Determinants of continued breast self-examinationpractice in women 40 years and over after personalized
instruction
Valerie Clarke, David Hill, Judy Rassaby, Victoria White and Sheila Hirst
Abstract
Variables related to continued breast self-examination (BSE) practice were assessed at theconclusion of small group BSE training sessionsand again 12 months later, using a sample of 1134women 40 years and over who had attended ateaching session between February 1985 andSeptember 1987 in Victoria, Australia. Theteaching program was successful in increasing thefrequency with which the participants practicedBSE, and in reducing the impact of emotionalbarriers to BSE practice, but it did not achievethe level of regular practice advocated. Thefrequency of BSE practice was increased,especially among those who reported some BSEpractice before attending the training sessions.Although participants recognized the benefits ofBSE, failure to practice it regularly was attributedprimarily to the practical problem of forgettingand to lack of perceived self-efficacy in relationto BSE practice. The need for further work todevelop techniques for reminding women to doBSE was noted and the provision of the oppor-tunity to attend a further training session wasrecommended.
Introduction
Breast self-examination (BSE) is an effective methodof detecting breast cancer in the early stages (Hillet al., 1988). It is particularly effective if women
Centre for Behavioural Research in Cancer, Anti-CancerCouncil of Victoria, 1 Rathdowne Street, Carlton South,Victoria 3053, Australia
have been personally trained in BSE techniques(Mant et al., 1987) and their examinations areregular and systematic. The aim of this paper is toevaluate the effectiveness of a comprehensive BSEteaching program, called Mammacheck, which isdescribed by Rassaby et al. in the Programme Paperssection of this issue of Health Education Research.This evaluation, based on data collected both at theconclusion of the BSE teaching session (Time 1) andagain 12 months later (Time 2), will assess the impactof that program on reported frequency of doing BSEand identify the factors which facilitate and inhibitcontinued BSE practice. As the immediate benefitsof BSE are less likely to be found in women under40 years of age (Baines, 1988), the focus of this studyis upon women 40 and over.
The Mammacheck program was based on a modelof behaviour change derived from the theory ofreasoned action (Ajzen and Fishbein, 1980), thehealth belief model (HBM) (Becker and Maiman,1975) and social learning theory (Bandura, 1977).The model proposed that BSE practice depends onone's intention to practice BSE. These intentions areinfluenced by six perceptions: self-efficacy in rela-tion to BSE practice, behavioural beliefs about BSE,barriers to doing BSE, personal susceptibility tobreast cancer, efficacy of breast cancer treatment,severity of breast cancer and cues to doing BSE(Hirst etal., 1989).
In developing our model, we accepted the postulatederived from Ajzen and Fishbein (1980) thatbehavioural intention to do BSE would be causallyrelated to subsequent BSE practice, but we did notinclude the social referent variables as previousresearch has shown that they add little (Hill et al.,1985; Hill and Shugg, 1989) or nothing (Rutledge,
1986) to the prediction of BSE practice, a fact whichhas been attributed to the private nature of BSE. Weadopted four concepts from the HBM: perceivedsusceptibility, perceived benefits, perceived costs andperceived severity. After carefully evaluating therelevance of severity, it was retained in the model,but neither directly targetted in the training sessionsnor assessed as our previous research has shown itslack of ability to predict BSE behaviour (Hill et al.,1985), a finding which has been attributed to thealmost universal acceptance among women thatbreast cancer is serious. It was considered preferableto focus on the role of early detection of breast cancerin its successful treatment, rather than to focus onits severity which might lead to an unwillingness tocarry out BSE or to a denial of the potential dangerof any abnormality. Although other modifiedversions of the HBM have included health motiva-tion (Prentice-Dunn and Rogers, 1986), the impactof this variable on BSE is dubious. Hill et al. (1985)found that it was not a significant predictor of inten-tions to do BSE, while Champion (1987) reportedthat it neither added significantly to the predictionof BSE practice nor discriminated between BSE prac-ticers and non-practicers. The model adopted theconcept of self-efficacy (Bandura, 1977; Strecheret al., 1986), recognizing that a woman was morelikely to practice BSE if she felt competent to do so.The central aim of the training program was to drawon social learning principles to enhance self-efficacyboth by furthering the participants' knowledge ofBSE and breast cancer, and by developing their skillsthrough modelling and reinforced practice.
The program involves a single session of60—90 min in which 8-12 women meet in acomfortable, relaxed atmosphere with a femalefacilitator. The session begins with a talk by thefacilitator designed to explain the rationale underlyingBSE and to increase knowledge of the basicphysiology of the breast, the skills participants candevelop and knowledge of breast cancer. Emphasisis placed on the success of treatment if diagnosis isearly. A short video provides a demonstration of theBSE technique. Then women practice on themselves,over their clothing, and on artificial breast modelswhich contain a number of lumps of different sizes
and at different depths. During this practice thefacilitator checks for correctness of the technique andoffers encouragement and assistance. The sessionconcludes with a discussion of barriers to BSE prac-tice, answers to questions, the development of actionplans and the distribution of materials to assistparticipants at home.
Following our composite model of healthbehaviour which provided the theoretical basis forthe Mammacheck program, three hypotheses weretested: first, that there will be an increase in BSEpractice between Time 1 and Time 2; second, thatthe best predictor of BSE practice at Time 2 will bestated intentions to practice BSE as assessed at Time1 and, third, that BSE practice as assessed 12 monthsafter the training session will be predicted by acombination of previous BSE practice, self-efficacy,barriers, benefits and susceptibility.
Method
SampleThis evaluation is based on the responses of 1134women over 40 years of age who attended aMammacheck training session between February1985 and September 1987, and completed a follow-upquestionnaire one year later. Of the 6035 womenwho had attended a session, 37% were 40 years ofage or older. Of these 2130 women, 1134 (53%) hadcompleted a second questionnaire when the programwas evaluated. Of those women eligible to completea second questionnaire, 11 % of them indicated thatthey did not wish to do so. Of the remaining women84% completed and returned the questionnaire. The16% who did not return the questionnaire includedthose who had moved and could not be contacted andthose who were overseas.
QuestionnairesIn formulating the questionnaires, items weredeveloped to assess the attitudes and beliefs whichthe training program had been designed to influence.The questions had been tested in previous studies(Hill etal., 1985; Hill and Shugg, 1989). TheTime 1 questionnaire included pre-coded questionsassessing demographic variables, previous
experience with BSE, perceived susceptibility tobreast cancer, perceived efficacy of breast cancertreatment, beliefs about BSE (perceived benefits,barriers and self-efficacy) and intentions to practiceBSE. The Time 2 questionnaire included pre-codedquestions assessing practice of BSE sinceparticipating in the program and measuring attitudesto BSE (perceived confidence, competence andbarriers to the practice of BSE).
Scales and analysesWhere response categories fell along somecontinuum in order of intensity, numerical valueswere assigned to responses so that group resultswould be condensed for simplicity of presentationand analysis. The Statistical Package for the SocialSciences (SPSSX) was used for all analyses. Alphawas set at 0.001. A high alpha level was selectedto reduce the possibility of Type 1 errors when usingquestionnaire measures that might not be strictlyindependent.
Results
Entry characteristics of participantsFifty-four per cent of the sample were 40—49 yearsof age, 29% were 50-59 and 17% were 60 yearsor over. Reports of marital status indicated that 5 %of the sample had never married, 79% were marriedor living in a de facto relationship, 8% were divorcedor separated and 8% were widowed. In response toa question asking about highest level of education,17% had completed a tertiary qualification, 9% hadsome post-secondary education, 29% had completedsecondary schooling and only 7% had not reachedsecondary schooling. Relative to the Australianpopulation, the sample contained more women whowere married (79% sample, 64% population), morewomen with some tertiary education (26% and 12%)and fewer who had not reached secondary schooling(7% and 15%) (ABS, 1989).
Experience of breast cancerPersonal experience of breast cancer was very low.Only 1 % of the sample (23 women) had actually hadbreast cancer, although 17% of them had found a
breast lump on one occasion and a further 8% ontwo or more occasions.
Beliefs and behaviours immediately afterBSE trainingIntentions to do BSEStated intentions to do BSE were assessed using twoitems, each with seven interval scaled responseoptions. Following Ajzen and Fishbein (1980), thefirst question was cast as a subjective probability ofaction and asked ' 'How likely is it that you will doBSE monthly from now on?". In response to thisquestion over one-third of the participants (35%)responded that they were "certain to do BSEmonthly", a further 40% were "very likely to doBSE monthly" and another 22% were "likely to doBSE monthly", leaving only 3% whose responsesindicated that they were relatively unlikely to do BSEmonthly. The second question, somewhat later in thequestionnaire, asked participants how often theyexpected to do BSE within the next 12 months.Responses to this question reflected equally strongintentions to do BSE, with over half the sampleindicating that they would do it 11 —12 times (61 %),a further 21% about 9-10 times and another 9%5—6 times. The remaining 9% forecast less regularpractice.
Self-efficacySelf-efficacy (Bandura, 1977; Stretcher et al., 1986)was assessed at Time 1 using the summed score oftwo questions, one assessing confidence in doingBSE properly and one focusing on perceptions ofBSE performance. The mean scores of 4.5 and 4.6on the two seven-point scales (scored 1 -7) indicatedthat the respondents had a moderate level of self-confidence in relation to BSE practice. In fact, only7% of participants suggested that they were 'lackingconfidence' in their ability to do BSE.
Behavioural beliefs about BSEFour beliefs derived from previous work (Hill et al.,1985) were presented as personalized statements:doing BSE would (i) increase the chances of detect-ing breast cancer early, (ii) mean that any breastcancer found would be curable, (iii) give a sense of
relief to find that nothing was wrong and (iv) stirup worries about breast cancer. Participantsresponded using a seven-point scale ranging from"extremely unlikely" (scored 1) to "extremelylikely" (7). The three positive beliefs were held fairlyfirmly: that doing BSE increases the chances ofdetecting cancer in the early stages (mean=6.1); thatdoing BSE would give a sense of relief when it isfound that nothing is amiss (6.0); and that earlydetection increases the curability of cancer (5.7).Fortunately, participants tended not to believe thatdoing BSE would stir up worries about breast cancer(2.8). To obtain a composite score for use in lateranalyses, the scoring on the negative item wasreversed, then the item scores were summed.Barriers to doing BSE
The list of perceived barriers to action previouslydeveloped by Hill et al. (1985) were presented in theTime 1 questionnaire as possible barriers to doingBSE: forgetting, laziness, lack of time, fear offinding a lump, fear of losing a breast and denial ofthe threat of breast cancer, expressed as "it couldn'thappen to me". The potential impact of each of thesebarriers was assessed using a four-point scale rangingfrom "no obstacle" (scored 0) to "extreme obstacle"(3). The most important single barrier was forget-ting (0.9) followed by laziness (0.7), lack of time(0.4), fear of losing a breast (0.5), fear of findinga lump (0.4) and denial of the threat of breast cancer(0.3).Perceived efficacy of treatment
It was assumed that women who were overlypessimistic about the results of treatment of breastcancer would be unlikely to do BSE, so the programaimed to portray treatment as effective whendiagnosis was early. After the program, perceptionsof breast cancer treatment results reflected a highlevel of confidence, in that 62% of participants sawthe success of medical treatments as "good" or"extremely good", 32% as "fair", 4% as "poor"and only 1% as "very poor".Perceived susceptibilityBreast cancer was only seen as a serious threat by8% of participants who rated their chance of gettingbreast cancer as "high", while 52% saw their
chances as "moderate", 36% as a "small chance"and 4% as "no chance". It should be noted that theseestimates were made after receiving informationabout the risk factors associated with breast cancer,that is, statistics which show that one in 15 Australianwomen will have breast cancer, with it being moreprevalent among older women (Giles, 1987).
Use of cues to BSEWomen were able to purchase cassettes at a nominalcost and were provided with calendars to remindthem to do BSE. About a quarter (27%) of thewomen purchased a cassette.
Representatives of the Time 2 sampleTo assess the representatives of the sample of womenincluded at Time 2, the responses of those includedin the Time 2 sample and those not included in theTime 2 sample were compared for selected itemsfrom the Time 1 questionnaire. There were nosignificant differences between the two samples onthe demographic variables: age, marital status andhighest level of formal education. There were nodifferences between the two samples in their meanscores on the measures of perceived self-efficacy inrelation to BSE practice, susceptibility to breastcancer, treatment efficacy of breast cancer, practicalbarriers or emotional barriers to doing BSE, but thewomen who completed both the Time 1 and Time2 questionnaires were slightly more positive in theirstated intentions to do BSE, suggesting that those whowere included in the Time 2 sample were slightlymore motivated to practice BSE. However, thisdifference may realistically be described as barelysignificant, but not substantive.
Effect of the program on frequency ofdoing BSEThere was substantial support for the first hypothesiswhich predicted that BSE practice would be morefrequent at Time 2 than at Time 1 (see Table I). AtTime 1 only 15% of the sample could be describedas regular BSE practicers in that they did BSEmonthly, although a further 29% could be describedas adequate practicers, in that 14% had done BSEevery 2 months and 15% every 3 months. Nearly
half the sample (46%) did very little BSE: 17% hadnever done BSE; 14% had done it at some time, butnot in the last 6 months, and 25% had done it oncein the past 6 months.
At Time 2 there was a dramatic increase in thefrequency with which most women reported that theyhad done BSE (see Table I). Although only 22% ofthem complied with the recommendations to do BSEmonthly, most women (94%) said they had done BSEat least once in the year since attendingMammacheck: 17% had done it 9 -10 times, 17%7 - 8 times, 22% 5 - 6 times, 17% 3 - 4 times andonly 6% as little as once or twice. Table I also showsthat BSE practice at Time 2 was strongly related toreported BSE practice at Time 1 (x2 = 274.49, 25d.f., P<0.001). The women who reported someBSE practice before attending Mammacheck weremore likely to increase the frequency with which theypracticed BSE in the year after attending the trainingsession.
Barriers to doing BSEAdditional barriers were included in the Time 2 ques-tionnaire: embarrassing nature of BSE, lack ofconfidence in doing BSE properly and thecomplicated nature of BSE. These were included inthe follow-up questionnaire as experience with theprogram suggested that they might be important.Greater experience of doing BSE during theintervening 12 months appeared to influence theperception of the practical barriers. At Time 2 eachof the three practical barriers was seen as being agreater obstacle to doing BSE than at Time 1: therate of forgetting increased from 0.9 to 1.1 {t=5.39,1133 d.f., P<0.00l); lack of time increased from0.4 to 0.6 (t= 5.48, 1132 d.f., /><0.001); lazinessincreased from 0.7 to 1.7 (t = 26.49, d.f. = 1132,P< 0.001). However, none of the ratings of theemotional barriers changed significantly. The addi-tional barriers assessed at Time 2 indicated thatneither embarrassment (mean = 0.1) nor thecomplicated nature of BSE (0.2) were major barriersto BSE practice for many women.
It would be anticipated that different techniqueswould be required to reduce barriers in each of thethree areas: practical, emotional and cognitive.1
Drawing on the work of Folkman and Lazarus(1980), it would be expected that the practical andcognitive barriers would be associated with problem-focused appraisals and would require some type ofproblem-focused coping strategy, while the emotionalbarriers would be associated with emotion-focusedappraisals and coping strategies. Rather than treatthe barriers as a single construct, three scores werederived by summing the scores of the responses ofthe items loading on each of the two factors assessedat Time 1 and the three factors assessed at Time 2.Cues to action
Recall of having the items and their frequency of usewere assessed at Time 2. Of the 298 women (27%)who purchased a cassette, 76% had used it. Sixty-nine per cent of women recalled having been givena calendar, but 43% of these women did not use itat all.
Factors facilitating and inhibiting BSEpracticePredictors of BSE practiceBy following up the women 12 months after theyattended training sessions, it was possible to evaluatethe factors that predicted actual practice of BSE. Aregression analysis was completed using reportedBSE practice as assessed at Time 2 as the dependentvariable and seven independent variables: previousBSE practice, self-efficacy, benefits, practicalbarriers, emotional barriers, susceptibility and treat-
Table
Time 2
1 -2 times3-4 times5-6 times7-8 times9-10
times11-12 times
I. BSE practice at Time 2 as a JunctionBSE practice at Time 1
Predictors of BSE practiceprevious BSE practiceintentionspractical barriersself-efficacysusceptibilitytreatment efficacyemotional barriersbenefits
Predictors of BSE practiceprevious BSE practicepractical barriersself-efficacysusceptibilitytreatment efficacyemotional barriersbenefits
Correlates of BSE practicepractical barriersself-efficacycues to rememberingemotional barrierscognitive barriers
Table II. Summary of regression analyses
Total sample
r
0.38*0.31*
-0.23*0.19*
-0.040.10*
-0.070.03
R2=0.23F (8,1043) = 39.76*
(omitting intentions)0.38*
-0.23*0.19*0040.10*
-0.070.03
«2=0.21F(7,1044) = 39.28*
-0 .41*0.36*0.16*0.08
-0.17*R2=0.27F(5.818)=61.06*
0
0.32*0.18*
-0.15*0.11*
-0.050.03
-0.030.01
0.35*-0.19*
0.14*-0.02
0.040.020.01
-0.38*0.31*0.09*0.050.04
Non-practicers
r
-0.22*
-0.19*0.14
-0.090.06
-0.050.10
R2=0.09/r(7,314)=4.37*
--0.19*
0.14-0.09
0.07-0.05
0.16*/?2=0.07F(6,315)=3.85*
-0.28*0.37*0.19*
-0.06-0.15*
R2=0.22F(5,247) = 14.02*
-0.16*
-0.13*0.08
-0.11-0.01-0.03
0.08
--0.18*
0.12- 0 09
0 01-0.01
0.10
-0.28*0.35*0.110.040.03
*/><0.001.
ment efficacy. The results are presented in the firsttwo numerical columns of Table II. Results of theanalyses of data for initial non-practicers of BSE arepresented in the two right-hand columns and will bediscussed later. The table includes all variablesentered into the analyses. The asterisks indicate thosevariables which made a significant contribution tothe final equation.
Four variables accounted for 23% of the variance.The most important predictor was previous BSEpractice (beta = 0.32), but intentions to practice BSE(0.18), practical barriers (-0.15) and self-efficacy(0.11) also entered into the equation. Althoughperceived treatment efficacy correlated significantlywith BSE practice, it did not enter into the equation.
The analysis was repeated controlling for age andpersonal experience of breast cancer, both in termsof having had breast cancer and of having a primaryrelative who had breast cancer. These variables werepartialled out in the first step of a regression analysis,and then the remaining variables were entered in thesecond step. The variables entered at the first stepdid not yield a significant value (F=3.17, NS),showing that they did not influence BSE practice.The remaining variables still accounted for the sameproportion of the variance.
The initial analysis was repeated again omittingintentions from the predictor set. The remainingvariables accounted for 21% of the variance. Thesignificant predictors were similar to those in the
previous analysis: previous BSE practice (beta =0.35), practical barriers (—0.19) and self-efficacy(0.14).Correlates of BSE practiceParticipants' perceptions of their self-efficacy and theroles of practical and emotional barriers mightchange as they have greater experience of BSE.Hence another regression analysis was completed toidentify the correlates of BSE practice as assessedat Time 2. The set of independent variablescomprised the scores on the measures of practicalbarriers, emotional barriers, cognitive barriers, self-efficacy and use of aids to remembering BSE, all asassessed at Time 2. The dependent variable wasBSE practice as assessed at Time 2. Twenty-sevenper cent of the variance was explained by three ofthe five variables: practical barriers (beta= -0.38),self-efficacy (0.23) and cues to remembering to doBSE (0.09). Neither the emotional nor cognitivebarriers entered into the equation.
Given the major role of the practical barriers ininhibiting BSE practice, a further regression analysiswas completed to identify the major barrier. Thedependent variable was BSE practice as reported atTime 2 and the independent variables were the scoreson the three separate practical barriers as assessedat Time 2: forgetting, laziness and lack of time.Seventeen per cent of the variance was explained:R2 = 0.17, F(l,1071)=216.76, P<0.00\. Theonly variable to enter the equation was forgetting.
Non-practicers of BSEAs many of the women who attended theMammacheck sessions may have been attending tofurther their skills and knowledge about BSE ratherthan to embark on a totally new health behaviour,a further set of analyses was completed using the sub-set of 358 women who could be described as non-practicers of BSE at Time 1 to identify the factorsrelating to their BSE practice. The results of theseanalyses are also presented in Table II. For the non-practicer sub-sample, the predictors accounted forrelatively little of the variance (7%), partly due tothe absence of the most important single predictor,previous BSE practice.
Discussion
Effectiveness of the programPre-program measures were not included. Particip-ants were selected for inclusion in the sample whenthey attended a training session and hence could notbe assessed prior to the session. A request tocomplete a questionnaire at the beginning of thesession was not appropriate, having the potential toirritate or alienate women who were already feelingself-conscious. It would be unethical to form anequivalent control group by denying them the oppor-tunity to learn BSE for 12 months when they arewishing to learn it. Because there were no pre-program measures, it is not possible to be certainthat the program produced the intended changes inhealth benefits. However, a comparison of the scoreson the benefits and barriers items assessed for thesample at the conclusion of the training sessions withthose of a comparable control sample reported in aprevious study by Hill et al. (1985) suggests that theMammacheck program did have an educational effectin that Mammacheck participants perceived greaterbenefits and fewer barriers associated with BSEpractice than the women not trained in BSE.
Certainly, the program was effective in greatlyincreasing the frequency with which most womenreported that they had done BSE in the preceding12 months, although only a small proportion ofwomen complied with the recommendation that theyshould do BSE regularly once a month. However,as most of the participants (94%) had done BSE atleast three or four times during the year, this shouldbe sufficiently frequent for them to obtain objectivebenefit (Feldman et al., 1981). Further, the level ofcompliance in this study is comparable with thatreported for a more labour-intensive program inwhich hospitalized women were each provided with30 min of individualized instruction, as well as avideo and printed materials (Roche and Gosnell,1989), indicating the effectiveness of this moreeconomical mode of training. Although it may bepreferable to provide individual tuition (Haran et al.,1987), in the absence of the resources required forsuch intensive training, the Mammacheck programis greatly increasing BSE practice.
Although the intervention was very effective ingreatly increasing the frequency with which thesewomen were practicing BSE, it failed to attain thefrequency of practice advocated. The failure toachieve monthly BSE practice may be partlyattributed to the single session character of theprogram. Overseas research suggests that frequencyof BSE practice increases incrementally after thefirst, second and third learning opportunities (Baineset al., 1986). Higher levels of BSE practice mightbe obtained by increasing competence through theparticipation in further training sessions. This maybe particularly true for the women who had not prac-ticed BSE prior to attending the Mammachecksession, women for whom self-efficacy was animportant factor in predicting BSE practice.Certainly, the opportunity to attend an additionalsession is markedly more economical than the provi-sion of individual training.
Intentions as predictors of BSE practiceThe major predictors of BSE practice were previousBSE practice, intentions to practice BSE, practicalbarriers, self-efficacy and perceived treatmentefficacy. However, the hypothesized role of statedintentions to do BSE as the major predictor of BSEpractice was not substantiated. Although intentionsto practice BSE was shown statistically to be arelatively important contributor to the equationpredicting BSE practice, the omission of this variablefrom the predictor set only served to reduce theamount of explained variance from 24 to 21 %. Therole of the other predictors remained, with the mostimportant predictor being previous BSE practice,supporting the common finding that previousbehaviour is an excellent indicator of futurebehaviour. Practical barriers and self-efficacycontinued to be significant factors.
Correlates of BSE practice a year afterMammacheckThe importance of the practical barriers is highlightedby their major role as a correlate of BSE practicea year after attending the Mammacheck session.When assessed at Time 2 concurrently with BSEpractice, practical barriers accounted for a large 17%
of the variance. This evidence for the major role ofperceived practical barriers in inhibiting BSE prac-tice and the relatively minor role of perceivedbenefits in facilitating such practice is consistent withother recent research findings (Champion, 1984,1985; Rutledge, 1986; Hill and Shugg, 1989). It isinteresting to note the important contribution ofperceived practical barriers, despite their absoluterating being relatively low. Although the possibilityexists that expressed practical barriers are a coverfor more deep-rooted psychological barriers, wefound no evidence of this in the present sample ofwomen who were presumably motivated to learnBSE.
The identification of forgetting as the greatestsingle obstacle to doing BSE highlights the import-ance of finding techniques to increase the salienceof BSE so more women will remember to do it. Ata practical level motivational factors need to besupplemented by cues to action. In the present study,women were provided with calendars and stickersto serve as reminders to do BSE. Nearly half thewomen did not use these materials, and althoughothers claimed to have used them, their use did notcontribute to reported BSE practice. Other data castsdoubt on the utilization of calendars. In a recentCanadian study assessing their use as reminders todo BSE, Baines et al. (1988) found that calendarsdid not contribute to an improvement in BSEcompliance and consequently they were droppedfrom usage. In a comparison of the relative utilityof BSE calendars (self-management) and postcardreminders (external cues), Grady (1984) found thatpostcard reminders led to a significant increase incompliance, and calendars did not, but BSE prac-tice declined sharply when the postcards ceased. Theabsence of a discernible benefit arising from the useof calendars in both this and other research suggeststhat some more effective means of reminding womenis required. Although this might involve a systemof personal reminders, it will only be effective if itis able to continue on a long-term basis.
Possibly the most effective reminder is a personalone from each woman's general practitioner whenshe visits either for a regular check-up or for someother reason. Certainly, there is evidence suggesting
that the frequency of BSE practice can be increasedthrough the integration of regular BSE instructionand evaluation into regular medical examinations(Baines et al., 1986), but this assumes the coopera-tion of general practitioners. Although it was difficultto obtain their cooperation in distributing the inform-ation about Mammacheck (Rassaby et al., 1991),their continuing support in encouraging patients todo BSE should be fostered. An additional, long-termtechnique may be to engage the services of theannouncers on selected popular women's radioprograms to offer reminders, to increase the use ofconspicuously placed posters, or to conduct extens-ive mass media advertising campaigns (Hill et al.,1982).
The relatively important association of self-efficacywith both intentions to practice BSE and with sub-sequent BSE practice is inconsistent with Lauver's(1987) conclusion that there is a lack of empiricalsupport for the construct of self-efficacy in explainingBSE practice. Rather, it supports the view of Haranet al. (1987) that a woman's confidence in hercompetence to do BSE is an important contributorto continued BSE practice. The fact that this factorwas of considerable importance among the womenwho had not practiced BSE prior to attendingMammacheck indicates the need for careful trainingin BSE techniques to further develop women'sperceived self-efficacy.
Summary and conclusions
In this study we made eclectic use of behaviourtheories to structure and evaluate a widelydisseminated BSE teaching program. Although wedid not set out to formally test specific theories, somesupport was obtained for those providing the basisof the program. When women left the trainingsession with strong self-efficacy beliefs, positivebeliefs in the benefits of BSE and a lack of concernabout potential emotional barriers to BSE practice,long-term BSE practice was more likely. However,if the teaching program failed to overcome the prac-tical barriers, long-term BSE practice was unlikely.These findings not only direct attention to those parts
of the program that need most emphasis, but illustratethe value of theory in formulating and evaluating theteaching programs.
Notes
1. To evaluate our pre-conceived assumptions that the barrierscan be classified into practical, emotional and cognitive barriers,the scores of the items used in the Time 2 questionnaire weresubmitted to a principal components factor analysis. Threefactors accounted for 61 % of the variance. The patterns wereconsistent with our conceptualization of barriers as falling intothree categories: practical, emotional and cognitive. Using acut-off point of 0.6, the first factor loaded on three itemsrepresenting practical barriers: lack of time (loading 0.80),laziness (0.80) and forgetting (0.72). The second factor loadedcognitive items: lack of confidence in ability to do BSE (0.85)and the complicated nature of BSE (0.81). The remaining twocognitive items: lack of confidence in ability to do BSE (0.85),and the complicated nature of BSE (0.81). The remaining twobarriers did not load on any of the factors: fear of finding alump and the embarrassing nature of BSE. The high loadingsof each of the items weighting on each factor and the failureof any item to load on more than one factor can be taken asproviding statistical support for the theoretically-basedassumptions.
References
ABS (1989) Statistical tables provided by the Australian Bureauof Statistics, Canberra.
Ajzen.I. and Fishbein.M. (1980) Understanding Attitudes andPredicting Social Behaviour. Prentice-Hall, Englewood Cliffs,NJ.
Baines.C.J. (1988) Breast self-examination: The known and theunknown. In Day.N.E. and Miller.A.B. (eds), Screening forBreast Cancer. VICC Workshop, Hans Hauber, Helsinki.
Baines.CJ., Wall.C, Risch,H.A., Kuin.J.K. and Fan,I.J. (1986)Changes in breast self-examination behaviour in a cohort of 8214women in the Canadian National Breast Screening Study.Cancer, 57, 1209-1216.
Baines.C.J., Krasowski.T.P. and Wall.C. (1988) Incentives forbreast self-examination. Cancer Detection and Prevention, 13,109-114.
Bandura.A. (1977) Self-efficacy: toward a unifying theory ofbehavioural change. Psychological Review, 84, 191—215.
Becker,M.H. and Maiman.L.A. (1975) Sociobehavioural deter-minants of compliance with health and medical care recommen-dations. Medical Care, 13, 10-24.
Champion,V.L. (1984) Instrument development for health modelconstructs. Advance in Nursing Science, 6, 73-85.
Champion,V.L. (1985) Use of the health belief model in deter-mining frequency of breast self-examination. Resarch in Nurs-ing and Health, 8, 373-379.
Champion,V.L. (1987) The relationship of breast self-examinationto health belief model variables. Research in Nursing and Health,10, 375-382.
FeldmanJ.C, Carter.A.C, Nicastri,A.D. and Hosat,S.T. (1981)Breast self-examination, relationship to stage of breast cancerat diagnosis.Cancer, 47, 2470-2745.
Folkman.S. and Lazarus,R.S. (1980) An analysis of coping in amiddle aged community sample. Journal of Health and SocialBehaviour, 21, 219-239.
Giles,G.G. (1987) Victorian Cancer Registry 1983 StatisticalReport. Anti-Cancer Council of Victoria, Melbourne.
Grady.K. (1984) Cue enhancement and the long-term practice ofbreast self-examination. Journal of Behavioural Medicine, 7,191-205.
Haran,D., Pendleton.L.L. and Hobbs.P. (1987) Women's reac-tions to the teaching of breast self-examination: a qualitativestudy using group interview techniques. Public Education aboutCancer: Recent Research and Current Programmes. UICCTechnical Report Series, Vol. 80. International Union AgainstCancer, Geneva.
Hill,D., Rassaby,J. and Gray,N. (1982) Health education aboutbreast cancer using television and doctor involvement. PreventiveMedicine, 11, 43 -55 .
Hill,D.J, Gardner,G. and RassabyJ. (1985) Factors predisposingwomen to take precautions against breast and cervix cancer.Journal of Applied Social Psychology, 15, 59—79.
Hill,D.J. and Shugg.D. (1989) Breast self-examination practicesand attitudes among breast cancer, benign breast disease andgeneral practice patients. Health Education Research: Theoryand Practice, 4, 193-203.
Hill.D., White.V., JoUey.D. and Mapperson.K. (1988) Breast self-examination: Is it beneficial? British Medical Journal, 197,271-275.
Hirst.S., Rassaby,J. and Walsh.R. (1989) Mammacheck Train-ing Manual. Anti-Cancer Council of Victoria, Melbourne.
Hobbs.P. (1986) The behavioural aspects of breast and cervicalscreening. Radiography, 52, 287—290.
Lauver.R.C. (1987) Theoretical perspectives relevant to breast self-examination. Advances in Nursing Science, 9, 16—24.
Mant,D., Vessey,M.P., Neil,A., McPherson.K. and Jones,L.(1987) Breast self-examination and breast cancer stage atdiagnosis. British Journal of Cancer, 55, 207-211.
Prentice-Dunn,S. and Rogers.R.W. (1986) Protection MotivationTheory and preventive health: Beyond the health belief model.Health Education Research, 1, 153-161.
Rassaby,J., Hirst,S., Hill.D.J., Bennett,R. and Clarke.V. (1991)Introduction of a breast self-examination teaching program inVictoria, Australia. Health Education Research: Theory andPractice, 6, 291-296.
Roche,M. and Gosnell.D.J. (1989) Evaluation of a hospitalteaching program for breast self-examination. Patient Educa-tion and Counselling, 13, 31—41.
Rutledge,D.N. (1986) Factors related to women's practice of breastself-examination. Nursing Research, 36, 117—121.
Strecher,V.J., DeVellis,B.M., Becker.M.H. and Rosenstock,I.M.(1986) The role of self-efficacy in achieving health behaviorchange. Health Education Quarterly, 13, 73-70.
Received on November 9, 1989, Revised on June 6, 1990, AcceptedJuly 21, 1990
