Professional Documents
Culture Documents
ORIGINAL RESEARCH
The Asthma Control Test
TM
(ACT) as a predictor of GINA
guideline-defined asthma control: analysis of a
multinational cross-sectional survey
*Mike Thomas
a
, Stephen Kay
b
, James Pike
b
, Angela Williams
c
,
Jacqueline R Carranza Rosenzweig
c
, Elizabeth V Hillyer
d
, David Price
a
a
C entre of A cadem ic Prim ary C are, U niversity of A berdeen, Scotland, U K
b
A delphi Real W orld Products, Bollington, U K
c
G lobal H ealth O utcom es, G laxoSm ithKline R& D , Research Triangle Park, N orth C arolina, U SA
d
Respiratory Research Ltd, N orw ich, U K
Subm itted 28th M ay 2008; revised version received 18th Septem ber 2008; further revisions received 23rd O ctober 2008;
accepted 28th N ovem ber 2008; online 24th February 2009
Abstract
Aims: To evaluate w hether the A sthm a C ontrol Test
TM
(A C T) score is predictive of G lobal Initiative for A sthm a (G IN A ) guideline-defined
classification levels of asthm a control. The A C T is a validated, 5-item , patient-com pleted m easure of asthm a control w ith a recall period
of four w eeks.
Methods: C ross-sectional survey com paring A C T score and G IN A classification of asthm a control am ong 2949 patients attending prim ary
care physicians and specialists in France, G erm any, Italy, Spain, the U K, and the U SA .
Results: The area under the receiver operating characteristics curve for A C T score predicting G IN A control w as 0.84 (95% C I 0.820.85).
A n A C T score of <19 (not w ell-controlled asthm a) correctly predicted G IN A -defined partly controlled/uncontrolled asthm a 94% of the
tim e, w hile an A C T score of >20 predicted G IN A -defined controlled asthm a 51% of the tim e, w ith kappa statistic of 0.42, representing
m oderate agreem ent.
Conclusions: A n A C T score <19 is useful for identifying patients w ith poorly controlled asthm a as defined by G IN A .
2009 G eneral Practice A irw ays G roup. A ll rights reserved.
M Thom as, et al. Prim Care Resp J 2009; 18(1): 41-49.
doi:10.4104/pcrj.2009.00010
Keywords asthma, Asthma Control Test, control, GINA, exacerbations, multinational survey, patient-completed outcome assessment,
Europe, United States
* Corresponding author: D r M ike Thom as, C entre of A cadem ic Prim ary C are, U niversity of A berdeen, Scotland, A B25 2AY, U K.
Tel: +44 (0)1285 760671 Fax: +44 (0)1224 550683 E-m ail: m ikethom as@ doctors.org.uk
41
PRIMARY CARE RESPIRATORY JOURNAL
w w w .thepcrj.org
doi:10.4104/pcrj.2009.00010
Introduction
The 2006 update to the G lobal Initiative for A sthm a (G IN A )
guideline em phasises the im portance of evaluating asthm a
control, rather than asthm a severity, in order to guide asthm a
m anagem ent decisions. C lassification of disease severity is a
static m easure that, w hilst useful in initiating treatm ent, is less
helpful in guiding subsequent treatm ent.
1,2
G IN A guidelines
suggest that classification of asthm a control m ore directly
reflects the effectiveness of therapeutic interventions and
thus it m ay be m ore useful clinically. C urrent guidelines define
asthm a control as: no lim itations of activities; no nocturnal
sym ptom s; m inim al or no daytim e sym ptom s; m inim al or no
need for rescue therapy; norm al lung function; and no
exacerbations.
3-5
G uideline-defined asthm a control can be attained and
m aintained for the m ajority of patients eligible to participate
in a controlled trial setting,
6,7
but it is frequently not achieved
in real w orld practice.
8-11
Poorly controlled asthm a accounts
for a disproportionate share of the costs of asthm a and
represents a heavy socioeconom ic burden.
12-15
H ow ever, m any
patients w orldw ide have sub-optim ally controlled asthm a,
The full version of this paper, w ith A ppendices and online
Tables and Figures, is available online at w w w .thepcrj.org
Copyright GPIAG - reproduction prohibited
http://www.thepcrj.org
C
o
p
y
r
i
g
h
t
G
e
n
e
r
a
l
P
r
a
c
t
i
c
e
A
i
r
w
a
y
s
G
r
o
u
p
R
e
p
r
o
d
u
c
t
i
o
n
p
r
o
h
i
b
i
t
e
d
M Thomas et al.
42
and they are often not aw are that better control can be
achieved; m oreover, physicians m ay overestim ate levels of
control or the extent of im provem ent achieved w ith therapy,
often because of inadequate assessm ent.
16,17
In addition,
selection of asthm a control criteria not consistent w ith
current asthm a guidelines m ay ham per assessm ent.
18
The A sthm a C ontrol Test
TM
(A C T) w as developed as a
screening tool to address the need for a sim ple, rapidly-
com pleted assessm ent tool in clinical practice. The A C T is a
validated, patient-com pleted m easure of asthm a control
com prising five questions that assess activity lim itation, shortness
of breath, night-tim e sym ptom s, use of rescue m edication, and
patient overall rating of asthm a control over the previous four
w eeks (see A ppendix 1 online).
19-21
The questions are scored from
1 (w orst) to 5 (best), and the A C T score is the sum of the
responses, giving a m axim um best score of 25. A n A C T score of
19 dem onstrated the highest area under the receiver operating
characteristic (RO C ) curve, and thus a score of >20 is the optim al
cut-off point defining w ell-controlled asthm a over the previous
four w eeks
19-22
although, as N athan et al
19
describe, the cut-off
point can be chosen according to application.
It is im portant in clinical practice to identify patients
w hose asthm a is not w ell-controlled, since these patients
require review of their therapy as w ell as assessm ent of risk
factors for poor asthm a control.
23
The A C T w as designed for
use in daily practice as a supplem entary m easure to the
physicians assessm ent and/or lung function testing, but it has
not been validated as a predictor of G IN A -defined asthm a
control. The objective of this m ultinational cross-sectional
survey w as to evaluate w hether the A C T can predict G IN A -
defined asthm a control, w ith particular em phasis on the
binary split betw een G IN A -defined partly
controlled/uncontrolledasthm a versus controlledasthm a.
Methods
Disease Specific Programmes
The D isease Specific Program m es (D SP) are large,
m ultinational observational studies of clinical practice that are
conducted every 12 to 18 m onths by the A delphi G roup in
the U SA and in five European countries (France, G erm any,
Italy, Spain, and the U K).
24
D esigned to survey patients and
physicians on their perceptions of treatm ent effectiveness,
sym ptom s, and im pact of com m on chronic diseases, each
D SP is specific to a disease area and includes questionnaires
com pleted by up to 1000 physicians and 12,000 patients.
D escriptions of D SP m ethods for studying asthm a and allergic
rhinitis have been published.
25,26
The respiratory program m e w as initiated in 2000. H ere w e
sum m arise the m ethods specific to patients w ith asthm a
included in the sixth w ave (Respiratory D SP VI),
conducted in the first quarter of 2007.
Respiratory DSP VI survey participants and
procedures
Physicians recruited for the Respiratory D SP VI num bered 120
in France, G erm any, Italy, and Spain (50 prim ary care
practitioners, 50 pulm onologists or country equivalent, and
20 allergists in each country), 100 in the U K (50 prim ary care
practitioners and 50 chest specialists), and 180 in the U S (75
prim ary care practitioners, 75 pulm onologists, and 30
allergists). Larger sam ples of physicians w ere recruited from
m ore densely populated areas. Physicians w ere asked to
collect inform ation for their next six consecutive patients >12
years of age w ith physician-diagnosed asthm a, irrespective of
the reason for the consultation. The physician form s,
com pleted after the consultation w ith no direct input from
the patient, included patient dem ographic data, disease
sym ptom s and severity, diagnostic and treatm ent history, and
health resource utilisation m easures.
Patients included in the survey by their physicians could
then be invited to com plete a survey form im m ediately after
the consultation. C om pletion of the survey form w as not
obligatory, and physicians did not see or influence patient
responses. The study protocol follow ed ethical procedures
including verbal inform ed consent of all physicians and
patients for anonym ous and aggregated reporting of research
findings based on the questionnaires em ployed.
The patient survey included questions on asthm a
sym ptom s and severity, the im pact of asthm a and any
m easures taken by the patient to control their asthm a, as w ell
as satisfaction and com pliance w ith m edication, expectations
of therapy, and health resource utilisation. Physicians w ere
provided w ith an honorarium for study com pletion. Patient
com pensation depended on local country regulations (som e
w ere not com pensated, som e received vouchers, som e
received paym ent).
Outcome measures
Included in the Respiratory DSP VI survey w ere the ACT tool and
variables to identify asthm a control level as per G INA classification
(other survey data not reported here). The ACT w as included as
part of the patient-com pleted form , and G INA-defined asthm a
control w as determ ined prim arily from patient responses.
Translations of the A C T w ere validated. The translations of
the Respiratory D SP VI survey w ere m ade using forw ard-
backw ard translation and w ere not validated. Specific cognitive
debriefing w as not done before initiating the survey; how ever,
the form s w ere piloted to see how they w ere com pleted, given
that patients w ould be com pleting them alone.
The G IN A classification and our survey definitions of G IN A
asthm a control are sum m arised in Table 1, w ith differences
betw een the tw o noted (all relate to tim e fram e). Q uestions
relating to the first four item s in the G IN A classification
daytim e sym ptom s, lim itations of activities, nocturnal
PRIMARY CARE RESPIRATORY JOURNAL
w w w .thepcrj.org
Copyright GPIAG - reproduction prohibited
http://www.thepcrj.org
C
o
p
y
r
i
g
h
t
G
e
n
e
r
a
l
P
r
a
c
t
i
c
e
A
i
r
w
a
y
s
G
r
o
u
p
R
e
p
r
o
d
u
c
t
i
o
n
p
r
o
h
i
b
i
t
e
d
The ACT as a predictor of GINA-defined asthma control
43
sym ptom s, and need for rescue m edication w ere included
on both physician and patient form s w ith reference to the
previous four w eeks, as in the A C T (online A ppendices 1, 2,
and 3). Q uestioning on the fifth item in the G IN A
classification lung function w as included on the physician
form w ith reference to the previous 12 m onths.
Q uestioning about the sixth item in the G IN A control
definition asthm a exacerbations w as w ith reference to the
preceding 12 m onths on both physician and patient form s;
how ever, the tim ing of the exacerbation w as enquired about
on (and thus data derived from ) only the physician-com pleted
form s (see A ppendices 2 and 3 online). O f note, the G IN A
definitions for partly controlled and uncontrolled categories
are not m utually exclusive based on the exacerbation item
because the tim e period for an exacerbation is not specified
(see Table 1).
3
W e therefore elected to specify an exacerbation
in the preceding seven days as defining uncontrolled asthm a,
and an exacerbation w ithin the preceding year (but not in the
previous seven days) to define partly controlled asthm a in
term s of the G IN A classification.
Analysis
Physicians and their patients w ere linked by assigned study
num bers, and data included in the analyses w ere anonym ised
and restricted to those from m atched and fully com pleted
patient and physician form s. The prim ary analyses w ere
perform ed including all patients as w ell as by country.
The analyses evaluated the relationship betw een A C T
scores and G IN A -defined asthm a control, taking the G IN A
classification as the trueclassification and the A C T score as
the predictorclassification. O ur prim ary analyses evaluated
the relationship betw een A C T scores and G IN A -defined partly
controlled/uncontrolled versus controlled asthm a. In
PRIMARY CARE RESPIRATORY JOURNAL
w w w .thepcrj.org
Levels of A sthm a C ontrol
C haracteristic C ontrolled Partly controlled U ncontrolled
(A ll of the follow ing) (A ny m easure present
in any week)
D aytim e sym ptom s N one (tw ice or less/w eek) M ore than tw ice/w eek
Lim itations of activities N one A ny Three or m ore
N octurnal sym ptom s/aw akening N one A ny features of partly
N eed for reliever/rescue treatm ent N one (tw ice or less/w eek) M ore than tw ice/w eek controlled asthm a
Lung function (PEF or FEV
1
) N orm al <80% predicted or personal present in any week
best (if know n)
Exacerbations N one O ne or m ore/year* O ne in any week
FEV
1
= forced expiratory volum e in 1 second; PEF = peak expiratory flow .
*A ny exacerbation should prom pt review of m aintenance treatm ent to ensure that it is adequate.
Lung function is not a reliable test for children 5 years and younger.
Table 1A. Levels of asthma control according to the Global Initiative for Asthma (GINA).
3
Reprinted with permission.
Levels of A sthm a C ontrol
C haracteristic C ontrolled Partly controlled U ncontrolled
(A ll of the follow ing) (A ny m easure present
in previous 4 weeks )
D aytim e sym ptom s* N one (tw ice or less/w eek) M ore than tw ice/w eek
Lim itations of activities* N one A ny Three or m ore
N octurnal sym ptom s/aw akening* N one A ny features of partly
N eed for reliever/rescue treatm ent* N one (tw ice or less/w eek) M ore than tw ice/w eek controlled asthm a
present in any w eek
of the last 4 weeks
Exacerbations N one O ne or m ore in prior year O ne or m ore in the
previous 7 days
*D ata derived from patient-com pleted form s
Version 10
(StataC orp LP, C ollege Station, Texas, U S).
Results
Survey participants
Physicians com pleted asthm a patient record form s for 3503
patients in Europe and 1080 patients in the U SA . Eighty-five
percent of these form s (3877/4583) could be m atched w ith a
corresponding patient self-com pleted form . A total of 2949
of the 3877 (76% ) patient form s, corresponding to 64% of
physician form s, had com plete responses for the A C T- and
G IN A -related questions and w ere included in the analyses.
Patient num bers by country are sum m arised in Table 2.
Patients ranged in age from 12 to 93 years (m ean [SD ], 42
[17]), and 56% (1657/2941) w ere fem ale (a few dem ographic
data w ere m issing). M ost patients w ere w hite (2598/2930 or
89% ); patients of H ispanic (102, 3% ) and A frican-A m erican
descent (98, 3% ) w ere equally represented, w ith sm aller
num bers w ho described them selves as A fro-C aribbean (62,
2% ), A sian (55, 2% ), and other (15, <1% ). Tw elve percent
w ere current sm okers, 24% w ere ex-sm okers, and 64% had
never sm oked.
Respiratory DSP VI Survey results
Lung function data w ere available from only 539 of 2949
(18% ) m atched physician-com pleted form s and thus w ere
not included in the m ain analyses. H ow ever, analyses
incorporating lung function data gave results very sim ilar to
those obtained w hen the data w ere excluded, both for the
overall study population as w ell as for the subgroup w ith lung
function data (online Tables 13, see w w w .thepcrj.org).
GINA partly controlled/uncontrolled versus controlled
asthma as defined by ACT score
W ith a cut-off point of >20 for the A C T score defining w ell-
controlled asthm a, and a binary split for G IN A classification
(partly controlled/uncontrolled versus controlled asthm a), an
A C T score of <19 (not w ell-controlled asthm a) correctly
predicted G IN A -defined partly controlled/uncontrolled
asthm a 93.9% of the tim e (Table 3). A n A C T score of >20
predicted G IN A -defined controlled asthm a 51.3% of the
tim e. Table 4 sum m arises the positive and negative predictive
values, as w ell as sensitivity and specificity, of the A C T score
cut point of >20 for all patients and by country.
The area under the RO C curve for the A C T score
predicting the G IN A control classification w as 0.84 (95% C I
PRIMARY CARE RESPIRATORY JOURNAL
w w w .thepcrj.org
PRF Matched PSC Matched PSC
for analysis*
(n=4583) (n=3877) (n=2949)
France 741 732 (99% ) 697 (94% )
G erm any 720 710 (99% ) 495 (69% )
Italy 720 702 (98% ) 495 (69% )
Spain 726 573 (79% ) 395 (54% )
U nited Kingdom 596 200 (34% ) 154 (26% )
U nited States 1080 960 (89% ) 713 (66% )
*N o. form s w ith corresponding PRF as w ell as com plete responses for inclusion
in the analyses.