Do the Five A’s Work When Physicians

8/3/2019 Do the Five As Work When Physicians

1/6

FAMILY MEDICINE VOL.43,NO.3MARCH2011 179

ORIGINALARTICLES

Obesity is one o the most sig-nicant health problems ac-ing Americans, with more

than two thirds o adults over-weight and obese.1, 2 The US Preven-tive Services Task Force (USPSTF)recommends that physicians oer

intensive counseling to obese adultsto promote sustained weight loss.3Physicians oten routinely discussweight and weight loss strategieswith their overweight and obese pa-tients.4-6 Further, when physicianscounsel patients to lose weight,

patients are more likely to attempt

to lose weight, increase physical ac-tivity, and improve diet.7-11

What is lacking rom the USPSTFguidelines, however, is eective,concise weight loss counseling tech-niques. Given competing demands,physicians need simple mnemonics.Such a tool exists or smoking ces-sation counseling: the Five As (Ask,

Advise, Assess, Assist, and Arrange).Use o the Five As has been linked tohigher motivation and more quit at-tempts among smokers.12,13Research

has suggested that this techniquecould be useul or weight loss coun-seling,14-16 and physicians in prima-ry care settings currently are usingsome o these techniques.17 However,it is unknown whether the Five Asare eective or promoting weight-related change and actual weightloss. The aim o this study was toexamine the eect o the use o Five

As during weight loss counseling onpatients motivation, condence, and

Do the Five As Work When PhysiciansCounsel About Weight Loss?StewartC.Alexander,PhD;MaryE.Cox,MD,MHS;ChristyL.BolingTurer,MD,MHS; PaulineLyna,MPH;Trulsstbye,MD,PhD;JamesA.Tulsky,MD; RowenaJ.Dolor,MD,MHS;KathrynI.Pollak,PhD

BACKGROUND AND OBJECTIVES: More than two thirds o Ameri-

cans are overweight or obese. Physician counseling may help pa-

tients lose weight; however, physicians perceive these discussions

as somewhat utile and time-consuming. An eective and efcient

tool or smoking cessation is the Five As (Ask, Advise, Assess, As-

sist, and Arrange). We studied the eectiveness o the Five As in

weight-loss counseling.

METHODS: We audiorecorded primary care encounters between

40 physicians and 461 o their overweight or obese patients. All

were told the study was about preventive health, not weight spe-

cifcally. Encounters were coded or physician use o the Five As.

Patients motivation and confdence were assessed beore and im-

mediately ater the encounter. Three months later, we assessed

patient change in dietary at intake, exercise, and weight.

RESULTS: Generalized linear models were ft adjusting or patientclustering within physician. Physicians used at least one o the Five

As oten (83%). Physicians routinely Ask and Advise patients to

lose weight; however, they rarely Assess, Assist, or Arrange. Assist

and Arrange were related to diet improvement, whereas Advise

was associated with increases in motivation and confdence to

change dietary at intake and confdence to lose weight.

CONCLUSIONS: Similar to smoking cessation counseling, physi-

cians routinely Asked and Advised patients to lose weight; howev-

er, they rarely Assessed, Assisted, or Arranged. Given the potential

impact o using all o these counseling tools on changing patient

behavior, physicians should be encouraged to increase their use o

the Five As when counseling patients to lose weight.

(Fam Med 2011;43(3):179-84.)From the Durham VA Medical Center,Durham, NC (Drs Alexander, Tulsky, andDolor); Department o Medicine, DukeUniversity Medical Center (Drs Alexander,Cox, Tulsky, and Dolor); Division o GeneralPediatrics and Department o General InternalMedicine, University o Texas at SouthwesternMedical Center, Dallas, TX (Dr Boling Turer);Cancer Prevention, Detection, and ControlResearch Program (Drs Lyna and Pollack)and Department o Community and FamilyMedicine (Dr Ostbye and Pollack), DukeUniversity Medical Center; and Duke-NUSMedical School Singapore (Dr Ostbye).


2/6

180 MARCH2011VOL.43,NO.3 FAMILY MEDICINE

ORIGINAL ARTICLES

change in dietary at intake, physi-cal activity, and weight.

MethodsThis study analyzed audio record-ings rom Project CHATCommu-nicating Health: Analyzing Talk. The

project was approved by the DukeUniversity Medical Center Institu-tional Review Board.

RecruitmentPhysicians. Primary care physi-cians rom community-based practic-es (n=54) were told the study wouldexamine communication about pre-

ventive health topics, not weightspecically. Forty (74% o those ap-proached) gave written consent.Fourteen reused or the ollowing

reasons: new to practice, back romsurgery, not enough patients, leav-ing practice, concerned about patientfow, and do not support research.

Patients. Potential patients wereidentiied by review o scheduledappointments 3 weeks in advance.Eligible participants were at least18 years o age, English speaking,overweight or obese (BMI 25 kg/m2), cognitively competent, and notpregnant. Ater obtaining consent, a

research assistant accompanied thepatient to the exam room to start adigital audio recorder. Immediatelyollowing the encounter, the researchassistant administered a post-en-counter survey and assessed the pa-tients vital signs, including weightand other measures (eg, heart rate)to mask the ocus on weight (n=461).Three months later, the research as-sistant met with the participant toassess vital signs again and admin-ister a survey assessing changes

in dietary at intake and exercise(n=426).

Data CodingThe audiorecorded conversationswere transcribed, and the Five Aswere coded:13 (1)Asks about weight,nutrition, and/or exercise, (2) Ad-vises on topics o nutrition, physicalactivity, or weight, (3)Assesses readi-ness to change, (4)Assists in setting

goals, and (5)Arranges or ollow-upvia physician visit, nutrition visit, ortelephone contact (Table 1).

Two independent coders analyzedaudio recordings; 20% were doublecoded to assess inter-rater reliabil-ity. Disagreements were discussed

and nal decisions made by consen-sus. Inter-rater agreement was cal-culated using Cohens Kappa.18 Allcodes had substantial to near per-ect agreement (Ask=.87, Advise=.78, Assess=.87, Assist=.77, and Ar-range=1.0).

MeasuresDietary Fat Intake. Dietary at in-take was assessed using the 22-itemFat and Fiber-related Diet BehaviorQuestionnaire.19,20 Questions about

requency o ood selections includ-ed: When you eat dessert, how otendid you eat only ruit? and When

you ate chicken, how oten did youtake o the skin? Responses wereaveraged into a total score where1 refected higher ber, lower atood choices, and a score o 4 re-fected lower ber, higher at choic-es (a=0.74 at baseline and a=0.77 at

3-month ollow-up).

Physical Activity. Physical activitywas measured (baseline, 3 months)using the Framingham Physical Ac-tivity Index.21 Participants recalledthe average number o hours spentengaged in various daily activities(sleeping, working, at leisure) andthe level o activity (sedentary, slight,moderate, and heavy) or each. Thecomposite score accounts or activityduration and intensity.

Anthropometrics. Patient weight(baseline, 3 months) and height

Table 1: Five As: Defnitions and Actual Examples

Defnition Examples

Ask Physician asks the patientabout weight, nutrition, and/orexercise.

Do you exercise?Tell me what you typicallyeat or breakast.

Advise Physician provides the patientwith clear, strong advice.

You need to get 30 minutes oexercise a day, 5 days a week.I think you need to lose about20 pounds.

Because o your diabetes, it isimportant or you to exercise.

Assess Physician verbally assessespatients readiness to change.

Is losing weight somethingyou want to do in the nearuture?Do you see yoursel gettingmore exercise in the comingmonths?

Assist Physician provides brie counseling or sel-help materials.

How much to you want tolose weight?What might get in the wayo your plan to exercise threetimes a week?Have you tried a very low

carbohydrate diet beore?How are you eeling aboutbeing able to make thischange?Is your amily supportive oyour attempts to eat better?

Arrange Physician arranges or ollow-upwith physician or nutritionist.

Why dont you call me in 2weeks to let me know how theweight loss plan is going?I will schedule anappointment or you to see ournutritionist.


3/6


ORIGINAL ARTICLES

(baseline only) were measured bystudy personnel using a calibratedscale and portable stadiometer. Pa-tients were asked to remove theirshoes, outerwear, and belongingsrom their pockets.

Motivation. We measured par-ticipant motivation to lose weight,change diet, and increase exerciseusing a seven-level Likert scale be-ore and immediately ater the en-counter.

Confidence. Conidence to loseweightwas measured with a ve-level Likert scale beore and imme-diately ater the encounter, Howcondent are you that you can loseweight? Condence to change diet

and condence to increase exercisewere measured similarly.

AnalysisAnalyses were perormed using SAS(SAS Institute, Inc, Cary, NC). We as-sessed the association between theFive As (Ask, Advise, Assess, Assist,and Arrange) and the ollowing:weight loss, improvement in dietaryat intake behaviors, increase in ex-ercise, and increase in motivationand conidence to: (1) lose weight,

(2) change diet, and (3) increase ex-ercise rom baseline to post-visit.We t hierarchical models account-ing or both repeated measures bypatient and or clustering by physi-cian. We ound no signicant physi-cian clustering eect, and this eectwas dropped rom the models. PROCMIXED was used to t general lin-ear models (GLM); responses wereincorporated into these models romall participants that provided atleast one time-point. This modeling

ramework yields unbiased estimateswhen missing data are unrelated tothe unobserved variable.22

Primary predictors included (1)each o the Five As, (2) time sincebaseline visit, and (3) time by predic-tor interaction. All models includedapriori dened patient, physician,and visit-related covariates. Four-teen patient covariates that weretheoretically or empirically thought

to be related to changes in weight,physical activity, or dietary at in-take were gender, age, race, highschool education, economic securi-ty (enough money to pay monthlybills), overweight (BMI 25.029.9kg/m2) or obese (BMI 30 kg/m2),actively trying to lose weight, mo-tivated to lose weight, comortablediscussing weight, condent aboutlosing weight, and the comorbidities

o diabetes, hypertension, arthritis,and hyperlipidemia. Nine physiciancovariates were gender, race, yearssince medical school graduation, spe-cialty (amily versus internal medi-cine), sel-ecacy or and barriersor weight counseling, comort dis-cussing weight, insurance reimburse-ment concerns, and prior training inbehavioral counseling. Finally, two

visit-level covariates were minutes

Table 2: Physician and Patient Characteristics (n=461)

Physicians %or M (SD)

Patients %or M (SD)

Race 85%

White/Asian/Pacic Islander 15% 66%

Arican American 60% 35%

Female 47.3 (8.2) 66%

Age (missing=1)1 24.9 (4.0) 59.8 (13.9)

BMI (missing=1) 33.1 (7.1)

Education (missing=1)

Post high school education 67%

Income (missing=37)

$45,000 or less 48%

High nancial burden (missing=13)

Pay bills with trouble 14%

Diagnosed with:

Diabetes 31%

Hypertension (missing=1) 69%

Hyperlipidemia (missing=1) 56%

Arthritis 47%

New patient 4 %

Motivation, baseline4

Change dietary at intake 5.5 (1.9)

Increase physical activity 4.7 (2.2)

Lose weight 5.2 (1.9)

Condence, baseline5

Change dietary at intake 3.7 (1.2)

Increase physical activity 3.7 (1.1)

Lose weight 3.5 (1.1)Years since medical school graduation 22.1 (8.0)

Sel-ecacy to address weight2 4.0 (0.67)

Comort discussing weight with patient3 4.4 (0.86)

1Missing data at baseline

2Sel-ecacy to address weight (1=not at all condent to 5=very condent)

3Comort discussing weight (1=not at all comortable to 5=very comortable)

4Motivation (1=not at all motivated to 7=very motivated)

5Condence (1=not at all condent to 5=very condent)


4/6


ORIGINAL ARTICLES

spent addressing weight issues andvisit type (preventive versus chron-ic).

ResultsSample CharacteristicsO the 40 physicians who agreedto participate in the study, 19 wereamily physicians, and 21 were inter-nists. Compared to those who agreedto participate, Arican American e-male physicians were more like-ly to reuse than their white, malecounterparts (P=.005), and young-er patients were more likely to re-use (P


5/6


ORIGINAL ARTICLES

o the Five As with more motivatedpatients and with patients who re-ported less condence to lose weight.Our own work indicates that phy-sicians are reluctant to counsel pa-tients who do not want to change.25Findings rom the current study

suggest that physicians do more orthose who need more, those who areless condent and are heavier, butalso may choose to do more with pa-tients with whom they think theywill have the biggest impact, namelythose who are motivated.

Overall, use o the Five As seemedto infuence patients to be more mo-tivated to change, more condent tochange, and more likely to change.Improvements in motivation andconidence are important as they

have been correlated with weightloss, weight maintenance, and con-tinued behavior change over the longterm.26,27 When physicians providestrong, clear advice, patients mightbe able to recognize the importanceo weight as a health concern; theconverse might be true when physi-cians do not provide advice.

Indeed, results suggest that pa-tients whose physicians Assistedor Arranged showed improvementsin actual dietary at intake change

scores. This modest improvementrelects a change in one at-relat-ed eating behavior, like trimmingthe at o o meat oten instead osometimes or by eating bread/rollswithout butter or margarine lessoten, eating lower at cheeses lessrequently, or rom switching rom2% to non-at milk. These are mi-nor dietary adjustments but onesthat can reduce energy intake by100 or 200 kcal/day, a decit largeenough to result in weight loss in

some patients. Thus, the dierencelikely represents a clinically signi-cant improvement. This supports theuse o explicit planning by the physi-cian-patient team in improving nu-trition behavior. It may also refectreerrals or medical nutrition ther-apy, which are an important compo-nent o multi-disciplinary weight lossstrategy. No changes were ound oractual exercise, but this is not sur-prising as even intensive behavioral

interventions are oten unable to im-prove physical activity.28

Only one o the As was linked toactual weight loss. Patients whosephysicians Arranged a ollow-up visitwere more likely to have lost weight3 months ollowing the visit. This is

consistent with the notion that re-quency o contact is an importantelement or inluencing behaviorchange; it may refect the patientseeling accountable to their physi-cian. It is encouraging that physi-cians were more likely to Arrangea ollow-up visit or patients withhigher BMI.

A major strength o this studyis that these primary care patientswere not enrolled in a weight-losstrial and thereore were not sel-se-

lected to be highly motivated to loseweight. Another strength is the largeand ethnically diverse sample. Thestudy also has several limitations.First, the results may not generalizeto younger, lower income patients.Second, the study was observation-al. Though we adjusted or a broadset o patient, physician, and visitcovariates, unmeasured conound-ing variables may still account orat least part o the observed associ-ations. Third, multiple comparisons

were done, so signicant associationswithP values near 0.05 must be in-terpreted with caution. Fourth, therewere low requencies ound or As-sessing, Arranging, and Assisting. Al-though this is not surprising, the lowrequencies o these techniques makeit dicult to detect the eectivenesso these techniques on weight loss.Finally, the analysis is limited bythe use o sel-reported dietary atand ber intake and physical activ-ity measures. A ood diary and an

accelerometer may have been moreaccurate; however, such involvedmeasures could invoke changes inbehavior, which would have madethe interpretation o results morecomplicated.

This is the rst study to assess therelationship between actual use othe Five As in weight loss counselingin the primary care encounter andsubsequent weight-related behaviorchange. Physicians routinely Asked

and Advised patients about weight;however, they rarely Assessed, As-sisted, or Arranged. Next steps orthis work would be to examine moreclosely whether some o the As arequalitatively more important thanothers. Further, given these prelimi-

nary results, a randomized controlledtrial might be warranted to test anintervention that attempts to teachphysicians how to incorporate theFive As in their weight loss coun-seling.

ACKNOWLEDGMENTS:All authors had accessto the data and helped write the manuscript.The authors wish to acknowledge Justin R.E.Manusov or his assistance in data collectionand Rebecca J. Namenek Brouwer or her as-sistance in manuscript writing.

This work was supported by grantsR01CA114392, R01DK64986, and R01

DK075439. Dr Alexander is supported byHealth Services Research Career DevelopmentAward RCD 07-006 rom the Department oVeterans Aairs.

This study was presented at theInternational Conerence on Communicationin Healthcare, Miami, FL, October 47, 2009.

CORRESPONDING AUTHOR: Address corre-spondence to Dr Alexander, 2424 Erwin Road,Suite 602, Durham, NC 27705. 919-668-7220.Fax: 919-668-1300. [email protected].

Reerences1. Yun S, Zhu BP, Black W, Brownson RC. A

comparison o national estimates o obesityprevalence rom the behavioral risk actor

surveillance system and the National Healthand Nutrition Examination Survey. Int J Obes(London) 2006;30:164-70.

2. Prevention CDCa. State-specic prevalenceo obesity among adultsUnited States, 2005.MMWR Morb Mortal Wkly Rep 2006;55:985-8.

3. Force USPST. Screening or obesity in adults:recommendations and rationale. Am J Nurs2004;104:94-5.

4. Greiner KA, Born W, Hall S, et al. Discussingweight with obese primary care patients: phy-sician and patient perceptions. J Gen InternMed 2008;23:581-7.

5. Boardley D, Sherman C, Ambrosetti L, et al.Obesity evaluation and intervention duringamily medicine well visits. J Am Board Fam

Med 2007;20:252-7.

6. Stange KC, Flocke SA, Goodwin MA, KellyRB, Zyzanski SJ. Direct observation o rateso preventive service delivery in communityamily practice. Prev Med 2000;31:167-76.

7. Nawaz H, Adams ML, Katz DL. Physician-patient interactions regarding diet, exercise,and smoking. Prev Med 2000;31:652-7.

8. Sciamanna CN, Tate DF, Lang W, Wing RR.Who reports receiving advice to lose weight?Results rom a multistate survey. Arch InternMed 2000;160:2334-9.


6/6


ORIGINAL ARTICLES

9. Galuska DA, Will JC, Serdula MK, Ford ES.Are health care proessionals advising obesepatients to lose weight? JAMA 1999;282:1576-8.

10. Evans E. Why should obesity be managed? Theobese individuals perspective. InternationalJournal o Obesity & Related Metabolic Disor-ders: Journal o the International Associationor the Study o Obesity 1999;23 Suppl 4:S3-5;discussion S6.

11. Mehrotra C, Naimi TS, Serdula M, et al.Arthritis, body mass index, and proessionaladvice to lose weight: implications or clinicalmedicine and public health. Am J Prev Med2004;27:16-21.

12. Clinical guidelines on the identication, evalu-ation, and treatment o overweight and obesityin adults: the evidence report. Bethesda, MD:National Heart, Lung, and Blood Instituteand the National Institutes o Health, ObesityEducation Initiative Expert Panel, 1998;Oct.

13. Serdula MK, Khan LK, Dietz WH, SerdulaMK, Khan LK, Dietz WH. Weight loss counsel-ing revisited. JAMA 2003;289:1747-50.

14. Estabrooks PA, Glasgow RE, DzewaltowskiDA. Physical activity promotion through pri-mary care. JAMA 2003;289:2913-6.

15. Whitlock EP, Orleans CT, Pender N, Allan J.Evaluating primary care behavioral counselinginterventions: an evidence-based approach. AmJ Prev Med 2002;22:267-84.

16. Huang N. Motivating patients to move. AustFam Physician 2005;34:413-7.

17. Flocke SA, Clark A, Schlessman K, et al.Exercise, diet, and weight loss advice in the

amily medicine outpatient setting. Fam Med2005;37:415-21.

18. Landis JR, Koch GG. The measurement oobserver agreement or categorical data. Bio-metrics 1977;33:159-74.

19. Shannon J, Kristal A, Curry S, Beresord S. Ap-plication o a behavioral approach to measur-ing dietary change: the at- and ber-relateddiet behavior questionnaire. Cancer EpidemiolBiomarkers Prev 1997;6:355-61.

20. Kristal A, Shattuck A, Henry H. Patterns odietary behavior associated with selecting dietslow in at: reliability and validity o a behav-ioral approach to dietary assessment. J AmDiet Assoc 1990;90:214-20.

21. Kannel W, Sorlie P. Some health benets o

physical activity: The Framingham Study. ArchIntern Med 1979;139:857-61.

22. Little RJA, Rubin DB. Statistical analysis withmissing data. New York: Wiley, 2002.

23. Schroeder SA, Schroeder SA. What to do witha patient who smokes. JAMA 2005;294:482-7.

24. Melvin C, Ganey C, Melvin C, Ganey C.Treating nicotine use and dependence o preg-nant and parenting smokers: an update. Nico-tine Tob Res 2004;6 Suppl 2:S107-S24.

25. Alexander S, stbye T, Pollak K, et al. Physi-cians belies about discussing obesity: resultsrom ocus groups. American Journal o HealthPromotion 2007;21:498-500.

26. Teixeira PJ, Going SB, Sardinha LB, LohmanTG. A review o psychosocial pre-treatmentpredictors o weight control. Obesity Reviews2005;6:43-65.

27. Elhag K, Rossner S. Who succeeds in main-taining weight loss? A conceptual review oactors associated with weight loss mainte-nance and weight regain. Obesity Reviews2005;6:67-85.

28. Martin SB, Morrow JR Jr, Jackson AW, DunnAL. Variables related to meeting the CDC/ACSM physical activity guidelines. Med Sci

Sports Exerc 2000;32:2087-92.

Documents

Do the Five A’s Work When Physicians