fetal alcohol syndrome prevention using community-based

11
Fetal Alcohol Syndrome Prevention Using Community-Based Narrowcasting Campaigns Deborah Glik, ScD Michael Prelip, DPA, MPH, CHES Amy Myerson, MA Katie Eilers, MPH aimed to increase brand awareness, brand loyalty, product demand, and consumer goods consumption. Alcohol, tobacco, fast food, soft drinks, and highly processed foods are among the many products promoted that can have sig- nificant negative impacts on health. Two concurrent campaigns conducted in 2003 in Southern California used a narrowcast approach within two disadvantaged communities to warn women between the ages of 18 to 35 about the dangers of drinking alcohol during pregnancy. These communities have a high pro- portion of residents with low income, low educational achievement, low literacy rates, and ethnic minority status. The campaign goals were to prevent fetal alcohol syndrome (FAS) by countering the mixed messages women of childbearing age receive about alcohol con- sumption (Caswell, 2004; Center on Alcohol Marketing and Youth, 2002; Wallack, Grube, Madden, & Breed, 1990; Wyllie, Zhang, & Caswell, 1998). The purpose of this article is to describe the planning of these community campaigns, the participatory process used to develop campaign mes- sages, the distribution and tracking of materials, and the evidence that they were seen by the priority population. > BACKGROUND Narrowcasting is a marketing strategy that uses highly focused messages for specific priority populations using media that reaches only a specific group (Chae & Flores, 1998; Fetto, 2002). There is a dearth of research about nar- rowcasting in health promotion, yet this technique is now frequently used in commercial marketing, political cam- paigns, and social marketing (Chae & Flores, 1998, Dretzin, Goodman, & Soenens, 2004; Evans, 2001; Reitman, 1986; Smith-Shomade, 2004). In narrowcasting, the subject matter or message is designed to appeal to a particular demographic, to a social or lifestyle group, or to persons with specific characteristics or orientations (Dretzin et al., Preventing fetal alcohol syndrome (FAS) by encouraging pregnant women to abstain from drinking alcohol com- petes with commercial alcohol marketing. Two FAS- prevention campaigns using a narrowcast approach among young women of childbearing age in two disad- vantaged Southern California communities are com- pared. The design, implementation process, and degree to which campaigns reached the priority populations are the focus of this article. Formative research shows that young women in disadvantaged communities receive mixed messages about dangers of drinking during pregnancy. A social norms approach using positive role models was the most acceptable message strategy based on materials pretesting. Differences in campaign implementation and distribution strategies between communities were docu- mented through program monitoring. Survey research indicated the more viable messaging and implementation strategies. Findings show that low-cost community cam- paigns are feasible; however, variations in messaging, dis- tribution strategies, and saturation levels determine whether such campaigns succeed or fail to reach priority populations. Keywords: fetal alcohol syndrome; formative research; program monitoring; exposure; university– community collaboration; partnerships A challenge in health promotion practice at present is countering the constant drumbeat of commercial media marketing vying for consumers’ attention, Health Promotion Practice Month XXXX Vol. XX, No. XX, XX-XX DOI: 10.1177/1524839907309044 ©XXXX Society for Public Health Education 1

Upload: others

Post on 07-Nov-2021

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Fetal Alcohol Syndrome Prevention Using Community-Based

Fetal Alcohol Syndrome Prevention UsingCommunity-Based Narrowcasting Campaigns

Deborah Glik, ScDMichael Prelip, DPA, MPH, CHES

Amy Myerson, MAKatie Eilers, MPH

aimed to increase brand awareness, brand loyalty, productdemand, and consumer goods consumption. Alcohol,tobacco, fast food, soft drinks, and highly processed foodsare among the many products promoted that can have sig-nificant negative impacts on health.

Two concurrent campaigns conducted in 2003 inSouthern California used a narrowcast approach withintwo disadvantaged communities to warn women betweenthe ages of 18 to 35 about the dangers of drinking alcoholduring pregnancy. These communities have a high pro-portion of residents with low income, low educationalachievement, low literacy rates, and ethnic minoritystatus. The campaign goals were to prevent fetal alcoholsyndrome (FAS) by countering the mixed messageswomen of childbearing age receive about alcohol con-sumption (Caswell, 2004; Center on Alcohol Marketingand Youth, 2002; Wallack, Grube, Madden, & Breed, 1990;Wyllie, Zhang, & Caswell, 1998). The purpose of this articleis to describe the planning of these community campaigns,the participatory process used to develop campaign mes-sages, the distribution and tracking of materials, and theevidence that they were seen by the priority population.

>>BACKGROUND

Narrowcasting is a marketing strategy that uses highlyfocused messages for specific priority populations usingmedia that reaches only a specific group (Chae & Flores,1998; Fetto, 2002). There is a dearth of research about nar-rowcasting in health promotion, yet this technique is nowfrequently used in commercial marketing, political cam-paigns, and social marketing (Chae & Flores, 1998, Dretzin,Goodman, & Soenens, 2004; Evans, 2001; Reitman, 1986;Smith-Shomade, 2004). In narrowcasting, the subjectmatter or message is designed to appeal to a particulardemographic, to a social or lifestyle group, or to personswith specific characteristics or orientations (Dretzin et al.,

Preventing fetal alcohol syndrome (FAS) by encouragingpregnant women to abstain from drinking alcohol com-petes with commercial alcohol marketing. Two FAS-prevention campaigns using a narrowcast approachamong young women of childbearing age in two disad-vantaged Southern California communities are com-pared. The design, implementation process, and degree towhich campaigns reached the priority populations are thefocus of this article. Formative research shows that youngwomen in disadvantaged communities receive mixedmessages about dangers of drinking during pregnancy. Asocial norms approach using positive role models was themost acceptable message strategy based on materialspretesting. Differences in campaign implementation anddistribution strategies between communities were docu-mented through program monitoring. Survey researchindicated the more viable messaging and implementationstrategies. Findings show that low-cost community cam-paigns are feasible; however, variations in messaging, dis-tribution strategies, and saturation levels determinewhether such campaigns succeed or fail to reach prioritypopulations.

Keywords: fetal alcohol syndrome; formative research;program monitoring; exposure; university–community collaboration; partnerships

Achallenge in health promotion practice at presentis countering the constant drumbeat of commercialmedia marketing vying for consumers’ attention,

Health Promotion PracticeMonth XXXX Vol. XX, No. XX, XX-XXDOI: 10.1177/1524839907309044©XXXX Society for Public Health Education

1

Page 2: Fetal Alcohol Syndrome Prevention Using Community-Based

2004; Evans, 2001; New California Media, 2005). Definedas limited-reception broadcasting services that reach spe-cial interest groups (Broadcasting Services Act, 1992), nar-rowcast media research identifies specific beliefs, attitudes,and values of a particular group or segment and then framesmessages that reflect these values, a contrast to broadcastmedia, in which mainstream messages reach mass audi-ences (Mendelsohn & Nadeau, 1996). Broadcast media areextensive and penetration is shallow, whereas narrowcastmedia are intensive and penetration is presumed to be deep(Chae & Flores, 1998; Smith-Shomade, 2004). A relatedconcept is niche or concentrated marketing (Kotler,Roberto, & Lee, 2002); however, narrowcasting uses tar-geted media that reach only specific groups (Evans, 2001;Rimal & Adkins, 2003). As markets become increasinglysegmented, narrowcasting has become an important way toreach groups with unique beliefs, customs, or lifestyles orhard-to-reach groups (Freimuth, 1990; Hoff & Green, 2002;New California Media, 2005). Narrowcasting is also effec-tive in communities saturated with commercial advertising(Fetto, 2002; Glik & Halpert-Schilt, 2001; Mendelsohn &Nadeau, 1996). In health promotion, narrowcast campaignsare sometimes referred to as targeted or tailored social mar-keting campaigns or targeted behavior change communica-tion programs (Kreuter & Wray, 2003; Rimal & Adkins,2003). However, tailoring uses data on personal preferencesand designs messages accordingly, whereas narrowcastingis a form of targeted marketing using restricted media chan-nels (Kreuter & Wray, 2003; Rimal & Adkins, 2003).

Creating messages and materials for members of specificpriority populations increases the probability that groupmembers will be exposed to, will attend to, and will takeinterest in the messages presented (Fetto, 2002; Reitman,1986). In a media-saturated world, audiences may “tuneout” messages that have no cultural relevance (Andreasen &Kotler, 2003; Hoff & Green, 2002; Hornik, 2002; Stewart,

Pavlou, & Ward, 2002). If a participatory approach is used, members of the community participate in campaigncreation and implementation, thereby increasing interest by community members (Bryant et al., 2007; Bryant,McCormack, Brown, Landis, & McDermott, 2000).

Narrowcasting campaigns are based on social mar-keting principles of extensive audience, market, andformative research conducted within population seg-ments (Andreasen & Kotler, 2003; Bloom & Novelli,1981; Grier & Bryant, 2005; Kotler et al., 2002; Maibach,Rothschild, & Novelli, 2002). As people become depen-dent on media for setting social norms (McQuail, 1987;Rubin, 2002), campaigns achieve objectives if personsare exposed and pay attention to materials, a phenom-enon that can occur when a person has been exposed tomaterials three to six times (Andreasen & Kotler, 2003;Hornik, 2002; Paisley, 2001; Stewart et al., 2002).

Linked to the idea of market saturation and exposureis the concept of product placement. In narrowcasting,higher exposure is achieved by placing more materialsin circumscribed geographic areas or in media con-sumed by persons in relatively narrow or specializedmarket niches (Dretzin et al., 2004; Fetto, 2002; NewCalifornia Media, 2005). Community-based narrowcastcampaigns typically use low-cost print materials suchas posters, fliers, T-shirts, or take-one cards (Evans,2001), although new digital technology may be chang-ing this reality (Hoff & Green, 2002). Ideally, narrowcastmaterials are placed in public or private spaces such asdressing rooms in stores, examination rooms in clinics,and restrooms in theaters or shops or bars, placeswhere the audience member is alone and can focus andreflect on the meaning of the message.

Successful narrowcast campaigns reviewed by Glik,Prelip, Myerson, and Eilers (2005) include preventingsexual assaults in an ultra-orthodox Jewish community(Boehm & Itzhaky, 2004), promotion of emergency contra-ception in Mexico City (Ellertson et al., 2002; Heimburgeret al., 2002), communicating male and female condom use(Bull, Cohen, Ortiz, & Evans, 2002), and communicatingrisks of alcohol use during pregnancy to African Americanand Latina adolescent girls (Halpert-Schilt, Glik, & Zhang,2001). This article is a comparative case study of two suchcampaigns and details their development and implemen-tation using results from formative, process, and surveyresearch.

>>DRINKING DURING PREGNANCY

Drinking alcohol during pregnancy is a behavior linkedto increased probability that an alcohol-exposed infantwill be born with fetal alcohol spectrum disorders (Floyd,O’Connor, Sokol, Bertrand, Cordero, 2005; Jones, Smith, &

2 HEALTH PROMOTION PRACTICE / Month XXXX

The Authors

Deborah Glik, ScD, is a professor in the Department ofCommunity Health Sciences in the UCLA School of PublicHealth in Los Angeles, California.

Michael Prelip, DPA, MPH, CHES, is an adjunct associateprofessor in the Department of Community Health Sciences inthe UCLA School of Public Health in Los Angeles, California.

Amy Myerson, MA, is assistant director for grants coordi-nator at Wheeler Clinic in Plainville, Connecticut.

Katie Eilers, MPH, is the executive director of INMED/Mothernet LA in Compton, California.

Page 3: Fetal Alcohol Syndrome Prevention Using Community-Based

Ulleland, 1973; Stratton, Howe, & Battaglia, 1996). Themost severe manifestation of this is FAS, is a condition inwhich the infant has severe facial abnormalities and sig-nificant impairments in neurodevelopment and physicalgrowth (Floyd et al., 2005). Current recommendations arethat no amount or type of alcohol is safe to consume at anytime during pregnancy, as research suggests that womenand their unborn babies have different levels of tolerancefor alcohol (Stratton et al., 1996; Willford, Leech, & Day, 2004). Although full-blown FAS with severe physi-cal and mental defects is rare even among children ofalcoholic mothers, prenatal alcohol exposure often leadsto less severe problems, or “fetal alcohol effects” (FAE),such as deficits in learning and memory (Willford, Leech, &Day, 2004), impairments in academic achievement(Goldschmidt, Richardson, Stoffer, Geva, & Day, 1996), mildmental retardation, or birth defects (Floyd et al., 2005;Hankin, 2002; National Institute on Alcohol Abuse andAlcoholism, 1991).

Despite years of educational outreach, counseling, andmedia messages, many women continue to drink alcoholduring pregnancy (Floyd et al., 2005; Hankin, 2002). Forexample, in 1999, 12.9% of women drank alcohol duringpregnancy (Centers for Disease Control and Prevention[CDC], 2002), down from 16.3% reported by CDC in 1995 (CDC, 2002; Hankin, 2002). Prevalence of drinkingduring pregnancy varies by ethnicity, with rates highestfor Caucasian women (25.4%), followed by AfricanAmerican (12.2%), and Latina women (10.6%), with FASand FAE rates highest among groups with the most drink-ing. However, national rates may not reflect local realities.A recent study among low-income women in Los AngelesCounty found that 30.0% of Caucasian women, 29.4% ofAfrican American, and 20.9% of Latina women reporteddrinking during pregnancy, with more acculturated Latinawomen reporting rates (29.7%) that approached othergroups (O’Connor & Whaley, 2003).

These rates should not be surprising, as alcohol compa-nies are adept at creating and maintaining market share,with many current campaigns to increase consumptionamong youth, young women, and members of ethnic minor-ity groups (Caswell, 2004; Center on Alcohol Marketing andYouth, 2002; Roberts, Christenson, & Henriksen, 2002;Wallack, 1981; Wechsler et al., 2003; Wyllie et al., 1998).Comprehensive alcohol advertising campaigns normalizedrinking, portraying it as an integral aspect of Americansocial life linked to social acceptability, popularity, suc-cess, and relaxation (Grant & Dawson, 1997; Kuo,Wechsler, Greenberg, & Lee, 2003). Thus, it is no wonderthat alcohol consumption among young persons of colorand young women have increased (CDC, 2003; Center onAlcohol Marketing and Youth, 2002; Substance Abuse andMental Health Services Administration, 2003).

>>METHOD

Campaigns were conducted in two low-income areas inCalifornia, both of which have multiple ethnic groups. InBakersfield, a southern Central Valley city in Kern County,the campaign took place in eight contiguous zip codes, thesponsoring agency having had 20 years of experience inconducting community-outreach programs. In Compton, a low-income community in Los Angeles County, thecampaign was conducted in three contiguous zipcodes, sponsored by a community-based agency and acommunity coalition established to improve health andsocial disparities.

Campaigns described could not have occurred withoutpartnerships established with well-respected community-based agencies and a marketing company that was familiarwith social marketing. These partnerships were invaluablefor helping form locally based advisory committees madeup of community leaders and members of the target audi-ence to gain input from community members about thecampaign themes, materials, and suggested implementa-tion and distribution plans. These same partnershipsbecame a key to distribution of campaign materials, asbusinesses owned and run by community members hostedthe materials produced. The marketing company under-stood that the campaigns were community driven anddiverse. Thus, much effort went into educating marketingprofessionals about the communities and the issues. Ineach community, a series of steps led to campaign materialcreation, production, distribution, and evaluation.

Formative Research

Community advisory groups suggested audience seg-mentation based on ethnicity and language, as these attrib-utes were seen to influence the types of alcohol peopledrank and where they drank, a community perceptionsupported in the literature (Casswell, 2004). In Bakersfield,this resulted in four segments: Caucasian women, AfricanAmerican women, Latina English-speaking women, andLatina Spanish-speaking women. In our campaign inCompton, all these segments except Caucasian wereextant. Focus group interviews were conducted withineach segment to assess cultural and psychosocial cor-relates of drinking behaviors and concerns linked topregnancy. A total of six focus groups were conducted(two African American, two Caucasian, two Latina) in Bakersfield. In Compton, four initial focus groupswere conducted (two African American, two Latina).Community input based on focus group interviews andadvisory board input suggested that campaigns should begeared to women ages 18 to 35, as it was perceived by com-munity members that younger women drank more, that

Glik et al. / FETAL ALCOHOL SYNDROME PREVENTION 3

Page 4: Fetal Alcohol Syndrome Prevention Using Community-Based

women are responsible for their drinking behaviors, andthat younger women are the target for commercial alcoholadvertising (Center on Alcohol Marketing and Youth,2002; Wyllie et al., 1998)

In Bakersfield, participants were concerned about theconflicting messages they received about drinking duringpregnancy from health care providers, their family andfriends, prevention organizations, and alcohol advertis-ing. Women learned from many sources that they shouldnot drink during pregnancy, but they also hear that it issafe to drink certain types of alcohol at certain times, evenfrom health care providers, whose opinions they respect.Participants had a disconnect between knowing that preg-nant women should not drink and understanding thateven one glass of wine after dinner counts as drinking.They also wanted to know exactly how much alcohol wassafe to drink before it would affect the baby. Furthermore,many women knew a friend or relative who drank duringpregnancy and had a healthy child. Given the conflictingmessages they received, women wanted clear-cut and spe-cific messages.

There was general agreement that strong graphic mes-sages and eye-catching colors and phrases are important.There was less agreement about the campaign approach.Some wanted to shock people with realistic pictures ofbabies who had FAS. Others thought that this might beless effective and stressed that women do have controlover the health of their babies. There was also recognitionthat some women drink and have healthy babies, whereasother women who do not drink have sick babies.

In Compton, women were also concerned about themixed messages they were receiving about the dangers ofdrinking during pregnancy. They wanted clear, definitiveguidelines, and some felt that beer, wine, or wine coolerswere not “alcohol” and therefore did not have the sameeffects as alcohol, liquor, or distilled spirits. They toowanted to see images of strong motherly women whowere worried about their children’s health. Here too therewas ambivalence about showing the consequences ofactions through graphic depictions of sick babies, vulner-able fetuses, or retarded children. Some thought that theshock value would be effective, whereas others thoughtthat some would not relate to these types of messages.

Message Development and Pretesting

The goal of this was not to create duplicate campaignsin each community but to create and distribute materialsthat were uniquely suited to each community. Many dif-ferent picture and slogan prototypes were developed foreach community. These were pretested with communitymembers and community leaders through additionalfocus group interviews to solicit feedback and to refine

material, messages, and layouts. Community leaders weremembers of health coalitions and interest groups in eachcommunity who were representative of target populations.Participants were asked to prioritize and evaluate materi-als created, to assess words, phrases, and images thatwould make pregnant or potentially pregnant women cur-tail drinking, and to assess suitability for the community.

Despite suggestions in initial formative research thatcommunity members wanted fear-based messages withpictures of sick children, pretesting of materialsdemonstrated that audiences favored the social normsapproaches showing what “idealized” women shoulddo if they were pregnant or thought they were preg-nant. Thus, modeling of socially acceptable or “good”behaviors was more appealing than showing the conse-quences of inadvisable or “bad” behaviors: the sickchild depiction. This supports the basic social value ofbeing a “good” parent.

Other considerations were choosing role models. Theinitial impetus was to select models from people in thecommunity. However, materials pretesting revealed thatcommunity members wanted someone who looked likethem but was better looking. In Bakersfield, the actualslogan selected, suggested by focus group respondents,was “Missed your period, don’t drink period,” whichmirrored concerns about younger women being at riskand women not knowing they were pregnant and contin-uing to drink (see Figure 1). The Spanish version of thatcampaign, “Quien cree quee espera un bebé, no bebe” (“Ifyou think you are pregnant, do not drink”), was not a lit-eral translation of the English version, as we were able toproduce a Spanish version that had a play on words forbaby (bebé) and drinking (bebe). In the Bakersfield cam-paign, supplemental materials included 5 in. × 7 in. take-one cards that were smaller versions of the posters on thefront and that had detailed information regarding theimpact of drinking or doing drugs during pregnancy onthe back side, specifically citing risk factors and conse-quences of drinking during pregnancy.

Compton campaign slogans and images echoedassertiveness and empowerment, with the slogan “I’m amama-to-be, no drinks for me” (for all groups exceptAfrican American, who preferred “mother-to-be” to “mama-to-be”; see Figure 2). The message in Spanish, “Por mibebé, no beberé!” (“For my baby, I will not drink!”) was not a literal translation of the English. However, the con-notation (I will not drink during pregnancy) and feeling of empowerment in both posters were consistent. InCompton, supplemental materials included T-shirts forpregnant women distributed at clinics and businesses thathad urban poetry that echoed campaign themes.

In both campaigns, audiences favored strong, saturatedcolors and sharp, clear images. The tone was serious as

4 HEALTH PROMOTION PRACTICE / Month XXXX

Page 5: Fetal Alcohol Syndrome Prevention Using Community-Based

Glik et al. / FETAL ALCOHOL SYNDROME PREVENTION 5

FIGURE 1 Bakersfield Posters: Caucasian, African American, and Latina Spanish PostersNOTE: The Latina English poster is not shown. The image is the same as in the Latina Spanish poster but with the text ofthe other English-language posters.

FIGURE 2 Compton Posters: African American, Latina English, and Latina Spanish Posters

Page 6: Fetal Alcohol Syndrome Prevention Using Community-Based

communities had rejected humorous prototypes. Wefound that our audiences often had problems with spe-cific words, phrases, body language, or images. Creatingmaterials that were acceptable and impactful was a drawnout, iterative process. Although we did include men inmany of the prototype materials, African American andCaucasian women did not see their presence as an addedvalue to conveyed messages.

Implementation and Process Tracking

Effective and efficient distribution and maintenance ofcampaign materials are concerns in any campaign, but thisis particularly the case in a low-resource context. For thecampaigns described, a decision was made not to spendscarce project resources on paid placements but instead tosupport project staff in distributing materials among busi-nesses and organizations serving these communities.Thus, both in Bakersfield and Compton, a full-time staffperson from the community was hired by participatingcommunity-based organizations (CBOs) for helping toorganize community outreach. Potential placement loca-tions were determined first by comprehensive mapping ofbusinesses and organizations in the community.

In both campaigns, materials were placed wherewomen of childbearing age actually go, determined byobservations of outreach workers. Convenience stores,restaurants, beauty salons, laundromats, car-repair shops,check-cashing businesses, and clothing stores wereincluded. Materials were also placed in clinics and schoolsand at other nonprofit agencies. To achieve these place-ments, a great deal of personal “promotion” was necessary.In both Bakersfield and Compton, many business ownersor managers of for-profit ventures were not interested inposting campaign messages in their establishments whenfirst approached by community-outreach workers. Promo-tional strategies to encourage participation included hand-outs about the project, background on participatingagencies, and letters from other reputable organizations orbusinesses in the community supporting the campaignand the sponsoring organizations. Over time and as moreand more community-based establishments participated,much more openness to posting messages was experi-enced. Large chain stores and restaurants were averse toposting materials, whereas small, community-based andrun business establishments saw these materials as a wayto support the community.

Community-outreach workers not only negotiatedplacement but also tried to negotiate more visible place-ment. Prime space included store windows or wall spacethat was at eye level. Other locations favored were thosewhere the priority population is “captive,” such as wait-ing rooms, examination rooms, and restrooms. Multiple

sets of materials in one place were more conducive to cap-turing attention, a narrowcasting technique (Dretzin et al.,2004). Materials maintenance during the 8 months of thecampaign was paramount, as posters were often vandal-ized or removed. Outreach workers who had establishedgood relationships with businesses and organizations hadmore success replenishing materials in these sites.

Process-tracking data delineated where, when, andwhat materials were placed in participating communities.Community-outreach workers devised and used stan-dardized spreadsheets they developed to document theiroutreach activities in both Bakersfield and Compton:whether visits were initial or maintenance visits, date of the visit, contact name (e.g., the owner or manager) atthe location, name of the location, type of the location(e.g., nail salon, restaurant, clinic, pharmacy, school, etc.),address and telephone number, and number and type ofmaterials distributed at the location.

Assessing Exposure

Two-wave repeated cross-sectional survey data werecollected to evaluate the impact of these narrowcast cam-paigns and to assess changes over time in knowledge, atti-tudes, and behavior associated with drinking and, morespecifically, drinking during pregnancy. These outcomedata are reported elsewhere (Prelip, Glik, Myerson, Ang, &Eilers, 2006). Survey questionnaires based on previousresearch (Halpert-Schilt et al., 2001) and formativeresearch assessed reported exposure rates and factors thatpredict higher exposure among the priority population.Data for the surveys, which had both English and Spanishversions, were collected in two ways. A random-digit dial(RDD) survey of eligible women at baseline was repeated8 months later at the conclusion of the campaign. In addi-tion, survey data were collected at women’s health clinicsand physicians’ offices. Both types of surveys, which tookapproximately 15 minutes to complete, were adminis-tered by trained interviewers, were anonymous, and werecollected in Bakersfield and Compton. For this exposi-tion, data from the follow-up wave of surveys are used, asthese contained relevant exposure variables and variablespredicting exposure to campaign materials.

All respondents were women between the ages of 18and 35 residing in specific zip codes where campaignswere conducted. For this analysis, measures on surveyinstruments included the following: exposure to mate-rials, pregnancy status, education, motherhood status(whether participants had children), drinking behavior,location (Compton or Bakersfield), ethnicity, primarylanguage spoken, and age. Two interaction terms inanalysis were drink alcohol × pregnancy and drinkalcohol × education.

6 HEALTH PROMOTION PRACTICE / Month XXXX

Page 7: Fetal Alcohol Syndrome Prevention Using Community-Based

Univariate descriptive methods were used to summa-rize process tracking data. To compare exposure ratesbetween communities, bivariate analysis using nonpara-metric methods was conducted. Logistic regression wasused to determine characteristics of respondents that pre-dicted exposure to campaign materials.

>>RESULTS

Results reported are based on process monitoring andsurvey data standardized across sites. The rationale for the

analysis is that although campaign messages were different,the purpose of the campaigns was the same, the target audi-ences were the same (minority and or underprivilegedwomen), and the methods used to work with communities,place materials, and conduct process and survey researchwere the same. Unlike Compton, there was a subset ofCaucasian respondents in Bakersfield. As ethnicity did notinfluence outcomes, comparing across sites is justified andgives insight about what works in a narrowcast approach.

Table 1 results are based on process tracking data, com-paring materials placement between the two campaign

Glik et al. / FETAL ALCOHOL SYNDROME PREVENTION 7

Table 1Campaign Tracking: Frequency and Location of Materials Posted

Compton Bakersfield

# of posters 2,860 1,293Supplemental materials 5,517 T-shirts 65,063 take-one cardsInitial locations 170 540Additional maintenance visits 483 329# of ZIP codes 3 8

Frequency (%) Frequency (%)

Types of Locations Examples n % n %

Civic Child care centers, churches, community centers, public transportation hubs, parks, women, infants, and children centers 23 13.5 43 8.0

Medical Primarily physician clinics but also an optometrist, private physician offices, pharmacies, and substance abuse centers 23 13.5 63 11.7

Restaurants Most were fast food, including donut shops or taco stands; very few sit-down restaurants 15 8.8 40 7.4

Retail—Service Nail salons, cleaners, check-cashing centers 41 24.1 113 20.9

Retail—Food Food stores, liquor stores, markets, mini markets 24 14.1 200 37.0

Retail—Stores Clothing, appliances, furniture, video, flowers, music, pet, computer, jewelry, department stores 30 17.6 37 6.9

Retail—Discount Discount stores, such as a 99 cent store 7 4.1 31 5.7

Education High schools, community colleges, universities 4 2.4 7 1.3

Unknown 3 1.8 6 1.1Total locations 170 100.0 540 100.0

Page 8: Fetal Alcohol Syndrome Prevention Using Community-Based

communities. Most striking is that more than twice asmany posters were placed in Compton (2,860) than inBakersfield (1,293) in one third as many sites (170 forCompton, 540 for Bakersfield). It should be noted that the geographic area for Bakersfield was more spread out,composing eight contiguous zip codes for Bakersfield,whereas in Compton there were three contiguous zipcodes. Communities were similar in regard to where mate-rials were placed, with the most popular venues beingretail stores and service businesses. Very few materialswere placed in educational institutions, whereas somematerials were placed in medical and civic organizations.

Table 2 is an analysis of exposure rates based on the twodifferent survey instruments. As can be seen, there is verylittle exposure to the campaign noted in the RDD samples.Based on RDD results, only 11.2% of respondents in

Compton noted the campaign, compared to 7.2% inBakersfield. In contrast, the clinic-based survey inCompton showed that 54.2% of those surveyed recognizedthe campaign. Not so in Bakersfield, where clinic-baseddata revealed that only 11.2% of those questioned recog-nized the campaign, a rate of exposure only slightly higherthan that for the RDD sample.

With such low reported exposure rates in Compton andBakersfield using RDD data, clinic data from Comptonand Bakersfield were used to determine characteristics ofpersons more likely to have been exposed to the cam-paigns. First, data were analyzed separately (i.e., withincommunities). In Compton, women who had graduatedfrom high school were more likely to see the poster (χ2 =6.823, p = .009) and to see the T-shirt (χ2 = 4.589, p = .032).There were no differences in exposure associated withethnicity (Latina, African American), age group, preg-nancy status, or motherhood status. In Bakersfield, bycontrast, drinkers were more likely than nondrinkers to beexposed to our campaign (χ2 = 6.697, p = .010). For thisgroup, there were no differences in exposure with ethnic-ity (Latina, African American, Caucasian), pregnancystatus, education, age group, or motherhood status.

Data were aggregated across sites after exploratoryanalysis, suggesting similar populations. In Table 3, multi-variate logistic regression was used to assess those variablesmost associated with exposure. In forced-entry forwardanalysis and then in backward stepwise elimination, threevariables were implicated—education level, drinking, andbeing a resident of Compton—confirming bivariate analy-ses. No interaction effects attained significance.

8 HEALTH PROMOTION PRACTICE / Month XXXX

Table 2Exposure Rates

Compton (%) Bakersfield (%)

Clinic dataOur campaign 54.2 11.3Our poster 42.7 11.3Our T-shirt 32.9 —Random-digit dial dataOur campaign 11.4 7.2Our poster 9.9 7.2Our T-shirt 1.9 —

Table 3Odds Ratios and 95% Confidence Intervals (CIs) for Logistic Regression Models of Predictors of Campaign Exposure

Forced Entry Backwards Stepwise

Variable Odds Ratio Lower CI Upper CI Odds Ratio Lower CI Upper CI

Pregnant 1.172 0.613 2.238 —Education 1.972* 1.056 3.680 1.630* 1.009 2.678Have kids 1.235 0.691 2.205 —Drink 3.382* 1.045 10.949 2.162* 1.078 4.334Compton 9.488*** 5.189 17.348 7.761*** 4.601 13.092Age 1.022 0.970 1.077 —Latina 0.861 0.360 2.059 —African American 0.538 0.198 1.460 —Drink × pregnant 0.464 0.157 1.367 —Drink × education 0.526 0.168 1.641 —

*p < .05. **p < .01. ***p < .001.

Page 9: Fetal Alcohol Syndrome Prevention Using Community-Based

>>DISCUSSION

In the two community-driven campaigns described,print media placed in retail and civic establishments in thecommunities of interest had to vie for both placement andthe attention of consumers in the current crowded mediamarketplace. Even so, in one community, Compton, morethan 50% of a clinic-based audience reported exposure.These are women who may be more sensitized to the issue,as RDD surveys of a more general population segmentrevealed low exposure rates. Although developing,printing, and distributing thousands of posters ischeaper than developing paid advertisements for tele-vision, radio, or outdoor advertising, a selected group ofresidents responded to the campaign, those who are attend-ing health clinics. Messages produced did not reach all res-idents at risk, suggesting that this approach may be mostsuccessful in very targeted, “primed” groups.

The finding that exposure rates were higher inCompton than in Bakersfield also sheds light on whatmakes a difference in this type of campaign. Communitieschose their own messages, and community-outreachworkers distributed materials based on the configurationof patterns of daily activity in their communities. Oneobvious difference between communities was the level ofsaturation attained: In Compton, this was much higher, asthere were more materials placed in a smaller geographicarea, which may have increased potential exposure. Asecond unanticipated difference was how materials wereplaced. In Bakersfield, not only were materials more geo-graphically spread out, but in most venues there was only one poster. In contrast, in Compton, the community-outreach coordinator posted more materials in fewerplaces, which in itself is an advertising technique usedtoday with print materials for narrowcast campaigns(Dretzin et al., 2004). More materials clustered togethertends to draw more attention, as persons are more likelyto process information if the poster looks familiar or ifthey have just seen it in another place, the déjà vu phe-nomenon (Stewart et al., 2002).

That messages and materials might have been morememorable in Compton than they were in Bakersfield mayalso be attributable to colors, images, and messages. InBakersfield, stock footage was used on a medium bluebackground that tended to recede into the background, andhead shots made models look more passive. In contrast,posters in Compton were produced in highly saturatedreds, oranges, and bright blues, which tend to advancetoward the viewer, making them more noticeable, andbody images selected were those of assertive and deter-mined women with their hands outreached to the viewer,as if asking the viewer to stop and pay attention. The

Compton message, “I’m a mama-to-be,” promotes mother-hood and empowerment, whereas the Bakersfield mes-sage, “Missed your period?” elicits feelings of “I messedup. Not only am I pregnant, but now I can’t drink,” whichmay have been less appealing to the audience. Both cam-paigns did use social norm messages, which has beenfound to be effective in alcohol prevention (Rimal & Real,2003), tobacco control, lifestyle campaigns, and nutritioncampaigns.

In relatively contained geographic areas, limitedresources can be leveraged in community narrowcastcampaigns by soliciting free or low-cost placements formaterials in local venues. This in turn means that much ofthe effort and resource expenditure at the local level is forstaff and volunteers who can place and maintain materi-als and for the cost for printing the materials. Moreover, inboth communities, CBO partners were welcoming of cam-paign activities, as they were relatively inexpensive, hada clear deliverable message, created a good public rela-tions image, and were seen to be an opportunity to facili-tate community input and outreach.

The two campaigns reported here are only one way tocounter constant reminders to young women to drink thatdiscount or omit the real dangers of drinking for theirunborn children. In a media saturated world, it is impor-tant to counter in as many ways as possible misinforma-tion or omissions of information that have serious healthimplications (Casswell, 2004). Narrowcasting is a type ofcampaign that makes sense for communities, as they aretypically contained and less expensive than larger efforts(Hoff & Green, 2002). Exposure levels can be increased ifcommunity members participate in their design, imple-mentation, and evaluation. Given the experiences of thesetwo campaigns, it can be said that narrowcastingapproaches are feasible and cost-effective ways communi-ties can leverage resources to promote positive healthbehaviors and dispel myths promoted by commercial mar-keting, if care is taken in the development of memorablematerials and if placement ensures sufficient exposure.

REFERENCES

Andreasen, A. R., & Kotler, P. (2003). Strategic marketing for non-profit organizations (6th ed.). Upper Saddle River, NJ: Prentice Hall.

Bloom, P. N., & Novelli, W. D. (1981). Problems and challenges insocial marketing. Journal of Marketing, 45, 79-88.

Boehm, A., & Itzhaky, H. (2004). The social marketing approach:A way to increase reporting and treatment of sexual assault. ChildAbuse and Neglect, 28, 253-265.

Bryant, C. A., Brown, K. R., McDermott, R. J., Forthofer, M. S.,Bumpus, E. C., Calkins, S. A., et al. (2007). Community-based pre-vention marketing: Organizing a community for health behaviorintervention. Health Promotion Practice, 8, 154-163.

Glik et al. / FETAL ALCOHOL SYNDROME PREVENTION 9

Page 10: Fetal Alcohol Syndrome Prevention Using Community-Based

Bryant, C. A., McCormack, F. M., Brown, K. R., Landis, D., &McDermott, R. J. (2000). Community-based prevention marketing:The next steps in disseminating behavior change. AmericanJournal of Health Behavior, 24, 61-68.

Bull, S. S., Cohen, J., Ortiz, C., & Evans, T. (2002). The POWER cam-paign for promotion of female and male condoms: Audience researchand campaign development. Health Communication, 14(4), 475-491.

Casswell, S. (2004). Alcohol brands in young people’s everydaylives: New developments in marketing. Alcohol and Alcoholism,39(6), 471-476.

Center on Alcohol Marketing and Youth. (2002, September).Overexposed: Youth a target of alcohol advertising in magazines.Washington, DC: Author.

Centers for Disease Control and Prevention. (2002). Alcohol useamong women of childbearing age—United States, 1991-1999.Morbidity and Mortality Weekly Report, 51, 273-276.

Centers for Disease Control and Prevention. (2003). Behavioral riskfactor surveillance system: Prevalence data Atlanta, GA: Author.

Chae, S., & Flores, D. (1998, March). Broadcasting versus narrow-casting. Information Economics and Policy, 10(1), 41-57.

Dretzin, R., Goodman, B. (Directors), & Soenens, M. (Writer). (2004,October). The persuaders [Television broadcast]. In D. Fanning(Executive Producer), Frontline. New York: Public BroadcastingService.

Ellertson, C., Heimburger, A., Acevedo-Garcia, D., Schiavon, R.,Mejia, G., & Corona, G. (2002). Information campaign and advo-cacy efforts to promote access to emergency contraception inMexico. Contraception, 66, 331-337.

Evans, T. (2001). Narrowcasting. Ventura, CA: Educational MessageServices.

Fetto, J. (2002). Targeted media. American Demographics, 24(7), 1-7.

Floyd, R. L., O’Connor, M. J., Sokol, R. J., Bertrand, J., & Cordero, J. F.(2005). Recognition and prevention of fetal alcohol syndrome.Obstetrics & Gynecology, 106(5), 1059-1064.

Freimuth, V. (1990). The chronically uninformed: Closing the know-ledge gap in health. In E. Ray & D. Lewis (Eds.), Communications andhealth (pp. 168-183). Hillsdale, NJ: Lawrence Erlbaum.

Glik, D. C., & Halpert-Schilt, E. (2001). Public education and outreach.In N. Halfon, E. Schulman, & M. Hochstein (Eds.), Building commu-nity systems for young children. Retrieved September 12, 2007, fromhttp://www.healthychild.ucla.edu/icecs/resource/materials/public/glik.public.education.pdf

Glik, D. C., Prelip, M., Myerson, A., & Eilers, K. (2005). Narrowcastcampaign guide for community programs: Creating health messagesfor targeted media campaigns. Retrieved September 12, 2007, fromhttp://www.ph.ucla.edu/chs/hmrg/

Goldschmidt, L., Richardson, G. A., Stoffer, D. S., Geva, D., & Day, N. L.(1996). Prenatal alcohol exposure and academic achievement at age six:A nonlinear fit. Alcoholism: Clinical and Experimental Research, 20,763-770.

Grant, B. F., & Dawson, D. A. (1997). Age at onset of alcohol use andits association with DSM-IV alcohol abuse and dependence: Resultsfrom the National Longitudinal Alcohol Epidemiologic Survey.Journal of Substance Abuse, 9, 103-110.

Grier, S., & Bryant, C. A. (2005). Social marketing in public health.Annual Review of Public Health, 26, 319-339.

Halpert-Schilt, E., Glik, D., & Zhang, W. (2001). Narrowcastingrisks of drinking during pregnancy among African American andLatina adolescent girls. Health Promotion Practice, 2(3), 222-232.

Hankin, J. R. (2002). Fetal alcohol syndrome prevention research.Alcohol Research and Health, 26(1), 58-65.

Heimburger, A., Acevedo-Garcia, D., Schiavon, R., Langer, A., Mejia, G.,& Cornoa, G. (2002). Emergency contraception in Mexico City:Knowledge, attitudes, and practices among providers and potentialclients after a 3-year introduction effort. Contraception, 66, 321-329.

Hoff, T., & Green, L. (2002). Shouting to be heard: Public service adver-tising in a new media age. Menlo Park, CA: Kaiser Family Foundation.

Hornik, R. (Ed.). (2002). Public health communications: Evidencefor behavior change. Mahwah, NJ: Lawrence Erlbaum.

Jones, K. L., Smith, D. W., & Ulleland, C. N. (1973). Pattern of mal-formation in offspring of chronic alcoholic mothers. Lancet, 1,1267-1271.

Kotler, P., Roberto, N., & Lee, N. (2002). Social marketing:Improving the quality of life (2nd ed.). Thousand Oaks, CA: Sage.

Kreuter, M. W., & Wray, R. J. (2003). Tailored and targeted healthcommunication: Strategies for enhancing information relevance.American Journal of Health Behavior, 27(Suppl. 3), S227-S232.

Kuo, M., Wechsler, H., Greenberg, P., & Lee, J. E. (2003). The mar-keting of alcohol to college students: The role of low prices andspecial promotions. American Journal of Preventive Medicine,25(3), 204-211.

Maibach, E. W., Rothschild, M. L., & Novelli, W. D. (2002). Socialmarketing. In K. Glanz, B. K. Rimer, & F. M. Lewis (Eds.), Healthbehavior and health education: Theory, research and practice(3rd ed., pp. 437-461). San Francisco: Jossey-Bass.

McQuail, D. (1987). Processes of media effects. Mass communi-cation theory (2nd ed.). Newbury Park, CA: Sage.

Mendelsohn, M., & Nadeau, R. (1996). The magnification andminimization of social cleavages by the broadcast and narrowcastnews media. International Journal of Public Opinion Research,8(4), 374-388.

National Institute on Alcohol Abuse and Alcoholism. (1991).Alcohol alert 13 (PH 297). Bethesda, MD: Author.

New California Media. (2005). Ethnic media in America: Thegiant hidden in plain sight. Available from the New CaliforniaMedia Web site, http://news.ncmonline.com/news/view_article.html?article_id=0443821787ac0210cbecebe8b1f576a3

O’Connor, M. J., & Whaley, S. E. (2003). Alcohol use in pregnantlow-income women. Journal of Studies of Alcohol, 64, 773-783.

Paisley, W. J. (2001). Public communications campaigns: TheAmerican experience. In R. E. Rice & C. K. Atkin (Eds.), Public com-munication campaigns (3rd ed., pp. 3-21). Thousand Oaks, CA: Sage.

Prelip, M., Glik, D. C., Myerson, A., Ang, A., & Eilers, K. (2006).Outcome evaluation of a community based narrowcasting cam-paign to prevent FAS. Manuscript submitted for publication.

Reitman, J. (1986, February). Narrowcasting opens up: Cable isexpanding its programming to win bigger numbers in the ratingsgame. Marketing and Media Decisions, 21, 70-75.

Rimal, R. N., & Adkins, A. D. (2003). Using computers to narrowcasthealth messages: The role of audience segmentation, targeting, and tailoring in health promotion. In T. L. Thompson, A. M. Dorsey,K. I. Miller, & R. Parrott (Eds.), Handbook Of health communication(pp. 497-513). Mahwah, NJ: Lawrence Erlbaum.

10 HEALTH PROMOTION PRACTICE / Month XXXX

Page 11: Fetal Alcohol Syndrome Prevention Using Community-Based

Rimal, R. N., & Real, K. (2003). Understanding the influence of per-ceived norms on behaviors. Communication Theory, 13(2), 184- 203.

Roberts, D. F., Christenson, P. G., Henriksen, L., & Bandy, E. (2002).Substance use in popular music videos. Washington, DC: Office ofNational Drug Control Policy.

Rubin, A. M. (2002). The uses-and-gratifications perspective of mediaeffects. In J. Bryant & D. Zillman (Eds.), Media effects: Advances intheory and research (2nd ed., pp. 525-548). Hillsdale, NJ: LawrenceErlbaum.

Smith-Shomade, B. E. (2004). Narrowcasting in the new worldinformation order: A space for the audience? Television & NewMedia, 5(1), 69-81.

Stewart, D. W., Pavlou, P., & Ward, S. (2002). Media influences on mar-keting communications. In J. Bryant & D. Zillman (Eds.), Media effects:Advances in theory and research (2nd ed., pp. 353-396). Hillsdale, NJ:Lawrence Erlbaum.

Stratton, K., Howe, C. V., & Battaglia, F. (Eds.). (1996). Fetal alco-hol syndrome: Diagnosis, epidemiology, prevention and treat-ment. Washington, DC: Institute of Medicine, Committee to StudyFetal Alcohol Syndrome.

Substance Abuse and Mental Health Services Administration.(2003). Results of the 2001 National Household Survey on DrugAbuse: Volume 1. Summary of findings. (SMA 03-3836, H-22).Washington, DC: Author.

Wallack, L. (1981). Mass media campaigns: The odds against find-ing behavior change. Health Education Quarterly, 8(3), 209-260.

Wallack, L., Grube, J. W., Madden, P. A., & Breed, W. (1990).Portrayals of alcohol on prime-time television. Journal of Studieson Alcohol, 51(5), 428-437.

Wechsler, H., Nelson, T. F., Lee, J. E., Seibring, M., Lewis, C., &Keeling, R. P. (2003). Perception and reality: A national eval-uation of social norms marketing interventions to reduce collegestudents’ heavy alcohol use. Journal of Studies on Alcohol, 64(4), 484.

Willford, J. A., Leech, S. L., & Day, N. L. (2004). Moderate prenatal alco-hol exposure and cognitive status of children at age 10. Alcoholism:Clinical and Experimental Research, 30(6), 1051-1059.

Wyllie, A., Zhang, J. F., & Casswell, S. (1998). Positive responsesto televised beer advertisements associated with drinking andproblems reported by 18 to 29 year olds. Addiction, 93, 749-760.

Glik et al. / FETAL ALCOHOL SYNDROME PREVENTION 11