RESEARCH ARTICLE
Sc eening o Ch onic Obs uc i e Pulmona y
Disease: Validi y and Reliabili y o a Po able
De ice in Non-Specialized Heal hca e
Se ings
C is ina Rep esas-Rep esas
1
*, Albe o Fe nández-Villa
1
, Albe o Ruano-Ra iña
2
,
Ana P iegue-Ca e a
1
, Ma ibel Bo ana-Rial
1
, s udy g oup o “Validi y o COPD-6 in non-
specialized heal hca e se ings”
¶
1Pulmonology Depa men , NeumoVigoI+i Resea ch G oup, Uni e si y Hospi al Complex o Vigo (CHUVI),
Es uc u a O ganiza i a de Xes ión In eg ada de Vigo (EOXI Vigo), BiomedicalResea chIns i u e o Vigo
(IBIV), Vigo, Spain, 2Depa men o P e en i e Medicine and PublicHeal h, Uni e si y o San iago de
Compos ela, San iago de Compos ela, Spain, CIBER de Epidemiología y Salud Pública, CIBERESP,
Mad id, Spain
¶ Membe ship o he s udy g oup o “Validi y o COPD-6 in non-specialized heal hca e se ings”is p o ided in
he Acknowledgmen s.
*c is ina. ep esas. ep esas@se gas.es
Abs ac
In oduc ion and Objec i es
The unde diagnosis o ch onic obs uc i e pulmona y disease (COPD) could be imp o ed
h ough sc eening using po able de ices simple han con en ional spi ome e s in speci ic
heal hca e se ings o each a highe pe cen age o he a - isk popula ion. This s udy was
designed o assess he alidi y and eliabili y o he COPD-6 po able de ice o sc een o
COPD in non-specialized heal hca e se ings.
Me hods
P ospec i e coho s udy o alida e a diagnos ic es . Th ee coho s we e ec ui ed: p ima y
ca e (PC), eme gency se ices (ES) and communi y pha macies (CPh). S udy popula ion:
indi iduals wi h isk ac o s o COPD (>40 yea s, smoking >10 pack-yea s, wi h espi a o y
symp oms). The alues measu ed using he COPD-6 we e FEV1, FEV6 and he FEV1/
FEV6 a io. Subsequen ly, pa icipan s unde wen con en ional spi ome y a hospi al,
using a pos -b onchodila o FEV1/FVC alue <0.7 as he gold s anda d c i e ion o he
COPD diagnosis.
Resul s
437 pa icipan s we e included, 362 we e alid o he analysis. COPD was diagnosed in
114 pa ien s (31.5%). The a ea unde he ROC cu e o he COPD-6 o COPD sc eening
was 0.8.The bes cu -o poin o he FEV1/FEV6 a io was 0.8 (sensi i i y, 92.1%) using
spi ome y wi h he b onchodila o es as he gold s anda d. The e we e p ac ically no
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 1/11
OPEN ACCESS
Ci a ion: Rep esas-Rep esas C, Fe nández-Villa A,
Ruano-Ra iña A, P iegue-Ca e a A, Bo ana-Rial M,
s udy g oup o “Validi y o COPD-6 in non-specialized
heal hca e se ings”(2016) Sc eening o Ch onic
Obs uc i e Pulmona y Disease: Validi y and
Reliabili y o a Po able De ice in Non-Specialized
Heal hca e Se ings. PLoS ONE 11(1): e0145571.
doi:10.1371/jou nal.pone.0145571
Edi o : Sanjay Ha esh Cho i mall, Lee Kong Chian
School o Medicine, SINGAPORE
Recei ed: Sep embe 12, 2015
Accep ed: Decembe 4, 2015
Published: Janua y 4, 2016
Copy igh : © 2016 Rep esas-Rep esas e al. This is
an open access a icle dis ibu ed unde he e ms o
he C ea i e Commons A ibu ion License, which
pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided he o iginal
au ho and sou ce a e c edi ed.
Da a A ailabili y S a emen : All ele an da a a e
wi hin he pape .
Funding: This esea ch was pa ially suppo ed by
Fundación de la Sociedad Gallega de Pa ología
Respi a o ia (SOGAPAR), and unding om he
Eu opean Union Se en h F amewo k P og amme
[FP7/REGPOT-2012-2013.1] unde g an ag eemen
n° 316265, BIOCAPS. The unde s had no ole in
s udy design, da a collec ion and analysis, decision o
publish, o p epa a ion o he manusc ip .
di e ences in he COPD-6 pe o mancein he di e en se ings and also ega ding age, gen-
de and smoking s a us.
Conclusions
The COPD-6 de ice is a alid ool o COPD sc eening in non-specialized heal hca e se -
ings. In his con ex , he bes cu -o poin o he FEV1/FEV6 a io is 0.8.
In oduc ion
Ch onic obs uc i e pulmona y disease (COPD) is a e y common disease, wi h a p e alence
o 10.2% in Spain, highe in men han in women, and i s equency inc eases wi h age. How-
e e , unde diagnosis can be as high as 73%[1,2]. The abili y o pe o m spi ome y—essen ial
es o COPD diagnosis—in p ima y ca e (PC) cen es would imp o e his si ua ion, and such
a ecommenda ion has been made by he Spanish Socie y o Pneumology and Tho acic Su ge y
(SEPAR)[1] and by he Na ional COPD S a egy o he Spanish Na ional Heal h Sys em[2].
Howe e , in PC we ind no only a limi ed a ailabili y o spi ome y bu also ha he es is
inadequa ely pe o med in a la ge numbe o cen es[3–5], hough his a ies conside ably
be ween he di e en Spanish egions[6,7].
One o he causes o an insu icien quali y o spi ome y is di icul y in ob aining he o ced
i al capaci y (FVC) co ec ly. An al e na i e pa ame e , he o ced expi a o y olume in six
seconds (FEV6), has he e o e been accep ed as a subs i u e[8,9]. This simpli ies spi ome y
echnique and hus imp o es accu acy in he diagnosis o ai way obs uc ion in non-specialis
se ings[6]. In addi ion, i ou aim is o imp o e ou -o -hospi al sc eening o COPD o educe
unde diagnosis, a key elemen would be o employ po able de ices ha a e easie o ope a e
han con en ional spi ome e s, such as he Vi alog aph COPD-6 de ice[10].
In iew o he ecognised bene i s o he ea ly diagnosis o COPD, bo h he Spanish COPD
S a egy[2] and GesEPOC ( he Spanish COPD Guidelines)[1] ecommend p oac i e sc eening
in indi iduals a - isk. PC would appea o be ideal o he implemen a ion o hese p og ams
[1,2,11], bu es ing could also be ex ended o hose se ings in which he e a e heal h p o es-
sionals wi h basic aining in pe o ming espi a o y unc ion es s, and easily accessible o he
gene al public, such as communi y pha macies (CPh), whe e many use s epo espi a o y
symp oms (wi hou ha ing consul ed hei doc o ), and eme gency se ices (ES), whe e
pa ien s may a end o an exace ba ion. Spi ome y wi h a b onchodila o es , he gold s an-
da d, should hen be used o con i m a suspec ed diagnosis.
The COPD-6 de ice measu es FEV1 and FEV6 and calcula es i s a io. Ou g oup has
al eady alida ed his de ice o he de ec ion o obs uc i e espi a o y disease in a hospi al
se ing[12]. The nex s ep would be o demons a e whe he he COPD-6 de ice could also be
use ul as a sc eening ool o COPD in he ou -o -hospi al se ing, adminis e ed by s a who
does no ha e ex ensi e expe ience in pe o ming spi ome ies.
The objec i e o his s udy was o es ablish he alidi y and eliabili y o he COPD-6 de ice
o sc een o COPD among indi iduals a high COPD isk in non-specialized heal hca e se -
ings (PC, ES and CPh). Con en ional spi ome y wi h a b onchodila o es was used as he
gold s anda d o compa ison. This objec i e alls wi hin he scope o COPD esea ch p oposed
in he o icial decla a ion o he Ame ican Tho acic Socie y (ATS) and he Eu opean Respi a-
o y Socie y (ERS)[13].
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 2/11
Compe ing In e es s: The au ho s ha e decla ed
ha no compe ing in e es s exis .
Ma e ial and Me hods
P ospec i e, mul i-coho s udy o alida e a diagnos ic es in a ious se ings. The s udy was
ca ied ou be ween 2011 and 2014 in he Vigo Uni e si y Hospi al Complex (CHUVI) in Gali-
cia, Spain, using consecu i e coho s: 1) a coho d awn om PC pa ien s (PC coho ), wi h he
pa icipa ion o 8 Heal h Cen es in he Vigo Heal h Dis ic in Galicia, Spain; 2) a coho
ec ui ed om use s o Communi y Pha macies(CPh coho ), wi h he pa icipa ion o 15
pha macies in he Vigo a ea, wi h he collabo a ion o he Pon e ed a O icial College o Pha -
macis s; and 3) a coho ec ui ed om pa ien s a ending he eme gency se ices (ES coho ),
unde aken in o he CHUVI (Xe al Hospi al and Meixoei o Hospi al) and in he Redondela
and Po iño eme gency se ices o PC.
Ten PC doc o s, 13 hospi al and ou -o -hospi al ES doc o s and 15 pha macis s pa icipa ed
in he s udy. All we e p e iously ained in he use o he COPD-6 de ice in a 2-hou heo e i-
cal and p ac ical session gi en by ches physicians and nu ses who we e pa icipa ing in he
s udy.T aining ime equi ed o pe o m con en ional spi ome ies adequa ely is longe .
S udy popula ion
Indi iduals who a ended a PC consul a ion ( o a espi a o y o o he p oblem), who used a
pa icipa ing CPh o who p esen ed o he hospi al o ou -o -hospi al ES wi h symp oms o a
espi a o y exace ba ion, we e asked o pa icipa e in he s udy i hey sa is ied he inclusion
c i e ia.
Inclusion c i e ia. The same inclusion c i e ia we e used o all 3 coho s: pa ien s aged
o e 40 yea s, smoke s o ex-smoke s o mo e han 10 pack-yea s, and symp oms sugges i e o
COPD (dyspnea, cough, spu um p oduc ion). These c i e ia we e consis en wi h he ecom-
menda ions o he Spanish COPD S a egy[2] and GesEPOC[1], hough i should be ecog-
nised ha he GesEPOC ecommenda ions we e published a e ini ia ion o he i s phase o
ou s udy and lowe ed he age c i e ion o he high- isk g oup o 35 yea s.
Exclusion c i e ia. The ollowing exclusion c i e ia we e applied: he indi idual declined
o pa icipa e in he s udy, was unable o a end he hospi al o pe o m he es o diagnos ic
con i ma ion, o p esen ed any o he absolu e con aindica ions o pe o ming spi ome y
[14]. Indi iduals who had al eady been diagnosed wi h a espi a o y disease we e also
excluded.
S udy pa icipan s we e ec ui ed consecu i ely in a non-sys ema ic manne a e he pa ic-
ipa ing doc o o pha macis (in PC, ES o CPh) had checked ha hey sa is ied he inclusion
c i e ia and had ob ained w i en in o med consen a e explaining he na u e o he s udy.
The s udies o each coho we e app o ed by he Galician E hics Commi ee (dossie num-
be : 2010–73, 2011–284, 2012–389).
P ocedu es
Measu emen o he unc ional pa ame e s, FEV1, FEV6 and FEV1/FEV6 a io, we e pe -
o med using he po able COPD-6 de ice (model 4000, Vi alog aph L d., Ennis, Co. Cla e, I e-
land). The pa ien was equi ed o pe o m manoeu es simila o hose used in o ced
spi ome y wi h he di e ence ha , in his case, i was only necessa y o main ain exhala ion
o 6 seconds. Measu emen s we e epea ed un il a leas 2 eliable measu emen s we e
achie ed, and he bes alues ob ained o each pa ien we e eco ded. The pa icipa ing PC
physicians and pha macis s ( o he PC and CPh coho s) hen con ac ed he ches physicians
o he esea ch eam by mobile phone o schedule he second phase o he s udy, which was ca -
ied ou wi hin he ollowing 24 hou s. In he case o he ES coho , he pa icipa ing doc o s in
he ES emi ed he pa ien s’da a o he esea ch eam by in e nal pos ; as he pa ien s in his
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 3/11
coho had a ended because o an exace ba ion, hey we e scheduled o he hospi al phase o
he s udy one mon h a e he eme gency consul a ion, du ing a clinically s able phase.
In he hospi al, blinded o he esul s o he es s pe o med wi h he COPD-6 de ice, he
esea ch s a pe o med con en ional spi ome y using a Da ospi 120 spi ome e (Sibelmed,
Ba celona, Spain), eco ding he FEV1, FVC and FEV1/FVC a io. The es was pe o med in
acco dance wi h SEPAR guidelines[14]. When an FEV1/FVC a io less han 0.7 was de ec ed,
he s udy was comple ed wi h a b onchodila o es , wi h measu emen o he same pa ame e s
a e he adminis a ion o 400 μg o salbu amol ia a space de ice.
S a is ical analysis
A desc ip i e analysis o he a iables was pe o med o each coho and o he o e all sam-
ple. Sensi i i y (S), speci ici y (Sp), posi i e p edic i e alue (PPV), nega i e p edic i e alue
(NPV) and he posi i e and nega i e likelihood a ios (PLR/NLR) we e calcula ed o he di -
e en cu -o poin s o he FEV1/FEV6 a io.
The a ea unde he ecei e ope a ing cha ac e is ic (ROC) cu e was de e mined o he
FEV1/FEV6 a io calcula ed using he COPD-6 de ice employed o sc een o COPD. An
FEV1/FVC a io less han 0.7 ob ained on spi ome y a e he b onchodila o es was used as
he gold s anda d c i e ion o he diagnosis o COPD. In addi ion, he a eas unde he ROC
cu es adjus ed o a se ies o a iables we e compa ed using he Chi-squa ed es o homoge-
nei y o a eas, aking he median alue as he cu -o poin o age and o pack-yea s o he
cumula i e smoking his o y.
The analysis was pe o med using he S a is ical Package o Social Sciences e sion 15.0
(SPSS, Chicago, Illinois, USA).
P esen a ion o he s udy was based on STARD me hodology o epo ing s udies o diag-
nos ic accu acy.
Resul s
A o al o 437 indi iduals we e included in he h ee coho s; 362 subjec s we e alid o analy-
sis (75 we e wi hd awn because hey p esen ed exclusion c i e ia o did no a end he hospi al
isi ).
The cha ac e is ics o he pa icipan s in each o he coho s and in he o e all sample a e
shown in Table 1. The mean age o he pa icipan s in he o e all s udy popula ion was 55
yea s, 61.9% men, and 72.4% ac i e smoke s.
COPD was diagnosed in 114 pa ien s, a p e alence o 31.5%; he majo i y (96 pa ien s,
84.2%) we e in Global Ini ia i e o Ch onic Obs uc i e Lung Disease (GOLD) s age I o II.
The ROC cu es o he FEV1/FEV6 a io calcula ed by he COPD-6 a e shown in Fig 1 o
each o he coho s and o he o e all pa ien sample, in which he a ea unde he cu e
(AUC) was 0.8 (95% con idence in e al, 0.752–0.847).
A compa ison was also pe o med o he a ea unde he COPD-6 ROC cu es as a unc ion
o sex, age, smoke s a us and cumula i e smoking his o y; no signi ican di e ences we e
ound be ween hese g oups. These esul s a e shown in Table 2. We also obse ed ha COPD
6 is alid o sc eening e en in mo e se e e ca ego ies o COPD. Wi h he cu o poin o 0.8,
no pa ien wi h se e e COPD was alse nega i e in he sc eening.
Tables 3–6show he sensi i i y (S), speci ici y (Sp), posi i e p edic i e alue (PPV), nega i e
p edic i e alue (NPV),posi i e likelihood a io (PLR) and nega i e likelihood a io (NLR) o
he di e en cu -o poin s o he FEV1/FEV6 a io in he PC, CPh and ES coho s and in he
o e all s udy popula ion. Based on he o e all popula ion, a cu -o poin o 0.8 o he FEV1/
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 4/11
FEV6 a io had a S o 92.1%, Sp o 52.8%, NPV o 93.6% and PPV o 47.3% o he de ec ion o
COPD.
Table 1. Cha ac e is ics o he pa ien s in he h ee coho s and in he o e all sample.
PC Coho CPh Coho ES Coho O e all Sample
Pa icipan s 167 143 127 437
Excluded/ Los o ollow-up 17 33 25 75
Valid o analysis 150 110 102 362
Age, yea s, mean (SD) 56.8 (9.7) 53.1 (8.6) 56 (11.1) 55.4 (9.9)
Men, n (%) 106 (70.6%) 44 (40%) 73 (71.6%) 224 (61.9%)
Ac i e smoke s 112 (74.6%) 85 (77.3%) 65 (63.7%) 262 (72.4%)
Cumula i e obacco consump ion in pack-yea s, mean (SD) 39 (21.6) 32.5 (17) 31.8 (19.1) 35 (19.8)
COPD 60 (40%) 22 (20%) 32 (31%) 114 (31,5%)
GOLD s age I 12 (20%) 8 (36.4%) 13(40.6%) 33 (29%)
GOLD s age II 37 (61.7%) 11 (50%) 15 (46.9%) 63 (55.2%)
GOLD s age III 10 (16.7%) 3 (13.6%) 4 (12.5%) 17 (15%)
GOLD s age IV 1 (1.7%) 0 0 1 (0.8%)
Abb e ia ions: COPD, ch onic obs uc i e pulmona y disease; CPh, communi y pha macies; ES, eme gency se ices; GOLD, Global Ini ia i e o Ch onic
Obs uc i e Lung Disease; PC, p ima y ca e.
doi:10.1371/jou nal.pone.0145571. 001
Fig 1. ROC cu es o he FEV1/FEV6 a io measu ed using he COPD-6 de ice o sc een o ch onic
obs uc i e pulmona y disease in p ima y ca e ( op le ), communi y pha macies ( op igh ),
eme gency se ices (bo om le ) and in he o e all sample (bo om igh ). Abb e ia ions: AUC, a ea
unde cu e; CI, con idence in e al; CPh, communi y pha macies; ES, eme gency se ices; GOLD, Global
Ini ia i e o Ch onic Obs uc i e Lung Disease; PC, p ima y ca e; ROC, ecei e ope a ing cha ac e is ic.
doi:10.1371/jou nal.pone.0145571.g001
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 5/11
Discussion
This s udy is he i s o show ha he COPD-6 is a alid and eliable ool o he de ec ion o
COPD in di e en non-specialized heal hca e se ings. The alidi y indexes ob ained a e e y
good and show no signi ican di e ences acco ding o sex, smoking his o y o age. Fu he -
mo e, he esul s ha dly di e when he COPD-6 is applied in di e en se ings and o di e -
en condi ions such as gende , age g oup o smoking s a us. These esul s ein o ce he alidi y
and eliabili y o he COPD-6 es .
COPD sc eening is impo an because he ea ly de ec ion o COPD and in e en ion o
smoking cessa ion can delay lung unc ion decline, educe he bu den o COPD symp oms,
educe heal hca e cos s, and imp o e pa ien ´s quali y o li e[2]. In a p e ious s udy in he hos-
pi al se ing, ou g oup has alida ed he COPD-6 o he de ec ion o obs uc i e pulmona y
disease[12]. The esul s o he p esen s udy a e highly ele an because hey p o ide e idence
o he eliabili y o simple spi ome ic manoeu es wi h an easy- o-use de ice o COPD
Table 2. Compa ison o he a ea unde he ROC cu es o he FEV1/FEV6 a io measu ed by he
COPD-6 de ice, acco ding o sex, age and smoking his o y.
A ea Unde he ROC Cu e P
Sex Male 0.803 (95% CI, 0.744–0.862) 0.99
Female 0.803 (95% CI, 0.721–0.885)
Age <55 y 0.816 (95% CI, 0.746–0.886) 0.48
55 y 0.782 (95% CI, 0.714–0.849)
Smoking Ac i e smoke 0.837 (95% CI, 0.776–0.897) 0.27
Ex-smoke 0.785 (95% CI, 0.713–0.856)
Cumula i e obacco consump ion <30 pack-yea s 0.822 (95% CI, 0.754–0.889) 0.34
30 pack-yea s 0.777 (95% CI, 0.711–0.842)
Abb e ia ion: CI, confidence in e al.
doi:10.1371/jou nal.pone.0145571. 002
Table 3. P ima y Ca e coho . Values o sensi i i y, speci ici y, posi i e p edic i e alue, nega i e p edic i e alue, posi i e likelihood a io and nega i e
likelihood a io o di e en cu -o poin s o he FEV1/FEV6 a io de e mined du ing sc eening o COPD using he COPD-6 de ice in PC (gold s anda d c i e-
ion o compa ison: pos -b onchodila o spi ome y FEV1/FVC <0.7).
FEV1/FEV6 (COPD-6) S Sp PPV NPV PLR NLR
0.70 38.3% 93.3% 79.3% 69.4% 5.75 0.66
0.71 40% 92.2% 77.4% 69.7% 5.14 0.65
0.72 45% 91.1% 77.1% 71.3% 5.06 0.6
0.73 48.3% 87.8% 72.5% 71.8% 3.95 0.59
0.74 55% 85.6% 71.7% 74% 3.81 0.53
0.75 60% 82.2% 69.2% 75.5% 3.38 0.49
0.76 68.3% 78.9% 68.3% 78.9% 3.24 0.4
0.77 75% 75.6% 67.2% 81.9% 3.07 0.33
0.78 83.3% 71.1% 65.8% 86.5% 2.88 0.23
0.79 85% 64.4% 61.4% 86.6% 2.39 0.23
0.80 91.7% 60% 60.4% 91.5% 2.29 0.14
0.81 95% 55.6% 58.8% 94.3% 2.14 0.09
Abb e ia ions: NLR, nega i e likelihood a io; NPV, nega i e p edic i e alue; PLR, posi i e likelihood a io; PPV, posi i e p edic i e alue; S, sensi i i y;
Sp, specifici y.
doi:10.1371/jou nal.pone.0145571. 003
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 6/11
sc eening, pe o med by s a who ha e no been ained in pulmona y unc ion es ing and
who ha e only ecei ed a b ie heo e ical-p ac ical cou se on he use o he de ice.
The use o he FEV1/FEV6 a io as a subs i u e o he FEV1/FVC a io has al eady been
accep ed[8,9].Acco ding o he PLATINO longi udinal s udy da a, FEV1/FEV6 is a be e indi-
ca o o ai low obs uc ion, likely due o compa ing olumes a ixed imes o he expi a o y
manoeu e and a oiding inconsis encies due o changes in he quali y o he spi ome ies,
especially in o ced expi a o y ime ac oss di e en echnicians, cen es, o along ime[15].
Pos b onchodila o FEV6 and FEV1/FEV6 as pe cen age o p edic ed a e also independen
p ognos ic ac o s in s able ou pa ien s wi h COPD[16]. In any case, he e is no consensus on
he FEV1/FEV6 cu -o poin conside ed o de ine obs uc ion[8,9,12]. The elec ion o he cu -
Table 4. Communi y pha macies coho . Values o sensi i i y, speci ici y, posi i e p edic i e alue, nega i e p edic i e alue, posi i e likelihood a io and
nega i e likelihood a io o di e en cu -o poin s o he FEV1/FEV6 a io de e mined du ing sc eening o COPD using he COPD-6 de ice in CPh (gold
s anda d c i e ion o compa ison: pos -b onchodila o spi ome y FEV1/FVC <0.7).
FEV1/FEV6 (COPD-6) S Sp PPV NPV PLR NLR
0.70 68.2% 84.1% 51.7% 91.4% 4.29 0.38
0.71 68.2% 80.7% 46.9% 91.0% 3.53 0.39
0.72 68.2% 80.7% 46.9% 91.0% 3.53 0.39
0.73 72.7% 78.4% 45.7% 92.0% 3.37 0.35
0.74 77.3% 73.9% 42.5% 92.9% 2.96 0.31
0.75 77.3% 72.7% 41.5% 92.8% 2.83 0.31
0.76 86.4% 69.3% 41.3% 95.3% 2.81 0.20
0.77 86.4% 69.3% 41.3% 95.3% 2.81 0.20
0.78 90.9% 64.8% 39.2% 96.6% 2.58 0.14
0.79 95.5% 58% 36.2% 98.1% 2.27 0.08
0.80 95.5% 51.1% 32.8% 97.8% 1.95 0.09
0.81 95.5% 44.3% 30% 97.5% 1.71 0.10
Abb e ia ions: NLR, nega i e likelihood a io; NPV, nega i e p edic i e alue; PLR, posi i e likelihood a io; PPV, posi i e p edic i e alue; S, sensi i i y;
Sp, specifici y.
doi:10.1371/jou nal.pone.0145571. 004
Table 5. Eme gency se ices coho . Values o sensi i i y, speci ici y, posi i e p edic i e alue, nega i e p edic i e alue, posi i e likelihood a io and
nega i e likelihood a io o he di e en cu -o poin s o he FEV1/FEV6 a io de e mined du ing sc eening o COPD using he COPD-6 de ice in ES (gold
s anda d c i e ion o compa ison: pos -b onchodila o spi ome y FEV1/FVC <0.7).
FEV1/FEV6 (COPD-6) S Sp PPV NPV PLR NLR
0.70 62.5% 77.1% 55.6% 81.8% 2.73 0.49
0.71 68.8% 77.1% 57.9% 84.4% 3.0 0.40
0.72 75% 77.1% 60% 87.1% 3.28 0.32
0.73 75% 75.7% 58.5% 86.9% 3.09 0.33
0.74 78.1% 75.7% 59.5% 88.3% 3.21 0.29
0.75 81.3% 72.9% 57.8% 89.5% 3.0 0.26
0.76 84.4% 68.6% 55.1% 90.6% 2.69 0.23
0.77 84.4% 62.9% 50.9% 89.8% 2.27 0.25
0.78 84.4% 61.4% 50% 89.6% 2.19 0.25
0.79 87.5% 51.4% 45.2% 90% 1.80 0.24
0.80 90.6% 44.3% 42.6% 91.2% 1.63 0.21
0.81 93.8% 42.9% 42.9% 93.8% 1.64 0.14
Abb e ia ions: NLR, nega i e likelihood a io; NPV, nega i e p edic i e alue; PLR, posi i e likelihood a io; PPV, posi i e p edic i e alue; S, sensi i i y;
Sp, specifici y.
doi:10.1371/jou nal.pone.0145571. 005
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 7/11
o poin o he FEV1/FEV6 a io mus ake in o accoun ha his es is being used o sc een-
ing, and a high sensi i i y is hus desi able. We he e o e p opose aising he cu -o poin o
0.8 o achie e his high sensi i i y and a high nega i e p edic i e alue, e en wi h a lowe speci-
ici y. Al hough he e we e small di e ences in he 3 heal hca e se ings in which he sc eening
s udies we e pe o med, i would appea ha cu -o poin s close o 0.8 o he FEV1/FEV6
a io a e hose wi h g ea es alidi y o sc eening. On his basis, indi iduals a isk o COPD
who ha e an FEV1/FEV6 a io g ea e han o equal o 0.8, as de e mined by he COPD-6,
would ha e a much lowe p obabili y o being diagnosed wi h COPD o , i his we e no
de ec ed, hey would ha e mild disease. In hese cases, spi ome y could heo e ically be
de e ed, and ollow-up scheduled. When he esul is below 0.8, spi ome y should be pe -
o med wi h a b onchodila o es o con i m o exclude he diagnosis o COPD. An AUC
alue o 0.8 seems also accep able o sc eening es s pe o med in ambula o y se ings o by
s a no ha ing ex ensi e aining in lung unc ion es s.
In he p esen s udy we diagnosed COPD in a signi ican numbe o a - isk pa ien s: 31.5%
in he o e all g oup, and 40% in he PC coho . The majo i y o hese pa ien s p esen ed a mild
o mode a e deg ee o obs uc ion, hough mo e han 16% o pa ien s had se e e obs uc ion.
This would suppo a ecommenda ion o p oac i e sc eening o suscep ible indi iduals in
o de o ini ia e app op ia e he apeu ic s a egies a he ea lies oppo uni y[1].
A cu ious inding in ou esul s was he di e ence be ween he pe cen age o male pa ien s
in he CPh coho (40%) compa ed wi h he o he 2 coho s (70.6% and 71.6%). We belie e
ha his simply e lec s ha women a e mo e likely han men o a end communi y pha ma-
cies.Se e al epo s ha e been published on he use o po able de ices o sc een o COPD in
non-specialized se ings. Mi a i lles e al[17] analysed he diagnos ic yield o he FEV1/FEV6
a io measu ed using he COPD-6 de ice o sc een o ch onic ai ways obs uc ion, concluding
ha i was use ul o his pu pose and he bes cu -o poin was 0.75. Bu he au ho s did no
speci y who pe o med he COPD-6 measu emen s, a ac o ha could a ec he esul s. In ou
s udy, he measu emen s in he PC coho we e made by he pa icipa ing doc o s who
ec ui ed he pa ien s. This design would p o ide esul s ha we e a mo e eliable e lec ion o
daily p ac ice, in which all doc o s a e equi ed o be able o use po able de ices as sc eening
Table 6. O e all sample. Values o sensi i i y, speci ici y, posi i e p edic i e alue, nega i e p edic i e alue, posi i e likelihood a io and nega i e likeli-
hood a io o he di e en cu -o poin s o he FEV-1/FEV6 a io de e mined du ing sc eening o COPD using he COPD-6 de ice in non-specialis heal h-
ca e se ings (gold s anda d c i e ion o compa ison: pos -b onchodila o spi ome y FEV1/FVC <0.7).
FEV1/FEV6(COPD-6) S Sp PPV NPV PLR NLR
0.70 50.9% 85.9% 62.4% 79.2% 3.61 0.57
0.71 53.5% 84.3% 61% 79.8% 3.41 0.55
0.72 57.9% 83.9% 62.3% 81.3% 3.6 0.5
0.73 60.5% 81.5% 60% 81.8% 3.27 0.48
0.74 65.8% 79% 59.1% 83.4% 3.13 0.43
0.75 69.3% 76.6% 57.7% 84.4% 2.96 0.4
0.76 76.3% 73% 56.5% 87% 2.83 0.32
0.77 79.8% 70.2% 55.2% 88.3% 2.68 0.29
0.78 85.1% 66.5% 53.9% 90.7% 2.54 0.22
0.79 87.7% 58.9% 49.5% 91.3% 2.13 0.21
0.80 92.1% 52.8% 47.3% 93.6% 1.95 0.15
0.81 94.7% 48.4% 45.8% 95.2% 1.84 0.11
Abb e ia ions: NLR, nega i e likelihood a io; NPV, nega i e p edic i e alue; PLR, posi i e likelihood a io; PPV, posi i e p edic i e alue; S, sensi i i y;
Sp, specifici y.
doi:10.1371/jou nal.pone.0145571. 006
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 8/11
ools, e en i hey ha e no ecei ed ex ensi e speci ic aining on pulmona y unc ion. F i h
e al[18] alida ed ano he po able de ice, he PiKo-6, using a spi ome y-measu ed pos -
b onchodila o FEV1/FVC a io <0.7 as he gold s anda d diagnos ic c i e ion. They demon-
s a ed he use ulness o he FEV1/FEV6 a io o COPD sc eening in PC (a ea unde he cu e
0.85, simila o ou esul s), and p oposed a cu -o poin o 0.75 o he FEV1/FEV6 a io (S
81%; Sp71%). In ou opinion, i we wish o use hese de ices o sc eening, we mus look o a
high sensi i i y, e en wi h a lowe speci ici y. Ou cu -o poin o 0.8 has also been p oposed
by o he au ho s[9,19].The e a e o he s udies ha ha e obse ed di e en AUCs such as he
ones pe o med by Van den Bem e al[20] o Tho n e al[21]. The eason o he obse ed di -
e ences migh be due ha hey s udied popula ions wi h di e en cha ac e is ics o o he ac
ha mic ospi ome y was pe o med by s a wi h highe expe ience in lung unc ion assess-
men compa ed wi h ou s a , who ecei ed a 2-hou session aining.
Wi h ega d o o he heal hca e se ings, ew s udies ha e been published. In he Fa maEPOC
s udy[22], pa ien s a isk o COPD (selec ed by a clinical ques ionnai e) we e sc eened in he
pha macy by con en ional spi ome y pe o med by s a ained on a 4-day cou se. An expe in
spi ome y elema ically e iewed he quali y o he s udies. Only hose pa ien s wi h an FEV1/
FVC <0.7 (wi hou b onchodila a ion) we e e e ed o PC o comple e he s udy. The esul s in
PC we e only epo ed o he pha macies in 15% o cases, which is a signi ican limi a ion. Fu -
he mo e, i was no explained how he diagnosis o COPD was eached in PC. This sc eening
s a egy would equi e pha macies o be equipped wi h spi ome e s, leng hie aining o he
s a pe o ming he s udies, and ime dedica ion by an expe heal hca e p o essional o e iew
he quali y, and i would he e o e be much mo e cos ly and di icul o apply in he majo i y o
CPh. Fo ha eason, in ou CPh coho , we op ed o use a po able de ice, he COPD-6, much
simple and cheape han con en ional spi ome y, an essen ial aspec i we wish o ex end i s
use widely o CPh. Solido o e al[23] in I alian pha macies also used a po able de ice, he PiKo-
6, o de ec obs uc i e and es ic i e lung disease, no speci ically COPD; u he mo e, in con-
as o ou s udy, hose au ho s did no ake in o accoun whe he pa ien s had a p e ious diag-
nosis no did hey de ine he s anda d compa a o es used o con i m he diagnosis o COPD.
The e a e e y ew da a on he esul s o COPD sc eening p og ammes in ES. Recen ly, a
F ench g oup[24] published a s udy in which hey pe o med COPD sc eening in an ES, using
he Nèo6 de ice. Spi ome y was pe o med in hose who had a ques ionnai e sugges i e o
COPD and an FEV1/FEV6 a io <0.8, he same as he cu -o poin ha we p opose. In he
g oup o pa ien s s udied by spi ome y, only 8,91% p esen ed COPD, a su p isingly low p e a-
lence o COPD compa ed wi h he esul s om ou ES coho (31.4%).
One o he s eng hs o ou s udy is ha we ook he measu emen s using a baseline COPD-
6 es , wi hou b onchodila o , which is how he s udy would be pe o med in daily p ac ice in
any o he non-specialized se ings, and we compa ed his wi h spi ome y a e b onchodila a-
ion, which is he gold s anda d es o he diagnosis o COPD. This could be one o he ea-
sons o he di e ence in esul s compa ed wi h ou s udy pe o med in he hospi al se ing
[12]. A u he s eng h is he sample size, which was o e 350 indi iduals om di e en se -
ings; his inc eases he alidi y o he esul s ob ained. Fu he mo e, he s udy was pe o med
on an unselec ed popula ion ha we belie e e lec s easonably and accu a ely he cha ac e is-
ics o he gene al popula ion. Finally, he ac ha he spi ome y was pe o med blinded o
he esul s o he COPD-6 elimina es obse e bias and inc eases he alidi y o ou esul s.
Mo eo e , ou esea ch alls wi hin he scope o COPD esea ch p oposed in he o icial decla-
a ion o he Ame ican Tho acic Socie y (ATS) and he Eu opean Respi a o y Socie y (ERS)
[13], ecommending s udies ha e alua e case- inding s a egies using mini-spi ome e s.
Howe e , he e a e ce ain limi a ions o ou s udy, as he use o a pos -b onchodila o
FEV1/FVC a io <0.7 as he gold s anda d, c i icised and ques ioned o possible inaccu acy
COPD Sc eening in Non-Specialized Heal hca e Se ings
PLOS ONE | DOI:10.1371/jou nal.pone.0145571 Janua y 4, 2016 9/11
