Preventive Medicine 57 (2013) S83–S86

Contents lists available at SciVerse ScienceDirect

Preventive Medicine

j ourna l homepage: www.e lsev ie r .com/ locate /ypmed

Applying the Transtheoretical Model to evaluate the effect of a call–recall program inenhancing Pap smear practice: A cluster randomized trial☆,☆☆

Fauziah Abdullah a,b,⁎, Tin Tin Su a

a Centre for Population Health (CePH), Department of Social & Preventive Medicine, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysiab Ministry of Health, Malaysia

☆ Trial registration: This trial was registered at the Ir(IRCT) with registration number 201103186088N1.☆☆ Funding: This study was financially supported by thof University of Malaya, Malaysia.

⁎ Corresponding author at: Centre for Population Heal& Preventive Medicine, Faculty of Medicine, University oMalaysia. Fax: +60 3 7967 4975.

E-mail addresses: abdullah_fauziah@yahoo.com, fau(F. Abdullah).

0091-7435/$ – see front matter © 2013 Elsevier Inc. Allhttp://dx.doi.org/10.1016/j.ypmed.2013.02.001

a b s t r a c t

a r t i c l e i n f o

Available online 12 February 2013

Keywords:Cervical screening programCall–recall systemTranstheoretical ModelStages of changeMalaysia

Objective. The objective of this study was to evaluate the effect of a call–recall approach in enhancing Papsmear practice by changes of motivation stage among non-compliant women.

Methods. A cluster randomized controlled trial with parallel and un-blinded design was conductedbetween January and November 2010 in 40 public secondary schools in Malaysia among 403 female teacherswho never or infrequently attended for a Pap test. A cluster randomization was applied in assigning schoolsto both groups. An intervention group received an invitation and reminder (call–recall program) for a Paptest (20 schools with 201 participants), while the control group received usual care from the existing cervical

screening program (20 schools with 202 participants). Multivariate logistic regression was performed todetermine the effect of the intervention program on the action stage (Pap smear uptake) at 24 weeks.

Results. In both groups, pre-contemplation stage was found as the highest proportion of changes in stages.At 24 weeks, an intervention group showed two times more in the action stage than control group (adjustedodds ratio 2.44, 95% CI 1.29–4.62).

Conclusion. The positive effect of a call–recall approach in motivating women to change the behavior ofscreening practice should be appreciated by policy makers and health care providers in developing countriesas an intervention to enhance Pap smear uptake.

© 2013 Elsevier Inc. All rights reserved.

Introduction

The Papanicolaou test has been recognized as a primary screening toolfor the early detection of cervical cancer since the 1940s, imple-mentedworldwide througheither opportunistic or organized approaches(Nygård, 2011; Papanicolaou, 1948). In opportunistic screening, theservice is provided to women who request it or who attend a healthfacility. This approach has been employed in Malaysia since 1960s. Yet,cervical cancer ranked as the second commonest cancer in Malaysianwomen with an incidence rate of 12.2 per 100,000 in 2006 and 80% ofpatients presented late (Chye et al., 2008; Omar et al., 2006; Othmanet al., 2009). In an organized program, women are regularly invitedfor screening tests, which have been practiced in many developed andfew middle-income countries in Latin America and Eastern Europe.

anian Registry of Clinical Trials

e postgraduate research grant

th (CePH), Department of Socialf Malaya, 50603 Kuala Lumpur,

ziah3@gmail.com

rights reserved.

Thus, these countries experience lower cervical cancer burden (Gakidouet al., 2008; Miles et al., 2004).

To our knowledge none of the studies have evaluated the impactof a call–recall screening practice on changes in the stages of cervicalscreening behavior among women. Hence, a cluster randomized trialwas undertaken to determine this. The trial was chosen for practicalreasons and to prevent contamination by preferences of teachers.

Methods

Trial design and participants

The study was a two armed, parallel group, un-blinded, and cluster ran-domized controlled trial. The clusters were national secondary schools with84 schools in Kuala Lumpur which divided into four zones with an averageof 20 schools per zone. 10 schools were selected at random from each ofthese zones and randomized (1:1) to the intervention or non-interventionarms of the trial. Female secondary teachers who were either naïve to Papsmear or had their last test more than three years were invited to participatein the study (n=403).

Sample size

Sample size was calculated using OpenEpi (Dean et al., 2012); based on a1:1 ratio of case to control participants, the study would need approximately

S84 F. Abdullah, T.T. Su / Preventive Medicine 57 (2013) S83–S86

200 participants (inclusive of an additional 20% to account for non-respondentsand design effect of two (Adamchak et al., 2000; World Health Organization,2012) without cluster effect) in the intervention and comparator arms tohave over 80% power at the 95% confidence level (alpha 0.05, two sided) todetect an effect size of 31% (Eaker et al., 2004).

Data collection

Data were collected from January to November 2010. A pre-testedself-administered baseline questionnaire was sent to all participants whofulfilled the inclusion criteria and consented to participate in the study. Thebaseline survey covered five areas: i) demographic and socioeconomic factors;ii) reproductive history; iii) healthy lifestyle or risk behavior; iv) attitudes andbeliefs of cervical screening practice, and v) stages of cervical screening behav-ior change by applying the Transtheoretical Model (TTM) (Glanz et al., 1997;Prochaska and DiClemente, 1983). A post-intervention questionnaire wasadministered at 24 weekswhere the information on stages of cervical screeningbehavior change was collected as main outcome.

Randomization

Using a computer generated simple randomization method in SPSSv15,each of the 40 public secondary schools which had agreed to participatewas randomized into either the intervention (n=20) or control (n=20)groups whereby all teachers from the same school (cluster) were assignedto the same group. Randomization was revealed after recruitment of thefinal school to ensure concealment of allocation (Fig. 1).

Intervention

A call–recall program was introduced to the intervention group, whichincludes a personal invitation letter with an information pamphlet of cervicalcancer screening, and followed by a telephone reminder with counselingafter four weeks that was performed once per participant. The controlgroup received usual care from the existing program.

Analysis

Follow-up

Intervention

Randomisation

INTERVEN

20 schools a

n=201 parti

(Invitation letter, inforon cervical cancer

additional telephone redescription of importa

test reiter

n=2 lost to fo

(both un-con

199participainclude

analys

Fig. 1. Flow of participants in the intervent

Statistical analysis

Baseline characteristics were compared using a t-test for continuousvariables and a chi-squared test for categorical data.Multivariate logistic regres-sion was applied to calculate odds ratios (ORs) and 95% confidence intervals(CIs) for the effect of the intervention on the action stage (Pap smear uptake).Variables from univariate analysis (irrespective of the intervention status)which had a p-value≤0.25were selected for the multivariate model. Collinear-ity among the variables in themultivariatemodel was checked and determinedabsent. The value from the Hosmer–Lemeshow goodness-of-fit test of themultivariate model was 3.74 (df=8, p-value=0.880) indicating a wellcalibrated model. SPSSv15 was used for all analyses.

Results

Twenty secondary schools were allocated to receive the interven-tion (n=201 eligible participants) and 20 schools (n=202 eligibleparticipants) were assigned to the standard (opportunistic) cervicalscreening arm. Both groups were followed up after a period of24 weeks, at which point subsequent uptake of Pap examination(action stage) was examined. Two participants from the interventionarm (both of whom were un-contactable) and three from the controlarm (one un-contactable, one refused participation and one could notparticipate due to medical leave) were lost to follow-up, leaving 20schools with 199 participants in each arm for analysis. By baselinecharacteristics, both groups were equally comparable to each other(Table 1).

The highest proportion of changes in stages in both groups wasnoted among those that belong to pre-contemplation stage, withslightly higher in the intervention group (Table 2). Chi-square testvalue of 19.893 showed that in all the stages, there was a significantdifference in the changes of stage at 24 weeks in both groups(p-valueb0.05).

40 selectedschools randomly

allocated tointervention

N=403

TION

llocated

cipants

mation pamphlet& pap test andminder with shortnce of Pap smearated)

llow-up

tactable)

ntsd inis

CONTROL

20 schools allocated

n=202 participants

(opportunistic screeningonly)

n=3 lost to follow-up

(one un-contactable, oneparticipant refused and one

was on medical leave)

199participantsincluded in

analysis

ion and control arm (Malaysia, 2010).

Table 1Baseline characteristics of women in the intervention and control group (40 selectedpublic secondary schools in Malaysia, 2010).

Variables Interventiongroup(n=201), %

Controlgroup(n=202), %

p-Value

DemographicsMean age (years) (±SD) 36.1±8.0 36.5±7.3 0.455a

Ethnicity: Malay 85.1 83.2 0.376a

Religion: Muslim 85.1 84.7 0.572a

Marital status: married 98.5 98.5 1.000b

Socioeconomic statusEducational level 0.939a

Diploma 2.5 2.5Graduate degree 90.0 89.1Master degree 7.5 8.4

Mean household monthlyincome (MYR) (±SD)

7967.0±8068.4 6777.8±3469.1 0.486c

Reproductive historyNumber of pregnancies(including abortion)

0.539a

1–4 66.2 71.3≥5 21.5 18.8Never pregnant 11.9 9.9

Mean number of pregnancies(±SD)

2.9±2.0 2.9±1.9 0.784c

Practicing contraceptivecurrently: yes

30.3 32.7 0.615a

Having any genital symptomscurrently: no

82.1 82.7 0.878a

Ever had a Pap test (≥3 years ago) 34.3 41.6 0.133a

Physical healthBody mass index (kg/m2) 0.944a

Normal (18.5–22.9) 31.3 32.2Underweight (b18.5) 5.0 5.4Overweight & obese (≥23) 63.7 61.9

Ever smoke: no 100.0 99.5 1.000b

Practice physical activity asrecommended: yes

10.9 9.4 0.609a

Perform breast self-examinationmonthly: yes

45.8 48.0 0.651a

Perform mammogram asrecommended

0.202a

Applicable 5.0 2.5Not applicable 67.2 63.4No 27.9 34.2

Had medical check-up in last3 years: yes

49.8 53.0 0.518a

Having any chronic diseases: yes 19.4 22.8 0.407a

Social healthNumber of marriages: once 99.0 98.5 1.000b

Mean age of first marriage(years) (±SD)

25.5±2.7 25.9±2.9 0.082c

Mean age of first sexualintercourse (years) (±SD)

25.6±2.5 26.0±2.9 0.123c

Age of first pregnancy (years) 0.091a

Never pregnant 11.9 9.9≤25 34.3 25.7>25 53.7 64.4

Mean age of first pregnancy(years) (±SD)

23.2±8.9 24.4±8.6 0.063c

Ever had a sexual health education: yes 34.8 35.6 0.864a

Practicing conservative methods toprevent cervical cancer:Safe sex 94.5 96.0 0.474a

Self circumcised 83.6 82.2 0.708a

Husband/partner circumcised 86.1 86.6 0.869a

Prohibit extra-marital sex 95.5 96.0 0.796a

Having health insurance: yes 66.2 70.3 0.413a

Notes: ±SD standard deviation, MYR denotes Malaysian Ringgit (where 1 MYR equals0.33 USD).

a Chi-square test.b Fisher's exact test.c t test.

Table 2Distribution of participants by stages of cervical screening behavior change at pre- andpost-intervention phase, and progress of changes in stage at post-intervention phase(Malaysia, 2010).

Stage of change Interventiongroupn (%)

Control groupn (%)

I. Distribution of participants bystages of change:Pre-intervention phase (0 week): N=201 N=202Pre-contemplation 72 (35.8) 53 (26.2)Contemplation 88 (43.8) 112 (55.4)Preparation 41 (20.4) 37 (18.3)

Post-intervention phase (24 weeks): N=199 N=199Pre-contemplation 33 (16.6) 40 (20.1)Contemplation 85 (42.7) 100 (50.2)Preparation 45 (22.6) 39 (19.6)Action (had a Pap test) 36 (18.1) 20 (10.1)

II. Progress of changes in stage at 24 weeks: N=199 N=199Improvement in stage of change⁎ 94 (47.2) 74 (37.2)No change or reduction in stage of change 105 (52.8) 125 (62.8)

Notes:

i) pre-contemplation describes individuals who for many reasons do not intend tochange within the next six months,

ii) contemplation refers to those individuals who are thinking about changing theirproblem behavior within the next six months,

iii) preparation refers to those individuals who committed to change their problembehavior soon usually within the next 30 days,

iv) action refers to those individuals who have changed their problem behavior withinthe past six months, and

v) maintenance refers to the individuals who have changed their problem behavior forat least sixmonths or relapse refers to those individualswho failed tomaintain the be-havior change over time and may or may not be contemplating taking action again.

In accordance to our study, only four stages were applied (pre-contemplation,contemplation, preparation and action stage).⁎ p-Valueb0.05.

S85F. Abdullah, T.T. Su / Preventive Medicine 57 (2013) S83–S86

After 24 weeks, a higher proportion of participants in the interventionarm (18.1%) were found to be in action stage than the control group(10.1%), representing an absolute increase in Pap smear uptake of 8%between the intervention and control group. Upon multivariate analysis,the action stage was twice as high among those women in the interven-tion arm compared to the control arm, OR 2.44 (95% CI 1.29, 4.62). It isestimated that 12 women would need to receive the intervention inorder for one woman to benefit from cervical cancer screening (Table 3).

Discussion

Our trial demonstrated that women's cervical screening behavioris determined by the type and nature of a screening program offered

Table 3The action stage (Pap smear uptake within 24 weeks of follow-up) in the interventionand control group (Malaysia, 2010).

Action stage (had a Paptest within 24 weeks),n (%)

Univariatemodeling

Multivariatemodelinga

OR 95% CI OR 95% CI

Interventiongroup(n=199)

36 (18.1) 1.98 1.1–3.5⁎ 2.44 1.29–4.62

Control group(n=199)

20 (10.1) 1.0 Ref. 1.0 Ref.

Notes:OR denotes odds ratio.CI denotes confidence interval.Ref. denotes reference.

a Multivariate model was adjusted for the demographic characteristics, socioeco-nomic status, reproductive history and lifestyle factors shown in Table 1, where thep-value in univariate analysis was ≤0.25.⁎ p-Valueb0.05.

S86 F. Abdullah, T.T. Su / Preventive Medicine 57 (2013) S83–S86

to them. This explained why women in developed countries whichpracticed call–recall program have better screening practice andexperience lower cervical cancer burden (Gakidou et al., 2008; Palenciaet al., 2010).

The finding highlighted the promising effect of a call–recallapproach for having a greater potential in reaching and motivatingnon-compliant women who are categorized as “hard-to-reach” groupto change towards beneficial health screening behavior (Kitcheneret al., 2006; Miles et al., 2004). The “hard-to-reach” group defined bynon-adherence to Pap smear practice is of greater concern becausethey present a greater risk of developing cervical cancer than others(Andrae et al., 2008).

Moreover, pre-contemplation stage was noted to have a remark-able drop at 24 weeks, with more in the intervention group. This isa source of great concern because pre-contemplates are often charac-terized as resistant, unmotivated to match their needs or as not readyfor health promotion programs (Glanz et al., 2008). Therefore, theeffort in motivating women to practice Pap test should be efficient.With regards to opportunistic program, the approach might facedifficulties due to lack of incentives in tracking and inviting womenfor regular Pap test performance. Thus, this situation further empha-sizes the importance of the nature of screening approach, whichshould be taken into consideration in order to promote cervicalscreening seeking behavior (Abdullah and Su, 2010; Abdullah et al.,2011).

The strengths of our study include an original approach of applyingthe Transtheoretical Model in evaluation of the effect of a call–recallprogram and intervention study design. The results of this study maybe generalized to the urban setting of other middle-income countriesin the Asian region. However, the result could be influenced by self-acquired awareness of cervical screening test by knowing themselvesas being studied and assessed in connection with the outcome mea-sured. This was considered as a limitation of our study that was identi-fied as being difficult to overcome due to natural event.

In conclusion, invitation and follow-up care have the ability toinduce our non-compliant participants for Pap smear practice. There-fore, a single adaptation of existing screening program could be benefit-ed to non-compliant women by changing their screening behavior.Ultimately, this new strategy should be highly valued by policy makersand health care providers in developing countries as a promising alter-native approach for reducing cervical cancer burden.

Ethics approval

Ethics approval was obtained from the University Malaya MedicalCentre (UMMC) Ethics Committee, Ministry of Education of Malaysiaand Ministry of Health of Malaysia. This trial was registered at theIranian Registry of Clinical Trials (IRCT) with registration number201103186088N1.

Conflicts of interest

The authors declare that there are no conflicts of interests.

Acknowledgments

This study was financially supported by the postgraduate researchgrant of University of Malaya, Malaysia. The authors would like tothank the official staffs of the Ministry of Education of Malaysia,Department of Education of the Federal Territory of Kuala Lumpur,Malaysia, Ministry of Health of Malaysia, University Malaya MedicalCentre, selected schools and study participants for their valuablesupport during the data collection phase.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ypmed.2013.02.001.

References

Abdullah, F., Su, T.T., 2010. Enhancement of the cervical cancer screening program inMalaysia: a qualitative study. Asian Pac. J. Cancer Prev. 11, 1359–1366.

Abdullah, F., Abdul_Aziz, N., Su, T.T., 2011. Factors related to poor practice of Pap smearscreening among secondary school teachers in Malaysia. Asian Pac. J. Cancer Prev.12, 1347–1352.

Adamchak, S., Bond, K., MacLaren, L., Magnani, R., Nelson, K., Seltzer, J., 2000. A guide tomonitoring and evaluating adolescent reproductive health programs. FOCUS onYoung Adults.

Andrae, B., Kemetli, L., Sparén, P., et al., 2008. Screening-preventable cervical cancerrisks: evidence from a nationwide audit in Sweden. J. Natl. Cancer Inst. 100,622–629.

Chye, G.L.C., Rampal, S., Yahaya, H., 2008. Cancer incidence in Peninsular Malaysia,2003–2005. The Third Report of the National Cancer Registry, Malaysia.NationalCancer Registry, Kuala Lumpur.

Dean, A.G., Sullivan, K.M., Soe, M., 2012. OpenEpi: open source epidemiologic statisticsfor public health, version 2.3.1. http://www.OpenEpi.com (updated 2011/23/06).

Eaker, S., Adami, H.-O., Granath, F., Wilander, E., Sparen, P., 2004. A large population-based randomized controlled trial to increase attendance at screening for cervicalcancer. Cancer Epidemiol. Biomarkers Prev. 13, 346–354.

Gakidou, E., Nordhagen, S., Obermeyer, Z., 2008. Coverage of cervical cancer screeningin 57 countries: low average levels and large inequalities. PLoS Med. 5,0863–0868.

Glanz, K., Lewis, F.M., Rimer, B.K. (Eds.), 1997. Health Behaviour and Health Education:Theory, Research, and Practice. Jossey-Bass Publishers, San Francisco.

Glanz, K., Rimer, B.K., Viswanath, K., 2008. Health Behavior and HealthEducation: Theory, Research, and Practice, 4th ed. Jossey-Bass Publishers,San Francisco.

Kitchener, H.C., Castle, P.E., Cox, J.T., 2006. Achievements and limitations of cervicalcytology screening. Vaccine 24, s63–s70.

Miles, A., Cockburn, J., Smith, R.A., Wardle, J., 2004. A perspective from countries usingorganized screening programs. Cancer 101, 1201–1213.

Nygård, M., 2011. Screening for cervical cancer: when theory meets reality. BMCCancer 11, 240.

Omar, Z.A., Mohd.-Ali, Z., Ibrahim-Tamin, N.S., 2006. Malaysian Cancer Statistics— Dataand Figure Peninsular Malaysia 2006. National Cancer Registry, Ministry of HealthMalaysia.

Othman, N., Devi, B., Halimah, Y., 2009. Cervical cancer screening: patients' under-standing of screening for cervical cancer in major hospitals in Malaysia. AsianPac. J. Cancer Prev. 10, 569–574.

Palencia, L., Espelt, A., Rodrıguez-Sanz, M., et al., 2010. Socio-economic inequalities inbreast and cervical cancer screening practices in Europe: influence of the type ofscreening program. Int. J. Epidemiol. 39, 757–765.

Papanicolaou, G.N., 1948. A survey of the actualities and potentialities of exfoliativecytology in cancer diagnosis. Ann. Med. 30, 661–674.

Prochaska, J.O., DiClemente, C.C., 1983. Stages and processes of selfchange ofsmoking: toward an integrative model of change. J. Counsel. Clin. Psychol. 51,390–395.

World Health Organization, 2012. Selecting your sample. http://www.paho.org/English/AD/DPC/NC/steps-tng-6-sample-select.ppt.