot, KlinikHochried, Dr. Meierjuergen/Barmer Ersatzkasse.
ACKNOWLEDGMENT: The authors thank all patients andtheir relatives for their cooperation in this study.
References1 Weiland SK, von Mutius E, Hirsch T, et al. Prevalence of
respiratory and atopic disorders among children in the Eastand West of Germany five years after unification. Eur RespirJ 1999; 14:862–870
2 Lozano P, Sullivan SD, Smith DH, et al. The economicburden of asthma in US children: estimates from the NationalMedical Expenditure Survey. J Allergy Clin Immunol 1999;104:957–963
3 Graf von der Schulenburg JM, Greiner W, Molitor S, et al.Cost of asthma therapy in relation to severity: an empiricalstudy [in German]. Med Klin 1996; 91:670–676
4 Godard P, Chanez P, Siraudin L, et al. Costs of asthma arecorrelated with severity: a 1-yr prospective study. Eur RespirJ 2002; 19:61–67
5 Barnes PJ, Jonsson B, Klim JB. The costs of asthma. EurRespir J 1996; 9:636–642
6 Bousquet J. Global Initiative for Asthma (GINA) and itsobjectives. Clin Exp Allergy 2000; 30(Suppl 1):2–5
7 Clark NM. Asthma self-management education. Chest 1989;95:1110–1113
8 Guevara JP, Wolf FM, Grum CM, et al. Effects of educa-tional interventions for self management of asthma in chil-dren and adolescents: systematic review and meta-analysis.BMJ 2003; 326:1308–1309
9 Clark NM, Feldman CH, Evans D, et al. The impact of healtheducation on frequency and cost of health care use by lowincome children with asthma. J Allergy Clin Immunol 1986;78:108–115
10 Volmer T. Economic considerations of patient education [inGerman]. Pneumologie 1997; 51:850–857
11 Smith DH, Malone DC, Lawson KA, et al. A nationalestimate of the economic costs of asthma. Am J Respir CritCare Med 1997; 156:787–793
12 Expert committee report for the German Ministry of Health2000/2001 [in German]. Available at: www.svr-gesundheit.de/gutacht/gutalt/gutaltle.htm 2002. Accessed February 17, 2006
13 Wettengel R. Recommendations for the treatment of asthmain adults and children [in German]. Pneumologie 1998;52:591–601
14 Ravens-Sieberer U, Bullinger M. Assessing health-relatedquality of life in chronically ill children with the GermanKINDL: first psychometric and content analytical results.Qual Life Res 1998; 7:399–407
15 Szczepanski R, Konning J, Lob-Corzilius T, et al. Analysis ofthe current status of asthma education for children andadolescents in German-speaking countries [German]. Pneu-mologie 1993; 47:583–587
16 Szczepanski R, Gebert N, Hummelink R, et al. Outcome ofstructured asthma education in childhood and adolescence [inGerman]. Pneumologie 1996; 50:544–548
17 Scholtz W, Haubrock M, Lob-Corzilius T, et al. Cost-effectiveness studies of ambulatory educational programs forchildren with asthma and their families [in German]. Pneu-mologie 1996; 50:538–543
18 Runge C, Tews JT, Lecheler J, et al. Do young asthmaticsaccept patient education via internet? Findings of a study totest a new approach. Eur J Med Res 2003; 7(suppl 1):68
19 Drummond MF, O’Brien B, Stoddart GL, et al. Methods forthe economic evaluation of health care programmes. 2nd ed.Oxford, UK: Oxford University Press, 1997; 107–108
20 Woehl A. The usefulness of health economic evaluations inthe health care system: the example of a cost-benefit-analysis
592 Original Research
of an asthma education program [in German]. DiplomaThesis: 2000; 44
21 Official statistic yearbook 2001 [in German]. Stuttgart, Ger-many: Metzler-Poeschel, 2001; 442
22 Rote Liste 2002. Aulendorf, Germany: Editio Cantor VerlagGmbH, 2002
23 German recommendations for health care economic evalua-tion studies: revised version of the Hannover consensus [inGerman]. Med Klin 2000; 95:52–55
24 Hennekens CH, Buring JE. Measures of disease frequencyand association. In: Mayrent SL, ed. Epidemiology in medi-cine. Boston, MA: Little, Brown and Company, 1987; 54–98
25 Lewis CE, Rachelefsky G, Lewis MA, et al. A randomizedtrial of A.C.T. (asthma care training) for kids. Pediatrics 1984;74:478–486
26 O’Brien KP. Managed care and the treatment of asthma. JAsthma 1995; 32:325–334
27 Trautner C, Richter B, Berger M. Cost-effectiveness of astructured treatment and teaching programme on asthma.Eur Respir J 1993; 6:1485–1491
28 Weinstein AG, Mckee L, Stapleford J, et al. An economicevaluation of short-term inpatient rehabilitation for childrenwith severe asthma. J Allergy Clin Immunol 1996; 98:264–273
29 Vermeire PA, Rabe KF, Soriano JB, et al. Asthma control and
differences in management practices across seven Europeancountries. Respir Med 2002; 96:142–149
30 Lagerlov P, Veninga CC, Muskova M, et al. Asthma manage-ment in five European countries: doctors’ knowledge, atti-tudes and prescribing behaviour; Drug Education Project(DEP) group. Eur Respir J 2000; 15:25–29
31 Maziak W, von Mutius E, Beimfohr C, et al. The manage-ment of childhood asthma in the community. Eur Respir J2002; 20:1476–1482
32 Neri M, Migliori GB, Spanevello A, et al. Economic analysisof two structured treatment and teaching programs onasthma. Allergy 1996; 51:313–319
33 Hospital statistics: cost structure of the hospitals in 2001 [inGerman]. Available at: www.dkgev.de. Accessed November12, 2002
34 Bavarian Red Cross. Emergency medical services in Bavaria[in German]. Available at: www.brk-weilheim-schongau.de/rettungsdienst/RD-Start2.htm. Accessed November 12, 2002
35 Ministry of Economics. New tariffs for the county [in Ger-man]. Available at: www.nahverkehr.nrw.de. Accessed Octo-ber 7, 2002
36 Deutsche Telekom. New tariffs effective from 01/11/2002 [inGerman]. Available at: www.service.t-online.de. AccessedOctober 7, 2002
www.chestjournal.org CHEST / 129 / 3 / MARCH, 2006 593