484080

2013

TAR731753465813484080Therapeutic Advances in Respiratory DiseaseM Miravitlles, A Sicras

Therapeutic Advances in Respiratory Disease

Original Research

Costs of chronic obstructive pulmonary

disease in relation to compliance with
guidelines: a study in the primary care
setting

Ther Adv Respir Dis

(2013) 7(3) 139150
DOI: 10.1177/
1753465813484080
The Author(s), 2013.
Reprints and permissions:
http://www.sagepub.co.uk/
journalsPermissions.nav

Marc Miravitlles, Antoni Sicras, Carles Crespo, Maribel Cuesta, Max Brosa, Jordi Galera,
Raquel Lahoz, Marta Lleonart and Maria Isabel Riera

Abstract
Background: The aim of this study was to analyse the economic impact of nonadherence to
the Global Initiative for Obstructive Lung Disease (GOLD) guidelines in patients with chronic
obstructive pulmonary disease (COPD).
Methods: A retrospective analysis was carried out on a claim database. Patients aged at least
40 years with a diagnosis of COPD were eligible for this analysis. Demographics, medical data
and use of resources were collected and direct and indirect costs were analysed (from January
2008 to June 2009). A probabilistic multivariate sensitivity analysis of avoided costs was carried
out. All results are presented in annualized form and costs are expressed in Euros (2009).
Results: A total of 1365 patients were included, 79.5% were men. The mean age (standard
deviation) was 71.4 (10.3) years, the mean forced expiratory volume in 1 s (FEV1) was 65.3%
and they had a COPD history of 5.5 (2.9) years. Patients were divided into an adherent group
and a nonadherent group depending on whether therapeutic recommendations according to
severity defined in the GOLD guidelines (2007) were followed. Patients in both groups were
also classified as having stage II (FEV1 < 80% and < 50%) or stage III disease (FEV1 < 50% and
30%). The total annual drug cost per patient in the nonadherent group was 771.5 while it
was only 426.4 for the adherent group. The average direct cost per patient per year in the
nonadherent stage II group was 1465 (971) and it rose to 2942 (1918) for patients in the
nonadherent group with stage III disease. The potential saving from the implementation of the
GOLD guidelines in stage II COPD amounted to 758 per patient per year (68% saving on drug
cost). In contrast, the cost for patients with stage III disease was higher in the adherent group
versus the nonadherent group (2468).
Conclusions: The cost of COPD may vary according to compliance with the GOLD guidelines.
The cost observed for patients with stage II disease is higher than expected in patients
who adhere to treatment, but patients with stage III disease treated according to the GOLD
guidelines had significantly higher treatment costs.
Keywords: chronic obstructive pulmonary disease, cost, primary care, Spain

Introduction
The prevalence of chronic obstructive pulmonary
disease (COPD) in the general population is
believed to be between 1% and 10% [Halbert
etal. 2006; Chapman et al. 2006; Pea et al. 2000]
and it is estimated that around 10.2% of people

aged 4080 in Spain have this condition [Pea

etal. 2000; Miravitlles etal. 2009b]. COPD is the
fifth biggest cause of death among men in Spain
and the seventh among women, with annual rates
of 61 per 100,000 people and 20 per 100,000

Correspondence to:
Marc Miravitlles, MD
Pneumology Department,
Hospital Universitari
Vall dHebron, Pg. vall
dHebron 119129, 08035
Barcelona, Spain
marcm@separ.es
Antoni Sicras, MD
Direccin de Planificacin,
Badalona Serveis
Assistencials, Badalona,
Spain
Carles Crespo, MStats,
MHE, BSc
Department of Statistics,
University of Barcelona,
and Health Economics
and Outcome Strategies
Department, Oblikue
Consulting, Barcelona,
Spain
Maribel Cuesta, MSc,
Max Brosa, MSc
Health Economics and
Outcome Strategies
Department, Oblikue
Consulting, Barcelona, Spain
Jordi Galera, BSc,
Raquel Lahoz, MSc,
Marta Lleonart, MD,
Maria Isabel Riera, MSc
Novartis Farmacutica SA,
Barcelona, Spain

http://tar.sagepub.com 139

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

people respectively in 2005 [Centro Nacional de
Epidemiologa, 2009].
The high prevalence and associated healthcare
costs, especially those generated by hospital
admissions [Strassels etal. 1996; Grasso etal.
1998; Sullivan etal. 2000; Miravitlles etal. 2003],
have prompted guidelines to be established at an
international level to improve the management of
these patients. In fact, a recent audit of admitted
patients in Spain has provided information about
the great variability in patterns of treatment and
outcomes [Pozo-Rodrguez etal. 2012]. This variability may be more relevant in patients with high
risk of readmission, those with increased dyspnoea,
and impaired lung function and exercise capacity
[Garca-Polo etal. 2012; Alczar etal. 2012].
Currently, the Global Initiative for Obstructive
Lung Disease (GOLD) guidelines are the most
widely recognized [GOLD, 2011]. However, dissemination of and compliance with guidelines is
not optimal and treatment patterns do not always
comply with existing guidelines. This would
explain the variability in the use and costs of treatment for COPD [Miravitlles etal. 2003, 2009a;
Miravitlles and Figueras, 2001; de Miguel Dez,
2005; Masa etal. 2004] and, in turn, the increasing interest in identifying the actual patterns of
use in these patients [Miravitlles etal. 2008]. The
objective of this study was to evaluate the real
costs associated with COPD according to adherence to the GOLD recommendations.
Materials and methods
Design
A multicentre, retrospective, observational study
estimated the annual cost of treating patients with
COPD using the records in a medical database of
patients treated in primary care in Spain. The
patients were classified into two groups: those in
whom the GOLD guideline recommendations
were adhered to (adherent group) and those in
whom they were not (nonadherent group). The
analysis was carried out from a healthcare perspective but was broadened to reflect the social
perspective (direct and indirect costs). The follow-up period for the observational study as a
whole was 18 months (from January 2008 to June
2009), although the results were annualized. The
costs have been expressed in Euros (2009) and
the costs of all drugs analysed were reimbursed by
the National Security System.

Sample size
Approximately 61% of patients with COPD in
Spain are treated in the primary care setting, with
an average of 5.3 [standard deviation (SD) 12.4]
visits per year. Of these, 1.3 (SD 4.7) are to specialists, 0.7 (SD 3.0) to emergency services and patients
spend an average of 1 day in hospital (SD 5.3)
[Izquierdo-Alonso and de Miguel-Dez, 2004].
From these figures, it was estimated that a sample
of 624 patients would allow a difference of 2 days to
be detected in the calculation of the mean total
number of visits, assuming a maximum SD of 25.5
days and a confidence interval of 95%. The sample
was then overestimated by an additional 10% to
avoid any possible bias caused by assuming that the
resources followed a normal distribution.
Ethics Committee
The study was approved by the Ethics Committee
of the Hospital Clnic de Barcelona and patient
anonymity and data confidentiality were preserved. Data were extracted for analysis from the
medical records so that the patients being assessed
could not be identified.
Study population
The study population consisted of men and
women belonging to six primary care centres
(EAP Mart Juli, EAP La Riera, EAP Montigal,
EAP La Morera, EAP Nova Lloreda and EAP
Montgat-Tiana), managed by Badalona Serveis
Assistencials SA. These centres provide healthcare
coverage to 107,208 people, 15.6% of whom are
over the age of 64, in a medium-to-low socioeconomic, predominantly industrial, urban setting.
The patients had an established clinical diagnosis
of COPD plus a postbronchodilator forced expiratory volume in 1 s (FEV1) of less than 80% of the
predicted value. They were aged 40 years or above
and were registered in the centres long-term
repeat-prescription schemes. Patients with a history of asthma, allergic rhinitis or atopy were
excluded, as were those on treatment with antiallergy drugs (antihistamines, cromones).
The computerized medical records were created
using the OMI-AP tool (software package for primary care), which is a patient-centred, problem-orientated record model which allows patient progress
to be recorded according to episodes, reasons for
consultation or diagnoses for care. Demographic
data (date of birth and sex), FEV1 measurements
and certain elements of previous medical history

140 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

M Miravitlles, A Sicras et al.

(conditions at the time of inclusion) which might
affect the use of resources by patients with COPD
were obtained from the medical records at the different centres. The healthcare resources compiled were
primary care appointments, specialist referrals, diagnostic tests ordered, visits to emergency services,
hospitalizations (number of admissions and duration
of stay in days), days of temporary sick leave, prescribing of oxygen therapy and drugs administered
during the 18-month study period. This database
has been validated and used in previous studies
[Rubio Gil et al. 2011].
There was no information available on the database on the FEV1/forced vital capacity ratio or the
number of exacerbations for each patient. The
number of exacerbations was therefore defined as
the maximum number among the number of visits to the emergency services, hospital admissions
and the number of prescriptions for systemic corticosteroids or antibiotics.
Drug treatment for chronic obstructive
pulmonary disease according to the Global
Initiative for Obstructive Lung Disease
Although a new GOLD document was launched in
November 2011 [GOLD, 2011], the lack of information on Modified Medical Research Council
Dyspnea Scale and COPD Assessment Test scores
does not allow us to investigate adherence to these
recommendations. Moreover, the guidelines in place
at the time of data collection were the previous
GOLD guidelines [Rabe, 2007]. The GOLD 2007
recommendations were based on stepwise drug
treatment according to the FEV1, with the first treatment step in any stage being smoking cessation
[Rabe, 2007]. For patients with stage II disease,
when the FEV1 is between 50% and 80%, the recommended treatment is the addition of one or more
long-acting bronchodilators. In patients with severe
stage disease (III and IV), if they have more than two
exacerbations per year, the addition of inhaled corticosteroids is justified. For patients with stage III disease (FEV1 between 30% and 50%), theophylline is
the next bronchodilator to add and in patients with
stage IV disease, with FEV1 less than 30%, the
option is to combine all the therapeutic groups.
Distribution of treatment according to the
Global Initiative for Obstructive Lung Disease
Considering the pharmacological recommendations according to stage and the number of exacerbations, the patients with COPD were classified

as as per guidelines (adherent group) or not as

per guidelines (nonadherent group) as follows:
(1)
Patients with stage II disease were
included as adherent except when they
had prescriptions of fixed combinations
of 2-adrenergic agonists and inhaled
glucocorticoids or inhaled glucocorticoids alone.
(2) Patients with stage III disease with two or
fewer exacerbations per year were subject
to the same classification criteria as
patients with stage II disease.
(3) Patients with stage III disease with two
or more exacerbations per year were
included as adherent except when they
had prescriptions of systemic glucocorticoids or systemic antibiotics.
Costs
To assess direct costs, the drug cost [Consejo
General de Colegios Oficiales de Farmacuticos,
Base de datos del Medicamento, 2009] and the
unit cost applied to each resource [Gisbert and
Brosa, 2009] were used. The indirect costs were
calculated using the human capital method, which
consists of applying the gross earnings standardized by sex and age group using official data from
the Instituto Nacional de Estadstica (INE;
National Statistics Institute) 2007 salary structure survey for each day of work lost [INE, 2007].
The gross earnings refer to the income, in cash
and in kind, paid to workers for time worked or
work done, along with payment for periods of
time not worked, such as vacation leave and public holidays. The costs coming from other years
were updated to 2009 according to the Spanish
consumer price index [INE, 2009].
Although it was possible to identify patients with
oxygen therapy, as there was no record of the actual
amounts used, the annual consumption and corresponding cost had to be estimated on the basis of
the average cost of oxygen therapy per session
[Gisbert and Brosa, 2009], multiplied by the number of sessions in a year adjusted by the level of
compliance observed in the Spanish study by Prez
de Alejo and colleagues [Prez de Alejo et al. 2009].
Statistical analysis
Descriptive and inferential statistics were used to
analyse the data, with the 2 test or Fishers exact
test to compare qualitative variables and Students

http://tar.sagepub.com 141

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

Figure 1. Study flowchart. COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s;
GOLD, Global Initiative for Obstructive Lung Disease.

t test, analysis of variance, MannWhitney U test

or KruskallWallis test, depending on the application conditions, for quantitative variables. Prior to
the comparisons, the application of parametric
methods was studied (KolmogorovSmirnov),
and if this was not possible, nonparametric tests
were performed. To evaluate the cost comparisons, they were first normalized by BoxCox
transformation. The statistical analysis was carried out with the SPSS 15.0 statistics package and
p value less than 0.05 was chosen as the level of
significance a priori.

age of 71.4 years (SD 10.3) and time since COPD

diagnosis of 5.6 years (SD 3.0). The mean FEV1
was 65.3% (SD 11.0%), with 7.5% (n = 103) of
the patients having a FEV1 between 30% and
50% and none having an FEV1 less than 30%.
The majority of patients had some type of associated comorbidity (89.6%) in addition to COPD,
the most common being hypertension (52.7%)
and dyslipidaemia (41.2%). Among the patients
with stage III COPD, there was a higher trend in
the number with ischaemic heart disease (15.5%
versus 22.3%, p = 0.071).

A probabilistic sensitivity analysis was carried out

using the bootstrap simulation technique to be
able to incorporate the uncertainty of the model
parameters into the results of the study. This type
of simulation allowed us to validate the results by
generating 1000 samples of the same size (n = 500)
and obtaining a measure of the variability of our
results. We were thus able to carry out an internal validation (robustness) of our results, thereby
observing how modifying the type of patient
cohort studied would change the result.

Approximately 18.0% of the patients included

belonged to the Adherent group, which had 7.7%
more men than the nonadherent group (p = 0.007).
In terms of risk factors, it was observed that the
percentage of smokers in the nonadherent
group was significantly lower than in the adherent group (24.9% versus 34.2%, p = 0.003)
(Table 1).

Results
The study rationale and the patient characteristics
according to COPD stage are shown in Figure 1.
A total of 1365 patients, primarily men (79.5%),
were included (Table 1 and Figure 1) with a mean

Use of resources
The patients made a total of 16,824 visits per
year, with a mean of 8.3 (SD 4.4) and 7.8 (SD
4.2) visits per patient per year (to primary care,
interdisciplinary consultations and visits to specialists) for the nonadherent group and adherent
group patients respectively (p = 0.103) (Table 2).
The annual mean for visits to specialists was

142 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

M Miravitlles, A Sicras et al.

Table 1. Demographic characteristics of the population studied.
Variable

Overall (N = 1365)

Adherent (N = 246) Nonadherent (N = 1119)

p Value

Age, years (SD)

Sex, men % (SD)
FEV1 (%), mean (SD)
Time since COPD diagnosis,
years (SD)
Comorbidity, %
Hypertension, %
Diabetes mellitus,%
Dyslipidaemia, %
Current smoker, %
Alcohol, %
Ischaemic heart disease, %
Cerebrovascular accident,%
Heart failure, %
Dementia, %
Depression, %

71.4 (10.3)
79.5 (40.4)
65.3 (11.0)
5.6 (3.0)

71.5 (10.5)
85.8 (35.0)
63.6 (12.9)
5.1 (2. 8)

71.4 (10.3)
78.1 (41.4)
65.7 (10.5)
5.7 (3.0)

0.822
0.007*
0.081
0.005*

89.6
52.7
24.1
41.2
26.6
5.6
16.0
6.0
21.9
2.7
17.3

91.1
49.2
19.9
40.6
34.1
6.9
18.7
6.5
26.4
3.7
18.7

89.3
53.5
25.0
41.3
24.9
5.3
15.5
5.9
20.9
2.5
17.0

0.407
0.217
0.090
0.854
0.003*
0.310
0.210
0.717
0.058
0.312
0.518

COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s; SD, standard deviation. *statistically
significant.

Table 2. Annual use and unit costs of resources.

Resources

Unit cost ( 2009)

Adherent

Nonadherent

p Value

Primary care visits (n)

Interdisciplinary visits (n)
Visits to specialist (n)
Visits to emergency service (n)
Hospital admission (days)
Laboratory (n)
Radiology (n)
Spirometry (n)
Oxygen therapy (days)
Sick leave (days)

38.70
78.40
118.10
163.33
376.35
4.46
13.93
38.16
5.83
INE 2007

3.0 (2.6)
4.1 (2.4)
0.7 (1.5)
0.30 (1.0)
1.62 (5.8)
3.81 (2.4)
1.73 (1.4)
0.19 (0.3)
24.11 (60.3)
1.52 (11.9)

3.2 (2.5)
4.1 (2.3)
0.9 (1.9)
0.25 (0.8)
0.27 (1.7)
3.89 (2.4)
1.74 (1.4)
0.23 (0.3)
11.07 (42.5)
1.41 (10.7)

0.059
0.940
0.008*
0.835
<0.001
0.629
0.834
0.076
<0.001
0.088

Data presented as mean (standard deviation).

*p < 0.01
INE, Instituto Nacional de Estadstica (National Institute of Statistics).

higher in the nonadherent group than in the

adherent group (p = 0.008) (Table 2). A closer
look at patients with stage II disease showed that
the adherent group had less access to specialists
(18.8% versus 35.3%; p < 0.001) and made fewer
visits (p < 0.001).
The adherent patients required admission to hospital for an annual mean of 1.6 days. In contrast,
the annual mean for nonadherent patient admissions was 0.3 days (p < 0.001).

In terms of diagnostic tests, spirometry was only

performed an average of 0.19 times per year in
adherent patients while the figure for nonadherent patients was 0.23 (p = 0.076). The oxygen
therapy prescribing rate for adherent patients was
double that for nonadherent patients (24.1 versus
11.1, p < 0.001).
Moreover, it was found that 2.8% of adherent
patients and 5.5% of nonadherent patients took
time off work (p = 0.081), although no significant

http://tar.sagepub.com 143

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

Table 3. Description of treatments used (%) in the two study groups and in total.

2-Adrenergic agonists and inhaled

glucocorticoids
Formoterol and budesonide
Salmeterol and fluticasone
Others
2-Adrenergic agonists
Formoterol
Salbutamol
Salmeterol
Terbutaline
Anticholinergics
Ipratropium bromide
Tiotropium bromide
Inhaled glucocorticoids
Beclomethasone
Budesonide
Fluticasone
Xanthines (theophylline)
Systemic glucocorticoids
Deflazacort
Methylprednisolone
Prednisone
Cough and cold remedies
Systemic antibiotics
Others

Overall (%)

Adherent (%)

Nonadherent (%)

53.3

14.6

61.8

23.1
34.1
0.1
61.2
8.4
52.7
6.0
4.5
61.5
33.0
39.0
29.5
1.3
25.6
2.6
4.7
13.8
1.5
1.8
11.4
73.2
61.0
4.8

6.1
9.8
0.0
71.9
9.3
60.6
7.3
3.3
72.4
40.2
43.9
6.5
0.4
5.3
0.8
3.3
11.4
0.8
2.4
9.3
68.7
52.4
3.7

26.8
39.4
0.2
58.8
8.1
50.9
5.7
4.8
59.1
31.4
38.0
34.5
1.5
30.1
3.0
5.0
14.4
1.7
1.6
11.9
74.2
62.8
5.0

differences were observed between the two groups

with respect to the number of days on sick leave.
Nevertheless, while not significant, overall, the
mean number of days on sick leave was higher for
the adherent group than for the nonadherent
group (Table 2).
Drug treatment
COPD-related drug use is shown in Table 3. The
most widely used treatments were cough and cold
remedies (73.2%), anticholinergics (61.5%), 2adrenergic agonists (61.2%), systemic antibiotics
(61.0%) and 2-adrenergic agonists and inhaled
glucocorticoids (53.3%). In the Adherent group,
2-adrenergic agonists (71.9% versus 58.8%, p <
0.001) and anticholinergics (72.4% versus 59.1%,
p < 0.001) were prescribed most often. In contrast, inhaled glucocorticoids (6.5% versus 34.5%,
p < 0.001) were prescribed less. It was found that
14.6% of adherent patients were prescribed 2adrenergic agonists with inhaled glucocorticoids,

as opposed to 61.8% of nonadherent patients (p <

0.001), which was due in part to the way in which
patients were classified. Prescribing theophylline,
which according to the GOLD guidelines is only
recommended for patients with stage IIIIV disease,
was found for 4.8% of the nonadherent patients with
stage II disease.
Cost of clinical practice
The annual drug cost for nonadherent patients
was essentially due to the use of fixed combinations of 2-adrenergic agonists and inhaled glucocorticoids (404.97) and anticholinergics
(198.05). For adherent patients, the annual
drug cost was primarily due to anticholinergics
(200.77) and 2-adrenergic agonists (123.49).
Consequently, the nonadherent group annual
drug cost was 771.47 per patient/year (median
657.63) and that of the Adherent group was
426.40 per patient/year (median 153.29),
345.07 less (Table 4).

144 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

M Miravitlles, A Sicras et al.

Table 4. Mean drug costs per patient () (18 months).
Overall (N = 1365)
2-Adrenergic agonists
Formoterol
Salbutamol
Salmeterol
Terbutaline
2-Adrenergic agonists and inhaled
glucocorticoids
Formoterol and other antiobstructive
syndrome agents
Salmeterol and other antiobstructive
syndrome agents
Others
Anticholinergics
Ipratropium bromide
Tiotropium bromide
Inhaled glucocorticoids
Beclomethasone
Budesonide
Fluticasone
Xanthines: theophylline
Systemic glucocorticoids
Deflazacort
Methylprednisolone
Prednisone
Cough and cold remedies
Antibiotics, systemic use
Others
Drug cost per patient
18 months
Annual

Adherent (N = 246)

Nonadherent (N = 1119)

94.5 (232.8)
41.7 (178.4)
23.3 (50.8)
27.7 (141.9)
1.8 (12.5)
531.4 (766.3)

94.1 (190.2)
28.6 (114.6)
27.7 (69.2)
37.4 (148.6)
0.5 (4.1)
185.2 (529.7)

94.5 (241.3)
44.6 (189.5)
22.3 (45.7)
25.6 (140.4)
2.1 (13.7)
607.5 (789.1)

172.8 (477.6)

55.0 (273.2)

198.7 (508.1)

358.3 (675.6)

130.3 (449.4)

408.5 (706.1)

0.2 (5.0)
297.8 (397.3)
30.6 (87.3)
267.2 (400.5)
78.4 (208.7)
3.0 (28.9)
54.6 (153.4)
20.8 (147.8)
3.4 (18.2)
3.4 (24.0)
1.2 (19.3)
1.0 (13.2)
1.3 (5.6)
17.7 (38.2)
26.2 (51.2)
10.8 (88.2)

0 (0)
301.2 (377.5)
49.8 (136.8)
251.4 (374.5)
15.6 (96.7)
0.1 (1.2)
7.8 (45.2)
7.7 (86.2)
2.7 (17.0)
2.1 (10.0)
0.2 (2.0)
0.9 (7.1)
1.1 (6.1)
13.9 (29.7)
22.2 (53.5)
2.1 (27.5)

1063.9 (1044.0)
709.3 (696.0)

639.6 (852.6)
426.4 (568.4)

0.2 (5.6)
297.1 (401.7)
26.3 (71.3)
270.7 (406.0)
92.2 (223.7)
3.6 (31.8)
64.9 (166.4)
23.67 (158.0)
3.5 (18.5)
3.7 (26.0)
1.5 (21.3)
1.0 (14.2)
1.3 (5.5)
18.6 (39.7)
27.0 (50.6)
12.7 (96.5)

1157.2 (1059.3)
771.5 (706.2)

Data presented as mean (standard deviation).

The mean direct cost per patient and year for the
nonadherent patients with stage II disease was
1465 (SD 971; median 1291) and, in nonadherent patients with stage III disease it was
2942 (SD 1918; median 2595). The relative weight of resource use over the total costs
differed among the nonadherent patients according to the severity of the COPD. While the mean
direct costs in patients with stage II disease were
due to drugs (50.9%) and visits and diagnostic
tests (41.5%), for patients with stage III disease,
the direct costs mainly came from hospitalizations (29.7%), drugs (29.1%), and to a lesser
extent, from visits and diagnostic tests (22.6%)
(Figure 2).

The indirect costs for nonadherent patients represented 9.4% and 16.9% of the total cost for stages
II and III disease respectively. In Adherent
patients with stage III disease, the indirect costs
represented 12.9% of total costs (Table 5).
Cost differences
The results for the overall annual cost (direct plus
indirect costs) per patient in the defined scenarios
and the calculation of the difference in costs are
shown in Table 5. It can be seen that the mean
total annual cost per patient with stage II disease
was 859 (median 706) for the Adherent group
and 1617 (median 1336) for the nonadherent

http://tar.sagepub.com 145

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

Figure 2. Distribution of annual direct costs according to adherent stage and patient. COPD, chronic
obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s; adherent Global Initiative for Obstructive
Lung Disease.

group. The costs for the Adherent patients were

therefore 758 less than for the nonadherent
patients (68% of the savings were from drug
costs). Analysis of the cost of the Adherent
patients with stage III disease showed that the
direct costs for the nonadherent patients were
2294 less than those for the adherent patients
and the increase in total costs was 2468. This
difference was due to the increase in hospitalization costs observed (58%) in the adherent group.
In short, from the simulation of 1000 bootstrap
scenarios with 500 patients, it was found that with
both scenarios, greater adherence to the GOLD
guidelines would produce average savings of
approximately 501996 for stage II disease and
an increase in costs of 14963277 for stage III
disease (Figure 3).
Discussion
Strict adherence to the GOLD guideline treatment recommendations is associated with drug
savings for the health service when the patients
have stage II disease. However, we have to be
aware that there may be patients who received
more treatment than recommended because
their symptoms were more severe. In contrast,
when patients with stage III disease were treated
according to guidelines, this led to an increase in
total costs, mainly due to an increase in the

number of hospitalizations and hospitalization

days. The cost of extra exacerbations or admissions cannot therefore be attributed to being as
per guidelines or not as per guidelines. Because
this is not a randomized study, we cannot rule
out a bias due to indication, when patients with
more severe disease were treated according to
guidelines compared with patients with fewer
risk factors for admission who could have
received treatment that does not follow GOLD
guidelines. It has to be borne in mind that exacerbations are responsible for a large part of the
direct and indirect costs of COPD [Miravitlles
etal. 2003] and for the greater perception of
reduced quality of life for the patients [Miravitlles
etal. 2004].
To date, the only comparison made in our area
between the GOLD recommendations and clinical practice looked at drug costs using a theoretical method [Miravitlles etal. 2009a]. This study
identified that the mean annual cost of drug treatment for a patient with COPD was 12181314
and the potential savings derived from adherence
to the GOLD recommendations was on average
between 198 and 293. In contrast, in our
study, the drug cost from the retrospective analysis was lower (709) than previously estimated.
Moreover, when we considered the costs of treatment according to GOLD recommendations, an

146 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

M Miravitlles, A Sicras et al.

Table 5. Annual cost per patient () of the clinical practice, GOLD costs and nonadherent costs.

GOLD II (FEV1
5080%)

GOLD III (FEV1

3049%)

Overall

Adherent (n =202)
Nonadherent
(n = 1060)
Cost difference
Adherent (n = 44)
Nonadherent
(n = 59)
Cost difference
Adherent (n = 246)
Nonadherent
(n = 1119)
Cost difference

Direct costs

Indirect costs

Mean total cost

778.9 (436.2)
1464.7 (971.4)

80.0 (696.6)
152.0 (1202.7)

858.9 (816.3)
1616.7 (1537.0)

685.8 (601.7769.9)
5236.5 (3225.1)
2942.1 (1918.3)

72.0 ( 99.3 to 243.3)

772.6 (3499.5)
598.9 (2673.3)

757.8 (611.7903.8)
6009.1 (4534.5)
3541.1 (3239.8)

2294.3 (3385.9 to 1202.8)

1576.2 (2216.3)
1542.6 (1092.6)

173.7 (1379.8 to 1032.5)

203.9 (1618.1)
175.6 (1323.2)

2468.0 (3987.2 to 948.8)

1780.1 (2840.3)
1718.2 (1722.6)

33.6 (319.1 to 251.9)

28.3 (219.1 to 162.4)

61.9 (432.5308.6)

Data presented as mean (standard deviation) or mean (95% confidence interval).

Cost difference is nonadherent costs minus adherent costs.

Figure 3. Costs avoided according to forced expiratory volume in 1 s (1000 bootstrap simulations of 500
samples). Nonparametric bootstrapping draws samples (1000) from our sampled data. The figure shows
the distribution of avoided costs between the Adherent group and the nonadherent group for each bootstrap
sample generated. Adherence to the GOLD guideline produces savings for stage II and increases costs for
stage III. GOLD, Global Initiative for Obstructive Lung Disease.

additional saving of 283 patient/year on average

was seen.

2009a] may be due to the fact that, in our case,

92.5% of patients had an FEV1 above 50%

Although the potential drug savings are of the same

order, the difference in average drug cost between
our study and a previous analysis [Miravitlles etal.

Aside from the possible comparison with the

GOLD guidelines, various studies in recent years
have estimated the economic burden of this

http://tar.sagepub.com 147

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

disease in Spain, showing that the total annual
direct cost of COPD per patient varies between
1100 and 3400, with a mean of approximately
2000 [Miravitlles and Figueras, 2001; Miravitlles
etal. 2003; de Miguel Dez, 2005; Masa etal.
2004; Izquierdo-Alonso and de Miguel-Dez,
2004; de Miguel Dez et al. 2008]. In our study we
found the costs for patients with stage II disease to
be in line with previous estimations, but the direct
cost observed for patients with stage III disease
was around 1900 higher than the average. This
point should be taken into account when interpreting the results because of the small number of
patients included with stage III disease.
When looking at the cost distribution pattern,
our results confirm the well known fact that the
more severe the COPD, the higher the proportion of hospital costs and the lower the proportion of drug costs. In one initial study, also
performed in primary care, drug costs were
responsible for 42% of the costs associated with
mild COPD, 44% of those associated with moderate COPD and 40% of those associated with
severe COPD [Miravitlles et al. 2003]. The study
by Izquierdo-Alonso and de Miguel-Dez showed
that the drug costs represented 43% of the total
direct costs in patients with stage II COPD,
while this figure went down to 37.6% and 28.4%
for COPD stages III and IV respectively
[Izquierdo-Alonso and Miguel-Dez, 2004].
These data are similar to those from the recent
study by de Miguel Dez and colleagues which
found drug costs to be 26% [de Miguel Dez et
al. 2008]. In our study, it was observed that the
weight of the drug costs over the direct costs for
patients with stage II disease was similar to that
found by Izquierdo-Alonso and de Miguel-Dez
[Izquierdo-Alonso and de Miguel-Dez, 2004].
Part of this extra cost in GOLD II is due to the
inadequate prescription of inhaled corticosteroids (Izquierdo-Alonso and Rodrguez-Gonzlez
Moro, 2012). However, the relative drug cost for
patients with stage III disease in the nonadherent group in our study was closer to that observed
for patients with stage disease in the IzquierdoAlonso and Miguel-Dez study [IzquierdoAlonso and de Miguel-Dez, 2004].
We also found the prescribing rate for theophylline
to be lower among patients included in our study
compared with other studies in Spain, possibly due
to the fact that only 103 patients with FEV1 between
30% and 50% were included in the study and there
were no patients with FEV1 below 30%, these

patients being the main recipients of that treatment.

Along the same lines, looking at the figures from
the inhaled steroids in obstructive lung disease in
Europe (ISOLDE) [Jones et al. 2003] and IMpacto
de las agudizaciones de la EPOC (IMPAC)
[Miravitlles et al. 2004] studies, a higher percentage
of patients had two or more exacerbations (mean of
34.4%) than in our study (23.3%); however, these
studies showed similar average costs. One of the
limitations of our study is the fact that we were unable to identify exacerbations in the clinical records
and had to work with an estimate. This may have
led us to underestimate the frequency of exacerbations. Even so, in our study, the patients who were
admitted spent an average of 11.3 days in hospital,
somewhat longer than that observed in the recent
study by Almagro and colleagues where, in a population with an FEV1 of 43.2% and mean age of 73.7
years, patient admissions lasted an average of 9.8
days [Almagro et al. 2010].
Apart from those discussed above, another important limitation of this study was the unavoidably
subjective classification of patients with COPD
on the basis of the therapeutic recommendations
of the GOLD guidelines. Furthermore, while
essentially specifying drug recommendations, the
GOLD guidelines do not identify follow-up regimens, which would make it possible to define the
ideal use of healthcare resources in the management of COPD. This explains our decision to
identify and classify the patients according to
their drug use. Moreover, in comparison with the
GOLD guidelines, it was observed that one of the
first recommendations, namely smoking cessation
as a first treatment step to prevent exacerbations
in patients with COPD, was not adhered to. In
our population, although the percentage of smokers in the most severe cases was low, it was found
that, overall, 26.6% were still smoking. Since this
study was performed in a limited geographic area,
the results may not be fully extrapolated to the
rest of the country.
As a result, we believe that this type of study needs
to be promoted if, in the future, we are to have better information to allow decisionmakers to reliably
deal with COPD. As with other chronic smokingrelated diseases, the distribution of COPD basically
depends on exposure to smoking and the gradual
aging of the population [Soriano et al. 2010].
Initiatives such as ours, which are evaluating strategies to improve the management of patients and
reduce unnecessary costs, will therefore be of crucial importance for the future control of COPD.

148 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

M Miravitlles, A Sicras et al.

To conclude, our study shows that adherence to
the GOLD guideline recommendations could
lead to a reduction in drug use which would
directly translate into financial savings for the
Spanish Health Service of at least 758 per
patient per year in those with stage II COPD. The
new approach to treatment of COPD based on
clinical phenotypes [Miravitlles et al. 2012] may
help us to further improve the care of patients
with this devastating disease.
Acknowledgements
M. Miravitlles was involved in study design, analysis and interpretation of data, and writing the
manuscript. J. Galera, R. Lahoz, M. Riera and M.
Lleonart developed the idea for the study, supervised it in its entirety and were involved in its
design. A. Sicras provided data, analysis and a
critical review of the manuscript. M. Brosa, M.
Cuesta and C. Crespo were involved in study
design, carried out the research and the data analysis and drafted the manuscript. All the investigators contributed to the final version of the
manuscript.
Funding
The study was funded by a grant from Novartis.
Conflict of interest statement
J. Galera, R. Lahoz, M. Riera and M. Lleonart
are employees of Novartis Farmacutica SA
(Spain), the company sponsoring this study. M.
Brosa, M. Cuesta and C. Crespo worked in an
independent consultant company and they
received funds from Novartis. Marc Miravitlles
has received speaker fees from Boehringer
Ingelheim, Pfizer, AstraZeneca, Bayer Schering,
Novartis, Talecris, Takeda-Nycomed, Merck,
Sharp & Dohme and Novartis, and consulting
fees from Boehringer Ingelheim, Pfizer, GSK,
AstraZeneca, Bayer Schering, Novartis, Almirall,
Merck, Sharp & Dohme and Takeda-Nycomed.

References
Alczar, B., Garca-Polo, C., Herrejn, A., Ruiz, L.,
de Miguel, J., Ros, J. et al. (2012) Factors associated
with hospital admission for exacerbation of chronic
obstructive pulmonary disease. Arch Bronconeumol 48:
7076.
Almagro, P., Lpez Garca, F., Cabrera, F., Montero, L.,
Morchn, D., Dez, J. et al. and Grupo Epoc de
la Sociedad Espanola de Medicina Interna (2010)
Comorbidity and gender-related differences in

patients hospitalized for COPD. The ECCO study.

Respir Med 104: 253259.
Centro Nacional de Epidemiologa (2009) Instituto de
Salud Carlos III, Ministerio de Ciencia e Innovacin.
Available from: http://www.isciii.es/htdocs/centros/
epidemiologia/ epidemiologia_presentacion.jsp
(accessed 15 May 2009).
Chapman, K., Mannino, D., Soriano, J., Vermeire, P.,
Buist, A., Thun, M. et al. (2006) Epidemiology and
costs of chronic obstructive pulmonary disease. Eur
Respir J 27: 188207.
Consejo General de Colegios Oficiales de
Farmacuticos, Base de datos del Medicamento
(2009) Explotacin de bases del BOT. Available at:
http://www.portalfarma.com (accessed 15 November
2009).
De Miguel Dez, J. (2005) Health care costs of asthma
and chronic obstructive pulmonary disease. Arch
Bronconeumol 41: 239241.
De Miguel Dez, J., Carrasco Garrido, P., Garca
Carballo, M., Gil de Miguel, A., Rejas Gutierrez,
J., Belln Cano, J. et al. (2008). Determinants and
predictors of the cost of COPD in primary care: a
Spanish perspective. Int J Chron Obstuct Pulmon Dis 3:
701712.
Garca-Polo, C., Alczar-Navarrete, B., RuizIturriaga, L., Herrejn, A., Ros-Lucas, J., GarcaSidro, P. et al. (2012) Factors associated with high
resource utilisation among COPD patients. Respir
Med 106: 17341742.
Gisbert, R. and Brosa, M. (2009) Base de Datos
de Costes Sanitarios eSALUD (base de datos en
Internet). Barcelona: Oblikue Consulting, SL.
Available at: http://www.oblikue.com/bddcostes/
(accessed 15 November 2009).
Global Initiative for Chronic Obstructive Lung
Disease (GOLD) (2011) Global Strategy for the
Diagnosis, Management and Prevention of COPD.
Available at: http://www.goldcopd.org (accessed 11
November 2012).
Grasso, M., Weller, W., Shaffer, T., Diette, G. and
Anderson, G. (1998) Capitation, managed care, and
chronic obstructive pulmonary disease. Am J Respir
Crit Care Med 158: 133138.
Halbert, R., Natoli, J., Gano, A., Badamgarav, E.,
Buist, A. and Mannino, D. (2006) Global burden
of COPD: systematic review and meta-analysis. Eur
Respir J 28: 523532.
INE (2007) Encuesta de Estructura Salarial:
Ganancia media anual por trabajador segmentado
por sexo y edad. Instituto Nacional de Estadstica.
Available at: http://www.ine.es (accessed 15 June
2009).

http://tar.sagepub.com 149

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014

Therapeutic Advances in Respiratory Disease 7 (3)

INE (2009) ndice de Precios al Consumo. Instituto
Nacional de Estadstica. Available at: http://www.ine.
es (accessed 15 November 2009).
Izquierdo-Alonso, J. and de Miguel-Dez, J. (2004)
Economic impact of pulmonary drugs on direct costs
of stable chronic obstructive pulmonary disease.
COPD 1: 215223.
Izquierdo-Alonso, J. and Rodrguez-Gonzlez Moro, J.
(2012) The excessive use of inhaled corticosteroids
in chronic obstructive pulmonary disease. Arch
Bornconeumol 48: 207212.
Jones, P., Willits, L., Burge, P. and Calverley, P. on
behalf of the Inhaled Steroids in Obstructive Lung
Disease in Europe investigators (2003) Disease
severity and the effect of fluticasone propionate on
chronic obstructive pulmonary disease exacerbations.
Eur Respir J 21: 6873.

Pea, V., Miravitlles, M., Gabriel, R., Jimnez-Ruiz,

C., Villasante, C., Masa, J. et al. (2000) Geographic
variations in prevalence and underdiagnosis of COPD:
results of the IBERPOC multicentre epidemiological
study. Chest 118: 981989.
Prez de Alejo, R., Caete, C., Sacristn, J., PecesBarba, G., Ramos, I., Navarro, A. et al. (2009)
Utilizacin del VisionOx para el registro del
cumplimiento objetivo de oxigenoterapia domiciliaria.
Arch Bronconeumol 45: 103.
Pozo-Rodrguez, F., Lpez-Campos, J., AlvarezMartinez, C., Castro-Acosta, A., Agero, R., Hueto,
J. et al. (2012). Clinical audit of COPD patients
requiring hospital admission in Spain: AUDIPOC
study. PLoS One 7: e42156.

Miravitlles, M., Brosa, M., Velasco, M., Crespo, C.,

Gobartt, E., Diaz, S. et al. (2009a) An economic
analysis of pharmacological treatment of COPD in
Spain. Respir Med 103: 714721.

Rabe, K., Hurd, S., Anzueto, A., Barnes, P.,

Buist, S., Calverley, P. et al. (2007) Global strategy
for the diagnosis, management, and prevention
of chronic obstructive pulmonary disease. GOLD
executive summary. Am J Respir Crit Care Med 176:
532555.

Miravitlles, M. and Figueras, M. (2001) El coste de la

enfermedad pulmonar constructiva crnica en Espaa.
Opciones para una optimizacin de recursos. Arch
Bronconeumol 37: 388393.
Miravitlles, M., Murio, C., Guerrero, T. and
Gisbert, R. (2003) Costs of chronic bronchitis and
COPD: a 1-year follow-up study. Chest 123:
784791.

SAGE journals

Miravitlles, M., Soriano, J., Garca-Ro, F., Muoz,

L., Duran-Tauleria, E., Sanchez, G. et al. (2009b)
Prevalence of COPD in Spain: impact of undiagnosed
COPD on quality of life and daily life activities.
Thorax 64: 863868.

Masa, J., Sobradillo, V., Villasante, C., JimnezRuiz, C., Fernndez-Fau, L., Viejo, J. et al. (2004)
Costes de la EPOC en Espaa. Estimacin a partir
de un estudio epidemiolgico poblacional. Arch
Bronconeumol 40: 7279.

Miravitlles, M., Ferrer, M., Pont, A., Zalacain, R.,

Alvarez-Sala, J., Masa, F. et al. for the IMPAC Study
Group (2004) Effect of exacerbations on quality of
life in patients with chronic obstructive pulmonary
disease. A 2 year follow up study. Thorax 59: 387395.

Visit SAGE journals online

http://tar.sagepub.com

guidelines (GesEPOC). Pharmacological treatment of

stable COPD. Arch Bronconeumol 48: 247257.

Miravitlles, M., Murio, C., Tirado-Conde, G.,

Levy, G., Muellerova, H., Soriano, J. et al. (2008)
Geographic differences in clinical characteristics and
management of COPD: the EPOCA study. Int J
Chron Obstuct Pulmon Dis 3: 803814.
Miravitlles, M., Soler-Catalua, J., Calle, M., Molina, J.,
Almagro, P., Quintano, J. et al. (2012) Spanish COPD

Rubio Gil, E., Martnez Pastor, A., Lpez-Picazo

Ferrer, J., Leal Hernndez, M., Martnez Navarro, A.
and Abelln Alemn, J. (2011) A study of follow-up
records and control of risk factors in stroke patients
followed up in primary care by computerised clinical
history (OMI-AP). Ictus-OMI study. Aten Primaria
43: 209210.
Soriano, J., Miravitlles, M., Borderas, L., DuranTauleria, E., Garca Ro, F., Martnez, J. et al. (2010)
Diferencias geogrficas en la prevalencia de EPOC
en Espaa: Relacin con hbito tabquico, tasas de
mortalidad y otros determinantes. Arch Bronconeumol
46: 522530.
Strassels, S., Sullivan, S. and Smith, D. (1996)
Characterization of the costs of chronic obstructive
pulmonary disease (COPD) in the US. Eur Respir J
J9(Suppl. 23): 421S.
Sullivan, S., Ramsey, S. and Lee, T. (2000) The
Economic burden of COPD. Chest 117: 5S9S.

150 http://tar.sagepub.com

Downloaded from tar.sagepub.com by Dejan Zujovic on December 11, 2014