The Effect of a Behavioral Contract on Adherence to a Walking Program in Postmenopausal African American Women

lwwj162-10 September 27, 2005 21:22 Char Count= 0

Topics in Geriatric RehabilitationVol. 21, No. 4, pp. 332–342c© 2005 Lippincott Williams & Wilkins, Inc.

The Effect of a BehavioralContract on Adherence to aWalking Program in PostmenopausalAfrican American Women

Bernadette R. Williams, ScD, PT, GCS; Janet Bezner, PhD, PT;Steven Chesbro, DPT, EdD, PT, GCS; Ronnie Leavitt, PhD, MPH, PT

This study examined the effect of a behavioral contract on adherence to a walking programin postmenopausal African American women. A convenience sample of 43 postmenopausalAfrican American women was used. Participants were randomly assigned to an experimental orcontrol group. Those in the experimental group signed a behavioral contract. A pedometer, dailylog, 7-day physical activity recall, and qualitative analysis were used during a 7-week program.The contract group adhered more to the brisk walking goal (P = .006). A behavioral contract iseffective in increasing exercise adherence in postmenopausal African American women. Keywords: brisk walking, exercise, health promotion, pedometer, sedentary

DESPITE numerous studies supporting thebenefits of exercise in reducing long-

term chronic disease, rates of exercise amongadults remains low.1–3 This finding is espe-cially true for African American women, whoare at increased risk for heart disease, dia-betes, and obesity.4 The literature providessubstantial evidence that the incidence ofcoronary heart disease and diabetes is greaterin the African American population compared

From the Department of Physical Therapy, HowardUniversity, Washington, DC (Drs Williams andChesbro); the Rocky Mountain University of HealthProfessions, Provo, Utah (Drs Bezner and Chesbro);and the School of Allied Health, University ofConnecticut, Storrs (Dr Leavitt). Dr Williams was adoctoral student in the geriatric physical therapyprogram at Rocky Mountain University of HealthProfessions during this study.

Funding for this study was provided by Howard Univer-sity’s Humanities, Social Sciences and Education Grant,No. 05/02.

Corresponding author: Bernadette R. Williams, ScD, PT,GCS, Department of Physical Therapy, Howard Univer-sity, 6th & Bryant Sts, NW, Annex 1, Office B-24, Wash-ington, DC 20059 (e-mail: [email protected]).

to the white population.5–7 Older AfricanAmerican women are particularly suscepti-ble owing to the loss of estrogen duringmenopause.5

Diabetes has been shown to increase therisk of coronary heart disease and disability.8

In a national sample of older men andwomen, researchers found that African Amer-ican women had a significantly higher preva-lence of type II diabetes than white women.8

Other factors that increase the risk of diabetesand coronary artery disease in African Amer-ican women are high-fat diets, lack of regularexercise, and obesity.1,9,10 According to theAmerican Obesity Association, African Amer-ican women have the highest prevalence ofoverweight (78%) and obesity (50.8%) com-pared to all other groups.11 The prevalenceof overweight and obesity in white womenis 57.5% and 30.6%, respectively.11 The highprevalence of obesity is reported to be a con-tributing factor to the high prevalence of hy-pertension in minority populations, especiallyamong African Americans who have an ear-lier onset and run a more severe course ofhypertension.11 Researchers have shown that

332


BEHAVIORAL CONTRACT ON ADHERENCE TO A WALKING PROGRAM

people who exercise reduce their risk of dia-betes, coronary heart disease, and obesity, yetAfrican American women have lower rates ofregular physical activity than whites as well asother ethnic minorities.3,12,13

Increasing the rates of exercise participa-tion among African American women is ofcritical importance. A simple and effectivestrategy may be the use of behavioral con-tracts. Contracting is an active process inwhich the patient and the health professionalwork together to establish specific measur-able goals.14 The contract is based on eachpatient’s individual needs, and the health pro-fessional helps to identify potential barriersthat may need to be addressed.14,15 In addi-tion, contracting uses reinforcement as a wayto increase the likelihood that patients willfollow instructions in order to reach agreed-upon goals. Reinforcers may involve praise,recognition, attention, or more tangible fac-tors like dinner at a favorite restaurant or anew book.14,15

Behavioral contracting has its roots inadult education.16 Contract learning, or self-directed learning, became the focus of ed-ucational research for adults in the 1960s.This period was followed by the develop-ment of a comprehensive theoretical frame-work that incorporated the concepts of inde-pendent study, individualized instruction, andself-directed and lifelong learning. Andragogywas the term used to describe this theoreticalframework and means, “the art and scienceof helping adults learn.”16 Contract learningis built upon the foundation of the andra-gogy model and, therefore, shares the sameassumptions, namely, that adult learners needto

1. understand the necessity of learningsomething,

2. be self-directed,3. have their unique experiences taken

into account,4. have learning geared to their readiness

to learn,5. organize learning around life tasks or

life problems, and6. tap into intrinsic motivations.16

The success of learning contracts, as a tool

[AQ1]

to facilitate independent self-directed learn-ing, has been adopted by health profession-als as a means to help their patients develophealth behavior goals and follow throughwith the recommended actions. These “pa-tient contracts” or “behavioral contracts” aredesigned jointly by the patient and the healthprofessional.

In a study by Solanto et al,17 adolescentswith anorexia nervosa signed a behavioralcontract upon admission to a treatment fa-cility. The contract specified the amount ofweight gain to be accomplished within a 4-dayperiod. Depending on the time period of ad-mission, the girls either signed a contractspecifying a 0.8-lb weight gain (contract 1) ora 1.2-lb weight gain (contract 2). Both groupsmet their goals, although the girls who signedcontract 2 were more successful than thosewho signed contract 1. The researchers con-cluded that their results proved the efficacy ofbehavioral contracting.17

Behavioral contracting was shown to beeffective in the promotion of health behav-iors in beginning nursing students.18 Eighty-seven students in the Waddell and Stephens19

study were assigned to write a health pro-motion contract for themselves on the ba-sis of the results from a self-assessment tool.Following a 6-week intervention period, stu-dents were asked about the perceived bene-fits of behavioral contracting. Students whosuccessfully completed the contract (n = 27)identified increased feelings of self-worth, de-creased stress levels, improved physical andemotional health, and an overall sense of per-sonal empowerment.18

Neale19 examined the effectiveness of be-havioral contracting for 179 volunteer partici-pants in a cardiovascular risk reduction pro-gram. The study sample was predominantlyblack (73%) and mostly female (67%), with anaverage age of 40. All 179 received counsel-ing and were encouraged to sign a contractspecifying regular exercise participation, butonly 96 (43%) did. Following the 12-week in-tervention period, 41 of the 96 (43%) contrac-tors fully adhered to the contract, 21 (22%)

VOL. 21, NO. 4/OCTOBER–DECEMBER 2005 333


WILLIAMS et al

had partial adherence, and 34 (35%) had noadherence. The 83 participants who refusedto sign a contract did not attend any of the ex-ercise classes. The researcher concluded thatwillingness to sign a contract may be a goodpredictor of patient compliance and that con-tracting is a feasible, effective tool for healthpromotion activities.19

In summary, the literature provides evi-dence that contracts have been effective inpromoting health behaviors; however, there islittle research investigating the use of behav-ioral contracts in African American women.

Brisk walking and average daily steps

The Surgeon General’s Report on Physi-cal Activity and Health, the Centers for Dis-ease Control and Prevention, and the Amer-ican College of Sports Medicine recommendthat every adult engage in 30 minutes of mod-erate intensity exercise at least 4 days a weekfor health benefits.20–22 On the basis of theseguidelines, sedentary individuals are classifiedas those who do not engage in any leisure-timephysical activity or who participate in lessthan 30 minutes of moderate-intensity physi-cal activity daily.20–22

Walking, the physical activity used in thisstudy, has received a lot of attention in the re-cent literature. Brisk walking is considered amoderate-intensity exercise and is very pop-ular among adults as a means to meet thephysical activity recommendations.20 Mostpeople consider walking to be easy, afford-able, and even enjoyable. In one study, thehighest prevalence for walking was reportedby women between the ages of 45 and 54years.23

A popular alternative to the current recom-mendation of regular moderate-intensity ex-ercise (ie, brisk walking) is to monitor dailysteps with the use of a pedometer. News mag-azines, such as Time and Women’s Day, havesupported the now common recommenda-tion of 10,000 steps/d.24 This specific valueapparently has its roots in Japan, where thefirst pedometer was marketed in 1965.25 Thepedometer was called the manpo-kei, which

literally meant “ten-thousand-steps meter.”25

Aiming for 10,000 steps/d is now a commongoal for those who choose to use pedometers.

Until recently, the literature offered no sup-port for the relationship between moderate-intensity walking and 10,000 steps/d. Ina study examining the utility of the Digi-Walker pedometer to assess physical activity, [AQ2]

researchers reported that subjects who in-cluded 30 minutes of moderate-intensity activ-ity in a day were very likely (73% of the time)to achieve 10,000 steps on that same day.26

The subject sample consisted of 31 adult em-ployees of an aerobics center with a mean ageof 29 years.

In an effort to determine if the 10,000steps/d recommendation should be advo-cated as a physical activity target, Wilde et al24

conducted a study with 32 women betweenthe ages of 30 and 55 years. The participantswore a Digi-Walker pedometer for 4 consec-utive days and included a 30-minute walk on2 of the 4 days. Results indicated that meanstep counts for the 2 walking days exceeded10,000 steps but not for the 2 nonwalkingdays. Researchers concluded that the 10,000steps/d goal is a challenging target for seden-tary women, although it may be too high forvery inactive women.24

In this same study, researchers found thatsedentary women averaged ∼3100 steps dur-ing a 30-minute self-paced walk.24 In compar-ison, Welk et al26 reported an average of 3800to 4000 steps for a 30-minute brisk walk. Sev-eral studies have reported the number of aver-age daily steps accumulated during normal ac-tivity. These values range from a low of 6700steps/d to a high of 7439 steps/d. However,if one is conservative and combines the low-est value of 6700 steps for normal daily activ-ity with the lowest value of 3100 steps for a30-minute walk, the total is 9800 steps. Thesum of these 2 numbers is very close to the10,000 steps/d recommendation.

New evidence demonstrates health bene-fits of walking 10,000 steps/d.27,28 Hyperten-sive women who averaged 9700 steps/d re-duced their blood pressure and body massafter a 24-week walking program.28 Sugiura

334 TOPICS IN GERIATRIC REHABILITATION



et al27 found that women who increased theirsteps/d from approximately 6500 to 9000 sig-nificantly improved their cholesterol levels.These results are promising for all Americans,and especially African American women, whoare less likely to participate in other forms ofregular exercise.

African American women have a high in-cidence of diabetes, hypertension, heart dis-ease, and obesity. Exercise has been proven tobe effective in mitigating the disability causedby these diseases. Behavioral contracts havebeen shown to be useful in some settingsand with some populations to promote theadoption of healthy behaviors, but few studieshave investigated the effectiveness of behav-ioral contracts in an African American popula-tion. Therefore, the purpose of this study wasto investigate the effect of a physical activitycontract on adherence to a walking programin postmenopausal African American women.

METHODS

Participants

Characteristics of the study participants areshown in Table 1. The study population con-sisted of 43 women who self-identified asAfrican American and volunteered to partic-ipate in the study. Eight women did not com-plete the study secondary to reported mus-culoskeletal injury (3) or noncompliance (5).The 3 women who ceased participation as aresult of injury had previous histories of mus-culoskeletal problems and were obese. Twoof the women reported exacerbation of backpain and the other reported increased kneepain related to arthritis. Thirty-five womencompleted the study.

Inclusion criteria included cessation ofmenses for at least 1 year and no participationin a regular exercise program. All participantscompleted a health history questionnairecalled the Physical Activity Readiness Ques-tionnaire (PAR-Q) to screen for cardiovascu-lar disease and orthopedic injury.29 Exclusioncriteria included presence of an acute mus-culoskeletal injury, impaired ambulation or

Table 1. Demographic characteristics ofparticipants∗

n % Mean Range

Age (mean), y 57.6 50–68BMI (mean), kg m−2 32.5

Normal 7 16.3Overweight 9 20.9Obese 20 46.5Extremely obese 7 16.3

EducationHigh school diploma 12 27.9Some college 18 41.8College graduate 7 16.3Postgraduate 6 14.0

Household income<$10,000 1 2.3$21,000–$40,000 13 30.2$40,000–$60,000 10 23.3>$60,000 18 41.9No response 1 2.3

Marital statusMarried/partnered 24 55.8Single 7 16.3Divorced 12 27.9Widowed 0 0

Work statusHospital employee 17 39.5University employee 15 35.0Retired 11 25.5

EthnicityBlack (not of Hispanic 43 100.0

origin)Other 0 0

Country of originUnited States 43 100.0Other 0 0

∗BMI indicates body mass index.

mobility, classification as high-risk accordingto the American College of Sports Medicineguidelines for exercise participation,20 or his-tory of coronary artery disease. Age range ofthe participants was 50 to 68 years, with amean of 58. Participants were recruited frompostings around the campus of a historicallyblack university and hospital. Before partici-pating in the study, all participants read andsigned an informed consent form approvedby both the Rocky Mountain University of



WILLIAMS et al

Health Professions Institutional Review Boardand the Howard University Institutional Re-view Board.

PROCEDURES

Instruments

The Yamax Digi-Walker pedometer was[AQ3]

used to objectively measure the number ofdaily steps walked by each participant. A sys-tematic review of the literature by Tudor-Locke et al30 found that pedometers correlatestrongly with different accelerometers (me-dian r = 0.86). They also found that pedome-ters correlate strongly with time in observedactivity (median r = 0.82). Specifically, theYamax Digi-Walker was found to be the mostaccurate brand in a study comparing 5 well-known pedometers and was the most valid inmeasuring steps compared to 10 other elec-tronic pedometers.31

Body mass index

Body mass index (BMI) was calculated andclassified using the guidelines from the Na-tional Institutes of Health for weight andheight.32

Physical Activity ReadinessQuestionnaire

The PAR-Q is a screening tool that has beenrecommended as a minimal standard for entryinto a moderate-intensity exercise program.29

The PAR-Q was designed to identify the smallnumber of adults for whom physical activitymight be inappropriate owing to cardiovascu-lar risk factors, musculoskeletal disorders, orthose who should receive medical advice con-cerning the most suitable type of activity.

The 7-Day Physical Activity Recall

The 7-Day Physical Activity Recall (7-d PAR)is a standardized self-report measure of phys-ical activity. It was first developed and de-scribed by Blair et al33 and is one of themost widely used physical activity assess-ments in exercise science and epidemiologi-cal research. The 7-d PAR provides details re-

garding the duration, intensity, and volume ofphysical activity for a 7-day time frame. In astudy by Dishman and Steinhardt,34 conver-gent validity was examined by comparing the7-d PAR against a concurrent physical activ-ity questionnaire (r = 0.83–0.94). Constructvalidity was established when the 7-d PARwas found to be sensitive enough to detectchanges in energy expenditure that was asso-ciated with changes in maximal oxygen up-take (r = 0.33, P < .05).34

Walking logs

As a method of recording daily minutes ofbrisk walking and daily pedometer steps, allparticipants were given 7 one-page weeklywalking logs. These logs were kept in a folderand participants were instructed to record ona daily basis their walking activity and any per-tinent comments (ie, inclement weather, ill-ness, etc).

Postintervention semistructuredinterview

To collect qualitative data about percep-tions, thoughts, and feelings related to thewalking program, a series of 8 questions wereposed to each participant at the conclusion ofthe study. Each interview was conducted pri-vately and an audiotape recorded each partic-ipant’s responses.

Walking contract group

Each participant randomly assigned to theexperimental group met individually with theprimary investigator to mutually decide walk-ing goals and discuss potential barriers, strate-gies to achieve goals, reward for success-ful achievement, and motivating strategies tobe employed. They then signed this writ-ten agreement and received a copy for theirrecords.

Walking program group

Participants randomly assigned to thecontrol group participated in the samegoal-setting and discussion session as the




experimental group; however, they did notsign any agreement or contract.

INTERVENTION

All participants completed the PAR-Qhealth history questionnaire, the demo-graphic data form (Appendix A), and the 7-d[AQ4]

PAR. Each participant was asked to reach intoan envelope and pull out 1 folded piece ofpaper, each having an odd or even numberfrom 1 to 45. It had previously been deter-mined that those who picked an odd numberwould be assigned to the walking contract orexperimental group and those who pickedan even number would be assigned to thewalking program or control group.

Blood pressure, heart rate, height, andweight were assessed by the primary investi-gator according to standard guidelines. Heightand weight were used to determine BMI ac-cording to the National Institutes of Healthguidelines.32 All participants received writ-ten information about the benefits of exer-cise and health risks of inactivity (Canada’sPhysical Activity Guide to Health ActiveLiving). All participants were instructed inthe use of the Yamax Digi-Walker pedome-ter. Participants were instructed to recordthe number of daily steps walked in theirwalking logs during the first week to serveas their baseline. Participants were also in-structed to continue their normal activitylevel during this initial baseline period. An av-erage daily walking value was computed foreach participant and served as the baselinevalue.

After each participant was randomly as-signed into either the control group or exper-imental group, she met with the principal in-vestigator for a counseling session to identifywalking goals, potential barriers, motivationalstrategies, and rewards for a successful out-come. One goal was based on the AmericanCollege of Sports Medicine–Centers for Dis-ease Control and Prevention minimum phys-ical activity recommendation of 30 min/d ofmoderate-intensity exercise most days of the

week. This goal was the same for all partici-pants. Participants were instructed on utiliz-ing the BORG Rating of Perceived Exertion tomonitor the intensity of their walking. Theywere instructed to achieve an intensity of“somewhat hard,” which is 12 to 14 on the6 to 20 scale and corresponds to moderate in-tensity or “brisk walking.”20 Participants weregiven a copy of the BORG Rating of PerceivedExertion Scale on a 2 × 5 laminated card andencouraged to carry it with them during theirbrisk walking activity. The other goal was apedometer goal and was negotiated accord-ing to the baseline average daily steps, the cur-rent pedometer guidelines of 10,000 steps/d,recommendation by the principal investiga-tor, and the participant’s suggested targetedvalue.

Participants in the experimental groupsigned a contract specifying the goals tobe achieved. Those in the control groupwere given a walking program following thegoal-setting session, but did not enter intoany agreement or sign a contract. All par-ticipants were given a copy of either thewalking contract (Appendix B) for those inthe experimental group or the walking pro-gram (Appendix C) for those in the controlgroup.

All participants recorded daily walking ac-tivity in their walking logs for 6 weeks af-ter the 1-week baseline period. To determineif the pedometer was serving as a cue toengage in walking activity, participants wereinstructed not to wear the pedometer af-ter the third week and these data were notrecorded on the walking logs during weeks 5through 7. At the end of each week, partici-pants faxed their walking logs to the principalinvestigator.

A research assistant, blinded to group as-signment, administered the 7-d PAR to eachparticipant on a weekly basis via telephone.At the end of the 6-week intervention pe-riod, a semistructured interview was admin-istered to the participants to collect qualita-tive data about their perceptions, thoughts,and feelings about the walking program(Appendix G).



WILLIAMS et al

DATA ANALYSIS

Statistical analyses were performed us-ing SPSS 10.0, graduate package. Descrip-tive statistics were calculated for the en-tire sample. The percentage of those whoachieved the established goals in each groupwas compared using a chi-square test. Becausedaily steps were not normally distributed, theWilcoxon signed rank test was used to an-alyze differences between groups for paireddata of average daily steps, and the Mann-Whitney test was used for independent groupassessment.

Audio recordings of the qualitative datacollected during the interviews were tran-scribed and coded. Coding was done by label-ing quotes according to themes, sorting thethemes into groups, and then collapsing simi-lar groups. Information was labeled accordingto the content, context, and meaning of thequote.

RESULTS

A Pearson chi-square analysis revealed a sig-nificant difference between the percentagesof participants in the walking contract groupwho met their brisk walking goal comparedto the walking program group (noncontract)(P = .006). Eighty-one percent of participants(n = 13) in the contract group met their briskwalking goal compared to 31% of the con-trol group (n = 6). No significant differencewas found for the average daily step goal be-tween the 2 groups. However, the Wilcoxonsigned rank test comparing the baseline aver-age daily steps (median = 4914; mean = 6245)and the postintervention average daily steps(median = 7103; mean = 8100) for the totalsample was statistically significant (P < .01).The median difference from baseline for allparticipants was 2189 steps. Almost the entiresample indicated during the postinterventioninterviews that the pedometer was a signif-icant motivator for increasing their walkinglevels. Ten participants met both goals (7 con-tract, 3 control). Fifty-seven percent, or 20

participants, met at least 1 goal. Forty-threepercent, or 15 participants, met neither goal(5 contract, 10 control). The Mann-Whitneytest found that the participants who met bothgoals, regardless of group assignment, aver-aged more daily steps at baseline (P = .04) andpostintervention (P = .02) than those who didnot meet both goals.

Eight of the 12 participants who met bothwalking goals were in the contract group.Of these 8 women, 6 stated that signing thecontract was a significant factor for stickingwith the walking program. One woman com-mented: “The fact that I signed the contractmade me determined to fulfill my goals.” An-other woman stated, “Knowing that I had tocome back to you and be accountable wasthe biggest factor for me.” In addition, sev-eral women in the contract group indicatedthat they did not feel they would continue thewalking program without the contractual rela-tionship. When asked, “Now that the study iscompleted, how confident do you feel aboutbeing able to maintain your walking program?”one woman commented, “I like to think Iwill, but this was a commitment.”Another re-sponded: “It’s going to be hard because noone is pushing me.”

During the postintervention interviews,92% of the participants indicated that the pe-dometer was the chief motivating factor forthem during the study. When asked the ques-tion, “What factors influenced your ability tocomplete the program?” one participant hadthe following comment: “I know the pedome-ter was a major factor . . . just having it on. Iknow I had set my goal for 10,000 and I triedto reach that goal.”Another participant stated,“I really like the pedometer . . . being able tosee what you were doing. It really helped tosee those numbers.”

They also expressed disappointment thatthe pedometer would only be worn duringthe first 4 weeks of the study and not theentire 7 weeks. Several participants com-mented that they lost some of their motiva-tion without the pedometer during those last3 weeks: “I was more motivated the fourweeks that I had the pedometer than the




weeks when I did not.” Another participantcommented, “Once you took the pedometeraway, I wasn’t as motivated.”

DISCUSSION

The results of this study indicate that par-ticipants who signed a behavioral contractwere more likely to achieve the brisk walk-ing goal than participants who did not. Thisfinding supports similar results by Neale,19

Solanto et al,17 and Feingold and Perlich,35

in which contracts were found to be benefi-cial in promoting positive behavioral change,and specifically reinforces Neale’s findingthat contracts can improve compliance withan exercise program in African Americanwomen.

One explanation for the success of con-tracts in promoting adherence to the briskwalking goal is the sense of commitmentparticipants felt to achieving the outcome.As noted in the comments recorded duringthe interviews, participants credited some oftheir success to the additional pressure theyfelt because they had agreed to something inwriting. These results suggest that signing acontract provides a source of external moti-vation that can increase one’s commitmentto engage in positive health behaviors likeexercise.

In contrast, the results from the presentstudy indicate that the behavioral contractdid not significantly influence adherence tothe average daily step goal. There are sev-eral possible explanations for this finding. Avery probable explanation is that the pedome-ter goals were set too high. The brisk walk-ing goal was the same for all participants,but the pedometer goal was negotiated ac-cording to a combination of factors, includingthe average daily steps achieved during base-line, a suggestion from the principal investi-gator, current recommended step goals, andthe participant’s input. Participants were in-formed of the popular goal of 10,000 steps,and this knowledge may have inadvertently in-fluenced the participants to aim for this goal.

Ultimately, the participant had the final say asto what the targeted goal would be.

Possibly, a better way to arrive at the stepgoal would have been to increase the base-line step average by a certain percentage.This method was used in a recent study byCroteau.36 In this 8-week study, participantswhose baseline daily steps were less than8000 had a goal of increasing their daily stepsby 10% every 2 weeks. These participantsincreased their daily steps by approximately40% over the 8-week period.36 Participants inthe present study increased their daily stepsby approximately 30% over only a 3-week pe-riod. While significant progress was achievedin a shorter time period in the present study,because the daily step goals were so muchhigher than those in the study by Croteau, asmaller percentage of the participants actuallymet their daily step goal.

Moreover, possibly related to the unrealis-tic step goals is the large body mass of mostof the participants in the present study. Accu-mulating 10,000 steps requires walking a dis-tance of approximately 5 miles. The activity ofwalking, especially brisk walking, may be un-comfortable and require more effort for par-ticipants with excess adipose tissue. In thisstudy, almost 63% of the participants wereclassified as obese or extremely obese accord-ing to their BMI. Other pedometer-based stud-ies have not enrolled such a high percent-age of obese participants.23,25,27,36 The goal of10,000 steps/d may not be realistic for thosewith BMIs that fall into the obese categories.Thompson et al23 found that the average BMIof women who accumulated 10,000 or moresteps/d was in the reference range (BMI < 25).This finding lends further support for the ra-tionale of using incremental or percentage in-creases as a method of setting truly obtainablegoals.

Secondly, participants may have set an un-realistic step goal owing to the effect of so-cial desirability.37 Because goals were set dur-ing a face-to-face interview, participants mayhave felt a desire to please the researcher byselecting a goal that they perceived to be moreacceptable.



WILLIAMS et al

A third possible explanation for the lackof a significant influence of the behavioralcontract on achievement of the pedometergoal was that the pedometer was a motivat-ing factor in and of itself. All of the partici-pants increased their average daily steps frombaseline with a mean increase of almost 1900steps, roughly equivalent to 1 mile. The par-ticipants indicated that wearing the pedome-ter was a source of motivation for them. How-ever, the amount of brisk walking recordedin the logs during the weeks without the pe-dometer was essentially unchanged from the3 weeks with the pedometer. Those whowere performing brisk walking regularly con-tinued to do so even after the pedometer wasremoved. The results of this study suggest thatthe pedometer served an important role inthe initiation of the walking program by pro-viding an incentive for participation. Clearly,the pedometer sparked initial interest in thewalking program and was a source of enjoy-ment for the participants. Other studies havefound that African American women considerpersonal enjoyment as a primary benefit ofexercising.38,39 If a simple and inexpensivedevice like a pedometer can motivate othersto initiate a new health behavior and add tothe enjoyment of exercise, its use should beencouraged.

This study has several limitations. Our sam-ple size was small, and the participants inthis study were not randomly chosen from thepopulation. They were a convenience sam-ple recruited from the campus of one histori-cally black university and hospital and may notbe truly representative of all postmenopausalAfrican American women.

The intervention period in this study of6 weeks was relatively short. While the resultsdid show that the contract group adheredmore to their brisk walking goal over a 6-weekperiod, would this have been true over a6-month time frame? Several researchers haveindicated that true behavioral change can oc-cur only if a person sticks to the desired be-havior for at least 6 months.40–42

While participants were randomly assignedto either the experimental or control group,

the principal investigator was not doubleblinded to group assignment. In addition,regardless of group assignment, all partici-pants attended a face-to-face goal-setting ses-sion with the principal investigator. There-fore, all participants received the benefit ofhaving negotiated goals, discussion of poten-tial barriers to goal achievement, discussion ofstrategies to achieve goals, and had the inputof a healthcare professional in the process. Agreater difference between groups may havebeen found if the control group had not par-ticipated in any such goal-setting session.

Future studies should use larger samplesizes, longer intervention periods, and otherforms of moderate intensity or vigorous ex-ercise. In addition, the researcher should bedouble blinded, and participants assigned tothe control group should not receive anyformal goal-setting session. More research isneeded on African American women whohave normal and high BMIs, so that the rela-tionship between exercise participation andBMI in African American women can be morethoroughly explored.

CONCLUSIONS

This study supports the use of a behavioralcontract to increase exercise adherence inAfrican American women. Healthcare profes-sionals need effective tools to motivate theirpatients or clients to adopt healthy behaviors,and the clinical utility of a behavioral contractis ideal. Even for the busy clinician, asking thepatient to sign an agreement of mutually es-tablished goals requires only few minutes, yetit increases the probability of a successful out-come. The use of a contract allows the health-care professional to be both a facilitator anda coach, thereby allowing clients to take re-sponsibility for their own success.

Furthermore, the results of this study sup-port the use of a pedometer as an incentiveand motivator for participation in the physicalactivity of walking. Pedometers are inexpen-sive, reliable, and easy to use, making themideal for adults who are not likely to engage in




other, more strenuous exercise. If a sufficientnumber of average daily steps are achieved, in-dividuals can meet the current physical activ-ity guidelines and improve their health. This

outcome is especially important for AfricanAmericans, who have alarmingly high ratesof obesity, hypertension, diabetes, and heartdisease.

REFERENCES

1. Railey M. Parameters of obesity in African-Americanwomen. J Natl Med Assoc. 2000;92(10):481–484.

2. Walcott-McQuigg J, Prohaska T. Factors influencingparticipation of African American elders in exercisebehavior. Public Health Nurs. 2001;18(3):194–203.

3. Wilcox S. Physical activity in older women of color.Top Geriatr Rehabil. 2002;18(1):21–33.

4. Covington J, Grisso J. Assessing cardiovascular dis-ease risk in women: a cultural approach. J Natl MedAssoc. 2001;93(11):430–435.

5. Weise SD. Women and Coronary Heart Disease:Exercise, Wellness, and Risk Modifications. Engle-wood, Colo: Sound Images Inc; 2002.

6. Rodriguez C, Pablos-Mendez A, Palmas W, LantiguaR, Mayeux R, Berglund L. Comparison of lipids andlipoprotein levels among African-Americans, Hispan-ics, and non-Hispanic Caucasians ≥ 65 years of ageliving in New York City. Am J Cardiol. 2002;89(2):178–183.

7. Keller C, Gargrove H. Health beliefs and cardiovas-cular health behavior in young African-Americanwomen. J Cardiopulm Rehabil. 1993;13:277–282.

8. Sundquist J, Winkleby MA, Pudaric S. Cardiovascu-lar disease risk factors among older black, Mexican-American, and white women and men: an analysisof NHANES III, 1988–1994. Third National Healthand Nutrition Examination Survey. J Am Geriatr Soc.2001;49(2):109–116.

9. Clark DO. Physical activity efficacy and effectivenessamong older adults and minorities. Diabetes Care.1997;20(7):1176–1182.

10. Yancey A, Jordan A, Bradford J, et al. Engaginghigh-risk populations in community-level fitness pro-motion: ROCK! Richmond. Health Promot Pract.2003;4(2):180–188.

11. American Obesity Association. Obesity in MinorityPopulations. Washington, DC: American Obesity As-sociation; 2004. Accessed October 4, 2004.[AQ5]

12. Masse C, Anderson C. Ethnic differences among cor-relates of physical activity in women. Am J HealthPromot. 2003;17(6):357–360.

13. Blair SN, Kohl HW III, Barlow CE, PaffenbargerRS Jr, Gibbons LW, Macera CA. Changes in phys-ical fitness and all-cause mortality. A prospec-tive study of healthy and unhealthy men. JAMA.1995;273(14):1093–1098.

14. Falvo D. Effective Patient Education. 2nd ed.Gaithersburg, Md: Aspen; 1994.

15. Rankin H, Stallings K. Patient Education: Issues,Principles, Practices. 3rd ed. Philadelphia: Lippin-cott Williams & Wilkins; 1996.

16. Knowles M. Using Learning Contracts. San Fran-cisco: Jossey-Bass; 1986.

17. Solanto M, Jacobson M, Heller L, Golden N, HertzS. Rate of weight gain of inpatients with anorexianervosa under two behavioral contracts. Pediatrics.1994;93(6):989–991.

18. Waddell D, Stephens S. Use of learning contracts ina RN-to-BSN leadership course. J Contin Educ Nurs.2000;31(4):179–184.

19. Neale A. Behavioural contracting as a tool to help pa-tients achieve better health. Fam Pract. 1991;8(4):336–342.

20. ACSM. ACSM’s Guidelines for Exercise Testingand Prescription. 6th ed. Philadelphia: LippincottWilliams & Wilkins; 2000.

21. US Department of Health and Human Services. Phys-ical Activity and Health: A Report of the SurgeonGeneral. Atlanta, Ga: Centers for Disease Control andPrevention; 1996.

22. Pate R, Pratt M, Blair N. Physical activity and pub-lic health: a recommendation from the Centers forDisease Control and Prevention and the AmericanCollege of Sports Medicine. JAMA. 1995;273:402–407.

23. Thompson DL, Rakow J, Perdue S. Relationshipbetween accumulated walking and body composi-tion in middle-aged women. Med Sci Sports Exerc.2004;36(5):911–914.

24. Wilde B, Sidman C, Corbin C. A 10,000-step count asa physical activity target for sedentary women. Res QExercSport. 2001;72(4):411–414.

25. Tudor-Locke C, Bassett D. How many steps/day areenough? Preliminary pedometer indices for publichealth. Sports Med. 2004;34(1):1–8.

26. Welk G, Differding J, Thompson R, Blair SN, Dziura J,Hart P. The utility of the Digi-Walker step counter toassess daily physical activity patterns. Med Sci SportsExerc. 2000;32(9):S481–S488.

27. Sugiura H, Kajima K, Mirbod S. Effects of long-termmoderate exercise and increase in number of dailysteps on serum lipids in women: randomized controltrial. BMC Women’s Health. 2002;2(1):3.

28. Moreau K, Degarmo R, Langley J, et al. In-creasing daily walking lowers blood pressure inpostmenopausal women. Med Sci Sports Exerc.2001;35(11):1825–1831.



WILLIAMS et al

29. Canadian Society for Exercise Physiology. PAR-Q &You. Gloucester, Ontario: Canadian Society for Exer-cise Physiology; 1994.

30. Tudor-Locke C, Williams J, Reis J, Pluto D. Utility ofpedometers for assessing physical activity: conver-gent validity. Sports Med. 2002;32(12):795–808.

31. Tudor-Locke C. A preliminary study to determine in-strument responsiveness to change with a walkingprogram: physical activity logs versus pedometers.Res Q Exerc Sport. 2001;72(3):288–292.

32. NIH, National Heart, Lung, and Blood Institute.Clinical Guidelines on the Identification, Evalua-tion and Treatment of Overweight and Obesity inAdults. US Dept of Health and Human Services, Pub-lic Health Service; 1998.

33. Blair S, Haskell W, Ho P, et al. Assessment of habitualphysical activity by a seven-day recall in a communitysurvey and controlled experiments. Am J Epidemiol.1985;122(5):794–804.

34. Dishman R, Steinhardt M. Reliability and concurrentvalidity for a 7-d re-call of physical activity in collegestudents. Med Sci Sports Exerc. 1988;20(1):14–25.

35. Feingold C, Perlich L. Teaching critical thinkingthrough a health-promotion contract. Nurse Educ.1999;24(4):42–44.

36. Croteau K. A preliminary study on the impact of apedometer-based intervention on daily steps. Am JHealth Promot. 2004;18(3):217–220.

37. Hebert J, Ma Y, Clemow L, et al. Gender differences insocial desirability and social approval bias in dietaryself-report. Am J Epidemiol. 1997;146(12):1046–1055.

38. Jones M, Nies M. The relationship of perceivedbenefits of and barriers to reported exercise inolder African-American women. Public Health Nurs.1996;13(2):151–158.

39. Nies M, Vollman M, Cook T. African Americanwomen’s experiences with physical activity in theirdaily lives. Public Health Nurs. 1999;16(1):23–31.

40. Bull FC, Eyler AA, King AC, Brownson RC. Stage ofreadiness to exercise in ethnically diverse women: aU.S. survey. Med Sci Sports Exerc. 2001;33(7):1147–1156.

41. Felton G, Ott A, Jeter C. Physical activity stages ofchange in African American women: implications fornurse practitioners. Nurse Pract Forum. 2000;11(2):116–123.

42. Marcus B, Selby V, Niaura R, Rossi J. Self-efficacy andthe stages of exercise behavior change. Res Q ExercSport. 1992;63(1):60–66.



Title: The Effect of a Behavioral Contract on Adherence to a Walking Program in Post-menopausal African American Women

Authors: Bernadette R. Williams, Janet Bezner, Steven Chesbro, and Ronnie Leavitt

Author Queries

AQ1: Check the running head provided.AQ2: Provide the manufacturer’s name and address.AQ3: Provide the manufacturer’s address.AQ4: Provide Appendix for this article.AQ5: Provide the URL for accessing this material.

The Effect of a Behavioral Contract on Adherence to a Walking Program in Postmenopausal African American Women

Documents