This article was downloaded by: [Universitario Miguel Hernandez]On: 06 March 2012, At: 04:30Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Psychology & HealthPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/gpsh20

Predicting uptake of a routine cervicalsmear test: A comparison of the healthbelief model and the theory of plannedbehaviourAlison Bish a , Stephen Sutton b & Susan Golombok ca Department of Clinical Genetics, Guy's Hospital, London, SE1 9RTb Health Behaviour Unit, University College London, London, WC1E6BTc Family and Child Psychology Research Unit, City University,London, EC1V 0HB

Available online: 19 Dec 2007

To cite this article: Alison Bish, Stephen Sutton & Susan Golombok (2000): Predicting uptake of aroutine cervical smear test: A comparison of the health belief model and the theory of plannedbehaviour, Psychology & Health, 15:1, 35-50

To link to this article: http://dx.doi.org/10.1080/08870440008400287

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representationthat the contents will be complete or accurate or up to date. The accuracy of anyinstructions, formulae, and drug doses should be independently verified with primarysources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand, or costs or damages whatsoever or howsoever caused arising directly orindirectly in connection with or arising out of the use of this material.

Psychology and Health, ZOOO, Vol. 15, pp. 35-50 Reprints available directly from the Publisher Photocopying permitted by license only

0 2000 OPA (Overseas Publishers Association) N.V. Published by license under

the Harwood Academic Publishers imprint, part of The Gordon and Breach Publishing Group.

Printed in Malaysia.

PREDICTING UPTAKE OF A ROUTINE CERVICAL SMEAR TEST: A COMPARISON OF THE HEALTH BELIEF MODEL AND THE

THEORY OF PLANNED BEHAVIOUR

ALISON BISH'?*, STEPHEN SUTTON2 and SUSAN GOLOMBOK3

'Department of Clinical Genetics, Guy's Hospital, London SEI 9RT

2Health Behaviour Unit, University College London, London WClE 6BT

3Farnily and Child Psychology Research Unit, City University, London ECI V OHB

(Received 24 July, 1998; in final form 24 January, 1999)

This paper reports on a study carried out to identify predictors of uptake of cervical screening among 142 women (59% response rate) in inner London. Two social cognition models were used. The Health Belief Model (HBM; Becker, 1974) and the Theory of Planned Behaviour (TPB; Ajzen, 1991) and in addition anticipated affect follow- ing non-attendance for screening was assessed. The TPB emerged as by far the superior model for predicting screening intentions, explaining 51% of the variance in comparison with only 4% explained by the HBM vari- ables. However, neither model was able to predict a significant amount of variance in uptake of screening three months later. Possible reasons for the poor prediction of this type of behaviour are discussed.

KEY WORDS: Health belief model, theory of planned behaviour, screening.

INTRODUCTION

Among countries in the European Community, the UK has the second highest incidence rate of invasive cervical cancer (Esteve, Kricker, Ferlay and Parkin, 1993). The detection of the pre-cursors of this cancer by use of a cervical smear test is the aim of a national UK screening programme, and understanding the factors that influence uptake of this type of screening is of vital importance. It is of particular interest to examine the predictors of adherence to screening programmes over time, as it is believed that it is repeated attend- ance which will ultimately lead to the reduction of mortality from this cancer (International Agency for Research on Cancer, IARC, 1986).

This paper describes a prospective study of uptake of cervical screening among a popu- lation of inner London women. In the present study, the decision to undergo screening is examined using two social cognition models; the Health Belief Model (Becker, 1974; Becker and Rosenstock, 1987) and the Theory of Planned Behaviour (Ajzen, 1991). These theoretical frameworks are briefly outlined below.

* Corresponding author. E-mail: alison@pollards.u-net.com

35

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

36 A. BISH ETAL.

Social Cognition Models

(a ) The Health Belief Model. The Health Belief Model (HBM) is the most widely used model of health behaviour. It has been applied in a number of contexts including use of pre- ventive screening, obtaining immunisations, compliance with medical regimens and response to illness symptoms (see Sheeran and Abraham, 1996, for a review; Harrison, Mullen and Green, 1992, for a meta-analysis). The HBM in its original form includes four main components: perceived susceptibility to, and perceived severity of, a disease; and per- ceived benefits and perceived costs of a preventive health action. Numerous different ver- sions of the HBM appear in the literature ( e g . some include a measure of self efficacy). In the present study a traditional version of the model using only the core constructs is employed, but the model is extended to include a measure of intention. Research suggests that the prediction of behaviour is improved when intention is included in the HBM (e.g., Quine, Rutter and Arnold, 1998). Intention may act as a mediator variable between the HBM variables and behaviour. This has been suggested previously by, amongst others, Calnan (1984) in relation to uptake of breast screening and King (1982) in relation to high blood pressure screening.

(b) Theory of Planned Behaviour. The Theory of Planned Behaviour (TPB) is an exten- sion of the Theory of Reasoned Action (TRA; Fishbein and Ajzen, 1975). It has been widely applied to a variety of behaviours - both health and non-health related (see Conner and Sparks, 1996, for a review and meta-analyses by Godin and Kok, 1996, and Sheppard, Hartwick and Warshaw, 1988). The TPB postulates that the proximal determinant of a behaviour is an intention to perform it. In turn, intentions are determined by three constructs: attitude towards the behaviour, subjective norm, and perceived behavioural control (PBC). Attitude toward the behaviour refers to the person’s overall evaluation of the behaviour, which may be positive or negative. Subjective norms involve perceptions of how other people think the individual should behave in relation to the particular behaviour in question (injunctive social norm). PBC refers to perceptions of personal control over carrying out the behaviour and also to perceptions of ease or difficulty in performing the behaviour .

Recently it has been suggested that the subjective norm construct could be extended to include measures of perceptions of how other people themselves behave - descriptive social norm (Cialdini, Kallgren and Reno, 1991). Arguably, though, the efficacy of this extension depends on the behaviour in question as some behaviours are less visible than others; having a smear test is a “private” behaviour. Both types of norms are explored in the present study with an expectation that injunctive norms will be more important than descriptive norms. Regarding PBC, Ajzen (1991) uses control to mean both internal and external constraints. However, it certainly seems possible that these may not be the same and therefore would not form one construct (see Terry and O’Leary, 1995). Internal con- straints seem closer to perceptions of self efficacy (Bandura, 1977) and to perceived costs of the behaviour which may affect the formation of intentions, whereas external constraints are perhaps more likely to reflect factors that may prevent a person from translating their intention into action. The possibility that the constructs are distinct, and that internal con- straints may be more influential for screening intentions and behaviour, is explored in this paper.

A criticism of social cognition models which has been particularly applied to the TPB is that they are unable to explain behaviour which may be under affective control as they do

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 31

not adequately take into account emotional factors in decision making. Individuals may be motivated to attend for screening to avoid feeling regret at not doing so since attending for a smear test is a behaviour which may lead to “bad news” (i.e., abnormal smear test result) or “good news” (i.e., normal smear test result). Consistent with this is the finding that efforts to encourage attendance for screening are more successful if women are given loss- framed messages - emphasising the risks of not having the screening - than gain-framed messages - emphasising the benefits of screening (Banks, Salovey, Greener and Rothman, 1995). The concept of anticipated affect stems from regret theory (Bell, 1981; Loomes and Sugden, 1982) and research suggests that it can be an important predictor of intentions in addition to attitudes to behaviour (e.g., van der Pligt, Zeelenberg, van Dijk, de Vries and Richard, 1998). The role of anticipated affect in predicting intentions to undergo a smear test and possibly to predict later behaviour has yet to be examined and is therefore invest- igated in this study as an extension to the TPB .

As can be seen from the above descriptions of these models, there is some overlap in the components they include. This has been noted amongst others by Sutton (1987) and Wein- stein (1993). Both have their theoretical basis in expectancy-value theory and are phrased in terms of the perceived consequences of action. Personal abilities to carry out a behaviour are explicitly included in the TPB (i.e., PBC) and arguably can be represented in the HBM in the costs element. The major differences between the models are that intentions and social norms are not included in the HBM as predictors of behaviour and the TPB does not include a measure of perceived threat of disease. Also the rules for combining constructs in the TPB are explicit, whereas there is no consensus regarding how the HBM constructs should be combined, if at all. A major aim of the present study was to establish whether a health specific model (i.e., the HBM) will be superior in terms of its ability to predict inten- tions and behaviour in this behavioural domain to a more general model (i.e., TPB). It has been argued that “it is uneconomical to entertain specific theories of health behaviour . . . unless the predictive success of these specific models is greater than that of general models” (Stroebe and Stroebe, 1995).

A number of studies have, in fact, compared the performance of models and attempted to combine constructs from these to achieve the best explanation of behaviour (e.g., Con- ner and Norman, 1994; Mullen, Hersey and Iverson, 1987; Oliver and Berger, 1979; Quine et al., 1998; Ronis and Kaiser, 1989). However, in the domain of cancer screening uptake, none of these studies included a measure of behaviour (Hennig and Knowles, 1990; Hill, Gardner and Rassaby, 1985; Seydel, Taal and Weigman, 1990; Seibold and Roper, 1979).

Uptake of cervical screening would seem to be a behaviour which readily lends itself to study with the application of decision making models. It is likely that owing to its non- habitual and infrequent nature individuals would make a decision each time about whether or not to be screened. In addition, participation in screening is voluntary; it is ultimately the individual’s choice as to whether or not she attends for a test. However, meta-analyses of applications of social cognition models to the prediction of intentions and behaviour have shown that screening behaviour is especially poorly predicted (e.g., Godin and Kok, 1996; Harrison et al., 1992). It should be noted, however, that the category of screening behavi- our used by both Godin and Kok (1996) and Harrison et al. (1 992) was very broad, including self examination behaviours, seeking medical care for symptoms, having mammograms, and asymptomatic check-ups for venereal disease. It has been pointed out that screening itself is not an homogeneous behaviour (e.g., Marteau, 1993) and that different types of screening may be predicted by different beliefs.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

38 A. BISH ETAL.

A review by Orbell and Sheeran (1 993) identifies a number of factors which may be associ- ated with cervical screening attendance. However, very few studies on the uptake of cervical screening have based their research design or the interpretation of their findings on a social cognition model, abd neither meta-analysis mentioned above included a study of uptake of cervical screening. Exceptions include Hill et al. (1985); Hennig and Knowles (1990); Man- dleblatt, Traxler, Lakin, Kanetsky, Kao and Harlem Study Team (1992); Mandleblatt, Traxler, Lakin, Kanetsky and Kao (1993); Murray and McMillan (1993); Orbell, Crombie and John- ston (1996); Orbell and Sheeran (1998); and Seibold and Roper (1 979). Some of these studies found that elements of the TPB and the HBM can predict intentions to be screened and past behaviour (e.g., Hennig and Knowles, 1990; Hill et al., 1985; Mandelblatt er al., 1992; Orbell et al., 1996). The TPB has not, thus far, been applied to uptake of cervical screening and there- fore the influence of perceived behavioural control on screening intentions and behaviour remains to be examined. Predictors of actual uptake of cervical screening have been under- researched; with the exception of Orbell and Sheeran (1998) most of the studies in this area have been cross sectional surveys with self reports of past behaviour. Such studies assume that the beliefs pre-dated the behaviour that they are supposed to predict. In fact, taking part in screening could, in itself, alter an individual's opinions of that behaviour, particularly if the behaviour is new to the individual. The current study, by contrast, used a prospective design to predict actual uptake of cervical screening. This allows more confident conclusions to be drawn concerning the relationship between cognitions and behaviour.

In summary, the aims of this study were: firstly, to identlfy the best predictors of inten- tions and uptake of a routine cervical smear test using the HBM and the TPB; secondly, to compare the ability of the two models to do this; and thirdly, to examine the role of anticip- ated negative emotion in predicting screening intentions and behaviour.

METHODS

Participants

The women who participated in the study were selected from two general practices in south-east London. Criteria for inclusion in the study were that the women were due to be asked to attend for a routine cervical smear test in the six weeks following their contact by the researchers.

A total of 283 questionnaires were sent to women identified on GP lists as being due for screening, together with an explanatory covering letter signed by a female GP and a free- post return envelope. Forty-two women were not eligible for the study, having left their address.' Thus, 241 questionnaires were sent to eligible women, One hundred and forty two questionnaires were completed before the women would have been sent an invitation letter to attend for screening. Two questionnaires were completed too late and were there- fore not included. The response rate was 58.9% (142/241).

The mean age of the women was 38.1 years (sd= 11.6). Most of the women (56%) were from social classes I or I1 [classified according to their current or previous occupation]. Eighty-three per cent of the women described themselves as white.

Although this percentage is quite high, representing 14.8% of the questionnaires sent, it transpired that the names of number of students who had been living in a hall of residence three years earlier when they registered with the general practice were included on the list. These women had all left the address and the questionnaires sent to them were returned by the Royal Mail.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 39

Measures

The measures included in the study questionnaire were developed from a pilot study and elicitation interview (see Bish, 1997, for details). Some HBM items were also adapted from a widely used reliable questionnaire developed by Champion (1984). TPB items were designed in accordance with the principle of correspondence (Fishbein and Ajzen, 1975; Ajzen and Timko, 1986), i.e. they were all phrased specifically as referring to attending for a smear test in the next three months to match the measure of behaviour. HBM items were not written in this way in order that new items would match those adapted from the previ- ously used questionnaire (Champion, 1984). Correspondence of measures is not a require- ment of the HBM and this issue is explored further in the discussion of this paper.

The components of the TPB and the HBM, and the measure of anticipated affect were assessed with multiple items which were summed to form scales. Most items used five- point bipolar scales with the exception of those used to assess PBC and anticipated affect which used four-point unipolar scales. The items were assessed for internal consistency using Cronbach’s a Only direct measures of attitude, subjective norm and PBC were assessed in the questionnaire to avoid the repetition of items which would have been covered by the items designed to measure HBM constructs (see Table 1 for items, Cronbach 06 and scoring details).

Table 1 Planned Behaviour, including Cronbach ma

Questionnaire items used to measure constructs from the Health Belief Model and the Theory of

Intentions (a= 0.86) If given the chance, do you intend to attend for a smear test in the next three months?

How likely is it that you will attend for a smear test in the next three months if given the chance?

Health Belief Model (strongly agree-strongly disagree) R = reversed item

Perceived susceptibility (a= 0.79) My physical health makes it likely that I will get cervical cancer My lifestyle makes it likely that I will get cervical cancer I believe that my chances of getting cervical cancer are high My chances of getting cervical cancer are small (R) I do not see myself getting cervical cancer in the next year (R) With my family history I am unlikely to get cervical cancer (R) There is a good possibility that I will get cervical cancer I worry a lot about getting cervical cancer I do not think that I am the sort of woman who would get cervical cancer Perceived severity (a= 0.61) Getting cervical cancer would interfere with my sex life If I got cervical cancer I would have problems which would last a long time If I got cervical cancer my whole life would change My feelings about myself would not change if I got cervical cancer (R) Getting cervical cancer would not he a problem for me (R) Perceived benefits (a= 0.71) There is nothing I can do to detect cervical cancer (R) I believe that a smear test will only find evidence of cervical cancer when it is too late to treat it (R) Cervical smear tests can detect abnormal changes before I would notice any symptoms If I have regular smear tests cervical cancer will be found before it is advanced Having a smear test would not give me peace of mind (R) Having regular smear tests is not a good idea (R) I have a lot to gain by having regular smear tests

(no definitely not-yes definitely)

(extremely likely-extremely unlikely).

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

40 A. BISH ETAL.

Table 1 (Continued)

I would be reassured about cervical cancer if I had smear tests regularly Cervical smear tests are no good at detecting cervical cancer in its early stages (R) Perceived costs (a= 0.78) I am very afraid of having a smear test Having a smear test is too inconvenient for me I am flustered whenever I have a smear test Getting a smear test does not interfere with my other activities (R) Getting a smear test is time consuming I don’t mind giving up my time to have a smear test (R) I am rarely embarrassed when I have a smear test (R) I find that smear tests are painful The way a smear test is performed causes me distress I am never made to feel uncomfortable when having a smear test (R) Theory of Planned Behaviour

Attitude (a= 0.85) Attending for a smear test in the next three months if given the chance would be: (5 items, extremely important,

Iqjunctive subjective norm (a= 0.87) Most people who are important to me would think I should attend for a smear test in the next three months if

Most people who are important to me would approve of me attending for a smear test in the next three months

Descriptive subjective norm Most women who are important to me attend for smear tests (strongly agree-strongly disagree). Self eficacy (a= 0.66) How difficult would it be for you to attend for a smear test in the next three months if given the chance? (4 point

If you were given the. chance, how confident are you that you could attend for a smear test in the next three

Perceived Behavioural Control How much control do you feel you have over whether or not you attend for a smear test in the next three months

Anticipated affect (a= 0.89) How would you feel if you did not attend for a smear test in the next three months when given the chance?

harmful, foolish, safe, good-extremely unimportant, beneficial, wise, unsafe, bad).

I am given the chance (strongly agree-strongly disagree).

if I am given the chance (strongly agree-strongly disagree).

scale, extremely dificult-not at all dificult)

months? (4-point scale, not at all conjident-extremely conjident)

if given the chance? (4-point scale, no control-complete control)

(4-items, 4-point scales, extremely tense, guilty, worried, regretful-not at all tense, guilty, worried, regretful)

a all 5-pint scales except where specified.

(a) Intentions. The women’s intention to be screened was assessed with two items.

(b) HBM. Nine items measured perceived susceptibility to cervical cancer and five meas- ured perceptions of its severity. Nine items measured the perceived benefits of cervical screening and ten measured perceptions of its costs.

(c) TPB. Five semantic differentials were used to provide a direct measure of attitude towards attending for a smear test in the next three months. Three items assessed subjective norms, two injunctive and one descriptive. It was found, as expected, that including the descriptive norm item reduced the reliability of the scale (from a= 0.87 to a= 0.74). It was therefore treated separately in the analysis. Three items assessed PBC. The items included measures of perceived control over attending for a test (external constraint), perceived dif- ficulty and perceived confidence (internal constraints). A reliability analysis revealed that the item for external control reduced the reliability of the scale (from a= 0.66 to a= 0.59)

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 41

and was therefore not combined with the two items measuring the ease and confidence with which a woman perceived she could attend for a smear test (i.e., self efficacy concepts). Subsequent analyses treat these variables separately.

(d) Anticipated affect. A measure of anticipated affect was derived from four items assess- ing how women anticipated they would feel if they did not attend for a smear test when invited to do so.

( e ) Past behaviour. Women were asked whether they had ever had a smear test and this information was cross checked with GP records.

Screening Attendance

Computerised GP records were checked to establish whether or not the women attended for a smear test at their own GP surgery or elsewhere in the three months following completion of their questionnaire.

RESULTS

Screening Intentions and Attendance

Only women who had ever had a smear test were included in the following analyses (n= 133) as there is evidence to suggest that predictors of first time and subsequent behavi- our may be different (e.g., Marteau, 1993; Norman and Conner, 1996). Only nine women had never been screened representing too small a group to be used in a separate analysis to explore predictors of first time attendance for screening.

One hundred and fourteen women (85.7%) reported that they would probably or def- initely attend for a smear test (scores of 2 to 4 on the intention scale - formed by summing the two intention items to give a possible range of scores of -4 to 4), 13 women (9.8%) reported they would probably or definitely not attend for a test (scores of -2 to -4) and six women were unsure (score of zero). Within three months of completing their question- naire, 69 women (5 1.9%) had attended for a smear test and 64 (48.1 %) had not.

Univariate Predictors of Intentions and Behaviour

Table 2 shows the means and standard deviations for the social cognition variables meas- ured in this study. Pearson's product moment correlation coefficients were calculated to explore the associations of the social cognition variables with intentions and behaviour and these are also reported in Table 2. Variables derived from the TPB tended to have the strongest relationships with reported intentions, in particular attitude towards the behavi- our ( r= 0.67) and injunctive subjective norms ( r = 0.54). In contrast, descriptive subjective norms were not correlated with intentions. Whilst perceptions of control over attending for screening were not correlated with intentions, the self efficacy measure was. Descriptive subjective norms and perceived behavioural control did not significantly increase the vari- ance explained by the models in the following multivariate analyses and are therefore not considered further. Variables derived from the HBM were less strongly related to intentions and in the case of perceptions of the severity of cervical cancer no significant correlation

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

P

w

Tab

le 2

M

ean

scor

es fo

r the

soc

ial c

ogni

tion

varia

bles

and

Pea

rson

’s co

rrel

atio

ns a

mon

g th

ese a

nd w

ith in

tent

ions

and

beha

viou

r

Scal

e M

ean

(sd)

Be

havi

our

Inte

ntio

n Su

scep

tibili

ty

Seve

rity

Be

nefit

s Co

sts

Attit

ude

Inju

nctiv

e D

escr

iptiv

e Se

lfef

icac

y Pe

rcei

ved

subj

ectiv

e su

bjec

tive

cont

rol

norm

no

rm

Inte

ntio

n 2.

5 (2

.3)

0.17

Su

scep

tibili

ty

23.7

(5.6

) -0

.02

0.28

**

Seve

nty

17.3

(3.4

) -0

.01

-0.1

3 -0

.27*

* Be

nefit

s 38

.6 (

4.4)

0.

15

0.18

* 0.

12

-0.1

1 co

sts

21.6

(6.3

) -0

.12

-0.1

9*

0.02

0.

25**

-0

.50*

**

Atti

tude

1.

4 (0.

6)”

0.22

**

0.67

***

0.27

**

-0.2

1*

0.37

***

-0.3

5***

In

junc

tive

2.6(

1.6)

0.

11

0.54

***

0.27

**

-0.0

7 0.

24*

-0.1

8 0.

71**

*

Des

crip

tive

1 .O (0

.7)

0.03

0.

08

0.29

**

-0.1

5 0.

26*

-0.0

8 0.

24*

0.34

**

Self

effic

acy

6.8(

1.3)

0.

09

0.49

***

-0.0

5 -0

.06

0.41

***

-0.4

3***

0.

30**

0.

19

0.24

* Pe

rcei

ved

cont

rol

3.4

(0.8

) 0.

09

0.13

-0

.07

-0.1

2 0.

28**

-0

.34*

**

0.28

**

0.14

0.

14

0.32

**

subj

ectiv

e nor

m

subj

ectiv

e nor

m

Ant

icip

ated

affe

ct

9.5

(3.6

) 0.

18

0.46

***

0.38

***

-0.1

5 0.

28**

-0

.20*

0.

67**

* 0.

63**

* 0.

36**

* 0.

18

0.21

*

a Si

nce t

here

was

som

e miss

ing d

ata o

n th

is va

riabl

e a m

ean

scor

e was

calc

ulat

ed an

d us

ed in

furt

her a

naly

ses.

*p<O

.O5;

**p

<O.O

l; **

*p<O

.OO

l. Tw

o-ta

iled

signi

fican

cele

vels.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 43

was found. For behaviour, the strongest and only significant correlation was observed for attitude towards screening ( r = 0.22). The correlation between intentions and behaviour was only marginally significant (r=0.17, p=O.O57). None of the HBM variables were found to be significantly correlated with later behaviour.

Multivariate Analyses of Predictors of Intentions and Behaviour

A series of hierarchical multiple regressions was carried out to identify the best predictors of intentions and uptake of cervical screening and to compare the ability of the theoretical models to explain these. Data from a subset of participants (n=96) who had no missing data on any of the variables were used in these analyses. For comparability with the analyses of prediction of intentions, linear regression was used to analyse predictors of behaviour. However, since the behaviour measure was dichotomous (i.e., attendancehon-attendance) the results were checked using logistic regression analysis. The results obtained were the same in both types of analyses.' Table 2 shows a correlation matrix of the modelled variables.

(a) HBM bused analyses. In the first analysis (see Table 3), intentions were regressed onto the four HBM variables. A non-significant proportion of the variance was explained (F=2.11; R2 = 0.08, adjusted R2 = 0.04). Perceived susceptibility was the only belief that significantly contributed to the variance explained (p= 0.25, p < 0.05), indicating that those women who feel susceptible to cervical cancer are more likely to intend to attend for a smear test.

In the next analysis (see Table 4, step 1) behaviour was regressed onto the HBM vari- ables. A non-significant proportion of the variance was explained (F=0.45; R2=0.02, adjusted R2=-0.02) and no variable significantly contributed to the variance. Adding intention to the HBM variables (see step 2 ) was expected to increase the variance explained in behaviour. However, again, a non-significant proportion of the variance was explained by the variables (F= 0.69; R2 = 0.04, adjusted R2 =-0.02), the change in R2 was not signi- ficant, and no individual variable made a significant contribution to the variance.

(b) TPB based analyses. Firstly (see Table 5, step l), intentions were regressed onto the TPB variables. The results of this analysis showed that attitudes towards screening were predictive of intentions (p= 0.60, p < 0.001) and a significant proportion of variance was explained. (F= 34.02; R2 = 0.53, adjusted R2 = 0.51; p < 0.001). Neither injunctive norm nor self efficacy made a significant contribution to the variance explained. Secondly (see Table

Table 3 using HBM variables to predict intentions

Results of the multiple regression analysis

Variable P R2 Adjusted R2

Benefits 0.03 costs -0.15 Susceptibility 0.25* Seventy 0.01 0.08 0.04

*p<O.O5.

'The original -2 log likelihood of the constant only model was 132.91. The HBM variables reduced this to 129.26, a non-significant reduction (2 = 3.66 (5 ) . ns.). No variables were significant and 59% of women were correctly classified. The TPB variables reduced the -2 log likelihood to 127.31, a non-significant reduction (xz = 5.61 (4), ns.). No variables were significant and 59% of women were correctly classified.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

44 A. BISH ETAL.

5, step 2 ) anticipated affect was added. The equation was significant (F=25.20; R2=0.53, adjusted R2 = 0.5 1 ; p < 0.001) but there was no significant change in R2. As in step 1, an independent predictive effect on intentions was obtained for attitude towards the behaviour @= 0.60, p < 0.0001) showing that controlling for affect in the equation did not modify the effect of attitude on intentions. Anticipated affect had no significant independent effect.

Table 4 Results of the multiple regression analysis using HBM variables and intentions to predict behaviour

Step Variable P R2 AdjustedR’ ChangeinR’

1 Benefits costs Susceptibility Seventy

2 Benefits costs Susceptibility Severity Intention

0.09 4 . 0 6 -0.01 4 . 0 1 0.02 4 .02 nla

0.09 -0.04 4 . 0 4 4 . 0 1

0.14 0.04 -0.02 0.02

Table 5 Results of the multiple regression analysis using TPB variables and anticipated affect to predict intentions

Step Variable P R2 Adjusted Change in R2

1 Attitude 0.60* Injunctive Norm 0.13 Self efficacy 0.07 0.53* 0.51 nla

Injunctive Norm 0.13 Self efficacy 0.07 Affect 0.01 0.53* 0.5 1 0.00

2 Attitude 0.60*

* p < 0.001.

Table 6 and anticipated affect to predict behaviour

Results of the multiple regression analysis using TPB variables

Step Variable p R2 Adjusted RZ Change in R2

1 Intention 0.15 0.03 0.01 nla

2 Intention 0.02 Attitude 0.21 Injunctive Norm 0.02 Self efficacy -0.13 0.06 0.01 0.03

3 Intention 0.02 Attitude 0.22 Injunctive Norm 0.02 Self efficacy -0.13 Affect -0.01 0.06 0.01 0.00

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 45

Behaviour was then regressed onto the TPB components and affect. Intention was added at the first step as it is considered to be the proximal determinant of behaviour (see Table 6, step 1). It was not significantly predictive of behaviour, and the proportion of variance explained was not significant ( F = 2.10; RZ =0.03, adjusted R2=0.01). Attitude, subjective norm and self efficacy were added at the second step, the equation was still not significant (F=1.36; Rz=0.06, adjusted R2=0.01) and no significant change was observed in Rz. Finally affect was added (see step 3). The equation was not significant ( F = 1.08; R”0.06, adjusted Rz = 0.01) and no significant change was observed in R2.

DISCUSSION

Comparison of Social Cognition Models

The study highlights that whether or not a measure of intention is included in the HBM, the ability of the model to predict behaviour is as good - or bad - as that of the TPB in that neither model was able to account for a significant proportion of variance in behaviour. However the study also shows that for the prediction of intentions the TPB explained considerably more variance in intentions than the HBM - 5 1 % of the variance in compar- ison to 4%. Therefore despite criticisms of the TPB for lacking emotional content, for example in a measure of perceived threat, it was found that this model performed better than the health specific HBM in predicting screening intentions.

In contrast to this study’s findings, the few studies which have compared the HBM with the TRA (none have compared the HBM with the TPB for uptake of cancer screen- ing) have found that variables derived from the HBM are able to account for more vari- ance in intentions to have a smear test (e.g., Hennig and Knowles, 1990; Hill er al., 1985). It is therefore debatable whether the better prediction of intentions from the TPB in the current study results from conceptual differences between the theories or differences in the way the model components were operationalised. On a conceptual level, the study provides evidence that although the TPB is not specifically designed to predict health behaviour, whereas the HBM is, its components predict more variance in intentions. Equally, the study shows that direct measures of attitudes are better at predicting inten- tions than the more detailed belief based measures used in the HBM. It could therefore be argued, as others have, that the TPB is a more economical model (e.g., Quine et al., 1998).

On a methodological level, however, the results show that specific measures are effect- ive for predicting specific intentions to carry out a behaviour, and the poorer prediction from the HBM in the current study may be due to the non-specific measures used rather than to the type of components it contains. An alternative explanation of the present find- ings then may be that the poorer prediction of intention afforded by the HBM variables could be attributable to the lack of correspondence between measures of its components and the measure of intention. This has implications for how the HBM should be measured as its application rarely follows the principle of correspondence (Ajzen and Timko, 1986). It remains possible that if the model was operationalised in this way it would be as effective as the TPB in predicting cervical screening intentions. However, this would apply chiefly to the benefits and costs elements of the HBM; perceived susceptibility and perceived severity cannot easily be operationalised in this way.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

46 A. BISH ETAL.

Predicting Intentions to have a smear test

The study examined the relative contribution of attitudinal, normative, control and threat influences on intentions to have a smear test. The results showed that women’s attitudes to the behaviour are the most important influence on intentions to be screened. A measure of descriptive subjective norm was not related to intentions or behaviour, showing that, as suggested, as the behaviour in question is not observable by others, women’s intentions and behaviour may not be so influenced by the perception of what other people are doing. Despite the significant correlation with intentions, when examined multivariately per- ceived self efficacy did not significantly contribute to the prediction of intentions, and perceived behavioural control was not correlated with intentions. It can therefore be concluded that for uptake of cervical screening the TRA would be as useful a model as the TPB. This is in contrast to a number of studies of other behavioural domains which have found that perceived behavioural control significantly increases the prediction of intention (e.g., Ajzen and Driver, 1992; Beale and Manstead, 1991).

Perceptions of the severity of cervical cancer were not predictive of intentions and although perceptions of susceptibility were correlated with intentions, and significant in the multiple regression, the overall model was not significant. These results could be inter- preted to imply that perceptions of threat are not important in the decision to undergo cervical screening. It is possible that this is because the women were familiar with this screening and it has been argued that threat perceptions may only be important for predict- ing the performance of novel behaviours (Rogers, 1983; Schwarzer, 1992). When an indi- vidual has carried out a behaviour in the past a perception of threat may be unnecessary to motivate them to form an intention to perform the behaviour again. The finding that perceptions of the severity of cervical cancer are not important for predicting intentions and behaviour is consistent with the literature. For example, Janz and Becker (1984) found perceptions of severity to be the least important HBM variable in their review of research using the model and Harrison et al. (1992) observed particularly low effect sizes for this variable in studies of uptake of screening.

For this group of women anticipated affect with respect to non-performance of the beha- viour was not as important in predicting intentions as attitudes towards the act itself. It is possible that this is due to the nature of the behaviour in question and the severity of the consequences of action or inaction. In previous research where anticipated affect has been important (e.g., Parker, Manstead, Stradling and Reason, 1992; Parker, Manstead and Stradling, 1995; Richard, van der Pligt and de Vries, 1995) the behaviours in question, i.e., committing a driving violation and having unsafe sex, may in themselves be pleasurable, in contrast to having a smear test which is not. Therefore for screening, attitudes towards undergoing the behaviour are likely to be more salient than anticipated affect with respect to not undergoing the behaviour. In addition, the consequences of unsafe sex and driving violations may be severe whereas not having a smear test does not preclude the opportunity of having one at a later stage, again diminishing the importance of the perception of affect following non-performance of the behaviour.

Predicting Screening Behaviour

Whilst most (86%) of the women who participated in the study reported that they would attend for a smear test, far fewer (52%) actually did attend within the three months follow- ing the reporting of their intentions. This is in line with the similar poor predictions of

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 41

screening behaviour reported in the literature ( e g , meta analysis by Godin and Kok, 1996; study of health screening uptake by Conner and Norman, 1994). Such findings suggest that social cognition models are inadequate to explain uptake of health screening.

There are a number of reasons why there may be a discrepancy between intentions and behaviour (Sutton, 1998), in particular screening behaviour. Firstly, intentions when assessed by questionnaire are a proxy measure of an actual decision. In the current study the wording of the intention measures was hypothetical “if given the chance”. A woman’s decision to attend for screening might be made after she receives an invitation letter (in the current study this would have been up to approximately 6 weeks after the measurement of intention), or the decision might be made when she is visiting a doctor on another matter and therefore would take place only a few minutes before having the test. Screening deci- sions are probably made afresh each time the opportunity arises and are unlikely to be an habitual response. Studies appIying social cognition models assume that participants have already made the decision or make it when they complete the questionnaire; however, this is arguably not the case. This assumption is a drawback of the way in which the models are measured, and may explain the relatively poor prediction of behaviour which is often observed.

Secondly, intentions may change after measurement, for example, due to the receipt of new information or the occurrence of unforeseen obstacles that prevent the behaviour being carried out. This may particularly occur if there is a long period of time between their measurement and the assessment of behaviour (Ajzen and Timko, 1986). Indeed it is pos- sible that, as the time to act approaches, the negative aspects of the behaviour may become more salient, causing intentions to alter. In the present study there was a gap of up to three months.

If intentions to attend for screening are not very strong, it certainly seems plausible that they might change. However, in the current study many women expressed strong intentions but still did not carry them out. A failing of the social cognition models is to distinguish between inclined actors (individuals who intend to act and do) from inclined abstainers (individuals who intend to act but do not). For example, Orbell and Sheeran (1998) carried out an investigation of uptake of cervical screening based on Protection Motivation Theory (PMT; Rogers, 1975 and 1983). They found that whilst PMT was able to distinguish women who attended for cervical screening from those who did not, the theory could not identify the processes by which an intention to act was translated into behaviour.

Recent theoretical frameworks have been devised to explore this ( e g , Bagozzi and Warshaw’s 1990 theory of trying and Gollwitzer’s 1993 implementation intentions). These perspectives argue that further cognitive activity is required to translate an intention into action. Individuals who have furnished their goal intentions (“I intend to do x”) with imple- mentation intentions (“I intend to do x when situation y is encountered”) - i.e. those who have carried out more “cognitive work” - should be, according to the theory, more success- ful in achieving their goal, e.g., having a smear test.

Gollwitzer argues that implementation intentions are particularly important when the behaviour is not habitual and where there is still conflict between acting here or there, now or later etc. Therefore it is plausible that implementation intentions would be a useful additional measure in social cognition models when these are being used to predict non- habitual and infrequent behaviours such as attending for screening. However, there has been little empirical work in the health arena to test this idea and future research examin- ing the role of implementation intentions in predicting routine screening behaviour may prove to be fruitful.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

48 A. BISH ETAL.

CONCLUSIONS

This study of inner city women shows that a positive attitude towards having a smear test is the best predictor of an intention to be screened, being considerably more influential than a perception of a threat to health or perceived social pressure. The TPB emerged as superior to the HBM for explaining these screening intentions. However, before practical implica- tions can be drawn from the findings it will be necessary to examine ways of improving the prediction of behaviour from intentions and beliefs.

References

Ajzen, I. (1991) The theory of planned behavior. Organisational Behavior and Human Decision Processes, 50,

Ajzen, I. and Driver, B.L. (1992) Application of the theory of planned behavior to leisure choice. Journal of Leis-

Ajzen, I. and Timko, C. (1986) Correspondence between health attitudes and behavior. Basic and Applied Social

Bagozzi, R.P. and Warshaw, P.R. (1990) Trying to consume. Journal of Consumer Research, 17, 127-140. Bandura, A. (1977) Self-efficacy: Towards a unifying theory of behavioral change. Psychological Review, 84,

Banks, S.M., Salovey, P., Greener, S. and Rothman, A.J. (1995) The effects of message framing on mammo- graphy utilization. Health Psychology, 14(2), 178-184.

Beale, D.A. and Manstead, A.S.R. (1991) Predicting mothers’ intentions to limit frequency of infants’ sugar intake. Journal of Applied Social Psychology, 21,409-43 1.

Becker, M.H. (Ed.) (1974) The Health Belief Model and personal health behaviour. Health Education Mono- graphs, 2(4) 324-508.

Becker, M.H. and Rosenstock, LM. (1987) Comparing social learning theory and the health belief model. In: W.B. Ward (Ed.) Advances in health education andpromotion (Vol .2, pp. 245-249). Greenwich CT: JAI Press.

Bell, D.E. (1981) Explaining utility theory paradoxes by decision regret. In: J. Morse (Ed.) Proceedings of the fourth international conference on multiple criteria decision making (pp. 28-39). New York Springer.

Bish, A. (1997) Predicting and explaining uptake of cervical screening: The role of social cognitions. Unpub- lished doctoral thesis. City University, London.

Calnan, M. (1984) The Health Belief Model and participation in a programme for the early detection of breast cancer. Social Science and Medicine, 19,823-830.

Champion, V.L. (1984) Instrument development of the Health Belief Model constructs. Advances in Nursing Science, 6,7345.

Cialdini, R.B., Kallgren, C.A. and Reno, R.R. (1991) A focus theory of nonnative conduct: A theoretical refine- ment and re-evaluation of the role of social norms in human behaviour. In: M.P. Zanna (Ed.) Advances in experimental social psychology (Vol. 23, pp. 201-234). San Diego: Academic Press.

Conner, M. and Norman, P. (1994) Comparing the Health Belief Model and the Theory of Planned Behaviour in health screening. In: D.R. Rutter and L. Quine (Eds.) Social psychology and health: European perspectives (pp. 1-24). Newcastle upon Tyne: Athenaeum Press Ltd.

Conner, M. and Sparks, P. (1996) The Theory of Planned Behaviour and Health Behaviours. In: M. Conner and P. Norman (Eds.) Predicting health behaviour: Research and practice with social cognition models (pp. 121-162). Buckingham: Open University Press.

Esteve, J., Kricker, A,, Ferlay, J., Parkin, D.M. in collaboration with the Steering Committee of the European Network of Cancer Registries (1993) Facts andfigures ofcancer in the European Community. IARC.

Fishbein, M. and Ajzen, I. (1975) Belief; attitude. intention and behaviour: An introduction to theory and research. Massachusetts: Addison-Wesley.

Godin, G. and Kok, G. (1996) The Theory of Planned Behavior: A review of its applications to health-related behaviors. American Journal of Health Promotion, 11(2), 87-98.

Gollwitzer, PM. (1993) Goal achievement: The role of intentions. European Review of Social Psychology, 4, 141-185.

Harrison, J.A., Mullen, P.D. and Green, L.W. (1992) A meta-analysis of studies of the Health Belief Model with adults. Health Education Research: Theory and Practice, 7( 1). 107-1 16.

Hennig, P. and Knowles, A. (1990) Factors influencing women over 40 years to take precautions against cervical cancer. Journal ofApplied Social Psychology, 20, 1612-1621.

Hill, D., Gardner, G. and Rassaby, J. (1985) Factors predisposing women to take precautions against breast and cervix cancer. Journal of Applied Social Psychology, 15,59-79.

179-21 1.

ure Research, 24,207-224.

Psychology, 7(4), 259-276.

1 9 1 -2 15.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

COMPARISON OF HBM AND TPB 49

International Agency for Research on Cancer (IARC) Working Group in Evaluation of Cervical Cancer Screening Programme (1986). Screening for squamous cervical cancer: Duration of low risk after negative results of cervical cytology and its implications for future screening policies. British Medical Journal, 293,659464.

Janz, N.K. and Becker, M.H. (1984) The Health Belief Model: A decade later. Health Educarion Quarterly, 11(1), 1 4 7 .

King, J. (1982) The impact of patient’s perceptions of high blood pressure on attendance at screening: An exten- sion of the Health Belief Model. Social Science andMedicine, 16(10), 1079-1091.

h o m e s , G. and Sugden, R. (1982) Regret theory: An alternative theory of rational choice under uncertainty. Eco- nomic Journal, 92,805-82s.

Mandleblatt, J., Traxler, M., Lakin, P., Kanetsky, P. and Kao, R. (1993) Targeting breast and cervical cancer screening to elderly poor black women. Who will participate? Preventive Medicine, 22(1), 20-33.

Mandleblatt, J., Traxler, M., Lakin, P., Kanetsky, P., Kao, R. and Harlem Study Team (1992) Mammography and papanicolaou smear use by elderly poor black women. Journal ofAmerican Geriatrics Society, 40,1001-1007.

Marteau, T.M. (1993) Health related screening: Psychological predictors of uptake and impact. In: S . Maes, H. Leventhal and M. Johnston (Eds.) lntemtional review of health psychology (Vol. 2, pp. 149-174). Chiches- ter: Wiley.

Mullen, P.D., Hersey, J.C. and Iverson, D.C. (1987) Health behaviour models compared. Social Science and Medicine, 24( 1 1). 973-98 I.

Murray, M. and McMillan, C. (1993) Health beliefs, locus of control, emotional control and women’s cancer screening behaviour. British Journal of Clinical Psychology, 32, 87-100.

Norman, P. and Conner, M. (1996) Predicting health check attendance among prior attenders and non-attenders. The role of prior behavior in the Theory of Planned Behavior. Journal of Applied Social Psychology, 26( I I), 1010-1026.

Oliver, R.L. and Berger, P.K. (1979) A path analysis of preventive health care decision models. Journal of Con- sumer Research, 6, 113-122.

Orbell, S., Crombie, I. and Johnston, G. (1996) Social cognition and social structure in the prediction of cervical screening uptake. British Journal ofHealth Psychology, 1,35-50.

Orhell, S . and Sheeran, P. (1993) Health psychology and uptake of preventive health services: A review of 30 years’ research on cervical screening. Psychology and Health, 8(6), 417-433.

Orbell, S. and Sheeran, P. (1 998) “Inclined abstainers”: A problem for predicting health-related behaviour. British Journal of Social Psychology, 37, 151-165.

Parker, D., Manstead, A.S.R. and Stradling, S.G. (1995) Extending the theory of planned behaviour: The role of personal norm. British Journal of Social Psychology, 34, 127-137.

Parker, D., Manstead, A.S.R., Stradling, S.G. and Reason, J.T. (1992) Intentions to commit driving violations: An application of the Theory of Planned Behaviour. Journal of Applied Psychology, 77( I), 94-101.

Quine, L, Rutter, D. and Arnold, L. (1998) Predicting and understanding safety helmet use among schoolboy cyclists: A comparison of the Theory of Planned Behaviour and the Health Belief Model. Psychology and Health, 13, 251-269.

Richard, R., van der Pligt, J. and de Vries, N. (1995) Anticipated affective reactions and prevention of AIDS. Brit- ish Journal of Social Psychology, 34,9-21.

Rogers, R.W. (1975) A protection motivation theory of fear appeals and attitude change. The Journal of Psycho-

Rogers, R.W. (1983) Cognitive and physiological processes in attitude change: A revised theory of protection motivation. In: J. Cacioppo and R. Petty (Eds.) Social psychophysiology (pp. 153-176). New York: Guildford Press.

Ronis, D.L. and Kaiser, M.K. (1989) Correlates of breast self examination in a sample of college women: Ana- lysis of linear structural relations. Juurnal of Applied Social Psychology, 19(13), 1068-1084.

Schwarzer, R. (1992) Self efficacy in the adoption and maintenance of health behaviours: Theoretical approaches and a new model. In: R. Schwarzer (Ed.) Selfeficacyr Thought control of action (pp. 217-242). London: Hemisphere.

Seibold, D.R. and Roper, R.E. (1979) Psychosocial determinants of health care intentions: Test of the Triandis and Fishbein models. In: D. Nimmo (Ed.) Communication yearbook 3 (pp. 625443). Brunswick N.J.: Transac- tion Books.

Seydel, E., Taal, E. and Weigman, 0. (1990) Risk appraisal, outcome and self-efficacy expectancies: Cognitive factors in preventive behaviour related to cancer. Psychology and Health, 4,99-109.

Sheeran, P. and Abraham, C. (1996) The Health Belief Model. In: M. Conner and P. Norman (Eds.) Predicting health behaviour: Research and practice with social cognition models (pp. 23-61). Buckingham: Open Uni- versity Press.

Sheppard, B.H., Hartwick, J. and Warshaw, P.R. (1988) The Theory of Reasoned Action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research, 15, 325-343.

logy, 91,93-I 14.

Stroebe, W. and Stroebe, M.S. (1995) Socialpsychology and health. Buckingham: Open University Press. Sutton, S. (1987) Social-psychological approaches to understanding addictive behaviours: Attitude-behaviour

and decision making models. British Journal of Addiction, 82,355-370.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012

50 A. BISH ETAL.

Sutton, S. (1998) Predicting and explaining intentions and behavior: How well are we doing? Journal of Applied

Terry, D. and O’Leary, J.E. (1995) The theory of planned behaviour: The effects of perceived behavioural control

van der Pligt, J., Zeelenberg, M., van Dijk, W.W., de Vries, N. and Richard, R. (1998) Affect, attitudes and deci-

Weinstein, N.D. (1993) Testing four competing theories of health protective behavior. Health Psychology, 12(4),

Social Psychology, 28,1318-1339.

and self efficacy. British Journal of Social Psychology, 34, 199-220.

sions: Let’s be more specific. European Journal of Social Psychology, 8,3346.

324-333.

Dow

nloa

ded

by [

Uni

vers

itari

o M

igue

l Her

nand

ez]

at 0

4:30

06

Mar

ch 2

012