Culturally and linguistically diverse population health social marketing ...

14 downloads 190 Views 386KB Size Report
of effective mainstream social marketing campaigns. They include clearly defined segmentation and analysis of target audiences.' .e.5.11 a consumer focus.
·,4i¥U Y" I,I.'MMmti b

Culturally and linguistically diverse population health social marketing campaigns in Australia: a consideration of evidence and related evaluation issues Andrew]. Milat, Tom f. Carroll and Jennifer J. Taylor

Introduction Social marketing campaigns havebeenused for manyyears as a means of pursuing healthimprovements in populations. They

formative research.t> channel analysis andmediaplanning,',4.5 strategic planning of the marketing mix,1,",5 includingdesignation of specific behavioural and communication objectives," and

have beenused as a tool to influence awareness, attitudes and health behaviour in bothindustrialised anddeveloping countries and havefocused on healthissues asdiverse as illicit drug usc,'

campaign tracking and evaluation.v! There is also growing consensus on the optimal method for evaluating mainstream social marketing campaigns. Even though mostcampaigns are

smoklng.s> teenage drinking." nutrttton.' physical activity/'

evaluated usinga pre-postdesign, apre-post intervention-control group design provides the clearest evidence of campaign effectiveness.s-" Despite thisgrowingconsensus for mainstream campaigns, little is known about the processes critical to the development and evaluation of effective social marketing campaigns targeting people from culturally and linguistically diverse (CLD} backgrounds within English-speaking countries,

malaria control," immunisation8.9 and HIV/AIDS prevention." There appears to be a growingconsensus on the components of effective mainstream social marketing campaigns. They include clearly defined segmentation and analysis of target audiences.' .e. 5.11 a consumer focus." complementary and reciprocal communications and community actons.t-v- use of

Abstract

Issue addressed: This paper describes a review of population health social marketing campaigns targeting culturally and linguistically diverse (CLD)communities in Australia in order to identify characteristics of effective

ClD campaigns.

Methods: literature on CLD population health social marketing was identified from electronic searches of databases in August 2004. At the same time, the grey literature wasexamined by searching the Internet and talking to Australian experts in the fields of ClD social marketing and CLD research. Results: Eight studies met the search criteria, fourfrom the published literature. Two studies that employed prepostevaluation designs provided tentative supportfor the potential efficacy of CLD social marketing strategies. The remaining studies did not allow for causal attribution as they used post-campaign only or process ~ evaluations. Studies did, however, show that CLD communitiesaccess campaign-related information from both mainstream and ethnic media channels. In addition, Vietnamese respondents were more likely to access campaign messages through ethnic radio and Chinese respondents through ethnic press. Conclusions: There is insufficientevidenceto clearly identify the characteristics of effective CLD campaigns. Campaign evaluation designs usedto evaluate socialmarketingstrategies targeting CLD communities in Australia are generally weak, but there is tentativeevidencesupportingthe potentia! efficacyof thesestrategies in some Australian settings. Key words: Social marketing, cultural and linguistic diversity, media campaigns, evaluation. Heallh ftomotion Journal of Australia 2005;16:20-5

So what? Health promotion policy makers, researchers and social marketers are encouraged to evaluate (LD social marketingcampaigns usingas robust methodologies as possible and to publish findings in order to build a CLD campaign effectiveness evidence base.

20

Health Promotion Journal of Australia 2005 : 16 (1)

Copyright ofFull Text rests with the original copyright owner and, except as permitted under the Copyright Act 1968, copying this copyright material is prohibited without the permission ofthe owner or itsexclusive licensee oragent orbyway of alicence from Copyright Agency Limited. For information about such licences contact Copyright Agency Limited on (02) 93947600 (ph) or (02) 93947601 (fax)

Research and Methods

Cultural and linguistic diversity has significant implications for health promotion practitioners who wish to implementsocial marketing interventions in a societysuch as Australia.'! In the 2001 Census, 13% of Australians were born in a non-English speakingcountry and 15% stated that they spoke a language other than English at home.lvlf social marketing strategies are identified as an appropriate part of a population health intervention, it is important that communications reach all membersofthe Australian community. Forthisreason,specific communication strategies are oftendeveloped inorder to reach people from CLD backgrounds. The premise underlying the development of specific strategies is that communication materials carried in the mainstream media may not reach the majority of non-English speakingpeopleor maynot be culturally

appropriate.' ] Theobjective ofthispaperwasto conducta review ofpopulation healthsocial marketing campaigns targeting CLD communities inAustralia in order to identify characteristics of effective CLD campaigns.

Methods literature on CLD population health social marketing was identified from electronic searches of Medline, PsyciNFO, Sociofile, Proquest and Cochrane databases in August 2004. All abstracts wereextracted fromdatabasesbya single researcher from Research and Marketing Group,then checked bya second researcher from the same group.Thesearchkeywordsincluded 'mass media', 'communication', 'social marketing', 'ethnic minorities', 'culturaldiversity', 'multicultural' and 'non-English speaking background'.Inaddition, the names of the top seven languages other than English spoken at home in Australian households from the 2001 Census" were included in the key word search, These were: 'Chinese' ('Cantonese' and 'Mandarin'), 'Italian', 'Greek', 'Arabic', 'Vietnamese', and 'Spanish'. Morethan halfof the people who reportedspeaking a language other than English at home inthe 2001 Census spoke at leastone of these languages." Thegreyliterature wasexamined by searching the Internetand talking to Australian expertsinthe fields ofCLD sodal marketing. After conducting the database and Internet search, a list of published and unpublishedreportsthat met the searchcriteria was compiled by the authors and sent to experts who were asked ifthey were aware of any additionalstudiesthat met the searchcriteria. Theywere alsoasked ifthe apparent absenceof evaluated CLD social marketing campaigns in the literature related to an absence of campaignactivity or evaluations.

Search criteria Articles identified throughthe database search, Internetsearch and expertadvice were reviewed ifthey metthe following search criteria:

Culturally and linguistically diversesocial marketing campaigns

Research wasconducted inAustralia between August 1990 and August 2004. Themajorintervention inthe studywasa populationhealth communication or social marketingcampaign. The communication intervention was targeted at the communitylevel. The communication intervention was multi-levelled. The target group for communication was a non-English speakinglanguage group. The communication intervention was evaluated. Social marketing campaigns were defined as multi-faceted behaviour change interventions targeted at the populationor community level and incorporating media communication components. Communications interventions were considered 'multi-levelled' if they utilised two or more communication methods to reach target audiences.

Results and Discussion Of the 1,745 articles identified in the peer-reviewed literature, fourstudiesmet the inclusion criteria.'5.17 Asearch of the grey literature identified four more campaign evaluation reports.we Table 1 summarises the reviewed campaign evaluations. The first of the published studies described the 'Good Heart, Good life' anti-smoking media campaign that targeted Greekspeakingadultsliving incentralSydneybetween July 1991 and December 1993. 15The mainstrategy involved ethnic radioand pressadvertising. Toevaluatethe campaign,a post-intervention telephonesurvey wasconducted withGreekspeakersaged 1 B· 69 years living in the Marrickville local government area in December 1989 and compared with data from a household survey conducted in 1993. While the prevalence of smoking was similar to that found in the baselinesurvey, a Significantly lowerage-specific rate was found in the post-campaign survey for males in the 45 to 64 years cohort. Additionally, daily consumption ofcigarettes decreased betwee~ the pre-and pescampaign surveys and positive changes in knowledge and attitudewereobserved.Theabsenceofanysimilar interventions duringthe study period and the highlevel of recallln the targeted groupslendssupport to the conclusionthat these changesmay be partially attributedto the campaign. In 1997, the 'First Aid for Scalds' campaign targeted parents and carersfromVietnamese, Chinese,and Arabic backgrounds living in New South Walesin order to increasethe proportion who had correct knowledge of first aid treatment for scalds." The primary strategy wasan ethnic media campaign, including radio and newspaper advertising. The evaluation involved random, population-based telephone surveys with each of the three language groups, before and after the campaign. After the campaign, there were significant increases inthe proportion

Health Promotion Journal of Australia 2005 : 16 (1)

21

Milat, Carroll and Taylor

Article

of people who reported the correct firstaid treatment for scalds.

generatea significant increase in screening in areas with a high

The association between campaign recall and an increase in

proportion of women of non-English-speaking background

correct knowledge, and the absence of any similar interventions

compared with areas with a low proportion of women of non-

during the campaign period, give credence to the authors'

English-speakingbackground. The authors concluded that paid

conclusion that the changes observed were a result of the

publicityon ethnic radio may be an effective strategy to increase

campaign.

cervical screening rates among women of ClD backgrounds.

The next study evaluated the effect of three waves of ethnic

However, as measures of campaign awareness, reach and

radio advertising and publicity on cervical screening rates in

intentions among CLD women in the target postcodes were not

intervention and control areas of Victoria." The CLD 'Pap Test

collected before and after each phase of activity, it is difficultto

Vietor!a' program radio intervention occurred during three

make any confident causal attribution to the Pap test media

discrete periods between 1992 and 1994. The numbers of women having Pap smears before and after the intervention

intervention. In addition, the study does not account for the

were compared in postcode areas with highand low percentages

messageson CLDaudiences. Thisisimportant, as there isevidence

possible effect of mainstream cervical screening recruitment

of residents of non-English-speaking background. Across the

that CLD community members are consumers of both mainstream

three intervention periods, the media publicity appeared to

and CLDmedia." A post-intervention survey indicatinga positive

'Table1: Summary of CLD campaign evaluations. Target audience

Study

Intervention

content area Culpin et el (1996)

Greek community inner western Sydney

Tobacco

King et af.(1999)

Chinese, Vietnamese and Arabic speakers inSydney

Scalds

Mitchell et at. (1997)

Victorian women from 11 language groups

Cervical screening

Camit (2002)

Arabic, Chinese and Vietnamese speakers inNSW

Injury prevention - smoke detectors

Redding et al. (2001)

Italian community in NSW

Drink driving

Commonwealth Dept. of Hea!th and Family Services (1998) Carroll et al.

Parents from 13 language groups. Evaluation with Arabic, Cantonese, vietnamese andSpanish speakers Male andfemale smokers from eight language groups. Evaluation with Greek, Vietnamese, Cantonese and Arabic speakers

Immunisation

(2000)

8ertram et al. (2003)

22

ClD parents of8-17yearaids from 16 language groups. Evaluation with Arabic, Cantonese, Mandarin, Spanish and Vietnamese speakers

Tobacco

Illicit drugs

Health Promotion Journal of Australia 2005 : 16 (1)

Social marketing methods (multi-levelled communication)

Evaluation method

Ethnic radio and press advertisements Print resources Advertisements insix Greek video releases Distribution ofQuitmanual through community networks Ethnic radio and press advertisements Publicity Media conference Print resources Community action Ethnic radio and press advertisements Ethnic radio publicity Radio prize competition Community television Ethnic radio and press advertisements Dubbed video Mailed correspondence Print resources Community action Ethnic radio publicity Pri nt publicity Print resources Community action

Pre-post surveys

Pre-post surveys

Pre-post intervention-eontrol screening rates Process evaluation

Process evaluation

Ethnic radio and press advertising Post-campaign survey Publicity strategy Print resources Community action Community television inone State (six languages) Post-campaign survey Ethnic radio andpressadvertising PubliCity stratem' Workplace education strategy Print resources Community action Ethnic radio and press advertising Fost-cempaign survey Publicity strategy Pr'rnt resources Community action

Research and Methods

relationship between proximal evaluation measures of CLD campaign awareness, recall and screening behaviour wouldhave assisted interpretation of campaign impact. Thefinal published study, on the 'Smokealarmswake you up if there is a fire' campaign, targeted Arabic, Chinese and Vietnamese speakers and landlords in New South Wales in 1998. 18 Discounted smoke alarms were sold through ethnic community organisations, community talks, street stalls and pharmacists. Smokealarms were promotedthrough ethnicpress and radioadvetlslng.Adubbed television commercial wasaired on communitytelevision intarget languages and an installation demonstration tape wasshownincommunitytalks. Letters from the NSW Fire Brigades Commissioner were also sent to 822 realestateagentsinareaswith large concentrations of the target audiences emphasising the importance of installing smokealarms in rental properties. NSW Fire Brigades also offered free installation of firealarmsfor target audiences. During the campaign period 3,433 smoke alarms were sold through campaign distribution channels. This represented a sizeablepenetrationofthe estimated 50,000 targethouseholds in NSW. 18 However, the number of smoke detectorsthat were actually installed in target audience households was not measured. As the campaign used only processmeasuresit was also not possible to determine the relative effect of campaign messages on the purchasing behaviourofCLD targetaudiences. The remaining studiesidentified in the greyliterature were the Italo-Australian 'Reducing the Risk' drink-driving campaign'? and CLD components of three Australian Government campaigns: '1997 Immunise Australia Campaign'r'? '1997 National Tobacco Campaign',21 and '2001 National Illicit Drugs Campaign'."Unfortunately, thesestudies did notallow forcausal attribution of any observed change to campaign intervention, as they used post-campaign only or process evaluations. However, the findings of these studies can still inform our understanding of communicating with a number of CLD communities.

ClO media consumption in Australia Thecultural and linguistic diversity of non-English speakershas resulted in very little research being conducted on the media habitsofthissegmentofthe Australian community"!However, a number of the reviewed studies do providesome valuable insights into media consumption of CLD communities. 1S•17,2o-22 For example, Australian-born respondents ofGreekorigin inthe 'Good Heart, Good Life' campaign evaluation wereSignificantly lesslikely to haveseen or heard of the projectthan Creek-born Australians. The'First Aid forScalds' evaluation foundthatArabic and Vietnamese respondents were more likely to access campaign messages through ethnic radio and Chinese respondents throughethnic press.

Culturally and linguistically diverse social marketing campaigns

Findings from recentnational population healthsocial marketing campaigns alsosuggest thatCLD communities accesscampaignrelated information from both mainstream and ethnic media channels. 2G-22 Notsurprisingly, it appears that language groups with lower levels of English language proficiency such as Vietnamese and Arabic speakers are more likely to source campaign information from ethnic media than groups with higher English language proficiency.2o.n Supporting the findings of the 'First Aid for Scalds' campaign, these evaluationshave also consistently found that Vietnameseadults are more likely to accesscampaign messages throughethnic radioand Chinese speakers through ethnic press. N -n Challenges for ClO campaign evaluation in Australia

The review clearly demonstrates that there isa paucity ofprimary research measuring the effectiveness of CLD social marketing campaigns inAustralia. However, expert opinion suggests that this paucity does not necessarily relate to a paucity of CLD communication interventions, ratherto an absenceof published evaluations. The absence of publishedevaluation studies in this area may also be explained by the limited resources available to target the many CLD communities in Australia. It is an ongoing challenge forhealth promotion policy makersto keepevaluation costsin proportion to interventions and this is particularly true for CLD interventions that target multiple communities. This issuehas been partially addressed in nationalpopulationhealth social marketing campaigns byconductingtracking surveys with a limited numberof language groups.For example,whilephase one ofthe National Illicit Drugs Campaigntargeted 16 language groups, tracking surveys were conducted with five language groupsbecause ofthe prohibitive cost ofevaluating the campaign with all of these communities. aa Vv'ith the exception of the 'Good Heart, Good Life' and 'First Aid for Scalds' campafgns.ute reviewed studies employed relatively weakevaluation designs. Social marketing campaigns shouldideally beevaluated using a pre-postintervention-control group design, as it providesthe clearest evidence of campaign effectiveness'v" (seeTable 2). Thisevaluationdesign has been used inthe United States,23 but hasyet to be appliedinAustralia. InAustralia, itisoftennotfeasible to usecontrol groupsinstudies that examine the effectiveness of national social marketing campaigns and mostmainstream campaigns are evaluatedusing a time series design in which measurements are conducted before and afterthe intervention.P InAustralian settings there are a number of geographic, resource and sampling issues that prove to be barriers to using pre-post evaluation designs for CLD campaigns and specifically to usingpre-postinterventioncontrol designs, and this may go some way to explaining the limited publishedresearch in the area.

Health Promotion Journal of Australia 2005 : 16 (1)

23

Milat, Carroll and Taylor

Article

First, it is not feasible to use a control group in conducting a national campaign when the whole nation is exposed to the intervention. Although it ispossible toconducta stagedcampaign where different geographical regions receive the campaign at different times,thisisnot usualpracticefor national campaigns in Australia. In the case of conductinga campaign in a single Stateor Territory, it isdifficult to find matched ClD community control areas with largeenough ClD populationsrepresenting matched risk and demographic profiles. Researchers seekingto evaluate ClD campaignsare also likely to be ~nfronted with issues of sampling and representativeness. Sampling issues arisefromattemptsto attainsufficient numbers of respondents from a particular language group. To attain a representative sample for a particular language group, a large number of attempts need to be made to find appropriate respondents Ie.g. people who speak the specific language and meet sampling requirements of the study}.13 This becomes particularly difficult when the targetaudience isa lessprevalent language group. A way of reducingthe impact of this barrier is to use surname-based algorithms to select households where there isa highlikelihood ofone or moreofthe residents speaking one of the target language/s.20.22 This method may decrease costsand increasethe likelihood of reachingpeople fromtarget language groups, but little is known about its effect on representativeness of samples and subsequently the generallsablllty offindings. Encouragingly, researchers havefound consistency in prevalenceestimatesin ClD communitieswhen thissamplingmethod was compared with clusterand electoral role samplingmethodologies." Inlight ofthesebarriers, the scopeforusing pre-postinterventioncontrolevaluationdesignsislimited in Australia because of the difficulty of findingmatched targetaudiences and control sites. Itis more feasible to employ pre-postevaluationdesignswhen targetinglargerClD communities,as used in the 'Good Heart, Good Ufe' and 'FirstAid for Scalds' campajgns."-" However, the issue of the representativeness of samples and costeffectiveness of such an undertaking remains unresolved. In

practical terms, the cost of conducting pre- as well as postsurveys is likely to be prohibitive for many agencies. Ifgeographic, financial or methodological constraintspreclude the use of pre-post intervention-control group or pre-post designs, post-campaign measures are the next best option for health promotion policy makersand researchers. Itisthen not surprising that three of the eight studies reviewed employed this evaluation design. Even though post-campaign measures are not the goldstandard forevaluatingcampaigns,they are far better than no evaluation measuresat all, as policy makersstill require information on which to base campaign planningand decision making. However, it is important to beclear on the limitations of postcampaignsurveys. Theycan only givean indicationof the self· reportedimpactor a campaignat a single point intime. Without pre-campaign measures or a control, they cannot objectively assesschange that occurs over the period of the campaign, or confidently make causalattributionof any change observed to the campaign. Inaddition,they are subjectto the samesampling issues as other evaluationdesigns. None the less, post-campaign surveys can givean indication of awarenessand recall of campaign messages, relative efficiency of mediachannelsinachievingthisawarenessand, importantly, can indicate how messages are being interpreted and whether they are beingaccepted by ClD targetaudiences.This feedback is vital to determine the reach of ClD campaigns, to media planning, and to the developmentofsalientcampaign messages. Essentially, decisionsregarding evaluationmethodologies within a definedbudgetreflect pragmatic trade-offs. The moresensitive and reliable the assessmentof the impact of social marketing strategies, the more resource intensive the evaluation process becomes. If limited resources are available, ClD campaigns can also be evaluated using qualitative research and process measuresas in the reviewed smoke alarm" and drink-driving campaigns.'? This usually involves consulting with key community informants to assesstheir perception of campaign

Ranking

Evaluationdesign

Measurement

limitations

1

Pre-post intervention-control surveys

Change in awareness, attitude, intention and behaviour, while controlling formanypotential biases Efficiency of media channels

Availability of control areas Demographic comparability of samples Resources required - veryhigh Sampling issues

Pre-post campaign surveys

Change in awareness, attitude, intention and behaviour Efficiency of media channels

Post-campaign surveys

Awareness and recall of campaign messages Perceived personal impact Relative efficiency of media channels

Resources required - high Sampling issues Resources required - moderate Sampling issues

Process evaluation

Perceived impact amongkey informants Number of resource, distributed

4

24

Health Promotion Journal of Australia 2005 : 16 (1)

Non-representative Non-inferential Resources required -Jow

Research and Methods

effectiveness and recordingdistribution of campaign materials. Clearly, thisapproach limits inferencesthat can be made about the effectof campaignson targetaudiences. Conclusions

Given the paucity of CLD evaluation studies, it is difficult to draw clear conclusions in relation to the characteristics of effective CLD population healthsocialmarketingcampaignsin Australia. However, there is tentative evidence supportingthe potential efficacy ofsocialmarketingcampaignsthat specifically targetCLD communities. The reviewedstudiessuggest that CLD communities access campaign-related information from both mainstream and ethnicmediachannels. Inaddition,Vietnamese respondents are more likely to access campaign messages throughethnic radio and Chinese respondents through ethnic press. Health promotion policy makers, researchersand social marketers are encouraged to undertake CLD campaign evaluations that are as robustas possible and to publish results. Only through sharingof reliable campaign evaluationfindings will a body of knowledgesurroundingeffective approaches to CLD socialmarketing develop. Acknowledgement

The authors would liketo thank the following people for their advice and assistance: Pine Migliorino (Cultural Perspectives Pty Ltd); Margaret Lurn (Cultural Partners Australia Pty Ltd); Michelle Toms (Drugs and Alcohol Multicultural Education Centre); Peter Todaroand MichaelCarnit(NSWMulticultural Health Communication Service); Eleni Karantzas (Centre for Culture, Ethnidty and Health); Jackie Van Vugt (VicHealth) and Richard Adezio (Victorian Department of Human Services), References Vo.brid HealthOrganization.A'evention of Psychoactive Subsl.1nce Use_ ~ Selected Review ofWh.lt 'Abrks in the ~ of ~tion. Geneva (CHE): WHO; 2002, 2. Sowden AJ, Arblaster l. Mass media interventionsfor pr~ntiog 5/TlOking in young people (Cochrane Review). In: TheCochrane Dataoase of SystematicReviews, ls5Ue 2, 2003. Oxford{UK): Update Software; 2003. 1.

3. Warnecke RB,laogenburg P, Gruder CL,FlayBR,ja~ LA. Factorsin smoking cessation among participation in a televised intervention. Prev Med 1989;18(6);833-46.

Culturally and linguistically diversesocial marketingcampaigns

4

Carroll TE. 5o(;jal Marketing and TeenageDrinking- An Evaluarjon of Australia's National Drug Otienstve Campaigns 1988-1995 and the Competitive Commun;c:;won Environment WithinWhich TheyOperated IPhDThesis!. Sydney (NSIN): The University of Sydney; 1998. Strategic Inter-govemmf'ntal Nutrition Alliance. An truervenuon Ponfolio to Promote Fruit and Vegetable Comumption - Pan 2 Review of Intervf'ntiom. Melbourne (\llCI: Natlonal Public Health Partnerships; 2000. BaumanA, BellewB,VitaP, BrownW, Owen N.GettingAustralia Active:Towards BetterPracticeforthe Prommion of Physical Activity. Melbourne (VIC): Natonal PublicHealth Partnership;1001. 7. Hanson K. Kikumhm N, Schellenberg JA, et al, Cost effectiveness of social marketing of msectcce-rreateo nets for malaria control in the United Republic of Ianzanta. Bull WHO 2003;81 (41:169-76. 8. Turnbull FM, Burgess MA, Mcintyre PB, er al. The Australian Measles CO/ltrol Program.BullWHO 2001 ;79(9):882-B.

9. Grilli R, Ram$iJYC, MillozziS.Massmedia interventions:effects OIlhealth sevlces utilisation tccchere Review). In; TheCochrane Diltab.Jseof Systf'matic Reviews, Issue 2, 2003. Oxford (UKl; Update Software; 2003. 10. Elkamel FM.Can massmedia pr~ntAIDS: the need for well-planned behaviour change communication programmes.&ill WHO 1996;2(3):449-SB. 11. RedlTl.;ln S, Spencer EA, Sanson-Fisher RW. The role of mass media in changing bealth-relsted behavior: a critical appraisal of two models. Health Promot Int 1990; 5:85-101. 12. BlandjM, Taylor JC, Norman CL,Anderson HR, RamseyjD, Volatile Substance Abuse:Possible Effect of a NatlOOal Glmpaign. London(UK!: StGeorge's Hospital Medical School, Department of Public Health Sciences, Universityof London; 1997. 13. Cramer P.An baminatKlflof Issues ReLued to the lJeo..e/opment Of Communicatiorl Strategies for People from Noo-Eng/ish 5peilkjng Backgrounds. Sydney (NSW): Researchand MarketingGroup: 1997. 14. AustralianBureau of Statistics. 2001 Census Bilsk CommunityProfile. Canberra (ACT): ABS; 2002. 1S. Culpin A,Gleeson S,ThomasM, Bekiaris J.Evaluation of the 'Good Heart, Good life' Project: A Three-year Campaign to Reduce Smoking ~mong the Creel< Communityof Sydney's loner West. Health PromotJ !.usr 1996;6(2):44_9. 16. KingL,ThomasM. Gateoby K,GeorgiouA, Hua M. first aid forscaldscampaign:

reachingSydney's Chinese, Vietnamese,and Arabicspeakingcommunities. Injury Prell1999;5(2);104-8. 17. MitchellH, Hirst S. MitchelllA, StaplesM. Toree/loN. Effectof ethnic media OIl cervicalcancer screening rates. Aust N Zj f\Jblic Healrh 1997;21(3):265-7. 1B. Camit M. Smoke alarms wake you up if there's a fire: a smoke alarm campaign

targeting Arabic,Chinese and Vietnamese communities in NSW. Soc Markf'ting

Q 2002;B(1);52-4. 19. Redding K, TomsM. Drink-drivingin the Il.1lo-Australiiln Community- Reducing the Risks. S)'rloey(N$INJ: Dreg and AlcoholMulticultural EducationCentre; 2001. 20. Commonwealth Department of Healthand Family Services.lmmunise~stralia: Community Education Glmpaign - Research with people from Culturally and

linguisriaJly~ Bad'grouods. Sydney(NSWl: Researchand MarketingGroup; 1998. 21. Carroll T, KatzC. Carvill K. The National Tobacco Campaign: targeting people fromooo-English speakingbackgrounds. In:Australia's NationalTob.lcco Campaign

- Evaluation Report.Volumeone. Canberra (ACT): Commonwealth Department of Health and Aged Care; 2000. 22. BertramS, BalbirN. BallJ,CarrollT. Evaluation of the Llunch Phase of the National Illicit Drugs Campaign Canberra (ACT); Commonwealth of ~ustralia; 2003. 23. !eokinsCN, McPhee 51, BirdJA,Pham GQ, Nguyt'n BH, NguyenT,et al. Effect of

a media-led education campaign on breast and cervicalcaocerscreeoingOlmoog Vietnamese-American women. Prev Med 1999;28(4):395-406. 24. Risse! C. Ward £IN,)orm L Estimatesof smoking and related bf'havioor in an immigrantLebanesecommunity; does s.urvey method matter? Aust N ZJ f\Jblic Health 1999;23(5);534-7.

Authors Andrew). Mi/at, Research and Marketing Croup, Australian Government Department of Health and Ageing, New South Wales Tom E. Carroll, Carroll Communications, New South Wales Jennifer J. Taylor, Research and Marketing Croup, Australian Government Department of Health and Ageing New South Wales

Corres ondence Mr Andrew). Milat, Research and Marketing Croup, Australian Covemmem Department of Health and Ageing CPO Box 9848, Sydney, New South Wales 2001. Tel: (02) 9263 3548; fax: (02) 9263 3549; e-mail: [email protected]

Health Promotion Journal of Australia 2005 : 16 (l)

25