Post on 06-Apr-2023
REVIEW
Physician weight loss advice and patient weight loss behaviorchange: a literature review and meta-analysis of survey dataSA Rose1,2, PS Poynter1, JW Anderson1, SM Noar3 and J Conigliaro4
Primary care providers (PCPs) can empower their patients to make health-promoting behavior changes. Many guidelinesrecommend that PCPs counsel overweight and obese patients about weight loss, yet few studies examine the impact ofprovider weight loss counseling on actual changes in patient behavior. We performed a systematic review and meta-analysis ofpublished studies of survey data examining provider weight loss counseling and its association with changes in patient weightloss behavior. We reviewed the published literature using keywords related to weight loss advice. We used meta-analytictechniques to compute and aggregate effect sizes for the meta-analysis. We also tested variables that had the potential tomoderate the responses. A total of 32 studies met criteria for the literature review. Of these, 12 were appropriate for themeta-analysis. Most studies demonstrated a positive effect of provider weight loss advice on patient weight loss behavior.In random effects meta-analysis, the overall mean weighted effect size for patient weight loss efforts was odds ratio (OR)¼ 3.85(95% confidence interval (CI) 2.71, 5.49; Po0.01), indicating a statistically significant impact of weight loss advice. There was nosignificant difference in the effectiveness of advice in studies using obese patients alone versus mixed samples (obese aloneOR¼ 3.44, 95% CI 2.37, 5.00; mixed sample OR¼ 3.98, 95% CI 2.53, 6.26, P¼ 0.63). PCP advice on weight loss appears to have asignificant impact on patient attempts to change behaviors related to their weight. Providers should address weight loss withtheir overweight and obese patients.
International Journal of Obesity (2013) 37, 118--128; doi:10.1038/ijo.2012.24; published online 27 March 2012
Keywords: physician; counseling; documentation; patient behavior; meta-analysis
INTRODUCTIONObesity is a major clinical and public health problem associated withan increased risk of morbidity and mortality and is related to manycomorbidities treated by primary care providers (PCPs).1,2 Intentionalweight loss can potentially mitigate this increased risk.3,4 Studies ontobacco and alcohol use counseling in the clinical settingdemonstrate that physicians can have a positive impact on changingpatient behaviors, even with brief counseling.5,6 Advocating similarmethods, multiple organizations recommend physician screeningand counseling for overweight and obesity, but most physicians donot appear to be advising their patients to lose weight.7 -- 12
Lack of physician counseling may be related to controversyregarding the effectiveness and realistic nature of weightcounseling. The United States Preventive Services Task Force(USPSTF) found good evidence to recommend that cliniciansscreen all patients for obesity through measurement of body massindex. However, they found fair to good evidence that only ‘high-intensity counseling and behavioral interventions’, defined asmore than one person-to-person session during at least the first 3months, were effective for treatment purposes, and that evidencefor less-intensive interventions that may be more feasible in theclinical setting was insufficient.8 The few studies that addresshealth care provider self-efficacy on the topic of weight find thatproviders face challenges with weight counseling, including a lackof tools, training, reimbursement, staffing and time, and relay alack of confidence in their own abilities and in the effectiveness ofweight loss strategies in general.10,13 -- 17
The purpose of this study was to examine whether PCPcounseling enhances patient engagement in weight reductionefforts. We aimed to systematically review the survey literatureregarding the effectiveness of PCP-provided advice related topatient weight loss behaviors, followed by a meta-analysisexamining the hypothesis that PCP counseling would enhancepatient participation in weight loss efforts.
MATERIALS AND METHODSLiterature review search strategyWe performed an extensive search related to the effect of provider weightloss advice on patient behavior. We conducted a review of publishedliterature in the English language using numerous keywords in combina-tion relating weight loss advice, including physician, counseling, over-weight, obese, weight loss, advice, diagnosis, documentation, recording,recognition, screening and cardiovascular risk. We performed searches inPubMed, Web of Science, PsycINFO, Cochrane Library, ERIC, Dissertationand Theses and WorldCat, and used PubMed MeSH headings andanalogous search strategies in other databases. No limitations were madeupfront; however, one relevant non-English language article was excludedfrom the coding. Searches included articles published through Novemberof 2011. An initial search using PubMed alone found five articles with theoutcome appropriate for inclusion in the meta-analysis,18 -- 22 with one laterremoved in favor of an article using the same data set with a larger, moregeneralizable population.21 The list of references for each of these articleswas searched for additional potential articles.
Received 18 August 2011; revised 17 January 2012; accepted 21 January 2012; published online 27 March 2012
1University of Kentucky College of Medicine, Department of General Internal Medicine and Geriatrics, Lexington, KY, USA; 2University of Kentucky College of Public Health,Department of Epidemiology, Lexington, KY, USA; 3University of North Carolina at Chapel Hill School of Journalism and Mass Communication, Chapel Hill, NC, USA and 4New YorkUniversity School of Medicine, Department of Medicine, New York, NY, USA. Correspondence: Dr SA Rose, University of Kentucky College of Medicine, Department of GeneralInternal Medicine and Geriatrics, 740 South Limestone, K507, Lexington, KY 40536-0284, USA.E-mail: sro226@email.uky.edu
International Journal of Obesity (2013) 37, 118 -- 128& 2013 Macmillan Publishers Limited All rights reserved 0307-0565/13
www.nature.com/ijo
Inclusion criteria included all of the following: patient impression of theefficacy of PCP weight loss advice or counseling on overweight status asnoted through survey or observational study data; evaluation of thediscussion of overweight status between a patient and his/her providerduring an office visit; inclusion of patients of all weights, provision or self-reported or chart-documented information about physician advice andweight history; and written in English. Because many studies did notspecify, studies involving any PCP, such as a physician, nurse practitioner orphysician assistant, were included in this review, and we use the termsprimary care provider or provider to mean any of these practitioners.
Article codingWe used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) criteria for reporting our meta-analysis.23 Articles werecoded on multiple dimensions by two coders including demographicssuch as age; clinical characteristics such as body mass index; surveycharacteristics such as method of survey participation (telephone,in-person or written), deliverer of weight loss advice (attending physician,resident physician or other provider), type of practice (internal medicine,family practice, pediatrics, other) and survey type and year, andmethodological characteristics including type of counseling receivedregarding weight loss and type of counseling received regarding physicalactivity or dietary change or decreasing comorbidities. Because some ofthe articles presented outcomes in adjusted form, we coded foradjustment. Satisfactory intercoder reliability was established with anaverage percent agreement across all categories of 86% (k¼ 0.75), withcoding differences resolved through discussion. To assess study quality, weused a standardized abstraction form by Zaza et al.,24 the purpose of whichis to assess the quality of the study across six categories, includingintervention and study descriptions, sampling, measurement, analysis,interpretation of results and other execution issues. Two coders calculatedthe quality of each article on each category, and differences were resolvedthrough discussion.
Effect size extraction and calculationThe odds ratio (OR), or the odds of an event occurring in an exposedpopulation divided by the odds of an event occurring in an unexposedpopulation, was used as the effect size indicator. We calculated effect sizesfrom data from each article using appropriate formulas. The primaryexposure was provider weight loss advice, and the primary outcome waspatient weight loss attempt. We used meta-analytic techniques tocompute and aggregate effect sizes for the 12 studies that met inclusioncriteria. We also carried out sensitivity analysis on variables believed tohave the most potential to moderate the responses.
Meta-analytic approachWe used Microsoft Excel to calculate totals, means and medians for ourdemographic and clinical characteristics. We used the ComprehensiveMeta-Analysis software program, version 2.0 (ComprehensiveMeta-Analysis, Englewood, NJ, USA), to calculate effect sizes and weightthem by sample size. We combined the effect sizes using standard randomeffects meta-analytic procedures and presented them with their 95%confidence intervals (CIs). We used the Q statistic and the I2 statistic to lookfor significant heterogeneity among the effect sizes. In testing categoricalvariables that had the potential to moderate the responses, we calculatedthem along with their 95% CIs, and compared them with one anotherusing the Qb and I2 statistics.25 Attempts were made to contact authorswhen additional raw data were needed.
RESULTSWe found 6041 nonduplicate articles for possible inclusion(Figure 1). A number of these articles had low relevancy to thestudy, whereas many were background information about theevolution of the obesity issue and guidelines or editorialsregarding recommendations for provider counseling and
treatment. There were also hundreds of articles dealing withinterventions that went beyond provider advice. A total of 32studies were appropriate for the literature review.18 -- 22,26 -- 52
Of these, 22 had the outcome appropriate for the meta-analysis.Of these 22, 10 were excluded for two reasons. One was use of asimilar data set and time point (Behavioral Risk Factor SurveillanceSurvey (BRFSS); n¼ 11 of 22, or National Health and NutritionExamination Survey (NHANES; n¼ 4 of 22)). Of those from thesame year, we used studies with a larger n and a moregeneralizable population. The other reason was inappropriatereporting of outcomes (n¼ 2) for calculation in the meta-analysis.In the literature review we included all studies that met inclusioncriteria in order to demonstrate the wealth of papers published onthis topic, as well as the consistency of findings through differenttime points and populations. In all, 12 studies were appropriate formeta-analysis. Our literature review assesses the effectiveness ofPCP-provided advice related to six different patient weight lossbehaviors, and our meta-analysis examines the correlationbetween PCP counseling and weight loss efforts.
Literature reviewThe literature review identified 32 reports with analyzabledata18 -- 22,26 -- 52 related to PCP advice regarding patient overweightand changes in behavior compared with patients not receivingadvice. These reports summarized advice from providersregarding weight loss, including the following exposures:(1) advice to lose weight by a provider; (2) specific PCP advicerelated to weight loss; (3) information or communication from thePCP regarding overweight; (4) and provider use of motivationalinterviewing to promote weight loss (Table 1). The majoroutcome focused on increased patient action toward weight loss,and specifically included: (1) patient engagement in weightloss efforts; (2) patient perception of weight loss advice, includingperceived success of provider weight loss advice, patient desirefor weight loss, and patient confidence in ability to lose weight;(3) patient stages of change, either for readiness to lose weight orfor readiness to change weight-related behavior; (4) patientweight maintenance attempt; (5) patient attempt at a specificbehavior change related to weight loss; and (6) actual patientweight loss (Table 1). Several studies contained more than oneoutcome.
(1) Patient engagement in weight loss efforts (Table 1,category 1)In all, 22 studies looked for an association between weight lossadvice and patient engagement in weight loss efforts (Table 1). Ofthe 22 studies, 11 used data from the BRFSS, 4 used NHANES data,1 used National Health Interview Survey (NHIS) data and 6 usedlocal study data (Table 1). Only one did not find a positiveassociation.29
The BRFSS is a standardized, nationwide telephone survey ofadults X18 years of age to which states can attach separatemodules asking questions focusing on specific comorbidities. Thesestudies focused on questions regarding weight management. A fewstudies evaluated subsets of the general US population, such aspatients with comorbidities, veterans, specific states or certain racesand ethnicities. All demonstrated a positive association between PCPweight loss advice and patient weight loss efforts (Table 1). NHANESis an annual national survey that assesses the health and nutritionalstatus of children and adults in the United States through in-personinterviews and physical examinations.53 All four studies20,26,45,47
using these data found positive associations between advice andweight loss efforts (Table 1).
(2) Patient perceptions of weight loss advice (Table 1, category 2)Eight studies evaluated the association between physician adviceand positive thoughts about weight loss by patients (Table 1). Sixstudies26,28,26,34,35,45,51 found some positive association between
Physician weight loss advice and patient changeSA Rose et al
119
International Journal of Obesity (2013) 118 -- 128& 2013 Macmillan Publishers Limited
PCP discussion of weight and patient confidence in the ability ordesire to lose weight (Table 1).
(3) Patient stage of change (Table 1, category 3)Four studies evaluated physician advice and patient stage ofchange, based on Prochaska’s transtheoretical model of the fivestages of change, which includes precontemplation, contempla-tion, preparation, action and maintenance (Table 1). This modelcan be used to determine a patient’s receptiveness to specificstrategies or actions. Despite differences in patient population anddefinition of positive stage of change, all but one52 found apositive association between physician advice and increased stageof readiness for weight loss (Table 1).
(4) Patient weight maintenance attempt (Table 1, category 4)All three studies assessing the correlation between PCP counsel-ing on weight loss or about overweight and weight lossmaintenance found a positive correlation between the two.20,27,47
Patient attempt at a specific behavior change related to weightloss (Table 1, category 5)
(5) Patient attempt at a specific behavior change related to weightloss (Table 1, category 5). In all, 11 studies examined the associationbetween provider weight loss advice and attempts by patients at
specific behavior changes related to weight. Of these 11, 8studies29,34,38,40,41,48,50,51 found a positive association betweenphysician weight loss advice and dietary change. All studies, exceptthose of Himmel et al.38 and Rodondi et al.,39 examined theassociation between weight loss advice and physical activity, butonly three found a positive association.28,48,51 Rodondi et al.39 founda positive association between physician weight loss advice andseeing a dietitian and setting a target weight, but not betweenadvice and reading self-help material to lose weight (Table 1).
(6) Actual weight loss by a patient (Table 1, category 6)A total of seven studies examined actual patient weight loss andits association to provider advice, and all found some positiveassociations between the two (Table 1). Singh et al.26 usedNHANES data and found a positive association between physiciandiagnosis of being overweight and intentional weight loss of X5%(prevalence ratio 1.87; 95% CI 1.61, 2.17). Singh et al.47 used thesame data set but looked specifically at patients with cardiovas-cular disease and central obesity, and found a positive associationbetween patient report of PCP recognition of overweight andintentional weight loss of X5% (OR 2.78, 95% CI 1.37, 5.63).
Himmel et al.38 surveyed overweight patients from a solophysician family practice in Northern Germany. Of those
Records after duplicates removed(n = 6041)
Records screened(n = 464)
Records excluded (n = 372)
Full-text articles assessedfor eligibility
(n = 87)
Inappropriate study type(n = 24)No link given betweenphysician recognition andpatient outcome (n=13)Advice provided forspecific comorbidity (n=5)No information oncomparison group (n=3)Abstract only (n=3)Foreign language (n=1)Thesis (n=4)Link between advice andoutcome inappropriatefor study (n=1)Inappropriately presentedoutcomes (n=1)Commentary (n=2)
Abstract published as complete study (n=1)Study re-reviewed and found appropriate for
literature review (n=1)
Studies included inquantitative synthesis
(meta-analysis)(n = 12 )
Studies included in qualitative synthesis (n = 32)
•
•
•
•
••••
•
•
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram of article search strategy.23
Physician weight loss advice and patient changeSA Rose et al
120
International Journal of Obesity (2013) 118 -- 128 & 2013 Macmillan Publishers Limited
Tabl
e1.
Studiesincluded
inliterature
review
andmeta-an
alysis
Ou
tco
me
cate
go
ryn
o.
Stu
dy
(au
tho
r)a
nd
pu
blic
ati
on
yea
r(R
efn
o.)
Surv
eyty
pe
Sam
ple
,n
Stu
dy
des
crip
tio
nEx
po
sure
(s)
Ou
tco
me
no
.(a
nd
spec
ific
act
ion
ifa
pp
lica
ble
),p
erce
nta
ge
an
dP
-va
lue
or
OR
(95%
CI)
for
pts
rece
ivin
gP
CP
ad
vice
vsn
ot
1Galuska
eta
l.18
BRFS
S19
96USOBad
ults,
n¼12
835
ComparisonofWLattempts
withan
dwithoutWLad
vice
1(1)79
%vs
57.6%
(OR2.79
,2.53
,3.08
)
1Meh
otra
eta
l.21
BRFS
S20
00USOBad
ultswith
arthritis,
n¼79
69ComparisonofWLattempts
withan
dwithoutWLad
vice
1(1)80
.5%
vs48
.9%
(adjusted
OR2.8,
2.5,
3.1)
1Abid
eta
l.30
BRFS
S19
94,19
96,19
98,20
00USOBad
ults,
n¼24
179
(year20
00)
Assessm
entofan
associationbetwee
nWL
advice
andWLattemptusingBRFS
S20
001
(1)OR2.8(2.5,3.2)
1Felix
eta
l.31
BRFS
S20
00an
d20
05USOBad
ults,20
00n¼55
25an
d20
05n¼11
569
ComparisonofBRFS
S20
00an
d20
05data
regardingWLco
unselin
gan
dptWL
attempt
1(1)OR20
00:2.9(2.4,3.6);OR20
05:3.9(3.2,4.7)
1Fo
ntaine
eta
l.32
BRFS
S20
05USOW
andOBad
ults
witharthritis,
n¼10
144
Assessm
entofPCPco
unselin
gonWLto
controlarthritissxsin
OBpts
1(1)OR3.75
(3.39,
4.15
)
1Fo
rdet
al.4
2BRFS
S20
00USAdultsofallwts
with
asthma,
n¼10
955
Assessm
entofPCPco
unselin
gonWLto
controlasthmamorbidity
1(1)82
.9%
vs63
.8%
(Po0.00
1)(OBpts
only)
1Wan
get
al.3
3BRFS
S20
00USOW
andOBad
ults,
n¼17
873
5non-VAusers
n¼33
91VA
users
Assessm
entofassociationbetwee
nPCP
wtco
ntrolad
vice
andptWLattempt
1(1)80
%vs
59%
1Th
ande
eta
l.22
Localwritten
ptquestionnaires
2006
USad
ultsofallwt
categories,
n¼25
6Assessm
entofPCPWLco
unselin
gpractices
1(1)OR18
.03(6.63,
49.04)
1Zap
kaet
al.4
4Lo
calwritten
survey
2005
USOW
andOBad
ults,
n¼81
3Evaluationofptwt,wtperceptionan
dparticipationin
WLpractices
1(1)Stratifie
dbyrace/ethnicity
WhiteOR1.18
(1.03,
1.35
)Black
OR2.72
(1.52,
4.84
)1,
2Po
stet
al.4
5NHANES
2005
--200
8USOW
andOBad
ults,
n¼54
74Correlationbetwee
nPCPreco
gnitionofpt
OW
andptWLattemptan
dbeh
avior
chan
gerelatedto
weightloss
3(1)OR2.84
(2.55,
3.18
)(2)Desireto
weighless:97
.4%
vs86
.1%
(Po0.00
1)
1,2,
5Kab
eer
eta
l.34
BRFS
San
dCVD-TI19
96USOW
andOBad
ultsin
Missouri,
n¼31
01Assessm
entofassociationbetwee
nPCP
WLad
vice
andptWLattempt
1(1)78
%vs
32%
(OR4.6;
2.8,
7.5)
(2)desireto
lose
wt:98
%vs
59%
(OR26
.9;7.6,
95.3)
(5)attemptat
dietary
chan
ge:
91%
vs89
%(OR
1.7;
2.8,
7.5)
(5)attemptat
PA:65
%vs
75%
(OR0.9;
0.5,
1.6)
1,2,
5Po
llak
eta
l.28
Localan
dpre-an
dpost-office
visitTelephonesurvey
andin-
personoralsurvey
atthetimeof
theoffice
visit
USOW
andOBfemale
adults,
n¼25
Exam
inationofassociationofMI
tech
niques
withptWLattempts
4(1)Significantlyco
rrelated
withMIad
heren
tbeh
avior
(Spearm
anassociation¼0.42
)(2)Confid
ence
inlosingwt:nosignificantassociation
(2)Readinessto
lose
wt:significantlyco
rrelated
withwt
discu
ssion(0.52),timespen
tonwt(0.42)
andMIspirit(0.63)
(5)Attem
ptat
dietary
chan
ge:
nosignificantassociation
(5)Chan
gein
exercise
programs:significantlyco
rrelated
with
empathy(0.50)
andMIspirit(0.47)
1,2,
6Singhan
dLo
pez-
Jimen
ez26
NHANES
1999
--200
4USOW
adults,
n¼87
67Assessm
entofassociationbetwee
nphysicianad
vice
aboutOW
andptWL
attempt
1(1)PrR1.49
(1.61,
2.17
)(2)Desireto
weighless:PrR1.28
(1.25,
1.31
)(6)intentional
weightch
angeofX5%
:PrR1.87
(1.61,
2.17
)1,
3Huan
get
al.1
9Oralexitinterview
September
2001
--Jan
uary20
02USOW
adults,
n¼21
0EvaluationofPCPan
dptperceptionofWL
counselin
g,ptperceptionofPCPWL
reco
mmen
dationsan
dcu
rren
tptWL
activities
1,2
(1)OR7.25
(3.05,
17.21)
(3)increasedread
inessforWL:
91%
vs64
%P¼0.00
1
1,4
Naw
azet
al.2
7BRFS
S19
94USOW
adults,
n¼32
0Assessm
entofWLco
unselin
gan
dWL
attemptan
dmaintenan
ce1
(1)89
%vs
52%
(P¼0.00
1)(4)WLorwtmaintenan
ceattempt:95
%vs
81%
(P¼0.00
2)1,
4,5
Kan
tan
dMiner
20
NHANES
1999
--200
0an
d20
01--
2002
USOW
teen
agers
(16--1
9yearsofag
e),
n¼71
6
Assessm
entofprofessional
counselin
gab
outOW,WLattemptan
dspecific
subject
dietary
andPA
barriers
3(1)OR2.7(1.7,4.4)
(4)OR2.3(1.5,3.6)
(5)attemptat
dietary
chan
ge:
OR1.9(0.9,4.3)
(5)attemptat
physical
activity:OR1.2(0.5,2.8)
Physician weight loss advice and patient changeSA Rose et al
121
International Journal of Obesity (2013) 118 -- 128& 2013 Macmillan Publishers Limited
Tabl
e1
(Continued
)
Ou
tco
me
cate
go
ryn
o.
Stu
dy
(au
tho
r)a
nd
pu
blic
ati
on
yea
r(R
efn
o.)
Surv
eyty
pe
Sam
ple
,n
Stu
dy
des
crip
tio
nEx
po
sure
(s)
Ou
tco
me
no
.(a
nd
spec
ific
act
ion
ifa
pp
lica
ble
),p
erce
nta
ge
an
dP
-va
lue
or
OR
(95%
CI)
for
pts
rece
ivin
gP
CP
ad
vice
vsn
ot
1,4,
6Singh
eta
l.47
NHANES
1999
--200
4USOW
adultswith
cardiovasculardisease
andcentral
obesity,
n¼90
7
Correlationbetwee
nPCPreco
gnitionof
ptOW
andptWLattempt,weight
maintenan
cean
dactual
WL
1(1)OR2.31
(1.33,
4.00
)(4)OR2.49
(1.72,
3.62
)(6)intentional
WLofX5%
:OR2.78
(1.37,
5.63
)
1,5
Bish
eta
l.41
BRFS
S20
00USad
ultsofallwts,
n¼16
418
7ComparisonofWLattempts
withan
dwithoutWLad
vice
1(1)Women
OR6.24
(5.5,7.09
)Men
OR10
.13(8.53,
12.03)
(5)attemptat
dietary
chan
ge:
Women
OR1.25
(1.15,
1.37
)Men
OR1.25
(1.11,
1.41
)(5)attemptat
PA:
Women
OR0.94
(0.86,
1.04
)Men
OR0.92
(0.81,
1.04
)1,
5Dorsey
and
Songer
48
NHIS
2006
USOW
andOBad
ults,
N¼20
05(n¼56
3with
prediabetes
and
n¼14
42withdiabetes)
ComparisonofWLattempts
withan
dwithoutWLad
vice
1,2
(1)Pred
iabetes
men¼notcalc
Pred
iabetes
women¼notcalc
Diabetes
men¼9.1(5.4,15
.4)
Diabetes
women¼4.1(2.5,6.6)
(5)Dietary
chan
ge:
Pred
iabetes
men¼1.7(0.8,3.9)
Pred
iabetes
women¼2.8(1.5,5.3)
Diabetes
men¼11
.3(6.7,19
.0)
Diabetes
women¼5.7(3.7,8.6)
Physical
activity
increase:
Pred
iabetes
men¼3.6(1.5,8.7)
Pred
iabetes
women¼3.0(1.5,5.9)
Diabetes
men¼4.6(2.9,7.4)
Diabetes
women¼2.8(1.8,4.4)
1,5
Sciaman
na
eta
l.40
BRFS
S19
96USad
ultsofallwt
categories,
n¼10
187
Assessm
entofassociationbetwee
nPCP
advice
andptWLattempt
1(1)
Po0.00
1(5)attemptat
dietary
chan
ge:
61%
vs48
%,
Po0.01
(BMI25
--27kg
m--2)
(5)attemptat
PA:notsignificantforan
ywtcategory
1,5
Saelen
set
al.2
9Lo
calwritten
questionnaires
USnorm
al,OW
andOB
adolescen
ts(12--1
8years),
n¼11
0
Search
forassociationbetwee
nPCPwt-
relatedco
unselin
gan
dad
olescen
twt
controlefforts
2(1)
P¼0.28
(5)eatingless
food/few
ersnacks:86
%vs
71%
(5)eatingmore
fruits/ve
getab
les:86
%vs
80%
(5)gettingmore
exercise:87
%vs
89%
(5)eatingfoodslower
infat:83
%vs
85%
(5)co
untingcalories/w
rite
downfoods:83
%vs
55%
(5)takingmed
ication/dietpills:67
%vs
17%
(5)skippingmeals:10
0%vs
55%
(5)vo
miting/purging:NA%
vs10
0%1,
5,6
Him
mel
eta
l.38
Localwritten
survey
German
yOW
andOB
adults,
n¼74
0Assessm
entofreco
gnition,co
unselin
gan
dtreatm
entofOW
from
PCPan
dpt
view
point
2,3
(1)40
%triedto
lose
weightwhorankedad
vice
asgood,
compared
with8%
whorankeditas
averag
eorpoor
(5)d
ietary
treatm
ent:54
%triedto
chan
getheirdietwhoranked
advice
asgood,co
mpared
with42
%whorankeditas
averag
e,an
d4%
whorankeditas
poor
(6)28
%reported
actual
WLof4
3kg
whorankedad
vice
asgood,co
mpared
with50
%whorankeditas
averag
e,an
d12
%whorankeditas
poor
2Evan
s35
Localwritten
survey
UKOBad
ultpts
whohad
lost
atleast45
kg,n¼37
2EvaluationofsuccessofdoctorWLad
vice
inen
couragingpts
tolose
wt
1(2)Advice
reported
tobesuccessful:6%
;Rep
orted
tobeindirectlyeffective:
23%
2Jay
eta
l.49
Localin-personsurvey
USOBad
ultpts,
n¼13
7Correlationbetwee
nqualityof
physicianWLco
unselin
gan
dpt
motivationforWLan
d
1,2
(2)OR0.77
0(0.204
,2.90
2)
Physician weight loss advice and patient changeSA Rose et al
122
International Journal of Obesity (2013) 118 -- 128 & 2013 Macmillan Publishers Limited
Tabl
e1
(Continued
)
Ou
tco
me
cate
go
ryn
o.
Stu
dy
(au
tho
r)a
nd
pu
blic
ati
on
yea
r(R
efn
o.)
Surv
eyty
pe
Sam
ple
,n
Stu
dy
des
crip
tio
nEx
po
sure
(s)
Ou
tco
me
no
.(a
nd
spec
ific
act
ion
ifa
pp
lica
ble
),p
erce
nta
ge
an
dP
-va
lue
or
OR
(95%
CI)
for
pts
rece
ivin
gP
CP
ad
vice
vsn
ot
2,5
Cox
eta
l.50
Localin-personsurvey
(Project
CHAT---CommunicatingHealth---
AnalyzingTalk)
USOW
andOBad
ults,
n¼46
1Correlationbetwee
nPCPco
unselin
gtech
niques
andptmotivationan
dco
nfid
ence
toch
angeWLbeh
aviors
4(2)Motivation
PCPMIco
nsisten
tbeh
aviorOR0.8(0.3,1.8)
PCPMIspiritOR0.9(0.4,2.0)
Confid
ence
PCPMIco
nsisten
tbeh
aviorOR1.9(0.8,4.4)
PCPMIspiritOR0.8(0.3,1.8)
(5)Nutrition
Motivation
PCPMIco
nsisten
tbeh
aviorOR1.0(0.4,2.6)
PCPMIspiritOR0.8(0.3,1.9)
Confid
ence
PCPMIco
nsisten
tbeh
aviorOR2.6(1.2,5.7)
PCPMIspiritOR0.6(0.3,1.5)
Exercise
Motivation
PCPMIco
nsisten
tbeh
aviorOR0.9(0.4,2.1)
PCPMIspiritOR1.2(0.5,2.5)
Confid
ence
PCPMIco
nsisten
tbeh
aviorOR1.8(0.8,4.0)
PCPMIspiritOR0.8(0.3,1.8)
2,5,
6Alexander
eta
l.51
Localin-personsurvey
(Project
CHAT---Communicating
Health---A
nalyzingTalk)
USOW
andOBad
ults,
n¼46
1Correlationbetwee
nPCPuse
of5A
’sco
unselin
gtech
niques
andptco
nfid
ence
tolose
weight,ptdietary
andPA
chan
ge
andactual
ptWL
1,2,
3(2)Po
sitive
associationbetwee
npatientco
nfid
ence
tolose
weightan
dPCPad
vise
(P¼0.05
),an
dassess
(P¼0.05
)(5)Chan
gediet:positive
associationbetwee
ndietch
angean
dPCPassist
(P¼0.00
1),an
darrange(P¼0.00
1)Chan
gePA
:nosignificantassociationbetwee
nPCPuse
of5A
’san
dPA
chan
ge(6)Po
sitive
associationbetwee
npatientweight
loss
andPCPArran
ge(P¼0.05
)3
Fallo
net
al.3
6Lo
caltelephone
survey
2002
USad
ultsofallwts,
n¼57
2Evaluationofrate
andco
rrelates
of
physicianWLad
vice
andofthe
associationbetwee
nad
vice
andptstag
eofch
ange
1,2
(3)Stag
eofch
angeforen
gag
emen
tin
WLefforts:OR1.86
(1.13,
3.06
)(3)stag
eofch
angeformoderateto
vigorousPA
:ORnotgiven
(P¼0.07
)(3)stag
eofch
angeforfruitan
dve
getab
leco
nsumption:
OR2.25
(1.4,3.62
)3
Jay
eta
l.52
Localin-personsurvey
USOBad
ultpts,
N¼15
2Correlationbetwee
nPCPuse
of5A
’san
dad
vancedobesityco
unselin
gpractices
andptstag
eofch
ange
Nosignificantrelationship
foundbetwee
npercentofpts
inactionofmaintenan
cestag
ean
dhigher
PCPuse
of5A
’s(s.d.
b¼0.11
,R2¼0.1%
)orad
vancedobesityco
unselin
gbyPCP(s.d.
b¼0.09
,R2¼0.3%
)3
Wee
eta
l.37
Localtelephonesurvey
Nove
mber
2001
--June20
09USad
ultpts
ofallwt
categories,
n¼36
5Evaluationofaassociationbetwee
nPCP
motivationofpts
andptim
provemen
tin
WL,
diet,an
dexercise
2(3)Stag
eofch
angeforread
inessto
lose
wt:53
%vs
36%,
Po0.05
5,6
Rondoni
eta
l.39
Localin-personinterview
and
written
questionnaire
emailedto
homeat
1year
Switzerlan
dOW
andOB
adults,
n¼52
3EvaluationofPCPco
unselin
gofOW
and
OBpts
duringclinicvisits
andeffect
onpt
wtan
dwtco
ntrolefforts1year
post
visit
1,2
(5)Se
ttingatarget
wt:56
%vs
36%
(Po0.00
1)(5)modifyingdietto
lose
wt:47
%vs
38%,
P¼0.09
(5)visitingadietitian
:23
%vs
10%,
P¼0.00
1(5)read
ingself-helpmaterial:24
%vs
34%,
P¼0.09
5)X1pt
beh
aviorto
controlwt:82
%vs
62%
Po0.00
1(6)meanwtch
ange:�1kg
(s.d.5.0)
vs+0.3(s.d.5.0)
P¼0.02
6Levy
and
Williamson43
Telephonesurvey
1980
--198
1USOW
adults,
n¼56
Determinationoffactors
relatedto
WL
andco
nsidered
importan
tbythose
who
had
lost
wt
3(6)
P¼0.01
Physician weight loss advice and patient changeSA Rose et al
123
International Journal of Obesity (2013) 118 -- 128& 2013 Macmillan Publishers Limited
who ranked their communication with their physician as ‘good’,28% had lost at least 3 kg as compared with 20% who did not rankcommunication as good. Of those ranking communication as‘average’, 50% reported weight loss as compared with 55% whodid not rank it as average, and 12% ranking communication as‘poor’ reported losing weight when compared with 25% whodid not rank it as poor. Rodondi et al.39 conducted in-personinterviews and questionnaires with patients from Swiss university-affiliated primary care clinics. At the end of 1 year, questionnairecompleters (n¼ 407, 78%) who had received any form ofcounseling reported an average weight loss of 1 kg as comparedwith a gain of 0.3 kg (P¼ 0.002) in those who had not receivedcounseling. After adjustment, counseling with each additionalweight loss strategy was associated with a mean weight loss of0.2 kg (95% CI 0.03, 0.4, P¼ 0.02). Levy and Williamson43 surveyed56 obese adult patients from a midwestern university familypractice clinic, and found that those who lost weight (n¼ 21),defined as a loss of at least 2.3 kg over a 5-year period, reportedmore physician attention to their obesity.
Alexander et al.51conducted audio recordings of 40 physiciansand 461 of their overweight or obese patients looking for examplesof use of the 5A’s (Ask, Advise, Assess, Assist and Arrange) in weightloss counseling, and then conducted a post-visit survey andanthropometric measurement. They found a positive associationbetween 3-month weight loss and physician arrangement forfollow-up with physician or nutritionist regarding weight loss(arranged: mean (pre)¼ 101.4 kg (s.e. 3.49) versus mean(post)¼ 99.9 kg (s.e. 3.51); did not arrange: mean (pre)¼ 91.1 kg(s.e. 0.76) versus mean (post)¼ 91.2 kg (s.e. 0.76), P¼ 0.05). Pollaket al.46 used the same population as Alexander et al.51 to assess thedifference in patient weight loss between patients receivingphysician weight loss counseling versus not, and physician use ofmotivational interviewing techniques. They found a statisticallysignificant weight loss in those patients whose physicians exhibitedmotivational interviewing spirit (�1.6 kg (�2.9, �0.3), P¼ 0.02) andexhibited reflection or understanding of the patient’s statements(�0.9 (�1.8. �0.1), P¼ 0.03), but no statistically significant weightchange for other physician behaviors.
Meta-analysisStudy characteristics. The 12 studies included a total of 207 226individuals, and were published between 1999 and 2011 (Tables 1and 2). All studies were conducted in the United States. Excludingtwo articles that did not give demographic information ongender29,31 and one that did not give demographic informationon race,31 60% of patients were female and 37% were white. Twostudies evaluated a pediatric population.20,29 There were 39 310(19%) patients who reported receiving advice to lose weightor being told that they were overweight. Two studies18,31 lookedonly at an obese population, whereas the rest includedoverweight patients; two studies22,41 included normal-weightpatients in their sample. Eight studies used national and fourused local survey data. One study did not provide the OR andinstead gave a prevalence ratio.26 Satisfactory quality wasestablished with an average percent agreement of 89% acrossall categories (k¼ 0.77).
Efficacy of intervention: weight loss attempt. All studies exceptone29 demonstrated a positive effect of PCP advice on patientengagement in weight loss efforts (Figure 2). The overall randomeffects mean weighted effect size for weight loss efforts includingall studies was OR¼ 3.85 (95% CI 2.71, 5.49; Z¼ 7.47, Po0.01),indicating a statistically significant impact of weight loss advice onefforts at weight loss. Removal of studies with different popula-tions such as children20,29 and removal of the study with theoutlying OR size for the outcome22 did not change the resultssignificantly.
Tabl
e1
(Continued
)
Ou
tco
me
cate
go
ryn
o.
Stu
dy
(au
tho
r)a
nd
pu
blic
ati
on
yea
r(R
efn
o.)
Surv
eyty
pe
Sam
ple
,n
Stu
dy
des
crip
tio
nEx
po
sure
(s)
Ou
tco
me
no
.(a
nd
spec
ific
act
ion
ifa
pp
lica
ble
),p
erce
nta
ge
an
dP
-va
lue
or
OR
(95%
CI)
for
pts
rece
ivin
gP
CP
ad
vice
vsn
ot
6Po
llak
eta
l.46
Localin-personsurvey
(Project
CHAT---CommunicatingHealth---
AnalyzingTalk)
USOW
andOBad
ults,
n¼42
6Difference
inactual
ptWLove
r3months
betwee
npts
receivingPCPWLco
unselin
gan
duse
ofMIco
unselin
gtech
niques
1,4
(6)WLco
unselin
g:+0.1kg
(�0.7,
0.8;
P¼0.84
)MIco
unselin
gtech
niques
MIspirit:�1.6kg
(�2.9,�0.3;
P¼0.02
)Refl
ection:�0.9kg
(�1.8,�0.1;
P¼0.03
)Open
questions:+0.1kg
(�0.8,
0.9;
P¼0.86
)Em
pathy:�1.0kg
(�2.8,
0.8;
P¼0.26
MIco
nsisten
tbeh
aviors:�1.1kg
(�2.3,
0.1;
P¼0.07
)
Abbreviations:5A
’s,Assess,Advise,A
gree,Assistan
dArran
ge;BMI,bodymassindex;B
RFS
S,Beh
avioralR
iskFactorSu
rveillance
System
;95%
CI,95
%co
nfid
ence
interval;C
VD-TI,CardiovascularDisease
Targeted
Initiative
;MI,motivationalinterviewing;n
,number;N
HANES
,NationalHealthan
dNutritionExam
inationSu
rvey
;NHIS,N
ationalHealthInterview
Survey
;OB,o
bese;OR,o
ddsratio;O
W,o
verw
eight;PA
,physical
activity;PCP,primarycare
provider;PrR,p
revalence
ratio;p
t,patient,Ref,referen
ce;sxs,sym
ptoms;VA
,Veteran
sAdministration;W
L,weightloss;w
t(s),w
eight(s).Exp
osures:(1)P
CPweightloss
advice,either
for
gen
eral
healthorto
improve
aspecificco
morbidity;
(2)specificPCPad
vice
relatedto
weightloss;(3)
inform
ationorco
mmunicationfrom
thePCPregardingove
rweight;(4)PCPuse
ofMIto
promote
weight
loss.O
utcomes:(1)
Patien
ten
gag
emen
tin
weightloss
efforts;(2)patientperceptionofweightloss
advice,includingperceived
successofprovider
weightloss
advice,p
atientdesireforweightloss
andpatient
confid
ence
inab
ility
tolose
weight;(3)patientstag
esofch
ange,
either
forread
inessto
lose
weightorforread
inessto
chan
geweight-relatedbeh
avior;(4)patientweightmaintenan
ceattempt;(5)patient
attemptat
aspecificbeh
aviorch
angerelatedto
weightloss;(6)actual
patientweightloss.
Physician weight loss advice and patient changeSA Rose et al
124
International Journal of Obesity (2013) 118 -- 128 & 2013 Macmillan Publishers Limited
Tabl
e2.
Characteristicsofstudiesincluded
inthemeta-an
alysisevaluatingassociationbetwee
nprovider
advice
andpatientweightloss
attempt
Stu
dy
(ref
eren
cen
o.)
Sam
ple
Wei
gh
tca
teg
ori
esin
clu
ded
Ho
wsu
rvey
del
iver
edSu
rvey
typ
e(y
ear)
Effe
ctsi
ze(O
R,
95%
CI)
Bish
eta
l.41
16418
7ad
ultsfrom
all50
states,
DC,an
dPu
erto
Rico
Norm
al-w
eight,ove
rweightan
dobesead
ults
Telephone
BRFS
S,national
(200
0)Male¼10
.13(8.53,
12.03)
Z¼26
.40
Po0.01
Female¼6.24
(5.5,7.09
)Z¼28
.26
Po0.01
Dorsey
and
Songer
48
Nationwidesample
of20
05ad
ults
(n¼56
3withprediabetes
and
n¼14
42withdiabetes)
Ove
rweightan
dobesead
ultswith
prediabetes
anddiabetes
In-person
NHIS,national
(200
6)aPred
iabetes
M¼notcalc
aPred
iabetes
F¼notcalc
Diabetes
M¼9.1(5.4,15
.4)
Z¼8.26
Po0.01
Diabetes
F¼4.1(2.5,6.7)
Z¼5.70
Po0.01
Felix
eta
l.31
Nationwidesample
of11
569ad
ults
Obesead
ults
Telephone
BRFS
S,national
(200
5)4.17
(3.74,
4.65
)Z¼25
.70
Po0.01
Galuska
eta
l.18
Nationwidesample
of12
835ad
ults
Obesead
ults
Telephone
BRFS
S,national
(199
6)2.85
(2.63,
3.09
)Z¼25
.47
Po0.01
Huan
get
al.1
921
0ad
ultsfrom
twoprimarycare
clinicsaffiliatedwithLo
uisianaState
University
Ove
rweightan
dobesead
ults
In-person
Local(200
1--2
002)
6.17
(2.70,
14.10)
Z¼4.32
Po0.01
Kan
tan
dMiner
20
716teen
sag
ed16
--19yearsfrom
all
50states
At-risk
foroverw
eightan
dove
rweightad
olescen
ts(BMIX
85th
percentile)
In-person
NHANES
,national
(199
9--2
000an
d20
01--2
002)
2.34
(1.74,
3.16
)Z¼5.56
Po0.01
Naw
azet
al.2
732
0(25%
oftotalsample)ad
ults
from
Connecticu
tOve
rweightad
ults
Telephone
BRFS
S,national
(state
of
Connecticu
tonly)(199
4)7.51
(3.71,
15.21)
Z¼5.60
Po0.01
Post
eta
l.45
Nationwidesample
of54
74ad
ults
Ove
rweightan
dobesead
ults
In-person
NHANES
,national
(200
5--2
008)
2.84
(2.55,
3.18
)Z¼18
.53
Po0.01
Saelen
set
al.2
974
adolescen
tsag
ed12
--18years
from
RhodeIslandan
dCalifo
rnia
Ove
rweightad
olescen
ts(BMIX85
thpercentile)
In-person
Local(200
2)1.58
(0.68,
3.70
)Z¼1.05
P¼0.29
Singhan
dLo
pez-
Jimen
ez26
Nationwidesample
of87
67ad
ults
Ove
rweightad
ults
In-person
NHANES
,national
(199
9--2
004)
PR¼1.49
(1.42,
1.57
)Z¼15
.57
Po0.01
Than
de
eta
l.22
256ad
ultpatients
from
aninternal
med
icineclinic
atColumbia
University,New
York
Norm
al-w
eight,ove
rweightan
dobesead
ults
In-person
BRFS
S-based
,local(200
3)18
.03(6.63,
49.04)
Z¼5.67
Po0.01
Zap
kaet
al.4
481
3ad
ulthospital
employees
from
ahealthcare
system
incentral
Massach
usetts
Ove
rweightan
dobesead
ults
In-person
Local(200
5)White1.18
(1.03,
1.35
)Z¼2.40
P¼0.02
African
American
2.72
(1.53,
4.84
)Z¼3.40
P¼0.00
1
Abbreviations:BMI,bodymassindex;B
RFS
S,Beh
avioralR
iskFactorSu
rveillance
System
;CI,co
nfid
ence
interval;F,fem
ale;
M,m
ale;
NHANES
,National
Healthan
dNutritionExam
inationSu
rvey
;NHIS,N
ational
HealthInterview
Survey
;OR,oddsratio.aNotcalculatedbecau
seofsm
allsample
sizes.
Physician weight loss advice and patient changeSA Rose et al
125
International Journal of Obesity (2013) 118 -- 128& 2013 Macmillan Publishers Limited
Heterogeneity and intervention moderators. We found significantheterogeneity of the effect sizes among the studies (Q¼ 1124.09,I2¼ 98.76, Po0.01), and therefore looked at between-studymoderators that could impact the association of advice andengagement in weight loss efforts. Using random effects analysis,we found no statistically significant difference in the effectivenessof advice in studies using obese patients alone versus mixedsamples (OR¼ 3.44, 95% CI 2.37, 5.00 versus mixed sampleOR¼ 3.98, 95% CI 2.53, 6.26, P¼ 0.63), studies using nationalversus local survey data (OR¼ 4.19, 95% CI 2.75, 6.38 versus 3.33,95% CI 1.36, 8.15, P¼ 0.65) or from data published before the 1998NHLBI (National Heart, Lung, and Blood Institute) guidelines on theidentification, evaluation and treatment of overweight and obesitycompared with after publication (OR¼ 4.33, 95% CI 1.69, 11.09versus 3.81, 95% CI 2.48, 5.85, P¼ 0.81). We did find a significantdifference in patient report of weight loss attempt if advised bytheir PCP in adults versus children (OR¼ 4.24, 95% CI 2.90, 6.22versus 2.24, 1.69, 2.98, P¼ 0.01), and in patients who weresurveyed by telephone compared with in-person survey(OR¼ 5.50, 95% CI 3.46, 8.72 versus 3.00, 95% CI 2.15, 4.17,P¼ 0.04).
DISCUSSIONIn this literature review and meta-analysis we assessed studiesaddressing communication between PCPs and patients regardingweight. Our literature review consistently demonstrated thatphysician advice can have a positive effect on patient actionregarding weight loss. Our meta-analysis indicated that providerweight loss advice has a statistically significant impact on patientattempt to lose weight, suggesting that brief counseling by theprovider may play a role in patient weight management.
In our meta-analysis, comparison of data before and after the1998 NHLBI guidelines did not find a significantly positive impactof guideline publication on the association between provideradvice and patient weight loss efforts. Caution should be takenwith these results as there were only two studies that used databefore the 1998 guidelines, both of them larger studies usingnational or state-based data as opposed to smaller local studies.Caution should also be taken with our finding of a significantdifference in patient report of weight loss attempt if advised bytheir PCP in adults versus children (OR¼ 4.24, 95% CI 2.90, 6.22versus 2.24, 1.69, 2.98, P¼ 0.01), because of the small number andtotal n of studies including children. Further study is needed tobetter evaluate and identify a true significance in the differencewe found in patient report of weight loss attempt between
patients who were surveyed by telephone compared with in-person (OR¼ 5.50, 95% CI 3.46, 8.72 versus 3.00, 95% CI 2.15, 4.17,P¼ 0.04). In-person surveys included both large national studies(NHANES and NHIS) that are conducted in person and smallerlocal surveys, whereas telephone surveys consisted solely of largenational surveys, and these two may not be easily compared.
We found no significant difference in the correlation betweenprovider advice and weight loss attempt in obese-only versusmixed-weight samples in our meta-analysis. Again, there were asmall number of obese-only studies, requiring caution withinterpretation. Rodondi et al.39 and Sciammana et al.40 bothreport that obese patients, especially those with an obesity-relatedcomorbidity, are more likely to receive weight loss advice fromtheir physicians compared with overweight patients. However,Sciammana et al.40 found that patients in lower overweight bodymass index categories were more likely to act on such advicecompared with obese patients with a body mass indexof X30 kg m -- 2. Bish et al.41 had similar findings to those ofSciammana et al.40 and advise caution with these findings, aspatients may have lost weight since receiving physician advice, ormay have sought assistance with their weight from a medicalprofessional because of feelings about their weight, regardless oftheir actual weight.39
Only studies assessing the correlation between provider weightloss advice and specific behavior change attempt related tophysical activity demonstrated a negative correlation. This may berelated to providers’ overall discomfort and limited knowledge ofthese topics.54,55
Importantly, we found that all studies that looked at theassociation between provider weight loss advice and actualpatient weight loss found a positive association between the two.This positive association was found in studies of different sizes andpopulations. Furthermore, it was demonstrated in populationswho received both advanced counseling and simple PCPrecognition or diagnosis of overweight and obesity.
LimitationsThis study contains several limitations. One is the small number ofstudies included in the meta-analysis. Because of repeat use ofsimilar data sources or inappropriate outcome data, we wereunable to use many potential studies. However, reliability isdemonstrated by the relative consistency of the effect sizes andpositive direction in nearly all studies. The meta-analysis containsstudies of differing sizes and potentially of differing quality,possibly affecting outcomes. This is demonstrated further in themoderator analysis of national versus local studies, whichessentially separates out studies with the smallest and largestsample sizes, but from which we found no significant difference.Meta-analysis calculations attempt to take into account study sizewhen calculating effect size but cannot discern quality. However,all were published in peer-reviewed publications, many used well-validated surveys and techniques for obtaining subjects and ourquality calculation revealed adequate agreement.
Inability to obtain raw data led to use of studies with bothadjusted and unadjusted data. However, adjusted ORs will tend tobe smaller, underestimating the effect of the exposure rather thanoverestimating it. The use of the prevalence ratio instead of theOR to display effect size in one study26 tends to underestimatethe effect size compared with the OR, giving little concern thatthis study would cause overestimation of the effect of theexposure on the outcome. As groups of studies were divided intovarious categories, the number of studies in each group (k)became small, which can lead to reduction in statistical power todetect significant differences, particularly in the case of randomeffects analyses.25
This study evaluates the effect of PCP advice on behaviorchange, potentially considered an intermediate outcome, whereas
Study name Odds ratio and 95% CI
Bish et al. (Male)Bish et al. (Female)Dorsey et al. (Diabetic Male)Dorsey et al. (Diabetic Female)Felix et al.Galuska et al.Huang et al.Kant et al.Nawaz et al.Post et al.Saelens et al.Singh et al.Thande et al.Zapka et al (White)Zapka et al (African-American)
0.01 0.1 1 10 100
No provider advice Provider advice
Total
Figure 2. Odds ratios for the effect of provider advice on patientweight loss attempt for each study and overall in the meta-analysis.
Physician weight loss advice and patient changeSA Rose et al
126
International Journal of Obesity (2013) 118 -- 128 & 2013 Macmillan Publishers Limited
only five of seven studies evaluating the outcome of actual weightloss used independent data for meta-analysis. Because of thissmall number, the heterogeneity of these studies and thedifference in their outcomes, meta-analysis of these studies wouldnot provide meaningful data. However, examining patientbehavior change in the context of provider advice is an outcomeworthy of evaluation and the overwhelmingly positive outcomesshould encourage more providers to talk to their patients aboutweight.
ConclusionsThe rapidly growing prevalence of obesity and its associated riskshave increased the importance of the role that the PCP can play.This study assesses the importance of weight loss advice by thePCP. We found that provider weight loss advice positively affectspatient weight loss behavior in all outcomes addressed. Morestudies and interventions are needed to clarify the role ofproviders in this interaction and to overcome barriers to providerself-efficacy regarding counseling during the clinic visit. Policy-makers should recognize the provider’s need for additional timeand training in advising their patients as well as improvements inreimbursement and provision of tools such as dietitians. Perhapsmost importantly, PCPs need to recognize their influence when itcomes to advising their patients about their weight.
CONFLICT OF INTERESTDr Anderson currently has research support from Health Management Resources(HMR), and is a consultant to HMR.
ACKNOWLEDGEMENTSWe thank James Meigs and Elisabeth Watkins for review and editing of thismanuscript. We thank Heather Bush and Daisuke Sugimoto for assistance withbiostatistical analysis and editing.
REFERENCES1 Bray G. Risks of obesity. Endocrinol Metab Clin North Am 2003; 32: 787 -- 804.2 Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific excess deaths
associated with underweight, overweight, and obesity. JAMA 2007; 298:2028 -- 2037.
3 Will JC, Williamson DF, Ford ES, Calle EE, Thun MJ. Intentional weight loss and13-year diabetes incidence in overweight adults. Am J Public Health 2002; 92:1245 -- 1248.
4 Williamson DF, Pamuk E, Thun M, Flanders D, Byers T, Clark H. Prospective study ofintentional weight loss and mortality in never-smoking overweight US whitewomen aged 40 -- 64 years. Am J Epidemiol 1995; 141: 1128 -- 1141.
5 Smith P, Sellick S, Brink P, Edwardson A. Brief smoking cessation interventions byfamily physicians in northwestern Ontario rural hospitals. Can J Rural Med 2009;14: 47 -- 53.
6 Fiore M, Jaen C, Baker T, Bailey WC, Benowitz NL, Curry SJ et al. Treating TobaccoUse and Dependence: 2008 Update.. Publisher: United States Department of Healthand Human Services, Public Health Service: Rockville, MD, 2008. Available at:www.ncbi.nlm.nih.gov/books/bv.fcgi?rid¼hstat2.chapter.28163.
7 HEDIS 2009 Public Comment, 2008. (Accessed 30 July 2008, at http://www.ncqa.org/tabid/661/Default.aspx ).
8 US Preventive Services Task Force. Screening for obesity in adults: recommenda-tions and rationale. Ann Int Med 2003; 139: 930 -- 932.
9 Executive summary of the clinical guidelines on the identification, evaluation, andtreatment of overweight and obesity in adults, 1998. (Accessed 29 May 2007, athttp://www.nhlbi.nih.gov/guidelines/obesity/ob_gdlns.pdf ).
10 Brotons C, Bjorkelund C, Bulc M, Ciurana R, Godycki-Cwirko M, Jurgova E et al.Prevention and health promotion in clinical practice: the views of generalpractitioners in Europe. Prev Med 2005; 40: 595 -- 601.
11 Screening and interventions to prevent obesity in adults, 2003. (Accessed 14 May2008, at http://www.ahrq.gov/clinic/USpstf/uspsobes.htm ).
12 Body mass index (BMI) assessment for adults (BAA), 2008. (Accessed 25 April2008, at http://www.ncqa.org/Portals/0/PublicComment/HEDIS2009/BAA_Spec-s_and_Workup_PDF.pdf ).
13 Jackson J, Doescher M, Saver B, Hart L. Trends in professional advice to loseweight among obese adults. J Gen Intern Med 2005; 20: 814 -- 818.
14 Epstein L, Ogden J. A qualitative study of GPs’ views of treating obesity. Br J GenPract 2005; 55: 750 -- 754.
15 Whittemore R, Melkus G, Wagner J, Dziura J, Northrup V, Grey M. Translatingthe diabetes prevention program to primary care: a pilot study. Nurs Res 2009; 58:2 -- 12.
16 Douglas F, Torrance N, van Teilingen E, Meloni S, Kerr A. Primary care staff’s viewsand experiences related to routinely advising patients about physical activity. Aquestionnaire survey. BMC Public Health 2006; 6: 138.
17 Hudon E, Beaulieu M, Roberge D. Integration of the recommendations of theCanadian Task Force on preventive health care. Fam Pract 2004; 24: 11 -- 17.
18 Galuska D, Will J, Serdula M, Ford E. Are health care professionals advising obesepatients to lose weight? JAMA 1999; 282: 1576 -- 1578.
19 Huang J, Yu H, Marin E, Brock S, Carden D, Davis T. Physicians’ weightloss counseling in two public hospital primary care clinics. Acad Med 2004; 79:156 -- 161.
20 Kant A, Miner P. Physician advice about being overweight: association with self-reported weight loss, dietary and physical activity behaviors of US adolescents inthe National Health and Nutrition Examination Survey, 1999 -- 2002. Pediatrics2007; 119: e142 -- e147.
21 Mehotra C, Naimi T, Serdula M, Bolen J, Pearson K. Arthritis, body mass index, andprofessional advice to lose weight: implications for clinical medicine and publichealth. Am J Prev Med 2004; 27: 16 -- 21.
22 Thande N, Hurstak E, Sciacca R, Giardina E. Management of obesity: a challengefor medical training and practice. Obesity 2009; 17: 107 -- 113.
23 Moher D, Liberati A, Tetzlaff J, Altman D. Preferred reporting items for systematicreviews and meta-analyses: the PRISMA statement. PLOS Med 2009; 6: e1000097.
24 Zaza S, Wright-De Aguero LK, Briss PA, Truman BI, Hopkins DP, Hennessy MH et al.Data collection instrument and procedure for systematic reviews in the guide tocommunity preventive services. Task force on community preventive services. AmJ Prev Med 2000; 18 (1 Suppl): 44 -- 74.
25 Lipsey M, Wilson D. Practical Meta-Analysis (Chapter 6). Sage Publications Inc.:Thousand Oaks, CA, 2001.
26 Singh S, Lopez-Jimenez F. Medically diagnosed overweight and weight loss in aUS national survey. Prev Med 2010; 51: 24 .
27 Nawaz H, Adams M, Katz D. Weight loss counseling by health care providers. Am JPublic Health 1999; 89: 764 -- 767.
28 Pollak K, Ostbye T, Alexander S, Gradison M, Bastian L, Brouwer R, Lyna P.Empathy goes a long way in weight loss discussions. J Fam Pract 2007; 56: 1031 --1036.
29 Saelens B, Jelalian E, Kukene D. Physician weight loss counseling for adolescents.Clin Pediatr 2002; 41: 575 -- 585.
30 Abid O, Galuska D, Kettel Khan L, Gillespie C, Ford E, Serdula M. Are healthcareprofessionals advising obese patients to lose weight? A trend analysis. Med GenMed 2005; 7: 10.
31 Felix H, West D, Bursac Z. Impact of USPSTF practice guidelines on clinician weightloss counseling as reported by obese patients. Prev Med 2008; 47: 394 -- 397.
32 Fontaine K, Haez S, Bartlett S. Are overweight and obese adults with arthritisbeing advised to lose weight? J Clin Rheumatol 2007; 13: 12 -- 15.
33 Wang A, Kinsinger L, Kahwati L, Das S, Gizlice Z, Harvey R et al. Obesity and weightcontrol practices in 2000 among veterans using VA facilities. Obes Res 2005; 13:1405 -- 1411.
34 Kabeer N, Simoes E, Murayi T, Brownson R. Correlates of overweight and weightloss practices in Missouri. Am J Health Behav 2001; 25: 125 -- 139.
35 Evans E. Why should obesity be managed? The obese individual’s perspective. IntJ Obes 1999; 23 (Suppl 4): S3 -- S6.
36 Fallon E, Wilcox S, Laken M. Health care provider advice for African Americanadults not meeting health behavior recommendations. Prev Chronic Dis 2006; 3:1 -- 11.
37 Wee C, Davis R, Phillips R. Stage of readiness to control weight and adopt weightcontrol behaviors in primary care. J Gen Int Med 2005; 20: 410 -- 415.
38 Himmel W, Stolpe C, Kochen M. Information and communication aboutoverweight in family practice. Fam Pract Res J 1994; 14: 339 -- 351.
39 Rodondi N, Humair J, Ghali W, Ruffieux C, Stoianov R, Seematter-Bagnoud L et al.Counselling overweight and obese patients in primary care: a prospective cohortstudy. Eur J Cardiovasc Prev Rehabil 2006; 13: 222 -- 228.
40 Sciammana C, Tate D, Lang W, Wing R. Who reports receiving advice to loseweight? Results from a multistate survey. Arch Intern Med 2000; 160: 2334 -- 2339.
41 Bish C, Blanck H, Serdula M, Marcus M, Kohl H, Khan L. Diet and physical activitybehaviors among Americans trying to lose weight: 2000 behavioral risk factorsurveillance system. Obes Res 2005; 13: 596 -- 607.
42 Ford E, Mannino D, Redd S, Mokdad A, Galuska D, Serdula M. Weight loss practicesand asthma: findings from the behavioral risk factor surveillance system. Obes Res2003; 11: 81 -- 86.
Physician weight loss advice and patient changeSA Rose et al
127
International Journal of Obesity (2013) 118 -- 128& 2013 Macmillan Publishers Limited
43 Levy B, Williamson P. Patient perceptions and weight loss of obese adults. J FamPract 1988; 27: 285 -- 290.
44 Zapka J, Lemon S, Estabrook B, Rosal M. Factors related to weight loss behavior ina multiracial/ethnic workforce. Ethn Dis 2009; 19: 154 -- 160.
45 Post RE, Mainous AG, Gregorie SH, Knoll ME, Diaz VA, Saxena SK. The influence ofphysician acknowledgement of patients’ weight status on patient perceptions ofoverweight and obesity in the United States. Arch Intern Med 2011; 171: 316 -- 320.
46 Pollak KI, Alexander SC, Coffman CJ, Tulsky JA, Lyna P, Dolor RJ et al. Physiciancommunication techniques and weight loss in adults: project CHAT. Am J PrevMed 2010; 39: 321 -- 328.
47 Singh S, Somers VK, Clark MM, Vickers-Douglas K, Hensrud DD, Korenfeld Y et al.Physician diagnosis of overweight status predicts attempted and successfulweight loss in patients with cardiovascular disease and central obesity. Am Heart J2010; 160: 934 -- 942.
48 Dorsey R, Songer T. Lifestyle behaviors and physician advice for change amongoverweight and obese adults with prediabetes and diabetes in the United States,2006. Prev Chronic Dis 2011; 8: A132.
49 Jay M, Gillespie C, Schlaie S, Sherman S, Kalet A. Physicians’ use of the 5As incounseling obese patients: is the quality of counseling associated with patients’motivation and intention to lose weight? BMC Health Serv Res 2010; 10: 159.
50 Cox ME, Yancy WS, Coffman CJ, Ostbye T, Tulsky JA, Alexander SC et al. Effects ofcounseling techniques on patients’ weight-related attitudes and behaviors in aprimary care clinic. Patient Educ Couns 2011; 85: 363 -- 368.
51 Alexander SC, Cox ME, Boling Turer CL, Lyna P, Ostbye T, Tulsky JA et al. Do the 5Aswork when physicians counsel about weight loss? Fam Med 2011; 43: 179 -- 184.
52 Jay M, Schlair S, Caldwell R, Kalet A, Sherman S, Gillespie C. From the patient’sperspective: the impact of training on resident physician’s obesity counseling.J Gen Intern Med 2010; 25: 415 -- 422.
53 National health and nutrition examination survey, 2010. (Accessed 6 July 2010, athttp://www.cdc.gov/nchs/nhanes.htm ).
54 Walsh JM, Swangard DM, Davis T, McPhee SJ. Exercise counseling by primary carephysicians in the era of managed care. Am J Prev Med 1999; 16: 307 -- 313.
55 Ward R. Talking with your patients about dietary cholesterol, diet and nutrition:best practices for family physicians. Int J Clin Pract 2009; 63 (Suppl 163): 22 -- 26.
Physician weight loss advice and patient changeSA Rose et al
128
International Journal of Obesity (2013) 118 -- 128 & 2013 Macmillan Publishers Limited