community health workers and environmental interventions for children with asthma: a systematic...
TRANSCRIPT
Journal of Asthma, 46:564–576, 2009Copyright C© 2009 Informa Healthcare USA, Inc.ISSN: 0277-0903 print / 1532-4303 onlineDOI: 10.1080/02770900902912638
ORIGINAL ARTICLE
Community Health Workers and Environmental Interventionsfor Children with Asthma: A Systematic Review
Julie Postma, Ph.D.,1,∗ Catherine Karr,2 and Gail Kieckhefer1
1University of Washington School of Nursing, Family and Child Nursing, Seattle2Pediatrics, Environmental & Occupational Health Sciences, University of Washington, Seattle
Community health worker (CHW)–delivered, home-based environmental interventions for pediatric asthma were systematically reviewed. SevenPubMed/MEDLINE listed randomized controlled trials that encompassed the following intervention criteria were identified: (1) home-based; (2)delivered by a CHW; (3) delivered to families with children with asthma; and (4) addressed multiple environmental triggers for asthma. Details ofresearch design, intervention type, and setting, interventionist, population served, and the evaluated outcomes were abstracted.Outcome assessment was broad and non-uniform. Categories included direct mediators of improved health outcomes, such as trigger-related knowl-edge, trigger reduction behaviors and allergen or exposure levels, and asthma-related health outcomes: change in lung function, medication use,asthma symptoms, activity limitations, and health care utilization. Indirect mediators of health outcomes, or psychosocial influences on health, weremeasured in few studies.Overall, the studies consistently identified positive outcomes associated with CHW-delivered interventions, including decreased asthma symptoms,daytime activity limitations, and emergency and urgent care use. However, improvements in trigger reduction behaviors and allergen levels, hypoth-esized mediators of these outcomes, were inconsistent. Trigger reduction behaviors appeared to be tied to study-based resource provision.To better understand the mechanism through which CHW-led environmental interventions cause a change in asthma-related health outcomes, infor-mation on the theoretical concepts that mediate behavior change in trigger control (self-efficacy, social support) is needed. In addition, evaluating theinfluence of CHWs as clinic liaisons that enhance access to health professionals, complement clinic-based teaching, and improve appropriate use ofasthma medications should be considered, alongside their effect on environmental management. A conceptual model identifying pathways for futureinvestigation is presented.
Keywords asthma, allergens, patient education, community health aides, review
IntroductionPediatric asthma is a significant public health problem.
In the United States, approximately 6.5 million children(younger than 18 years of age) have asthma, with rates dispro-portionately affecting low-income and minority populations(1). Based on evidence that controlling exposure to indoorenvironmental triggers reduces asthma symptoms and ex-acerbations, the National Heart, Lung and Blood Institute(NHLBI) guidelines highlight control of environmental fac-tors among the four core components of comprehensive pedi-atric asthma management (2). While the majority of asthmamanagement education occurs in the clinical setting, increas-ingly, multifaceted environmental interventions are deliveredin the home setting. Home visitation offers an opportunityto assess and address indoor environmental triggers whilereinforcing clinical education.
The NHLBI’s guidelines cite several randomized con-trolled trials (RCTs) of interventions using clinical special-ists (e.g., medical doctors, registered nurses) or communityhealth workers (CHWs) to deliver the interventions (3–8).CHWs are members of the communities they serve and arewell-positioned to establish rapport with families and provideculturally appropriate services (9). Previous studies supportthe effectiveness of CHWs in improving diabetes knowledge,management of hypertension, and appropriate health care uti-
∗Corresponding author: Julie Marie Postma, Ph.D., Washington StateUniversity College of Nursing, PO Box 1495, Spokane, WA 99210; E-mail:[email protected]
lization (10–12). Here we provide a systematic review of theeffectiveness of CHW-delivered environmental interventionsfor pediatric asthma and include a conceptual model to guidefuture research in this area.
MethodsWe identified RCTs that encompassed the following inter-
vention criteria: (1) home-based; (2) delivered by a CHW; (3)delivered to families with children with asthma; and (4) ad-dressed multiple environmental triggers for asthma. CHWswere defined as interventionists who were specifically trainedto deliver the intervention but had no formal professional orparaprofessional training in healthcare (13).
Studies were identified in a four-step PubMed/MEDLINEsearch using the terms: (1) asthma (major); (2) air pollutionOR environmental exposure OR allergens OR environmen-tal trigger OR environmental pollution; (3) patient educa-tion OR intervention OR skills OR remediation OR healtheducation OR education; and (4) community health aidesOR community health worker OR lay educator OR com-munity health services. Searches were then combined withthe following limits: Humans, Clinical Trial, Meta-Analysis,Randomized Controlled Trial, English. This yielded 40 pub-lications. No other publications were identified in a five-stepCINAHL search using the search terms: (1) asthma (majorconcept); (2) environmental pollution OR antigens; (3) com-munity health workers OR health educators OR allied healthpersonnel; (4) education OR health services; (5) clinical tri-als. All CINAHL search terms were exploded and used as
564
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
COMMUNITY HEALTH WORKERS & ENVIRONMENTAL TRIGGERS 565
exact subject headings or key words. Unpublished data werenot considered.
An abstraction form was used to document research de-sign, population served, intervention type and setting, andinterventionist. Eight publications met the criteria, one ofwhich reports allergen and exposure results from a subset ofhomes in the study by Krieger et al. Exclusions included (ina hierarchical order) interventions that were not RCTs (12),focused on adults (2), addressed a single trigger (9), deliv-ered in a non-home setting (3) (e.g., school or emergencyroom), and/or were delivered by non-CHWs (6). (One pub-lication included in the review focused on outcomes relatedto cockroaches, although the intervention addressed multi-ple triggers. Related outcomes will be presented in futurepublications.) In addition to the primary publications, fivecompanion papers and a project website were used to gatheradditional detail about the studies (14–19).
We systematically evaluated the state of the evidencefor the practice of CHWs in environmental management ofchildhood asthma. Given the heterogeneity of the outcomemeasures, statistical methods were not used to combine theresults. Thus, results are qualitatively synthesized and pre-sented in a table format (20).
ResultsStudy Characteristics
All of the studies were published within the last 5 years andconducted in urban settings in the U.S. (Table 1). One studywas conducted at numerous sites (6). Sample sizes rangedfrom 100 (7) to 937 participants (6) and included childrenwith current asthma 2 to 16 years of age (21). With theexception of one study whose participants were screened forasthma using a questionnaire (22), eligibility requirementsincluded either a physician diagnosis of asthma or a recentvisit to a clinic, emergency room, or hospital for asthma.Two studies also required at least one positive skin test to anallergen (6, 8).
Recruitment occurred through self- and physician refer-rals (21), school-based programs (7, 8, 22), community andpublic health clinics, local hospital and emergency rooms (5,23), and community agencies and residents (5). One studydid not report recruitment methods (6).
Selection of control groups and the type of “usual care”they received varied. Four studies used parallel control andintervention groups (5, 7, 8, 23) with provision of the in-tervention in total or in a condensed form at the end of thestudy. One study had three arms: (1) an observation-onlyor “attention control” group, (2) a treatment group, (3) anda case-matched control group in which asthma hospitaliza-tions and emergency department visits were recorded (21).A two-by-two factorial design to evaluate environmental andphysician-feedback interventions in the same study popula-tion was used in one study with a traditional non-treated con-trol group (6). A staggered design, such that the groups thatserved as controls in “Wave 1” received the full interventionin “Wave 2” was used in one study (19). Community BasedParticipatory Research (CBPR) principles were cited in fourstudies (5, 7, 22, 23) as providing the basis for using either alow-intensity control group versus a usual care group and/orproviding intervention materials to the control group at the
end of the study (15, 16, 22). Length of follow-up rangedfrom 4 months (8) to 2 years (6), with five studies followingparticipants for 1 year after the start of the intervention (5, 7,21–23). There was no blinding in five studies, only the datacollectors were blinded in the one study (7) and blindingwas not reported in another study (22). In three studies, theinterventionists were also the data collectors (8, 21, 23).
In general, study subjects were between the ages of 5and 9 years, low-income, and ethnic minorities (Table 2). Inparticular, subjects were all or primarily African Americanin three studies (7, 21, 23) and almost all Hispanic in an-other study (8). African Americans and Hispanics made upthe majority of participants in two studies (6, 22). Kriegeret al. enrolled the most ethnically diverse sample with sub-jects reporting being African American, Hispanic, Viet-namese, and white.
Five studies reported the atopic status of their participantsat baseline. Between 28.2% of participants in one study (18)and 65% and 75% of the control and treatment group par-ticipants in another study, respectively, had more than onepositive skin test for an allergen at baseline (7).
Characteristics, Training, and Supervision of CommunityHealth Workers
The CHWs came from, lived, or worked in the same com-munity as study subjects, with the exception of one study(23) (Table 3). One study reported that the CHWs either hadasthma or had close family members with asthma (16). Onlyone author reported the minimum educational backgroundrequired to be a CHW (22).
Three studies quantified training requirements for theCHWs (16, 18, 22) and three others discussed training pro-grams in which the CHWs had participated (8, 15, 23). Interms of quality control, the use of standardized protocolswas reported in three studies and materials posted on anotherstudy’s website suggest the same. Supervision was only re-ported in one study, whereby CHWs met with the primaryinvestigator every other week and with the project’s steeringcommittee quarterly (16).
InterventionsSocial cognitive theory (SCT) and its precursor social
learning theory was cited most frequently as guiding the in-tervention approach, with five of seven studies incorporatingcomponents of SCT into their intervention delivery (Table 4).For example, to improve caregivers’ self efficacy (confidence)to reduce exposures to triggers, CHWs promoted behavioralcapability (the knowledge and skills to perform a behavior),via role modeling (7, 16, 18, 19, 21).
The health belief model was used extensively by Parkeret al. to guide their intervention, from the educationalmessages developed to the resources and referrals provided.Organizational and community level theories were alsosynthesized in their comprehensive “ecological stressmodel,” which was presented, but not tested, in their pub-lished study. Social networks and social support conceptswere cited by two authors to validate the use of CHWsas appropriate interventionists (16, 19). In one study, thetranstheoretical stages of change guided the collection ofdata on caregivers’ smoking status and the delivery of astage-specific intervention approach (16).
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
Tab
le1.
—St
udy
char
acte
rist
ics.
Stud
yC
hara
cter
istic
s
Prim
ary
auth
or(y
ear)
,St
udy,
Elig
ibili
tyC
ontr
olgr
oup
Ass
essm
ent/t
otal
leng
thst
udy
desi
gnn
Setti
ngre
quir
emen
tsR
ecru
itmen
tst
udy
prot
ocol
offo
llow
-up
inm
onth
s
Bry
ant-
Step
hens
(200
8)3
arm
RC
TTo
tal:
396
Phila
delp
hia
•2–
16ye
ars
ofag
e•
Liv
ein
Wes
tPhi
lade
lphi
azi
pco
de•
Rec
eive
care
atC
hild
ren’
sH
ospi
talo
fPh
ilade
lphi
a•
Atl
east
one
asth
ma-
rela
ted
hosp
italiz
atio
nor
two
asth
ma-
rela
ted
emer
genc
yde
part
men
tvi
sits
,in
year
befo
reen
rollm
ent
Self
-&
phys
icia
nre
ferr
als
Obs
erva
tion-
only
grou
p(O
BS)
:vi
site
dev
ery
mon
thto
colle
ctsy
mpt
omdi
arie
s,gi
ven
info
rmat
ion
abou
tast
hma
self
-man
agem
entc
lass
esin
the
com
mun
ity.
Cas
e-m
atch
edco
ntro
ls(C
MC
):no
care
Eve
rym
onth
,ou
tcom
esre
port
edat
6,12
mo
/12
mo
Egg
lest
on(2
005)
2ar
mR
CT
100
Bal
timor
e•
6–12
year
sof
age
•C
urre
ntas
thm
asy
mpt
oms
•Ph
ysic
ian-
diag
nose
das
thm
a•
No
othe
rch
roni
clu
ngdi
seas
e
Scho
ol-b
ased
asth
ma
educ
atio
npr
ogra
mC
ontr
olgr
oup
rece
ived
inte
rven
tion
atth
een
dof
the
stud
y.6,
12m
o/12
mo
Kri
eger
(200
5)2
arm
RC
T27
4Se
attle
/Tac
oma
•4–
12ye
ars
ofag
e•
Pers
iste
ntas
thm
adi
agno
sis
•In
com
ebe
low
200%
of19
96FP
Lor
child
enro
lled
inM
edic
aid
•C
areg
iver
verb
ally
profi
cien
tin
Eng
lish,
Span
ish,
orV
ietn
ames
e•
Chi
ldsp
enta
tlea
st50
%ni
ghts
inth
eho
use
•H
ouse
was
inK
ing
Cou
nty
Com
mun
ity&
publ
iche
alth
clin
ics
(65%
),lo
cal
hosp
ital&
emer
genc
yro
oms
(27%
),co
mm
unity
agen
cies
&re
side
nts
(8%
)
Low
-int
ensi
tyco
ntro
lgro
upre
ceiv
eda
sing
leC
HW
visi
t,w
hich
incl
uded
aho
me
envi
ronm
enta
las
sess
men
t,an
actio
npl
an,l
imite
ded
ucat
ion,
&pi
llow
&m
attr
ess
enca
sem
ents
.A
tthe
end
ofth
est
udy,
this
grou
pre
ceiv
edal
lres
ourc
es&
addi
tiona
led
ucat
ion.
12m
o/12
mo
18m
ofo
rhi
gh-
inte
nsity
grou
pon
ly
McC
onne
ll.(2
005)
2ar
mR
CT
150
Los
Ang
eles
•6–
14ye
ars
ofag
e•
Atl
east
3vi
sits
tom
obile
clin
icor
clin
ic•
Posi
tive
skin
test
tom
ixed
cock
roac
hal
lerg
en•
Pers
iste
ntas
thm
adi
agno
sis
Scho
ol-b
ased
mob
ileas
thm
acl
inic
oral
lerg
ycl
inic
atun
iver
sity
med
ical
cent
er
Con
trol
grou
pre
ceiv
edco
nden
sed
vers
ion
ofth
ein
terv
entio
nat
the
end
ofth
est
udy.
4m
o/4
mo
Mor
gan
(200
4)2
×2
Fact
oria
lRC
T93
7B
osto
n,C
hica
go,D
alla
s,N
ewY
ork,
Seat
tle/T
acom
a,T
ucso
n
•5–
11ye
ars
ofag
e•
Res
iden
tsof
cens
ustr
acts
whe
reat
leas
t20%
ofho
useh
olds
had
inco
mes
belo
wth
eFP
L•
Phys
icia
ndi
agno
sis
ofas
thm
a•
Atl
east
one
asth
ma-
rela
ted
hosp
italiz
atio
nor
two
unsc
hedu
led,
asth
ma-
rela
ted
visi
tsto
the
clin
icor
emer
genc
yde
part
men
tdur
ing
the
prev
ious
6m
onth
s•
Posi
tive
skin
test
inre
spon
seto
atle
asto
neal
lerg
en
NR
Con
trol
grou
pre
ceiv
edvi
sits
only
for
eval
uatio
nof
heal
thou
tcom
esat
6-m
onth
inte
rval
sth
roug
hout
the
stud
y.(P
hysi
cian
feed
back
inte
rven
tion
incl
uded
bim
onth
lyre
port
sof
the
child
ren’
sas
thm
asy
mpt
oms
&us
eof
heal
thca
rese
rvic
esto
thei
rpr
imar
yca
reph
ysic
ians
.)
6,12
,18,
24m
o/24
mo
Park
er(2
007)
Stag
gere
d,2
arm
RC
T
298
Det
roit
•7–
11ye
ars
ofag
e•
Res
ided
inSo
uthw
esto
rE
asts
ide
Det
roit
•R
epor
ted
pers
iste
ntsy
mpt
oms
Scho
ol-b
ased
prog
ram
sC
ontr
olgr
oup
rece
ived
abo
okle
tof
basi
cin
form
atio
non
asth
ma.
Con
trol
grou
pre
ceiv
edin
terv
entio
nat
the
end
ofth
est
udy.
12m
o/12
mo
Will
iam
s(2
006)
2ar
mR
CT
161
Atla
nta
•5–
12ye
arof
age
•E
nglis
hsp
eaki
ng•
Res
ided
with
inth
ezi
pco
deof
the
Atla
nta
Em
pow
erm
entZ
one
•Pr
evio
uspr
esen
tatio
nto
the
emer
genc
yde
part
men
tof
the
loca
lmaj
orpu
blic
child
ren’
sho
spita
l
Em
erge
ncy
depa
rtm
enta
tlo
calc
hild
ren’
sho
spita
lC
ontr
olgr
oup
rece
ived
inte
rven
tion
atth
een
dof
the
stud
y.4,
8,12
mo/
12m
o
Abb
revi
atio
ns:C
MC
=ca
se-m
atch
edco
ntro
lgro
up;F
PL=
Fede
ralP
over
tyL
evel
;Mo
=m
onth
s;N
R=
notr
epor
ted;
OB
S=
obse
rvat
ion
grou
p;R
CT
=ra
ndom
ized
clin
ical
tria
l;T
=tr
eatm
entg
roup
.
566
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
COMMUNITY HEALTH WORKERS & ENVIRONMENTAL TRIGGERS 567
Table 2.—Characteristics of study population.
Characteristics of study population
Atopic statusAge, years Gender (% Race/ethnicity Household (% positive)
Primary author (year) (mean) [SD] C/T female) C/T (%) C/T Income (%) C/T at baseline C/T
Bryant-Stephens (2008) T: 6.1 [3.9]OBS: 5.6 [3.5]CMC: 5.3 [NR]
T: 40OBS: 34CMC: 43
T: AA 99OBS: AA 100CMC: AA 100
NR NR
Eggleston (2005) 8.3 [1.4]/8.5 [1.5] 60/48 98 AA/100 AA $<10,000/yr: 33/45$10,000–20,000: 42/34>$20,000: 25/21
DM: 36/23CR: 44/40Mouse: 8/10Cat: 26/19≥ 1+ : 65/75
Krieger (2005) 7.3 [NR]/ 7.4 [NR] 38.2/44.2 Non–Hispanic White21.3/12.3
AA 27.9/31.9Vietnamese 22.1/25.4Other Asian 5.2/9.4Hispanic 17.7/17.4Other 5.9/3.6
<100% FPL: 60.9/51.9100–149% FPL: 24.1/33.3150–200% FPL: 15/14.8
NR
McConnell (2005) NR 33/34 Hispanic 92/94Other 8/6
<$15,000/yr: 38/40$15,000–24,999: 33/31≥ 25, 000 : 29/29
CR: 40/61
Morgan (2004) 7.7 [SE 0.09]/7.6 [SE 0.09] 37.8/36.9 AA 41.5/40.3Hispanic 40/40.3Other 18.5/22
% with Income<$15,000/yr:60.9/59.8
DM: 63.3/62.8CR: 70.3/67.8Cat: 47.8/40.8Dog: 22.6/21.4Rodent: 33.6/33.3Mold: 48.1/51.8
Parker (2007) 8.8 [1.41]/9.01 [1.50] 41/43 AA 79/83Hispanic 10/11Caucasian 5/4Other 6/3
% with Income < $10,000/yr:46/37
DM: 35/41CR: 16/25Cat: 25/21Dog: 5/11Grasses: 30/29Ragweed: 23/24Mouse: 13/12Rat: 10/10Mold: 18/22
Williams (2006) 8 [NR]/8 [NR] 34/48 AA 100/99 <$500/mo: 39/33$500–1200: 33/36$1201–1600: 13/15>$1601: 8/6Did not answer: 7/10
DM: 60/57CR: 38/35Cat: 17/19Dog: 14/16Grasses: 51/47Trees: 31/23Mold: 13/15
AA = African American; C = control group; CMC = case-matched control group; CR = cockroach; DM = dust mite; FPL = Federal poverty level; NR = not reported; OBS =observation group; SE = standard error; SD = standard deviation; T = treatment group; Yr = year.
Education related to the relationship between triggers andasthma and ways to decrease or avoid triggers was deliveredin all seven studies. Most studies focused solely on environ-mental triggers for asthma, but three studies included generalinformation about asthma (5, 21, 22). One study also includededucation around medication use and devices (21).
All of the RCTs’ interventions included providing pillowand mattress encasements for dust mite avoidance. Other re-sources given to some or all participants (depending on theassessment and their need) included cockroach bait, pest re-mediation, rodent traps, boric acid and caulk, vacuum bags,vacuums with HEPA filters, HEPA air filters/purifiers, clean-ing kits, shower curtains, and commercial quality door mats.Referrals varied among studies but encompassed smokingcessation, professional house cleaning and extermination,skin testing, weatherization program assistance, and socialservice agencies.
All but one study provided individually tailored interven-tions, either in terms of (1) incorporating the caregivers’preferences in determining problems and solutions (8); (2)responding to individual subjects’ allergic sensitivity (skinprick testing); and/or (3) responding to triggers identified
during the home assessment (23). Three studies combined allthree of these approaches (5, 15, 19). Morgan et al. tailoredthe selection and intensity of the intervention modules toeach child’s allergic profile and their home assessments, butnot caregiver preference for particular problems or solutions.
The length of the intervention period varied from threeto nine home visits, although one study’s authors reportedthat some participants actually received 17 home visits (22).Visits were generally 60 minutes each but sometimes ranas long as 150 minutes (8). In addition to content coveredat the home visits, 2 of the studies also included follow-uptelephone calls (6, 7). The intervention period ranged from 6weeks to 1 year.
Four studies described the cost of intervention per child.Costs ranged from $492, which included encasements, theroom HEPA air filter, pest control visits, and CHW visits (7)to $1500–2000/child (not itemized in publication) (6).
Study FindingsOutcome categories among the studies were broad and
non-uniform. Categories included mediators of improvedhealth outcomes, such as trigger-related knowledge, trigger
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
568 J. POSTMA ET AL.
Table 3.—Characteristics, training, & supervision of community health workers.
Social congruence withPrimary author (year) study participants Training Quality control
Bryant-Stephens (2008) • From target community NR Use of protocolsEggleston (2005) Swartz (2004) • All lived, had lived, or had worked in
target community• All African American
Previously trained, but training notdescribed
NR
Krieger (2005) Krieger (2002) • All lived in target community• 54% shared ethnic background of
study participants• Most communicated in primary
language of participants• All either had asthma or had close
family members with asthma
40 hours, 10–20 hours of continuingeducation/yr
Met with the primary investigator everyother week; met with the project’ssteering committee quarterly, use ofprotocols
McConnell (2005) • From target community Participation in training program inurban health education, includedcontent specific to asthma
Use of protocols
Morgan (2004) Crain (2002) • “Culturally sensitive”• Bilingual
2 centralized training sessions Use of protocols
Parker (2007) • “Culturally similar”• Half were bilingual
4 weeks of “intensive training” &“ongoing” training
NR
Williams (2006) NR Participation in CHW TrainingProgram via local university
NR
NR = not reported; Yr = year.
reduction behaviors and allergen or exposure levels, as wellas asthma-related health outcomes, such as a change in lungfunction, medication use, asthma symptoms, functional lim-itations, and health care utilization (Table 5). Psychosocialinfluences on health, including measures of quality of life,social support, and depression, were measured in few studies.
Caregiver’s depressive symptoms significantly decreasedin the intervention group (p value not reported) but increasedin the control group (p = 0.028), whereas two measures ofsocial support, instrumental and emotional, did not improvein the one study that measured these outcomes (22). Childquality of life did not improve in the one study that measuredit (7), while caregiver quality of life did significantly improvewithin the low- and high-intensity groups and across-groups(0.58 points, 95% CI: 0.18-0.99, p = 0.005) in the one studythat measured it (5).
Knowledge was measured in two studies and results werenon-significant, although in the direction predicted. In one
study, knowledge significantly increased in both the low- andhigh-intensity groups but was greater in the high-intensitygroup (17). In another study, a summary score of knowledgeincreased in the intervention group but was not significantacross-groups (8).
Behavior change was measured in four studies and resultswere mixed, with most behavioral improvements tied to re-source provision (Figure 1). For example, in one study, twoout of four improved behaviors were related to the use ofpillow and mattress encasements (21). In a second study,self-reported behavior change significantly increased in boththe low- and high-intensity groups but was greater in thehigh-intensity group. Again, the across-group effect was in-significant when measured as a summary score. The fivebehaviors that did significantly improve across the groupsincluded vacuuming the child’s bedroom at least twice overthe past 2 weeks, vacuuming or removing cloth covered fur-niture at least twice over the past 2 weeks, using doormats
Figure 1.—Behavior improvement & resource provision.
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
Tab
le4.
—In
terv
entio
nch
arac
teri
stic
s.
Inte
rven
tion
char
acte
rist
ics
Prim
ary
auth
orT
heor
etic
alD
ose/
Len
gth
(yea
r)fr
amew
ork
Con
tent
Res
ourc
es/r
efer
rals
Mod
efr
eque
ncy
ofH
VC
ost/c
hild
Bry
ant-
Step
hens
(200
8)N
RH
V1:
No
educ
atio
ngi
ven
HV
2:A
sthm
aas
adi
seas
e,si
gns
&sy
mpt
oms
ofan
asth
ma
atta
ck,m
edic
atio
nre
view
,pes
tcon
trol
HV
3:In
door
trig
gers
,avo
idan
cete
chni
ques
,use
ofen
case
men
tsH
V4:
Rev
iew
ofas
thm
am
edic
atio
ns&
devi
ces,
cock
roac
h&
petd
ande
r,av
oida
nce
tech
niqu
esH
V5:
Med
icat
ion
actio
npl
anH
V6:
Rev
iew
ofm
edic
atio
ns&
devi
ces,
ques
tion
&an
swer
sess
ion
Pillo
w&
mat
tres
sen
case
men
ts,
bait,
clea
ning
supp
lies,
vacu
um&
tras
hba
gs,r
efer
rals
toco
ntra
ctor
s
HV
6H
Vw
ithin
11/ 2
mo
NR
Inte
rven
tion
&C
HW
sala
ry:$
675
Egg
lest
on(2
005)
Swar
tz(2
004)
Soci
alco
gniti
veth
eory
HV
1:So
urce
sof
indo
orai
rpo
lluta
nts,
way
sto
redu
cein
door
air
pollu
tant
s,av
oidi
ngex
posu
reto
seco
ndh&
smok
e,sm
okin
gce
ssat
ion
HV
2:In
door
alle
rgen
&ir
rita
ntco
ntro
l,cr
eatio
nof
pers
onal
ized
care
plan
HV
3:R
einf
orce
alle
rgen
&ai
rpo
llutio
nin
form
atio
n
Pillo
w&
mat
tres
sen
case
men
ts,
CR
&ro
dent
exte
rmin
atio
nif
child
was
alle
rgic
,HE
PAai
rfil
ter,
smok
ing
cess
atio
n,vi
sual
aids
,mat
eria
ls
HV
&Ph
one
calls
3H
Vw
ithin
5m
o60
min
Inte
rven
tion
&C
HW
sala
ry:$
492
Kri
eger
(200
5)K
rieg
er(2
002)
Soci
alco
gniti
veth
eory
Soci
alne
twor
ks,s
ocia
lsup
port
Tra
nsth
eore
tical
stag
esof
chan
ge
HV
1:A
sthm
aba
sics
,DM
s,ro
ache
sH
V2:
Dis
cuss
hous
ehol
dac
tion
plan
,DM
s,m
oist
ure,
toba
cco,
CR
HV
3:D
iscu
ssho
useh
old
actio
npl
an,D
M,h
ouse
hold
clea
ning
,hou
seho
ldto
xins
,moi
stur
e,pe
ts,r
oden
ts,
outd
oor
air
HV
4:D
iscu
ssho
useh
old
actio
npl
an,a
sthm
aba
sics
,ho
useh
old
clea
ning
,CR
sH
V5:
Dis
cuss
hous
ehol
dac
tion
plan
,DM
s,ho
useh
old
clea
ning
,hou
seho
ldto
xins
HV
6:D
iscu
ssho
useh
old
actio
npl
an
Pillo
w&
mat
tres
sen
case
men
ts,
bait/
trap
s,pe
stre
med
iatio
n,va
cuum
,sho
wer
curt
ain,
door
mat
,cle
anin
gki
t,w
eath
eriz
atio
npr
ogra
ms,
soci
alse
rvic
ere
ferr
als,
skin
test
ing
HV
5H
V,4
mor
eop
tiona
l,w
ithin
12m
oM
ean:
7H
V
60m
inIn
terv
entio
n&
CH
Wsa
lary
,ind
irec
tco
sts:
$112
4
McC
onne
ll(2
005)
NR
(NR
per
HV
)(L
AC
ASA
web
site
,200
8)E
ffec
tof
CR
onlu
ngs,
lifec
ycle
ofC
R,C
Rco
ntro
l,st
rate
gies
tore
duce
infe
stat
ions
(e.g
.sea
ling
crac
ks,c
lean
ing,
trap
s)
Pillo
w&
mat
tres
sen
case
men
ts,
bori
cac
id,c
aulk
,wri
tten
mat
eria
ls,c
olor
ing
book
HV
2H
Vw
ithin
4m
o60
–150
min
NR
Mor
gan
(200
4)C
rain
(200
2)So
cial
lear
ning
theo
ry/S
ocia
lco
gniti
veth
eory
Mod
ule
1:R
ole
ofal
lerg
ens
&ir
rita
nts
inas
thm
a,st
rate
gies
tom
inim
ize
DM
s&
crea
tea
safe
slee
ping
zone
Mod
ule
2.R
evie
wof
HV
1,st
rate
gies
tode
crea
seex
posu
reto
ET
SM
odul
e3.
Rev
iew
ofH
V1,
cock
roac
hbe
havi
or,s
trat
egie
sto
decr
ease
expo
sure
toco
ckro
ache
s(e
.g.c
lean
ing,
IPM
)M
odul
e4.
Rev
iew
ofH
V1,
peta
llerg
ens
&as
thm
a,st
rate
gies
toav
oid
pets
,pet
dand
erM
odul
e5.
Rev
iew
ofH
V1,
rode
ntbe
havi
or,s
trat
egie
sto
prev
entr
oden
ts(e
.g.c
oppe
rm
esh
aten
try
poin
ts)
Mod
ule
6.R
evie
wof
HV
1,so
urce
sof
mol
d/m
oist
ure,
Stra
tegi
esto
prev
ent&
clea
nm
old
Pillo
w&
mat
tres
sen
case
men
ts,
vacu
um,H
EPA
air
puri
fier,
pest
rem
edia
tion,
ifne
eded
HV
&Ph
one
calls
5H
Vw
ithin
12m
oM
ean:
4.7
HV
+/−
SD1.
09
<12
0m
inM
ean:
73.6
+/−
SD27
.59
$150
0–20
00
Park
er(2
007)
Soci
alco
gniti
veth
eory
Hea
lthbe
lief
mod
elO
rgan
izat
iona
l&co
mm
unity
leve
lthe
orie
s
HV
1:G
ener
alas
thm
ain
form
atio
n&
the
role
ofen
viro
nmen
talt
rigg
ers
HV
2:A
ctio
npl
ande
velo
ped
base
don
prio
ritiz
edlis
tof
trig
gers
HV
3–9:
Met
hods
tore
duce
trig
gers
(DM
,ET
S,C
R,p
ets,
rode
nts,
mol
d)
Pillo
w&
mat
tres
sen
case
men
ts,
vacu
um,c
lean
ing
supp
lies,
IPM
,clin
ic,&
soci
alse
rvic
ere
ferr
als,
ifne
eded
HV
9H
Vw
ithin
12m
oR
ange
:1–
17M
ean:
9.24
NR
NR
Will
iam
s(2
006)
Self
effic
acy
(NR
per
HV
)C
are
ofen
case
men
ts,w
ashi
ngsh
eets
&up
hols
tery
Use
ofC
Rba
it,fo
odha
ndlin
gpr
actic
esIm
pact
ofE
TS
onas
thm
aO
ther
topi
csas
need
ed:c
ontr
ollin
gm
oist
ure
&hu
mid
ity,
clea
ning
mol
d,pe
ts
Pillo
w&
mat
tres
sen
case
men
ts,
one
time
prof
essi
onal
hous
ecle
anin
g,ro
ach
bait
HV
3H
Vw
ithin
12m
oN
RN
R
CR
=co
ckro
ache
s;D
M=
dust
mite
s;E
TS
=en
viro
nmen
talt
obac
cosm
oke;
HE
PA=
Hig
h-ef
ficie
ncy
part
icul
ate
air;
HV
=ho
me
visi
t;IP
M=
inte
grat
edpe
stm
anag
emen
t;M
in=
min
utes
;Mo
=m
onth
s;N
R=
notr
epor
ted.
569
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
Tab
le5.
—Fi
ndin
gs.
Med
iato
rsof
impr
oved
heal
th-r
elat
edou
tcom
esH
ealth
-rel
ated
outc
omes
Tri
gger
redu
ctio
nE
nvir
onm
ent
Ast
hma
Act
ivity
Med
icat
ion
Urg
ent
Prim
ary
auth
or(y
ear)
beha
vior
s(a
llerg
en/e
xpos
ures
)sy
mpt
oms
limita
tions
use
care
use
Bry
ant-
Step
hens
(200
8)S↑
inus
eof
pillo
w(p
<0.
001)
&m
attr
ess
enca
sem
ents
(p<
0.00
1)N
Sch
ange
inre
duci
ngca
rpet
&pr
esen
ceof
ET
S
NS
chan
gein
CR
&ro
dent
prob
lem
sN
RN
RN
eed
for
dail
yus
eof
:R
escu
em
edN
S(N
R)
Con
trol
ler
med
NS
(NR
)
T&
OB
S:N
Sdi
ffer
ence
on3/
3m
easu
res
(Hos
p.p
<0.
95,E
Dp
<0.
99,C
linic
p<
0.26
)T
&C
MC
:S↓a
cros
s3/
3m
easu
res
(Hos
p.0.
31,
p<
0.05
,ED
0.97
,p
<0.
01,C
linic
0.48
p<
0.05
Egg
lest
on(2
005)
T:2
7%ha
dpi
llow
/mat
tres
sen
case
men
tsin
plac
eat
12m
o
S39
%↓i
nPM
10in
Tvs
.5%
↑in
C(p
<0.
001)
,S↓i
nPM
2.5
(p<
0.00
1)N
Sch
ange
inC
R(p
=0.
08),
NO
,cat
,DM
,Mou
se(N
R)
S↓i
nda
ytim
esy
mpt
oms/
2w
ks(p
<0.
05)
NS
chan
gein
othe
rca
tego
ries
NR
NR
NS
↓in
clin
ic,E
Rvi
sits
&ho
spita
lizat
ions
(NR
)
Kri
eger
(200
5)A
llerg
en&
expo
sure
anal
ysis
isfr
oma
subg
roup
of60
hom
es(T
akar
o,20
04)
NS
(p=
0.14
1)E
xpos
ures
T:S
↓in
pres
ence
ofC
R(p
=0.
003)
,Con
dens
atio
n(p
=0.
000)
,Moi
stur
e(p
=0.
003)
,Dus
twei
ght
(p=
0.00
8),C
ompo
site
trig
ger
scor
e(p
=0.
005)
C:N
SA
:Con
dens
atio
n(p
=0.
013)
,du
stw
eigh
t(p
=0.
042)
Ant
igen
surf
ace
load
ing
T:S
↓in
DM
(p=
0.02
5),d
og(p
=0.
034)
,fra
ctio
nof
antig
ens≥
cuto
ffpo
int(
p=
0.00
6)C
:S↓i
nca
t(p
=0.
020)
A:N
RA
ntig
enco
ncen
trat
ion
indu
stA
llgr
oups
:NS
#da
ysw
ith
sym
ptom
s/2
wks
NS
(p=
0.13
8)#
days
wit
hac
tivi
tyli
mit
atio
ns/2
wks
S(p
=0.
029)
#da
ysch
ild
mis
sed
scho
ol/2
wks
NS
(p=
0.10
5)#
days
care
give
rm
isse
dw
ork/
2w
ksN
S(p
=0.
890)
#da
ysus
ing
are
scue
med
/2w
ksN
S(p
=0.
781)
#da
ysus
ing
aco
ntro
ller
med
/2w
ksN
S(p
=0.
250)
#vi
sits
tour
gent
care
/2m
oS
(p=
0.02
6)T
:Gai
nsin
beha
vior
s&
heal
thou
tcom
esat
exit
pers
iste
d@
18m
ofo
llow
up,w
itha
tend
ency
for
furt
her
impr
ovem
enta
cros
sal
lou
tcom
es(d
ata
not
show
nin
orig
inal
pape
r)
McC
onne
ll(2
005)
Rep
orte
dpr
acti
ces
S↑i
nus
eof
stic
kytr
aps
(p<
0.05
),bo
ric
acid
(p<
0.01
),ca
ulki
ng(p
<0.
05)
NS
chan
gein
3be
havi
ors
(NR
)O
bser
ved
prac
tice
sS
↑in
use
ofen
case
men
ts(p
<0.
01),
bori
cac
id(p
<0.
01)
NS
chan
gein
3C
Rpr
actic
es&
4fo
odbe
havi
ors
S↓C
Rlo
adin
bedd
ing
(p<
0.05
),N
Sch
ange
inC
Rco
ncen
trat
ion
inbe
ddin
g,&
CR
load
,con
cent
ratio
nin
kitc
hen,
S↑i
n#
CR
trap
ped
(p<
0.05
)N
S↓i
nev
iden
ceob
serv
ed,p
robl
emre
port
ed,p
rese
nce
ontr
aps
NR
NR
NR
NR
Mor
gan
(200
4)N
R@
12m
oB
edal
lerg
ens
S↓D
M1
(p<
0.00
1),D
M2
(p=
0.00
7),c
at(p
<0.
001)
NS
chan
ges
inC
R(p
=0.
13)
&do
g(p
=0.
29)
Flo
oral
lerg
ens
S↓C
R(p
<0.
001)
,DM
1(p
=0.
004)
,S
↑cat
(p=
0.02
)N
Sch
ange
sin
DM
2(p
=0.
28)
&do
g(p
=0.
56)
@24
mo
Bed
alle
rgen
sS
↓DM
1(p
=0.
004)
Flo
oral
lerg
ens
S↓C
R(p
<0.
003)
#of
days
wit
hsy
mpt
oms/
2w
ks@
12m
oS
↓by
0.82
(p<
0.00
1)@
24m
oS
↓by
0.60
(p<
0.01
)
@12
mo
#of
nigh
tsca
reta
ker
wok
e/2
wks
S↓b
y0.
61(p
<0.
001)
#da
ysca
reta
ker
chan
ged
plan
s/2
wks
S↓b
y0.
31(p
<0.
001)
#sc
hool
days
mis
sed/
2w
ksS
↓by
0.17
(p<
0.00
3)@
24m
o2/
3ac
tivity
impr
ovem
ents
wer
esu
stai
ned
NR
@12
mo
#of
unsc
hed.
visi
ts/y
rE
D↓b
y0.
35(p
=0.
04),
Hos
pN
S(p
=0.
56)
@24
mo
#of
unsc
hed.
visi
ts/y
rE
R↓b
y0.
26(p
=0.
07),
Hos
pN
S(p
=0.
19)
(Con
tinu
edon
next
page
)
570
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
Tab
le5.
—Fi
ndin
gs.(
Con
tinu
ed)
Med
iato
rsof
impr
oved
heal
th-r
elat
edou
tcom
esH
ealth
-rel
ated
outc
omes
Prim
ary
auth
orT
rigg
erre
duct
ion
Env
iron
men
tA
sthm
aA
ctiv
ityM
edic
atio
nU
rgen
t(y
ear)
beha
vior
s(a
llerg
en/e
xpos
ures
)sy
mpt
oms
limita
tions
use
care
use
Park
er(2
007)
S↑i
nch
angi
ngsh
eets
wee
kly
(p=
0.00
4),
vacu
umin
g(p
<0.
0001
),us
eof
pillo
w(p
=0.
0006
)&
mat
tres
sen
case
men
ts(p
<0.
0001
)N
Sch
ange
indu
stin
g,w
ashi
ngbe
dco
vers
,re
mov
ing
mol
d,&
3E
TS
mea
sure
s
S↓i
ndo
g(p
<0.
001)
NS
chan
gein
CR
,DM
,cat
S↓i
n2/
8sy
mpt
oms:
Pers
iste
ntco
ugh
(p=
0.03
4),C
ough
with
exer
cise
(p=
0.01
7)N
Sch
ange
in6/
8sy
mpt
oms
NR
Not
ona
cort
icos
tero
idN
S↓(p
=0.
073)
Not
onan
yco
ntro
ller
S↓
(p=
0.00
4)
S↓i
nne
edfo
run
sche
d.M
edic
alca
re(p
=0.
004)
Will
iam
s(2
006)
NR
S↓in
DM
(p<
0.05
),C
R(p
<0.
05)
25%
↓in
FSS
(p<
0.01
)25
%↓i
nFS
S(p
<0.
01)
NS
(NR
)N
S(N
R)
Abb
revi
atio
ns:
A=
acro
ssgr
oup;
C=
cont
rol
grou
p;C
MC
=ca
se-m
atch
edco
ntro
ls;
CR
=co
ckro
ache
s;D
M=
dust
mite
s;E
D=
emer
genc
yde
part
men
t;E
TS
=en
viro
nmen
tal
toba
cco
smok
e;FS
S=
Func
tiona
lse
veri
tysc
ore;
Hos
p=
hosp
ital;
NR
=no
trep
orte
d;N
S=
nots
igni
fican
t;O
BS
=ob
serv
atio
ngr
oup;
S=
sign
ifica
nt;T
=tr
eatm
entg
roup
;Uns
ched
=un
sche
dule
d.
571
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
572 J. POSTMA ET AL.
or removing shoes, using allergy control devices on mattressand pillows, and having and using a working kitchen exhaustfan (5). Four of five of these behaviors were facilitated byresources (e.g., vacuums, pillow and mattress encasements,door mats) provided in the study. Behaviors that did notsignificantly change included dusting the child’s bedroomtwice a week, washing sheets in hot water weekly, removingpets from the child’s bedroom and/or home, having and usinga working bathroom exhaust fan, smoking in the home, andsmoking by the caregiver (5).
In a third study, 5 of 14 reported and observed behaviorssignificantly improved across-groups, with 4 of 5 related toresources provided in the study (e.g., mattress/pillow encase-ments, caulk, boric acid [use of boric acid counted as botha reported and observed behavioral outcome]) (8). A fourthstudy reported similar findings. Of 10 behaviors, 4 signifi-cantly improved, 3 of which were related to resource provi-sion (e.g., vacuums, pillow and mattress encasements) (22).
A variety of allergens and exposures were measured in thestudies. Results were mixed in terms of which allergens or ex-posures significantly decreased over time and how or wherethey were measured. One study found no significant across-group difference in reducing cockroach or rodent problems(21). Another study found no significant reduction in the lev-els of cockroach allergen, dust mite allergens, cat allergen, ornitric oxide concentration but did find significant reductionsin particulate matter 2.5 and particulate matter 10 at 1 year(7).
Exposure levels for a subgroup of 60 randomly selectedhomes from the Seattle-based Healthy Homes Project werereported in a separate publication (17). Overall, the high-intensity intervention group had significant reductions frombaseline-to-exit in the surface loading of dust mite and dogantigens and the fraction of cat, dog, dust mite, and roachantigens that exceeded an a priori ( ≥ 5 µg/m2) cutoff point.The surface loading of cat allergen was the only significantimprovement in the low-intensity group and the interventioneffect across-groups was not reported. Presence of roaches,condensation and moisture score, dust weight, and compos-ite trigger scores also significantly decreased in the high-intensity group, whereas there were no significant improve-ments in the low-intensity group, and only condensation anddust weight were significant across-groups. No significantchanges were demonstrated within or across-groups in termsof antigen concentration in dust (17).
McConnell et al. also demonstrated mixed results in termsof cockroach allergen; there was little difference across-groups in the proportion of homes at a 4- month follow-upwith trapped cockroaches, although there was a significantdifference in the geometric mean cockroach count (killed bymeans other than traps) at the follow-up visit. There was alsoa significant across-group reduction in the level of cockroachallergen in bedding, likely due to the mattress and pillowencasements, but not in the allergen samples obtained fromthe kitchen.
Morgan et al. found that at 1 year, the intervention waseffective in decreasing levels of dust mite and cat allergenin bed samples and cockroach, dust mite, and cat allergenobtained from floor samples across-groups. There were nosignificant differences in the levels of cockroach or dog al-lergens sampled from the bed or dog allergens sampled from
the floor. At 2 years, dust mite and cat allergen obtained frombed samples and cockroach and cat allergen in floor samplesremained significantly lower in the intervention group whencompared to control subjects. Across-group differences indog and cockroach allergens sampled from the bed remainedinsignificant as did dog allergen sampled from the floor. Thedecrease in cat allergen was not sustained at 2 years.
Parker et al. reported a significant intervention effect onlyin the reduction of concentration of dog allergen per gramof bedroom dust, but no effect in terms of cockroach, dustmite, or cat allergen concentration. Williams et al. found thatthe difference in the median pooled dust mite antigen levelsbetween the intervention group and the delayed interventiongroup was significantly different in the direction predicted;however, they found no difference across-groups in cock-roach antigen.
The lack of a consistent allergen reduction effect in twoof the studies may be explained by limited statistical powerto detect a difference (17) and differences in sampling andanalytic strategy (8). Nonetheless, Takaro et al. demonstratedcorrelations between frequent vacuuming and antigen levelsfound in dust. This supports the assumption that interven-tions directed at the sources of these exposures, as well asvacuuming itself, can reduce the surface loading of antigens.One study also demonstrated a dose-response relationshipbetween reduction in bedroom dust mite and cockroach al-lergen levels with decreases in the maximal number of dayswith symptoms, the number of hospitalizations, and the num-ber of unscheduled visits for asthma in both years of the study(6).
Asthma-related health outcomes were reported in six outof seven studies, including asthma symptoms, activity limi-tations, medication use, lung function, and urgent care use.Symptoms were defined similarly across three studies andincluded caregiver report of the child’s daytime symptoms,nighttime symptoms, and symptoms that interfered with theirchild’s activity or exercise over the past 2 weeks (5–7).Bryant-Stephens et al. collected information from caregiverson the frequency of nighttime wheezing, nighttime cough-ing, daytime wheezing, and daytime coughing over the past 2weeks. Parker et al. asked caregivers to recall similar symp-toms over the past year. Williams et al. asked caregivers torecall wheeze frequency, nighttime awakening symptoms,occurrence of a severe asthma attack, and limited homeand sports activities, every 4 months. This information wasincluded as the “functional severity” component of a totalasthma severity score that also included a medication and ur-gent care component. For the purposes of this review, thesesubcomponents will be considered separately.
There was a consistent and significant decrease incaregiver-reported asthma symptoms among interventionsubjects compared with control subjects, and, in one study,this decrease persisted for 1 year after the intervention wascompleted. Eggleston et al. reported a significant decrease inall categories of asthma symptoms across-groups at 6 months,but only daytime symptoms remained significantly differentin the intervention group at 1 year (7). Others reported signif-icantly fewer asthma symptoms during the intervention andfollow-up year on all measures (6, 23). (Note: Williams et al.report this decrease in symptoms as part of the functionalseverity score [FSS]).
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
COMMUNITY HEALTH WORKERS & ENVIRONMENTAL TRIGGERS 573
Two studies had insignificant findings across-groups, butone found a trend toward significance (21) and the otherdemonstrated a reduction in symptoms in both the highand low-intensity groups, thereby attenuating the differenceacross-groups (5). Parker et al. reported that although the in-tervention group reported a decrease in eight symptoms, thecontrol group reported a decrease in six of eight symptoms.Thus, only two of the eight symptoms (persistent cough andcough with exercise) significantly differed across groups.
In addition to a decrease in symptoms, there was a sig-nificant decrease in caregiver reports of activity limitationsamong intervention subjects compared to control subjects.Again, in one study, this decrease persisted for one year afterthe intervention was completed. Morgan et al. found signifi-cant across-group effects on all three measures (the numberof nights a caretaker woke up or changed plans because ofchild’s asthma and the number of school days missed due tochild’s asthma), two of which remained significant at the 2-year follow-up (number of nights caretaker wore up, schooldays missed) (6). Another study found that the number ofdays with activity limitations significantly decreased in boththe high- and low-intensity groups, and the difference acrossgroups was significant. However, missed work days did notimprove in either group, and missed school days improvedonly in the high-intensity group (5). Williams et al. reportedthat in ad hoc analyses and as a subcomponent of the FSS,caregiver-reported limitations in the child’s home and sportsactivities contributed to the significant across-group differ-ences in the FSS.
Four studies measured a change in medication use, and theeffect of the interventions on use of “rescue” and “controller”medications was mixed. One study reported that individualsin both the low and high-intensity groups significantly de-creased their use of beta-2 agonist “rescue” medications,but the difference across-groups was insignificant. In addi-tion, the “need for asthma controller medications” decreasedin the high-intensity group only, although it is not clear that aself-reported decrease in use of controller medication is an in-dication of improved asthma management (5). For example,another study reported a significant and positive interventioneffect in “reducing under treatment for active symptoms inthe category of children who should be on a controller butare not,” or increasing participants’ access to controller med-ications (22). Two studies found no significant across-groupdifferences in terms of medication use (21, 23).
Four studies assessed changes in lung function. Resultswere mixed and complicated by poor technique and com-pliance (21, 22). One study demonstrated, in a subgroup ofparticipants, significant improvement in daily nadir PF andFEV1, but no significant change in intraday variability inthe same measures (22). Neither of the other two studiesdemonstrated a significant across-group effect as measuredby spirometry or peak-flow monitoring (6, 7).
Six studies measured unscheduled asthma-related clinicalutilization. Results were, for the most part, statistically sig-nificant, with fewer unscheduled asthma-related visits associ-ated with the intervention group. Three studies demonstratedreduced urgent health care use across groups, including ur-gent care visits, emergency room visits, and hospitalizations(5, 21, 22). One study demonstrated a reduction in asthma-related visits to the emergency department or clinic, but no
reduction in asthma-related hospitalizations during the inter-vention year, although this study was not powered to detect adifference in this infrequent outcome. In the follow-up year,significant differences across-groups for urgent health careuse disappeared (6). Notably, Bryant-Stephens et al. foundthat there was no significant difference in urgent health careuse across the intervention and observation groups, but therewere differences between case-matched controls and the in-tervention group. That said, the observation-only group inthis study was given information on asthma self-managementclasses available in the community. (Use of those services bythe participants was not reported.) Another study found thatboth intervention and control groups made fewer visits to theemergency department and had fewer hospitalizations, butthese reductions were not statistically significant (7). Finally,when measured as part of the asthma severity score, urgenthealth care use was not significantly reduced in the study byWilliams et al.
DiscussionNotable similarities among the seven RCTs of CHW-
delivered environmental interventions include their focus onminorities with low-income and urban residence. Most usedresearch designs that eventually provided education and re-sources to all subjects, including “controls.” The majorityof studies used social cognitive theory (SCT) to guide inter-vention delivery, using role modeling and tailoring to delivereducational messages and build caregivers’ skills in avoidingallergens and other asthma triggers.
Assessment of intervention effectiveness differed amongthe studies. Only two measured indirect mediators, or psy-chosocial aspects, such as a change in depression, socialsupport, and (emotional aspects of) quality of life. Directmediators, including trigger-related knowledge acquisition,behaviors to reduce triggers, and allergen levels were avail-able for two, four, and six studies, respectively. Direct mea-sures of asthma status using lung function, symptoms, activ-ity limitations, and/or urgent health care use were reportedby at least three studies.
Overall, the studies support the effectiveness of CHW-delivered interventions, demonstrating decreased asthmasymptoms, lessened daytime activity limitations, and less-ened emergency and urgent care use, as indicated by thesolid line in Figure 2. The studies with higher intensity andhigher frequency (“higher dose”) interventions reported themost positive health outcomes. However, evidence of triggerreduction behaviors and improvements in allergen levels, hy-pothesized mediators of these interventions, were mixed, asindicated by the dashed lines in Figure 2. Nonetheless, onestudy demonstrated that trigger reduction behavior correlatedwith a reduction in antigens, and another study demonstrateda dose-response relationship between reductions in bedroomantigen levels with improvements in asthma-related healthoutcomes. Trigger reduction behaviors that did improve ap-peared to be tied to resource provision.
Sources of potential bias in these studies include baselinedifferences among intervention and control subjects, differ-ential drop out, nonblinding, and the Hawthorne effect. Over-all, there was little evidence for baseline differences in thereviewed studies. Two studies reported a higher dropout rate
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
Figu
re2.
—C
once
ptua
lMod
elof
Com
mun
ityH
ealth
Wor
ker-
Del
iver
edE
nvir
onm
enta
lInt
erve
ntio
ns
574
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
COMMUNITY HEALTH WORKERS & ENVIRONMENTAL TRIGGERS 575
among the intervention group, with Williams et al. noting thatpersons lost to follow-up differed on residential, staff, andlogistic factors (8, 23). One study found the opposite effect,that 5% of the intervention group and 9% of the control groupdropped out by one year (6). Another reported their overallattrition rate of 3% but did not report the rate per group (7).Parker et al. reported similar dropout rates across-groups;however, subjects that completed the study were more likelyto be male children and living in a home that their familyowned (22). Likewise, one study had similar dropout ratesbut those who completed the study were more likely to beAsian and less likely to have pets (5). The greatest sourceof bias was likely related to the lack of blinding in most ofthe studies. Nonblinding may have inflated the interventioneffect. In addition, questions asked about environmental trig-gers by a home visitor to control subjects may have motivatedcaregivers to pay attention to triggers, minimizing the effectof the intervention (Hawthorne effect).
Three pathways need to be explored in future studies to bet-ter determine the mechanism through which CHW-deliveredinterventions cause a change in asthma-related clinical out-comes, as indicated by the dotted lines and numbers inFigure 2. First, although most interventions were based onSCT, few studies operationalized theoretical concepts thatserve to mediate behavior change, thereby limiting assess-ment of the influence of this component of the causal path-way. No study measured participants’ “self-efficacy” in re-ducing environmental triggers via, for example, making andusing a weak bleach solution on mold, placing a vapor barrierto prevent moisture buildup, and applying principles of in-tegrated pest management. In addition, although two studiesmeasured “knowledge,” in keeping with SCT, “knowledge”includes elements of “behavioral capability,” the knowledgeand skills needed to perform a behavior, and “outcome ex-pectations,” the anticipatory outcomes of a behavior. Withoutthese intermediate variables, it is difficult to understand whataspects of the intervention promoted behavioral change anddelineate the pathways through which trigger reduction be-havior occurs (22, 24).
A second pathway in need of further understanding relatesto the therapeutic processes of the CHWs. Qualitative studiessuggest that CHWs are instrumental in providing social andemotional support (12). In this review, social support wasmeasured in only one study, and the change was insignificant(22). These authors did, however, find a significant decreasein levels of depressive symptoms among caregivers in inter-vention families, which suggests that CHWs facilitated cop-ing mechanisms among caregivers “through the information,assistance, and referrals given to them” by the CHWs (22).Notably, one study found that 74% of the enrolled caregivershad symptoms of depression, 44% of whom were severelydepressed per the CES-D scale (7). No subanalyses werereported in these studies testing for a moderator effect, al-though it could be hypothesized that a depressed caregiverwould lack the motivation to make behavior changes thatwould consequently lead to allergen reduction.
Other researchers have hypothesized that increased care-giver quality of life is related to an increase in self-efficacyand coping (4). Nine of 13 items on the Pediatric AsthmaCaregiver’s Quality of Life Questionnaire (QOL) concernemotional function and likely tap into the related construct
(25). In this review, caregiver quality of life significantly in-creased in one study (5), a finding that is consistent witha quasi-experimental CHW-delivered asthma interventionstudy using the same QOL tool (26) as well as a nurse-delivered environmental health intervention in an RCT (27).Better coping could lead to a decrease in depression and animprovement in asthma management (including trigger re-duction behaviors), which would likely contribute to betterclinical outcomes.
Finally, none of the studies assessed the role of the CHW asa liaison to health care providers, which may have led to a de-crease in urgent care visits due to appropriate medication useand better asthma management. Previous literature supportsCHWs in this role (9, 11, 12, 28–31). Alternative pathwaysto improved clinical outcomes are supported by findings thatdemonstrated a reduction in urgent care use by both the envi-ronmental intervention and observation-only (e.g., attentioncontrol) groups, with no significant difference found acrossthese groups (21). In another study, those subjects that had aprolonged relationship with a CHW demonstrated improve-ments in caregiver quality-of-life score, child’s asthma symp-toms, and reductions in urgent health services use, relativeto households receiving a single visit from a CHW and noresources other than bedding encasements. Those receivingthe low-intensity intervention showed smaller improvementsthat reached statistical significance for quality-of-life andsymptoms (5).
While evidence suggests that CHW-delivered interven-tions cause a change in asthma-related health outcomes,understanding the mechanisms through which this improve-ment occurs is critical to developing and evaluating asthmaintervention programs. Evidence from this review suggeststhat social, behavioral, and environmental dimensions ofasthma management are necessary to consider when design-ing and evaluating asthma intervention programs.
ConclusionIdentification and management of environmental asthma
triggers is a well-established cornerstone of optimal asthmacare. Preliminary evidence for the effectiveness of CHW-ledinterventions supports ongoing development of this CHWrole, which may be more cost effective and readily acceptedby at-risk communities. Overall, the studies consistentlyidentified positive outcomes associated with CHW-deliveredinterventions, including decreased asthma symptoms, day-time activity limitations, and emergency and urgent care use.However, improvements in trigger reduction behaviors andallergen levels, hypothesized mediators of these outcomes,were inconsistent. Trigger reduction behaviors appeared tobe tied to study-based resource provision. Knowledge gapsregarding the impact of CHW interventions on self-efficacyand other mediating factors such as depression and com-munity and individual stress, need to be addressed to betterunderstand the critical factors that link these interventions toimproved outcomes.
Future work in this area should include developing, testing,and/or incorporating measures that link the theory underlyingthe intervention to relevant outcomes, and measures that cap-ture complimentary social pathways through which CHWs
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.
576 J. POSTMA ET AL.
are effecting change in asthma morbidity, such as through in-creased access to health professionals and improvements ingeneral asthma management skills through coping and socialsupport and access to health care.
References1. Moorman JE, Rudd RA, Johnson CA, King M, Minor P, Bailey C, Scalia
MR, Akinbami LJ. National surveillance for asthma–United States, 1980–2004. MMWR Surveill Summ 2007; 56(8):1–54.
2. National Heart Lung and Blood Institute and National Asthma Educationand Prevention Program, Guidelines for the Diagnosis and Management ofAsthma 2007, U.S. Department of Health and Human Services, p. 440.
3. Carter MC, Perzanowski MS, Raymond A, Platts-Mills TA. Home inter-vention in the treatment of asthma among inner-city children. J Allergy ClinImmunol 2001; 108(5):732–737.
4. Klinnert MD, Liu AH, Pearson MR, Ellison MC, Budhiraja N, Robinson JL.Short-term impact of a randomized multifaceted intervention for wheezinginfants in low-income families. Arch Pediatr Adolesc Med 2005; 159(1):75–82.
5. Krieger JW, Takaro TK, Song L, Weaver M. The Seattle-King CountyHealthy Homes Project: a randomized, controlled trial of a communityhealth worker intervention to decrease exposure to indoor asthma triggers.Am J Public Health 2005; 95(4):652–659.
6. Morgan WJ, Crain EF, Gruchalla RS, O’Connor GT, Kattan M, Evans R III,Stout J, Malindzak G, Smartt E, Plaut M, Walter M, Vaughn B, Mitchell H.Results of a home-based environmental intervention among urban childrenwith asthma. N Engl J Med 2004; 351(11):1068–1080.
7. Eggleston PA, Butz A, Rand C, Curtin-Brosnan J, Kanchanaraksa S, SwartzL, Breysse P, Buckley T, Diette G, Merriman B, Krishnan JA. Home environ-mental intervention in inner-city asthma: a randomized controlled clinicaltrial. Ann Allergy Asthma Immunol 2005; 95(6):518–524.
8. McConnell R, Milam J, Richardson J, Galvan J, Jones C, Thorne PS,Berhane K. Educational intervention to control cockroach allergen exposurein the homes of hispanic children in Los Angeles: results of the La Casastudy. Clin Exp Allergy 2005; 35(4):426–433.
9. Health Resources and Services Administration. Community Health WorkerNational Workforce Study. 2007; [cited 2008; 7/30]; Available from:http://bhpr.hrsa.gov/healthworkforce/chw/default.htm#preface.
10. Norris SL, Chowdhury FM, Van Le K, Horsley T, Brownstein JN, ZhangX, Jack L Jr, Satterfield DW. Effectiveness of community health workers inthe care of persons with diabetes. Diabet Med 2006; 23(5):544–556.
11. Brownstein JN, Chowdhury FM, Norris SL, Horsley T, Jack L Jr, ZhangX, Satterfield D. Effectiveness of community health workers in the care ofpeople with hypertension. Am J Prev Med 2007; 32(5):435–447.
12. Andrews JO, Felton G, Wewers ME, Heath J. Use of community healthworkers in research with ethnic minority women. J Nurs Scholarsh 2004;36(4):358–365.
13. Lewin SA, Dick J, Pond P, Zwarenstein M, Aja G, van Wyk B, Bosch-Capblanch X, Patrick M. Lay Health workers in primary and communityhealth care. Cochrane Database of Systematic Reviews, 2005(1): Art. No.:CD004015. DOI: 10.1002/14651858.CD004015.pub2.
14. University of Southern California. LA CASA: Los Angeles ControllingAsthma by Stopping Allergens. 2008 [cited 2008 June 4]; Available from:http://www.usc.edu/schools /medicine /departments /preventive medicine/divisions/occupational/research/lacasa/.
15. Swartz LJ, Callahan KA, Butz AM, Rand CS, Kanchanaraksa S, Diette GB,Krishnan JA, Breysse PN, Buckley TJ, Mosley AM, Eggleston PA. Methodsand issues in conducting a community-based environmental randomizedtrial. Environ Res 2004; 95(2):156–165.
16. Krieger JK, Takaro TK, Allen C, Song L, Weaver M, Chai S, DickeyP. The Seattle-King County healthy homes project: implementation of acomprehensive approach to improving indoor environmental quality forlow-income children with asthma. Environ Health Perspect 2002; 110 Suppl2:311–322.
17. Takaro TK, Krieger JW, Song L. Effect of environmental interventions toreduce exposure to asthma triggers in homes of low-income children inSeattle. J Expo Anal Environ Epidemiol 2004; 14 Suppl 1:S133–143.
18. Crain EF, Walter M, O’Connor GT, Mitchell H, Gruchalla RS, KattanM, Malindzak GS, Enright P. Evans R III, Morgan W, Stout JW, Homeand allergic characteristics of children with asthma in seven U.S. urbancommunities and design of an environmental intervention: the Inner-CityAsthma Study. Environ Health Perspect 2002; 110(9):939–945.
19. Parker EA, Baldwin GT, Israel B, Salinas MA. Application of health pro-motion theories and models for environmental health. Health Educ Behav2004; 31(4):491–509.
20. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: synthesis of bestevidence for clinical decisions. Ann Intern Med 1997; 126(5):376–380.
21. Bryant-Stephens T, Li Y. Outcomes of a home-based environmental remedi-ation for urban children with asthma. J Natl Med Assoc 2008; 100(3):306–316.
22. Parker EA, Israel BA, Robins TG, Mentz G, Lin X, Brakefield-CaldwellW, Ramirez E, Edgren KK, Salinas M, Lewis TC. Evaluation of Commu-nity Action Against Asthma: A Community Health Worker Intervention toImprove Children’s Asthma-Related Health by Reducing Household Envi-ronmental Triggers for Asthma. Health Educ Behav 2008; 35(3):376–395.
23. Williams SG, Brown CM, Falter KH, Alverson CJ, Gotway-Crawford C,Homa D, Jones DS, Adams EK, Redd SC. Does a multifaceted environ-mental intervention alter the impact of asthma on inner-city children? J NatlMed Assoc 2006; 98(2):249–260.
24. Clark NM, Valerio MA. The role of behavioural theories in educationalinterventions for paediatric asthma. Paediatr Respir Rev 2003; 4(4):325–333.
25. Juniper EF, Guyatt GH, Feeny DH, Ferrie PJ, Griffith LE, Townsend M.Measuring quality of life in the parents of children with asthma. Qual LifeRes 1996; 5(1):27–34.
26. Primomo J, Johnston S, DiBiase F, Nodolf J, Noren L. Evaluation ofa community-based outreach worker program for children with asthma.Public Health Nurs 2006; 23(3):234–241.
27. Klinnert MD, Liu AH, Pearson MR, Tong S, Strand M, Luckow A, RobinsonJL. Outcome of a randomized multifaceted intervention with low-incomefamilies of wheezing infants. Arch Pediatr Adolesc Med 2007; 161(8):783–790.
28. Swider SM. Outcome effectiveness of community health workers: an inte-grative literature review. Public Health Nurs 2002; 19(1):11–20.
29. Nemcek MA, Sabatier R. State of evaluation: community health workers.Public Health Nurs 2003; 20(4):260–270.
30. Rosenthal EL. National Community Health Advisor Study: Weaving theFuture. 1998 [cited 2008 7/30]; Available from: http://www.rho.arizona.edu/Resources/Studies/cha-study/default.aspx.
31. Friedman AR, Butterfoss FD, Krieger JW, Peterson JW, Dwyer M,Wicklund K, Rosenthal MP, Smith L. Allies community health workers:Bridging the gap. Health Promot Pract 2006; 7(2 Suppl):96S–107S.
J A
sthm
a D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y M
ichi
gan
Uni
vers
ity o
n 10
/26/
14Fo
r pe
rson
al u
se o
nly.