American Journal of Respiratory and Critical Care Medicine

In this randomized controlled economic evaluation we compared guided asthma self-management with usual asthma care according to guidelines for Dutch family physicians. Nineteen family practices were randomized, and 193 adults with stable asthma (98 self-management, 95 usual care) were included and monitored for 2 years. We hypothesized that introducing self-management would not compromise asthma control and cost would be equal to or lower than in usual care. Patient-specific cost data were collected, preference-based utilities were assessed, and incremental cost per quality-adjusted life year (QALY) and successfully treated week gained was calculated. Self-management patients gained 0.039 QALY (95% confidence interval [CI], 0.003 to 0.075) and experienced 81 (95% CI, 78 to 84) successfully treated weeks in 2 years' time; the corresponding figures for usual care were 0.024 (95% CI, −0.022 to 0.071) and 75 (95% CI, 72 to 78). Total costs were €1,084 (95% CI, 938 to 1,228) for self-management and €1,097 (95% CI, 933 to 1,260) for usual care. Self-management patients consumed 1,680 (95% CI, 1,538 to 1,822) puffs of budesonide, usual care patients 1,897 (95% CI, 1,679 to 2,115). Mean productivity cost due to limited activity days was €213 lower among self-management patients. When all costs were included, self-management was cost-effective on all outcomes. The probability that self-management was cost-effective relative to usual care in terms of QALYs was 52%. We conclude that guided self-management is a safe and efficient alternative approach compared with asthma treatment usually provided in Dutch primary care.

Asthma is a chronic respiratory disease affecting people of all ages. In Western countries, 4 to 6% of the adult population have a physician-confirmed diagnosis of bronchial asthma (1, 2). Compliance with prophylactic inhaled corticosteroid treatment is poor in many patients with asthma, thus limiting its effectiveness (3, 4). Because the costs of asthma for society are largely due to the consequences of uncontrolled disease, it is presumed that the cost could be significantly reduced if disease control is improved (5).

Using a comprehensive approach generally consisting of education and training, written action plans, and periodic supervision, health professionals may try to improve knowledge, practical skills, decision-making responsibility and, ultimately, disease control in their patients with asthma (6, 7). A systematic review including 23 trials concluded that self-management programs are able to improve health outcomes in adult asthma if they include self-monitoring and are accompanied with written action plans and regular medical professional review (8). However, the trials included in this meta-analysis have been conducted mainly in selected (secondary care) patients.

When competing for scarce health care resources it is not sufficient to determine the effects of asthma self-management programs solely in terms of health outcomes. It is also important to analyze whether the costs of introducing self-management outweigh the—potential—subsequent savings in health care utilization and productivity (“indirect”) costs, the latter resulting from fewer days of limited activities and incapacity for work (9). If the savings do not outweigh the investments, it is essential to assess whether the additional—or incremental—costs of a self-management program can be justified by the health gains.

Meanwhile, several asthma guidelines recommend self-management (10, 11) and health professionals and patients with asthma themselves seem to appreciate the contemporary approach (1214). A number of economic evaluations of asthma education and self-management have been published (1527), but most authors have confined themselves to separate descriptions of costs and health effects without directly assessing their relationship by calculating summary ratios. Essential methodological shortcomings were the absence of a control group receiving an appropriate comparator treatment and a too-short duration of follow-up. None of the published economic studies included instruments to assess preference-based utilities (e.g., quality-adjusted life years [QALYs] or similar universal outcome measures) as is currently recommended for all economic evaluations (28, 29). Moreover, only a part of the studies used written action plans, which seems to be a prerequisite for a successful treatment result (8).

This article reports a state-of-the-art economic evaluation of a guided self-management program for adult patients with asthma treated in Dutch primary care. We compared the self-management program with the “best” generally available medical treatment for asthma (“usual care”) according to asthma treatment guidelines for family physicians (30, 31). Beforehand, we did not expect substantial differences in health outcomes because medical care for patients with asthma is already of a high standard in The Netherlands, with asthma-related hospital admissions and deaths almost becoming rare events (32). Therefore, the main objective of this evaluation was to investigate whether a family practice-based self-management program for adults with asthma provides an efficient treatment alternative in terms of health care utilization and absence from work, without asthma control being compromised.

Study Design

The study was a randomized controlled parallel group multicenter clinical trial. Nineteen (19) Dutch family practices (49 family physicians) were randomly allocated to guided self-management or usual care. Randomization was stratified on type of practice, number of patients with asthma initially identified from the practice records, and use of a computerized prescription system. Duration of follow-up was 2 years per patient. Self-management and usual care were fully pursued by the family physicians; no other health professionals were involved. The study protocol was approved by the medical ethics committee of the University Medical Center St. Radboud (Nijmegen, The Netherlands). Patients gave written informed consent before study entry. The first subject entered the study in March 1996, and the last subject completed the study in June 1999.

Participants

The 49 family physicians involved in the study selected subjects with asthma, aged 16–60 years, who were to be treated with inhaled steroids according to national guidelines (30, 31). Identification of subjects was based on the following information sources: problem list coding (International Classification of Primary Care: R96); prescription of inhaled steroids or bronchodilators from practice or pharmacy records; and the annual influenza vaccination campaign list. Subjects willing to participate were included if (1) their PC20-histamine (provocative dose of histamine causing a 20% drop in FEV1) was less than 8 mg/ml and/or their reversibility of FEV1 was greater than 9% of the predicted value after 800 μg of salbutamol was administered as an aerosol by spacer; (2) they had a smoking history of less than 15 pack-years; (3) they were not currently treated by a chest physician; and (4) they were able to communicate in the Dutch language. Eligible patients with an initial FEV1 of less than 80% of the predicted value were pretreated with budesonide 800 μg twice daily for 6 weeks.

Guided Self-Management and Usual Care

All participants were prescribed budesonide administered by multidose dry powder inhaler (Pulmicort Turbuhaler, 200 × 200 μg; Astra Zeneca, Zoetermeer, The Netherlands) by one of the investigators (B.T.). Participants received new budesonide inhalers and handed in used inhalers during half-yearly visits to the pulmonary function laboratory. The family physician of each patient was responsible for regulating the dosage scheme at study entry. Family physicians were not restricted in prescribing nonsteroid lung medication in either group, apart from cromoglycates and nedocromil being prohibited during the trial. Bronchodilators were preferably prescribed on an as-needed basis, if necessary at all.

Self-management patients received education and training of skills on an individual basis from their family physician. Training consisted of four visits to the practice scheduled within a period of 3 months. Subsequent control visits for the remaining follow-up period of 21 months were recommended, but it was left to the initiative of the family physician and patient if and when these visits took place. Training tools consisted of (1) a detailed manual for the physicians describing the educational topics to be discussed during the consecutive training sessions and instructions on how to teach patients self-management skills (i.e., peak flow measurement, proper inhalation technique, completing the self-management diary, and application of self-treatment guidelines); (2) checklists for patients and physicians to assess and record specific information needs of patients; (3) two booklets of the Dutch Asthma Foundation, one containing general information about asthma and the other containing information about asthma medication; and (4) diaries containing self-treatment guidelines, also used for data collection. Self-management patients were equipped with a portable peak flow meter (Asmaplan+; Vitalograph, Buckingham, UK) and instructed to measure morning and evening peak expiratory flow rates once a week and record the best of three attempts in their diary. Self-treatment guidelines were based on peak flow values and severity of respiratory symptoms (Figure 1)

. Detailed information about the exact contents of the education program and self-treatment guidelines have been published elsewhere (33). Usual care physicians were instructed to adhere to the asthma treatment guidelines issued by the Dutch College of Family Physicians in 1992 (30) and to the revised guidelines issued in 1997 (31). Usual care patients did not receive peak flow meters, nor were they instructed on how to adjust their dosage of budesonide.

Clinical Effectiveness

Clinical effectiveness was evaluated on the basis of asthma control parameters and quality of life. Asthma control was expressed as the number of successfully treated weeks in 2 years of follow-up, changes in postbronchodilator FEV1, changes in FEV1 reversibility as a percentage of predicted value, and changes in PC20-histamine (34). Asthma-specific quality of life was assessed with the interview-administered 32-question Asthma Quality of Life Questionnaire (AQLQ) (35). This instrument assesses four domains: (1) asthma symptoms, (2) limitation of activity, (3) emotional dysfunction, and (4) responses to environmental stimuli, respectively. An overall score as well as separate domain scores were calculated.

Economic Evaluation: Data Collection and Resource Valuation

A societal perspective was adopted for the economic evaluation. Patient-specific resource use was measured in natural units if possible. Resource use was valued in monetary terms by multiplying the units consumed by the cost per unit. Three major cost categories were distinguished: program implementation, direct health care, and productivity (indirect) costs.

Data regarding bronchodilators and other prescribed nonsteroid asthma medication, over-the-counter medication, and limited activity days were extracted from the diary cards. A limited activity day was defined as any day on which a patient could not perform his or her usual (paid or unpaid) daily activities. Consumption of budesonide was assessed by counting the remaining puffs in the inhalers returned and by registration on the diary cards. We considered the puff counts as the most reliable source of information for estimating budesonide consumption (36). Patient out-of-pocket cost on house dust mite allergen avoidance measures and smoking cessation attempts were assessed retrospectively by an ad hoc questionnaire. Family physicians reported details of asthma-related consultations, medication prescriptions, influenza vaccinations, referrals, and diagnostic procedures on study report forms. Completeness of consultation data was verified after a patient had completed study participation.

The first-choice source for resource unit valuation was the sum charged by family physicians to privately ensured patients (including value-added tax and a mark-up for administrative expenses). Secondary sources were annually updated drugs and diagnostic indexes (37, 38) and more recent recommendations regarding cost analysis (39) (all issued by the Dutch College of Health Insurance), study expense accounts, and patient questionnaires. The human capital approach (28) was adopted to value limited activity days. An individual hourly wage based on the gross monthly income and the number of hours of disbursed work was calculated for all participants in paid employment. The resultant average gross hourly wage (€9.53) was subsequently used to convert all limited activity days (8-hour workday) into monetary terms, regardless of the employment status or income of individual participants. All resources used were valued in Dutch guilders and converted to euros. For conversion to U.S. dollars, costs in euros should be multiplied by a factor of 0.912, based on the 2000 Purchasing Power Parities as issued by the Organisation for Economic Co-operation and Development (www.oecd.org). Purchasing Power Parities are the rates of currency conversion that equalize the purchasing power of different currencies, thus eliminating differences in price levels between countries. Neither costs nor effects were discounted for time preferences.

Cost-Effectiveness Analysis: Outcome Measures

We performed a “base case” cost–effectiveness analysis, as well as secondary cost–effectiveness analyses. With the term “base case” we refer to an analysis in which the direct health care cost, program implementation cost, and productivity cost of patients are included. Outcome for the base case analysis was defined in terms of QALYs. To calculate QALYs, preference-based utilities were assessed at baseline and half-yearly at the pulmonary function laboratory. An interval rating scale ranging from 0 to 1 was used for this purpose, 0 being equal to death and 1 being equal to perfect health (40). Participants first marked a standardized (hypothetical) reference health state on the rating scale and subsequently their own perceived health state.

The number of successfully treated weeks served as the main outcome for secondary cost–effectiveness analyses (9). Successfully treated weeks were defined on the basis of recorded scores for shortness of breath in the diaries (modified Borg interval scale scoring: 0 = no shortness of breath; 10 = maximal shortness of breath) (41). Any given week with a score higher than the individual's median score over the total follow-up was considered an unacceptably low level of control of asthma symptoms and therefore counted as unsuccessful. Subtracting this figure from the individual's total number of recorded weeks resulted in the proportion of weeks being treated successfully, which was eventually standardized to the number of successfully treated weeks per 2 years (104 weeks). Next to successfully treated weeks, the number of patients with a minimal clinically important difference (MCID) in quality of life between the baseline and final visit was studied as a secondary outcome. MCID was defined as a within-subject improvement of 0.5 unit on the overall AQLQ or domain scores (35).

Statistical Analysis

Patients were included in the intention-to-treat analysis if they had been present at the first follow-up visit at the pulmonary function laboratory after 6 months. Although distributions of resource units were skewed to the right for most cost components, arithmetic means and t test-based 95% confidence intervals (95% CIs) were calculated to compare self-management and usual care groups (42). Within-group cost differences between the first and second study year were analyzed by paired, between-group differences by unpaired t test. QALYs were determined by calculating the area under the curve (time × rating scale score) for each participant. Mean costs and effects were multiplied by a constant of 100 to standardize for inequalities of group sizes. Because of that, cost–effectiveness results reflect a situation in which 2 groups of 100 patients each would receive either self-management or usual care. Consequently, the cost–effectiveness ratios of the AQLQ data should be interpreted as the incremental cost or net savings to improve quality of life in one patient. A treatment was qualified to be “dominant” when this particular treatment was both more effective and less costly than the alternative (29). Secondary analyses were performed by calculating cost–effectiveness ratios with exclusion of the productivity cost.

The SAS statistical software package (release 6.12 for Windows; SAS Institute, Cary, NC) was used for statistical analyses. Regarding the incremental cost per successfully treated week, a 95% CI was determined on the basis of the Fieller theorem (43). To express uncertainty in the estimated incremental cost per QALY, DATA for Healthcare software (DATA Pro; TreeAge, Williamstown, MA) was used to generate graphical representations of the cost–effectiveness plane and accompanying two-dimensional 90 and 95% confidence intervals. This was done by nonparametric bootstrapping (Monte Carlo simulation): resampling with replacement from the patient-level cost and QALY data from the two comparator groups (1,000 random samples with size n = 100 each). Each point in the resulting scatter plot represents the incremental cost–effectiveness ratio of one iteration of the Monte Carlo simulation. A diagonal line intersecting the origin of the plot simplifies identification of points for which the incremental cost–effectiveness ratio of self-management versus usual care is less than, or equal to, an a priori specified societal “willingness-to-pay” limit (44) (λ) to gain one additional QALY. Arbitrarily, λ was set to €22,500. A graphic representation (“acceptability curve”) of the probability that a particular intervention is cost-effective over a range of increasing values for λ was generated (45). This bayesian approach of the stochastic analysis provides information relevant to health care decision making.

Study Population and Clinical Effects

Ninety-eight (98) self-management and 95 usual care patients were included in the intention-to-treat analyses (Figure 2)

. Treatment groups did not differ on general or clinical characteristics at baseline, apart from a higher proportion of patients reporting a recent episode of aggravated asthma symptoms and lower AQLQ scores in the self-management group (Table 1)

TABLE 1. Baseline characteristics of the study population by treatment group




Self-Management
 (n = 98)

Usual Care
 (n = 95)

p Value
General characteristics
Age, yr39.6 (11.2)39.3 (12.0)0.859
Sex, male/female34/6440/560.394
Employment status
Student, %54
Full-time or part-time job, %6666
Unemployed or retired, %29300.953
Smoking status
Never smokers45 (46%)55 (56%)
Former smokers31 (32%)21 (22%)
Current smokers22 (22%)21 (22%)0.254
Pack-years number5.8 (4.5)5.7 (4.5)0.881
Clinical characteristics
Duration of asthma,* yr21.0 (16.5)18.1 (14.3)0.232
Subjects with asthma attack(s) in previous 6 mo47 (48%)30 (32%)0.017
Allergy, number of positive skin prick tests
019 (23%)20 (26%)
1–337 (44%)29 (38%)
⩾ 428 (33%)27 (36%)0.735
Lung function parameters
FEV1, post-BD % of predicted value90.0 (12.1)92.6 (12.9)0.135
Median FEV1 reversibility, %5.0 (IQR 8.6)5.4 (IQR 6.8)0.930
PC20, geometric mean1.200.970.442
Utilities and quality of life
Utilities
Median “Own health state”0.80 (IQR 0.15)0.80 (IQR 0.16)0.668
Median “Reference health state”0.40 (IQR 0.30)0.40 (IQR 0.20)0.380
Quality of life
Overall AQLQ score5.4 (0.87)5.7 (0.77)0.013
AQLQ activities domain5.3 (1.03)5.6 (0.77)0.015
AQLQ emotions domain5.8 (1.01)6.2 (0.76)0.002
AQLQ symptoms domain5.3 (1.03)5.6 (0.90)0.074
AQLQ environment domain
5.3 (1.10)
5.5 (1.1)
0.165

*Missing in 17 self-management and 14 usual care patients.

Missing in 14 self-management and 19 usual care patients.

Difference between FEV1% predicted before and after bronchodilator.

Definition of abbreviations: AQLQ = Asthma Quality of Life Questionnaire; BD = bronchodilator; IQR = interquartile range.

. Fourteen self-management patients and 16 usual care patients did not use bronchodilator medication during the study. Twelve self-management and 5 usual care patients used a long-acting β2-agonist, and theophyllines were used by 3 self-management patients only. The course of the pre- and postbronchodilator FEV1 did not differ between groups, nor did FEV1 reversibility or PC20 (34). The mean number of successfully treated weeks in 2 years' time was 81 (95% CI, 78 to 84) for self-management and 75 (95% CI, 72 to 78) for usual care (Table 2)

TABLE 2. Average and incremental effects of self-management and usual care in adults with asthma



Average Effect

Outcome
Self-Management
Usual Care
Incremental* Effect
 of Self-Management
 in 100 Subjects
 Treated for 2 yr
QALYs (95% CI)0.039 (0.003, 0.075)0.024 (−0.022, 0.071)+1.5 (−1.4, 4.4)
Number of successfully treated weeks (95% CI)81 (78, 84)75 (72, 78)+600 (230, 970)
Proportion of subjects with MCID
AQLQ total score (95% CI)39 (29, 48)29 (20, 38)+10 (−3, 23)
AQLQ emotions domain (95% CI)42 (32, 51)25 (16, 33)+17 (10, 24)
AQLQ activities domain (95% CI)52 (42, 62)39 (30, 49)+13 (−1, 27)
AQLQ symptoms domain (95% CI)35 (26, 45)28 (19, 37)+7 (−6, 20)
AQLQ environment domain (95% CI)
42 (32, 51)
39 (30, 49)
+3 (−11, 17)

*Self-management minus usual care.

Final AQLQ measurement was missing in two self-management and six usual care patients.

Definition of abbreviations: AQLQ = Asthma Quality of Life Questionnaire; CI = confidence interval; MCID = minimal clinically important difference; QALY = quality adjusted life year.

Results of the base case analysis are printed in boldface. Increments are standardized to 100 subjects per group treated for 2 years.

. This corresponds with a statistically significant gain of six successfully treated weeks in 2 years in favor of self-management. In the self-management group, 17% (95% CI, 10 to 24) more participants showed an MCID on the AQLQ Emotions Domain compared with usual care. No statistically significant differences were observed for the Activities, Environmental, and Symptoms Domains, or the total AQLQ score.

Cost Analysis

The total implementation cost of the self-management program amounted to €189 (95% CI, €179 to €199) per patient (Table 3)

TABLE 3. Breakdown of the implementation cost of the self-management program for 98 adults with asthma treated for 2 years in eight family practices


Component of Cost

Source for
 Unit Valuation

Unit

Number
 of Units

Cost per Unit
 (€)

Total Cost
 (€)
Preparation
Prestudy training and instruction
     of family physicians*Hours2373.531,691
Application of self-management program
Educational and self-management aidsSet9814.251,396
Peak flow metersMeter9829.612,902
Education sessions
Family physician time*Hours15173.5311,075
Patient time§Hours151 9.531,436
Total implementation cost:18,500
Average implementation cost per patient (95% CI):



189 (179, 199)

*Source used for unit valuation: Guidebook for Cost Investigation (Dutch College of Health Insurance [39]).

Source used for unit valuation: retail prices (index year 2000).

Set: all materials necessary to educate and train one patient, that is, information brochures, self-management diaries, and information feedback forms.

§Source used for unit valuation: study-specific inquiry by questionnaire.

Based on the average gross hourly wage of all employed participants.

. Time invested by family physicians and purchase of peak flow meters constituted the major part of the implementation cost (60 and 16%, respectively). Mean budesonide usage was 1,680 puffs (95% CI, 1,538 to 1,822) or €414 for self-management and 1,897 puffs (95% CI, 1,679 to 2,115) or €467 for usual care, indicating a saving of 217 puffs or €53 per patient during the 2-year follow-up (Table 4)

TABLE 4. Mean and incremental program implementation, direct health care, and productivity cost of self-management and usual care per patient per 2 years




Self-Management (n = 98)

Usual Care (n = 95)

Component of Cost
Source for
 Unit Valuation
Units (95% CI)
Cost (€)
Units (95% CI)
Cost (€)
Incremental
 Cost (€)
Program implementation cost
a,b,c
98
189
0
0
+189
Subtotal implementation cost:
189 (179, 199)
0
+189
Direct health care cost
Drugs and other interventions
Budesonide, dosesd1,680 (1,538, 1,822)4141,897 (1,679, 2,115)467−53
Short-acting bronchodilators, dosesd469 (347, 591)84796 (526, 1,066)141−57
Long-acting bronchodilators, dosesd67 (10, 124)5130 (−6, 66)16+35
Theophylline, dosesd11 (−5, 26)200+2
Prednisone, coursesd0.33 (0.19, 0.46)30.22 (0.08 , 0.36)2+1
Antibiotics, coursesd0.28 (0.11, 0.44)10.40 (0.19, 0.61)2−1
Other asthma medication*b,d,eNA2NA6−4
Influenza vaccinations, numberd0.72 (0.55, 0.90)50.38 (0.25, 0.51)3+2
Physiotherapy, coursesa0.03 (0, 0.07)40.01 (−0.01, 0.03)1+3
Allergen avoidance measurescNA193NA109+84
Other resources
Family physician consultations, numbera2.2 (1.6, 2.8)372.4 (1.9, 2.9)40−3
Chest physician consultations, numbera0.09 (0.03, 0.15)50.01 (−0.01, 0.03 )1+4
Diagnostic proceduresfNA6NA60
Emergency room visits, number00000
Hospital admissions, number

0
0
0
0
0
Subtotal direct cost:
809 (683, 934)
798 (682, 914) +11
Productivity cost
Limited activity days, d§
c
1.2 (0.5, 1.9)
86
3.9 (2.5, 5.4)
299
−213
Subtotal productivity cost:
86 (35, 136)
299 (191, 406) −213
Total cost:
1,084 (938, 1,228)
1,097 (933, 1,260) −13

*Both prescribed and over-the-counter medication.

Purchase of house dust mite impermeable mattress covers, smooth floors, special vacuum cleaners and air cleaning equipment.

Cost of various pulmonary function and allergy tests, chest X-rays, and sputum cultures.

§Highest two patients were excluded in both groups (see text).

Self-management minus usual care.

NA = not applicable.

Sources used for unit valuation: [a] guidebook for Cost Investigation (Dutch College of Health Insurance, reference [39]); [b] retail prices (index year 2000); [c] study-specific inquiry by questionnaire; [d] sum charged by family physicians to privately insured patients, including administrative expenses; [e] Pharmacotherapeutic Compass (Dutch College of Health Insurance, reference [37]); [f] Diagnostic Compass (Dutch College of Health Insurance, reference [38]).

. Converted to the level of budesonide inhalers, 0.5 inhaler per year was saved by self-management patients. Costs of short-acting bronchodilators were significantly lower for self-management, but this difference was largely compensated by the higher cost of long-acting β2-agonists and theophyllines in this same group. During the study, 30 (31%) self-management and 10 (11%) usual care patients took domestic house dust mite avoidance measures (relative risk = 1.7; 95% CI, 1.3 to 2.2). Consequently, mean costs of domestic house dust mite allergen avoidance measures were significantly higher among self-management patients (€193 versus €109 for usual care, p = 0.0015). Although the cost of influenza vaccinations composed only a marginal proportion of the total direct cost, there were significantly more vaccinations in the self-management group (Table 4): 46 (47%) self-management and 27 (28%) usual care patients received at least one influenza vaccination during follow-up (relative risk = 1.5, 95% CI, 1.1 to 1.9). There were more referrals to chest physicians among self-management than among usual care patients: 9 (4.6%) and 1 (0.6%), respectively (p = 0.011). No asthma-related emergency unit visits or hospital admissions were reported. Mean direct health care cost aggregated to €809 (95% CI, 683 to 934) for self-management and €798 (95% CI, 682 to 914) for usual care (Table 4).

Sixty-two percent of self-management patients and 79% of usual care patients reported one or more limited activity days at some point during follow-up. The mean number of limited activity days was 1.9 (95% CI, 0.7 to 3.2) for self-management and 6.0 (95% CI, 2.6 to 9.4) for usual care, corresponding with mean productivity costs of €144 and €462, respectively. However, closer examination of the productivity cost data identified two distinct outliers in the usual care group, with a productivity cost of €10,831 (142 limited activity days) and €5,263 (69 limited activity days), respectively. One outlier had a period of several months with frequent but short episodes of sick leave due to asthma, and the other had a 3-month episode of uninterrupted sick leave. In both cases, irritant exposure in the workplace explained the high productivity cost. Because of the clear work-related cause and the disproportionate impact of these two outliers on the average productivity cost in the usual care group, we decided to exclude subjects above the 98th percentile of the productivity cost distribution from the final cost calculations in both groups. This resulted in an average number of limited activity days of 1.2 (95% CI, 0.5 to 1.9) for self-management and 3.9 (2.5 to 5.4) for usual care, corresponding to a €213 productivity cost saving for self-management (Table 4). We consider the productivity cost without the outliers as the main results.

The sum of direct health care and implementation costs amounted to a difference of €199 (95% CI, 70 to 328) in favor of usual care (Table 5)

TABLE 5. Average and incremental cost* during 2 years of self-management and usual care in adults with asthma


Cost Components

Average Total Cost (95% CI)

Direct
 Health Care
Program
 Implementation
Productivity
Self-Management
Usual Care
Incremental Cost of
 Self-Management per
 Subject Treated
1,084 (938, 1,228)1,097 (933, 1,260)−13 (−232, 206)
997 (871, 1,124)798 (681, 914)+199 (70, 328)
894 (751, 1,038)1,096 (933, 1,260)−202 (−420, 16)



809 (683, 934)
798 (681, 914)
+11 (−77, 99)

*In euros.

✓ = cost components included in the calculations of average costs.

Self-management minus usual care.

Definition of abbreviation: CI = Confidence interval.

Results of the base case analysis are printed in boldface. A plus sign indicates an expenditure due to the self-management program, a minus sign indicates a saving.

. The between-group difference in the total cost of €13 was not statistically significant (p = 0.906). Analyzing the cost for the first and second year separately showed that, as expected, the major part (91%) of the program implementation cost was spent during the first study year (Figure 3) . A significant reduction of the productivity cost from the first to the second year was observed for self-management (p = 0.036) but not for usual care (p = 0.487). During the second year the total cost per patient was €147 (p = 0.0013) lower in the self-management group.

Base Case Cost-Effectiveness Analysis

The course of rating scale scores is given in Figure 4

. The mean number of QALYs gained during the 2-year follow-up was 0.039 (95% CI, 0.003 to 0.075) for self-management and 0.024 (95% CI, −0.022 to 0.071) for usual care (Table 2). This would imply that in 100 patients with asthma, self-management is associated with a gain of 1.5 QALYs (95% CI, −1.4 to 4.4) relative to usual care. In terms of cost-effectiveness, self-management dominated usual care (Table 6)

TABLE 6. Incremental cost–effectiveness ratios for asthma self-management relative to usual care during 2 years of follow-up




Productivity Cost Included
 (95% CI)

Productivity Cost Excluded
 (95% CI)
Incremental cost per QALY gainedSelf-management dominant*,13,267
Incremental cost per successfully treated week gainedSelf-management dominant33 (4, 99)
Observed incremental cost for one patient to experience an MCID,
     when 100 patients are treated:
AQLQ total scoreSelf-management dominant§Self-management dominant§
AQLQ Emotions DomainSelf-management dominant§Self-management dominant§
AQLQ Activities DomainSelf-management dominant§Self-management dominant§
AQLQ Symptoms DomainSelf-management dominant§1
AQLQ Environment Domain
Self-management dominant§
328

*Dominant: treatment both more effective and less costly than the alternative treatment.

Uncertainty in the base case cost per QALY estimate is depicted in Figure 5.

Final AQLQ measurement was missing in two self-management and six usual care patients.

§Because the cost per MCID estimates are based on group mean cost, it should be noted that uncertainty exists around the ratio estimates (results not presented).

Definition of abbreviations: AQLQ = Asthma Quality of Life Questionnaire; MCID = minimal clinically important difference; QALY = quality-adjusted life year; 95% CI = 95% confidence interval.

Costs are in euros. Results of the base case analysis are printed in boldface.

. Uncertainty around the incremental cost per QALY point estimate is depicted in Figure 5 . This scatter plot shows that the uncertainty around the cost–effectiveness estimate is large. In other words, the dominance of self-management cannot be firmly established. This is supported by the cost–effectiveness acceptability curve (Figure 6) : regardless of the societal willingness to pay, the probability that self-management is cost-effective relative to usual care is about 52% when a prior probability of 50% is assumed.

Secondary Cost-Effectiveness Analyses

When productivity costs were excluded, the incremental cost per QALY of self-management relative to usual care was €13,267 (Table 6). Self-management dominated usual care with regard to successfully treated weeks and the proportion of patients with an MCID in quality of life. Without the productivity cost, the incremental cost–effectiveness ratio was €33 (95% CI, 4 to 99) to gain one successfully treated week due to self-management. Cost–effectiveness ratios based on the cost per patient with an MCID in quality of life preponderantly pointed to self-management as the dominant treatment, regardless of the inclusion or exclusion of productivity cost (Table 6).

This article reports the economic evaluation of a family medicine-based asthma self-management program, with “usual care” according to Dutch asthma treatment guidelines as the comparator treatment. In summary, the results were as follows. Net savings in favor of self-management were observed in some of the direct health care cost components (i.e., use of budesonide and short-acting bronchodilators) and productivity (“indirect”) cost. When all costs were included, a mean net saving of €13 in favor of self-management was observed (not statistically significant). Despite the investment necessary for program implementation, the total costs for the self-management group were significantly lower during the second year of follow-up. The base case cost–effectiveness ratio pointed to self-management as a cost-effective treatment option: self-management dominated usual care (i.e., was more effective and less costly). However, the graphic evaluation of uncertainty around the cost per QALY estimate showed that the observed dominance of self-management could not be firmly established. Overall, the secondary analyses based on successfully treated weeks and patients with a clinically important improved quality of life pointed to self-management as the dominant treatment option. When productivity costs were ignored, self-management was no longer dominant in the secondary analyses (€13,267 to gain 1 QALY and €33 to gain one successfully treated week).

Some comments on the methodology of the study need to be made before further discussing our findings. First, a disadvantage of using rating scales to value health states (and subsequently estimate QALYs) is that these instruments do not take risk avoidance and uncertainty about future health outcomes into account. Therefore, rating scale utilities tend to produce higher quality weights then other techniques such as time-tradeoff and standard gamble methods (46). Moreover, rating scale scores appear not to be a true interval scale of preference for certain health states. Unfortunately, in the current study we did not include a standard gamble or time-tradeoff instrument. The mean number of QALYs in both treatment groups may have been overestimated because of this, but the incremental difference between the groups is probably valid. However, this point should be kept in mind when comparing our QALY results with external information from other studies.

We did not randomize individual patients with asthma, but family practices. The reason for doing so was to avoid potential “contamination” of the usual care group by family physicians who had to practice both usual care and self-management simultaneously for different patients. Whereas in the clinical evaluation a multilevel analysis was used to address possible dependency in clustered observations induced by this kind of randomization (34), some influence on the cost data cannot be ruled out completely. For instance, prevailing habits and preferences in prescribing bronchodilators by family physicians may have biased the results for this cost component to an unknown extent. The same argument holds for the promotion of influenza vaccination among individuals with asthma.

The baseline level of quality of life scores was higher in usual care patients, possibly leaving less room for improvement in this group. The comprehensive clinical evaluation of the data showed that the differences in AQLQ scores existing at baseline gradually disappeared during the 2-year follow-up period, which may indicate that quality of life was maximized in both groups (34). However, the observation that self-management patients experienced significantly more successfully treated weeks implies that the self-management program also had an independent effect, regardless of the health status differences present at baseline.

As a consequence of our study design, we cannot be sure which component of the self-management program in particular was responsible for the observed effects and savings: the (expensive) educational efforts made by the family physicians or the (relatively inexpensive) guidelines for self-monitoring and self-treatment. There is some evidence that addition of self-treatment guidelines to an asthma education program does yield extra effects in terms of health outcomes (47).

We have previously looked at the generalizability of our study population (48). Evaluation of the recruitment process showed that patients who use a low or intermediate dosage of inhaled steroids were more likely to participate in the study than patients receiving a high dosage or patients who did not use inhaled steroids at all (although, according to our national treatment guidelines [30, 31], they should have). Moreover, patients in paid employment were more likely to refrain from participation than those not in paid employment.

Regarding the cost analysis, several points need to be addressed. The most important expenditure necessary to implement the self-management program was the time spent by family physicians to educate and train their patients with asthma (€113 per patient on average). Delegation of this task to, for instance, nurses specialized in respiratory care could reduce these costs considerably. Assuming delegation would not diminish program effectiveness, any reduction in the implementation cost would obviously affect cost–effectiveness ratios in favor of self-management. Another advantage of transferring the actual pursuance of self-management training to other professionals would be the diminished impact on the (already) high workload of family physicians. Targeting the self-management intervention to patients with a high likelihood of treatment success could also enhance overall efficiency, although at this time it is unknown how these patients could be identified beforehand.

One of the most remarkable findings in this study was that the introduction of self-management led to substitution of particular cost components with other components. For instance, the financial saving due to reduced budesonide use and fewer limited activity days in the self-management group was outweighed for the greater part by the extra out-of-pocket cost for domestic allergen avoidance measures, and, although to a much lesser extent, more influenza vaccinations and referrals to chest physicians. These favorable “side effects” of the self-management program are probably explained by the emphasis put on the importance of healthy behavior (i.e., allergen avoidance, influenza vaccination, and smoking cessation) during the education sessions. The higher out-of-pocket cost for domestic allergen avoidance measures in the self-management group may be due to specific contents of our educational program. “Nature, cause and prevention of allergy or allergic symptoms,” “Hyperreactivity and personal triggers,” and “Allergen avoidance measures at home” were 3 of the 31 educational topics the family physicians discussed with their participants. One previous study has reported that asthma education may be effective in promoting house dust mite avoidance measures in patients with moderate to severe asthma (49). The extra attention focused on self-management patients as a consequence of the intensified doctor–patient relationship may have influenced the higher referral rate observed in the self-management group.

We observed significant differences in the use of asthma medication between self-management and usual care patients, especially for budesonide. This difference suggests a more efficient use of prophylactic medication due to self-management, a finding inconsistent with previously reported higher compliance rates regarding the use of inhaled steroids after introducing self-management (4, 50). However, use of the term “compliance” may be inappropriate when it comes to evaluation of self-management in patients with asthma. After all, the essence of the approach is to fine-tune the use of inhaled steroids to the actual need as determined by self-monitoring, without a prescribed (fixed) daily dose. For this reason, we anticipated a reduced consumption of inhaled steroids in the intervention group beforehand, although it has been shown that self-management patients do not always adhere to their personalized self-treatment guidelines (4).

The main objective of any self-management program is to attain a long-wearing behavioral change in patients with regard to their disease. Once accomplished, this effect could be expected to persist for a longer period of time. Although in the current study we had to limit the time horizon to a maximum of 2 years, there was a tendency toward further productivity cost reduction during the second year of follow-up. Because we have no cost data from the years before the study at our disposal, we can only speculate about how the observed productivity cost for the first and second years related to the annual productivity cost before the study. However, both Muhlhauser and coworkers (20) and Trautner and coworkers (25) have shown that significant changes from the prestudy situation may indeed be achieved. Moreover, findings reported by Trautner and coworkers (25) agree with our observation of a progressive reduction of productivity cost between the first and second year in self-management patients: they observed a 5% reduction in the number of days of absence from work during the first year, but an 18% reduction during the third year. This suggests that savings in productivity cost resulting from asthma self-management are retained in the long term.

Several other authors have reported significantly lower productivity costs due to self-management as well (1719, 22, 23, 26, 27). The estimated savings from these studies range from 25 to 70% of the productivity costs observed in control patients. It should, however, be kept in mind that these studies were performed in populations with varying asthma severity, with diverse control groups, in different countries, and with different methods used for valuing productivity losses. Because there is no consensus in the literature as to what method is most suitable for valuing productivity losses, we applied the widely used human capital approach. An alternative method would have been the more advanced friction cost method as proposed by Koopmanschap and coworkers (51). The basic idea of this method is that the amount of production lost due to disease depends on the time span organizations need to restore the initial production level. This “friction period” is likely to differ by location, industry, firm, and category of worker, making the method rather complex. Had we used the friction cost method, our estimate of productivity cost would probably have been lower, as has been demonstrated for other health care programs (52).

It is generally recognized that a large proportion of the total cost of asthma is derived from treating the consequences of poor asthma control, such as emergency room use and hospitalizations (5). Therefore, improved asthma control is likely to reduce the number of acute asthma-related hospital admissions as well as the productivity costs resulting from the admission itself and recovery time after discharge. Although several authors have reported reductions in use of hospital services due to self-management, hospital admissions did not occur at all in our study and can therefore be no explanation for the lower number of sick days observed in self-management patients. Thus, the effect of self-management on asthma-related limited activity days appears to be more subtle in patients with mild asthma under adequate control, like the patients involved in the current study.

Although the base case cost–effectiveness analysis demonstrated a 52% probability of the self-management program being cost-effective relative to usual care, we conclude that guided self-management is a safe and efficient alternative approach compared with asthma treatment usually provided in Dutch primary health care.

1. Hedman J, Kaprio J, Poussa T, Nieminen MM. Prevalence of asthma, aspirin intolerance, nasal polyposis and chronic obstructive pulmonary disease in a population-based study. Int J Epidemiol 1999;28:717–722.
2. Hahn DL, Beasley JW. Diagnosed and possible undiagnosed asthma: a Wisconsin Research Network (WReN) study. J Fam Pract 1994;38:373–379.
3. Tettersell MJ. Asthma patients' knowledge in relation to compliance with drug therapy. J Adv Nurs 1993;18:103–113.
4. Van der Palen J, Klein JJ, Rovers MM. Compliance with inhaled medication and self-treatment guidelines following a self-management programme in adult asthmatics. Eur Respir J 1997;10:652–657.
5. Barnes PJ, Jonsson B, Klim JB. The costs of asthma. Eur Respir J 1996;9:636–642.
6. Lahdensuo A. Guided self management of asthma: how to do it. BMJ 1999;319:759–760.
7. Neville R. Patient education and guided self-management plans. Respir Med 1996;90:385–386.
8. Gibson PG, Coughlan J, Wilson AJ, Abramson M, Bauman A, Hensley MJ, Walters EH. Self-management education and regular practitioner review for adults with asthma [Cochrane review]. In: The Cochrane library. Issue 3. Oxford: Update Software; 2000.
9. Liljas B, Lahdensuo A. Is asthma self-management cost-effective? Patient Educ Couns 1997;32:S97–S104.
10. British Thoracic Society, General Practitioners in Asthma Group, British Association of Accident and Emergency Medicine, British Paediatric Respiratory Society, Royal College of Paediatrics and Child Health. British guidelines on asthma management: 1995 review and position statement. Thorax 1997;52(Suppl 1):S2–S21.
11. Anonymous. Global strategy for asthma management and prevention: NHLBI/WHO Workshop Report 95-3659. Bethesda, MD: National Heart, Lung, and Blood Institute; 1995.
12. Thoonen BPA, Jones KP, Van Rooij HA, Van de Hout AC, Smeele I, Grol R, Van Schayck CP. Self-treatment of asthma: possibilities and perspectives from the practitioner's point of view. Fam Pract 1999;16:117–122.
13. Barner JC, Mason HL, Murray MD. Assessment of asthma patients' willingness to pay for and give time to an asthma self-management program. Clin Ther 1999;21:878–894.
14. Partridge MR, Hill SR. Enhancing care for people with asthma: the role of communication, education, training and self-management. 1998 World Asthma Meeting Education and Delivery of Care Working Group. Eur Respir J 2000;16:333–348.
15. Drummond N, Abdalla M, Buckingham JK, Beattie JAG, Lindsay T, Osman LM, Ross SJ, Roy-Chaudhury A, Russell I, Turner M. Integrated care for asthma: a clinical, social, and economic evaluation. Grampian Asthma Study of Integrated Care (GRASSIC). BMJ 1994;308:559–564.
16. Wilson SR, Scamangas P, German DF, Hughes GW, Lulla S, Coss S, Chardon L, Thomas RG, Starr-Schneidkraut N, Stancavage FB, et al. A controlled trial of two forms of self-management education for adults with asthma. Am J Med 1993;94:564–575.
17. Taitel MS, Kotses H, Bernstein IL, Bernstein DI, Creer TL. A self-management program for adult asthma. II. Cost–benefit analysis. J Allergy Clin Immunol 1995;95:672–676.
18. Bolton MB, Tilley BC, Kuder J, Reeves T, Schultz LR. The cost and effectiveness of an education program for adults who have asthma. J Gen Intern Med 1991;6:401–407.
19. Sondergaard B, Davidsen F, Kirkeby B, Rasmussen M, Hey H. The economics of an intensive education programme for asthmatic patients: a prosprective controlled trial. Pharmacoeconomics 1992;1:207–212.
20. Muhlhauser I, Richter B, Kraut D, Weske G, Worth H, Berger M. Evaluation of a structured treatment and teaching programme on asthma. J Intern Med 1991;230:157–164.
21. Windsor RA, Bailey WC, Richards JM, Manzella B, Soong SJ, Brooks M. Evaluation of the efficacy and cost effectiveness of health education methods to increase medication adherence among adults with asthma. Am J Public Health 1990;80:1519–1521.
22. Lahdensuo A, Haahtela T, Herrala J, Kava T, Kiviranta K, Kuusisto P, Pekurinen M, Peramaki E, Saarelainen S, Svahn T, et al. Randomised comparison of cost effectiveness of guided self management and traditional treatment of asthma in Finland. BMJ 1998;316:1138–1139.
23. Kauppinen R, Sintonen H, Vilkka V, Tukiainen H. Long-term (3-year) economic evaluation of intensive patient education for self-management during the first year in new asthmatics. Respir Med 1999;93:283–289.
24. Neri M, Migliori GB, Spanevello A, Berra D, Nicolin E, Landoni CV, Ballardini L, Sommaruga M, Zanon P. Economic analysis of two structured treatment and teaching programs on asthma. Allergy 1996;51:313–319.
25. Trautner C, Richter B, Berger M. Cost-effectiveness of a structured treatment and teaching programme on asthma. Eur Respir J 1993;6:1485–1491.
26. Ghosh CS, Ravindran P, Joshi M, Stearns SC. Reductions in hospital use from self management training for chronic asthmatics. Soc Sci Med 1998;46:1087–1093.
27. Gallefoss F, Bakke PS. Cost-effectiveness of self-management in asthmatics: a 1-yr follow-up randomized, controlled trial. Eur Respir J 2001;17:206–213.
28. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine. New York: Oxford University Press; 1996.
29. Drummond MF, O'Brien B, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press; 1997.
30. Van der Waart MAC, Dekker FW, Nijhoff S, Thiadens HA, Van Weel C, Van Schayck CP, Helder M, Rosmalen CFH. [NHG guideline on CNSLD in adults: treatment]. Huisarts & Wetenschap 1992;35:437–443.
31. Geijer RMM, Van Hensbergen W, Bottema BJAM, Van Schayck CP, Sachs APE, Smeele IJM, Thiadens HA, Van Weel C, Van der Zwan AAC, Rosmalen CFH. [NHG guideline on asthma in adults: treatment]. Huisarts & Wetenschap 1997;40:443–454.
32. Wever-Hess J, Wever AMJ. Asthma statistics in the Netherlands 1980–94. Respir Med 1997;91:417–422.
33. Thoonen BPA, Schermer TRJ, Van Schayck CP. Self-managed asthma education diminishes the risk of underestimating information needs of asthma patients. Prim Care Respir J 2000;9:39.
34. Thoonen BPA, Schermer TRJ, Van Weel C, Grol R, Van Schayck CP. Clinical effectiveness of self management of asthma in general practice: a randomised controlled trial. Eur Respir J 2001;18:518S.
35. Juniper EF, Guyatt GH, Epstein RS, Ferrie PJ, Jaeschke R, Hiller TK. Evaluation of impairment of health related quality of life in asthma: development of a questionnaire for use in clinical trials. Thorax 1992;47:76–83.
36. Rand CS, Wise RA. Measuring adherence to asthma medication regimens. Am J Respir Crit Care Med 1994;149:S69–S76.
37. Anonymous. Pharmacotherapeutic compass 2000/2001. Amstelveen, The Netherlands: Dutch College of Health Insurance; 2000.
38. Anonymous. Diagnostic compass 1999/2000. Amstelveen, The Netherlands: Dutch College of Health Insurance; 1999.
39. Oostenbrink JB, Koopmanschap MA, Rutten FFH. Guidebook for cost investigation. Amstelveen, The Netherlands: Dutch College of Health Insurance; 2000.
40. Rutten-Van Molken MP, Custers F, Van Doorslaer EK, Jansen CC, Heurman L, Maesen FP, Smeets JJ, Bommer AM, Raaijmakers JA. Comparison of performance of four instruments in evaluating the effects of salmeterol on asthma quality of life. Eur Respir J 1995;8:888–898.
41. Bijl-Hofland ID, Folgering HTM, Van den Hoogen H, Cloosterman SGM, Van Weel C, Donkers JM, Van Schayck CP. Perception of bronchoconstriction in asthma patients measured during histamine challenge test. Eur Respir J 1999;14:1049–1054.
42. Thompson SG, Barber JA. How should cost data in pragmatic randomised trials be analysed? BMJ 2000;320:1197–1200.
43. Severens JL, De Boo TM, Konst EM. Uncertainty of incremental cost–effectiveness ratios: a comparison of Fieller and bootstrap confidence intervals. Int J Technol Assess Health Care 1999;15:608–614.
44. Willan AR, O'Brien BJ, Leyva RA. Cost–effectiveness analysis when the WTA is greater than the WTP. Stat Med 2001;20:3251–3259.
45. Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost–effectiveness acceptability curves. Health Econ 2001;10:779–787.
46. Nord E. The validity of a visual analogue scale in determining social utility weights for health states. Int J Health Plann Manage 1991;6:234–242.
47. Klein JJ, Van der Palen J, Uil SS, Zielhuis GA, Seydel ER, Van Herwaarden CLA. Benefit from the inclusion of self-treatment guidelines to a self-management programme for adults with asthma. Eur Respir J 2001;17:386–394.
48. Thoonen BPA, Schermer TRJ, Akkermans RP, Den Otter JJ, Grol RP, Van Schayck CP. Willingness of patients to perform self-management of asthma and the role of inhaled steroids. Scand J Prim Health Care 2002;20:60–64.
49. Cote J, Cartier A, Robichaud P, Boutin H, Malo JL, Rouleau M, Boulet LP. Influence of asthma education on asthma severity, quality of life and environmental control. Can Respir J 2000;7:395–400.
50. Gallefoss F, Bakke PS. How does patient education and self-management among asthmatics and patients with chronic obstructive pulmonary disease affect medication. Am J Respir Crit Care Med 1999;160:2000–2005.
51. Koopmanschap MA, Rutten FFH, Van Ineveld BM, Van Roijen L. The friction cost method for measuring indirect costs of disease. J Health Econ 1995;14:171–189.
52. Koopmanschap MA, Rutten FF. The impact of indirect costs on outcomes of health care programs. Health Econ 1994;3:385–393.
Correspondence and requests for reprints should be addressed to Tjard R. Schermer, M.Sc., University Medical Center St. Radboud, Department of General Practice/Family Medicine, 229-HSV, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail:

Related

No related items
American Journal of Respiratory and Critical Care Medicine
166
8

Click to see any corrections or updates and to confirm this is the authentic version of record