Preventive Medicine 67 (2014) 75–81

Contents lists available at ScienceDirect

Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed

Church-based social marketing to motivate older adults to take balance classes for fall prevention: Cluster randomized controlled trial☆ Carolyn G. DiGuiseppi a,⁎, Sallie R. Thoreson b, Lauren Clark c, Cynthia W. Goss a, Mark J. Marosits d, Dustin W. Currie a, Dennis C. Lezotte e a

Colorado Injury Control Research Center, Department of Epidemiology, Colorado School of Public Health, Campus Box B119, 13001 E 17th Place, Aurora, CO 80045, USA Colorado Department of Public Health and Environment, 4300 Cherry Creek Dr S, Denver, CO 80246, USA College of Nursing, University of Utah, 10 S 2000 E, Salt Lake City, UT 84112, USA d Worldways Social Marketing, 240 Thames St., Suite 200, Newport, RI 02840, USA e Department of Biostatistics and Bioinformatics, Colorado School of Public Health, Campus Box B119, 13001 E 17th Place, Aurora, CO 80045, USA b c

a r t i c l e

i n f o

Available online 12 July 2014 Keywords: Social marketing Exercise Postural balance Aged Accidental falls Accident prevention Injuries

a b s t r a c t Objective. Determine whether a church-based social marketing program increases older adults' participation in balance classes for fall prevention. Methods. In 2009–10, 51 churches (7101 total members aged ≥60) in Colorado, U.S.A. were randomized to receive no intervention or a social marketing program. The program highlighted benefits of class participation (staying independent, building relationships), reduced potential barriers (providing convenient, subsidized classes), and communicated marketing messages through church leaders, trained “messengers,” printed materials and church-based communication channels. Between-group differences in balance class enrollment and marketing message recall among congregants were compared using Wilcoxon Two-Sample Test and regression models. Results. Compared to 25 control churches, 26 churches receiving the social marketing program had a higher median proportion (9.8% vs. 0.3%; p b 0.001) and mean number (7.0 vs. 0.5; IRR = 11.2 [95%CI: 7.5, 16.8]) of older adult congregants who joined balance classes. Intervention church members were also more likely to recall information about preventing falls with balance classes (AOR = 6.2; 95% CI: 2.6, 14.8) and availability of classes locally (AOR = 7.7; 95% CI: 2.6, 22.9). Conclusions. Church-based social marketing effectively disseminated messages about preventing falls through balance classes and, by emphasizing benefits and reducing barriers and costs of participation, successfully motivated older adults to enroll in the classes. © 2014 Elsevier Inc. All rights reserved.

Introduction Falls are the leading cause of injury-related emergency visits, hospitalizations, and deaths in older adults (National Center for Injury Prevention and Control, 2014a). Each year about one-third of persons aged ≥ 65 years fall (Hausdorff et al., 2001; Hornbrook et al., 1994); 20–30% of falls cause serious injuries that reduce mobility and independence, increasing the risk of institutionalization and premature death (Alexander et al., 1992; Magaziner et al., 2000; National Center for Injury Prevention and Control, 2014b; Sterling et al., 2001).

☆ Clinical Trial Registration: Marketing Fall Prevention Classes to Older Adults in FaithBased Congregations; ClinicalTrials.gov Identifier: NCT00542360. Registered 9-October2007. ⁎ Corresponding author at: 13001 E. 17th Place, Campus Box B119, Aurora, CO, 80045, USA. Fax: +1 303 724 4489. E-mail address: [email protected] (C.G. DiGuiseppi).

http://dx.doi.org/10.1016/j.ypmed.2014.07.004 0091-7435/© 2014 Elsevier Inc. All rights reserved.

Balance and strength training reduces fall and fall injury risk in older adults (El-Khoury et al., 2013; Gillespie et al., 2012; Sherrington et al., 2011). However, few older adults participate (Hughes et al., 2005), due to fear of falling (Scheffer et al., 2008), fatalism about functional loss with age, or inexperience with organized exercise classes (Muse, 2005). In addition, few recreational facilities offer exercise programs for seniors due to perceived lack of interest (Hughes et al., 2005). Unless exercise programs are implemented and promoted in a manner appealing to older adults, participation is unlikely even if lack of availability and other barriers are addressed (Stevens, 2005; Stevens et al., 2010). New approaches are needed to motivate participation in balance and strength training. Social marketing, which promotes voluntary behavior change by increasing perceived advantages and reducing perceived and actual barriers (Kotler and Roberto, 1989; Maibach et al., 2002), has been shown to increase immunizations, cancer screening and exercise in older adults (Lipkus et al., 2003; McCaul et al., 2002; Reger et al., 2002; Van Harrison

76

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81

et al., 2003). After social marketing approaches were used to promote exercise classes for fall prevention, many community-dwelling older adults joined the classes (Li et al., 2008a, 2008b; York et al., 2011). A social marketing campaign to increase physical activity for fall prevention was followed by small increases in self-reported physical activity among older adults (John-Leader et al., 2008). However, the use of social marketing to promote balance classes has not been tested in controlled studies. Church-based programs have proven effective for increasing physical activity among middle-aged women (Peterson et al., 2005) and African American adults (Resnicow et al., 2005), and for increasing other health behaviors (DeHaven et al., 2004). Churches are potentially valuable settings for reaching older adults. In the U.S., 84% of adults aged ≥ 60 are affiliated with Christian denominations (Pew Forum on Religion and Public Life, 2008a); most older Christians attend church services at least weekly (Pew Forum on Religion and Public Life, 2008b). Given proven benefits of balance and strength training, churchbased social marketing holds promise for reducing falls and related injuries if it can be shown to motivate older adults' enrollment in classes. We hypothesized that older adults exposed to a targeted social marketing program at church would be more likely to join fall-prevention balance classes than older adults attending control churches. Methods Social marketing intervention The social marketing process incorporates key aspects of behavioral change theory (Janz et al., 2002) to understand why people behave as they do and what might be required to change their behavior (Maibach et al., 2002). A social marketing program was developed based on extensive formative research (Clark et al., 2013) to ensure a favorable response from the target audience (Aaker et al., 2000). It aimed to make balance classes more attractive to older adult church members by addressing product, price, place and promotion (Maibach et al., 2002). To address desired benefits (product), the program emphasized ‘gainframed’ messages (O'Keefe and Jensen, 2007) of staying independent and building social relationships. Monetary, emotional, psychological and time costs (price) were addressed by emphasizing a safe, comfortable atmosphere in marketing materials; demonstrating classes through videos; offering frequent classes at preferred schedules, coordinated through the marketing program; and, through provider contracts, subsidizing most of the course fee. A $20 fee was maintained to increase perceived value (Grier and Bryant, 2005) and long-term program sustainability. Place, i.e., where the product is offered to the target audience, was addressed by providing classes at local recreation facilities and churches. In contrast to health promotion programs, where place is primarily a setting for target audience engagement, place in social marketing is “an essential element of the marketing mix … i.e. where and when the target audience will perform the desired behavior (Griffiths et al., 2008).” The marketing program therefore set up classes at churches, ensuring both convenience and accessibility. Recreation facilities were chosen for their availability and were equally near intervention and control churches. Each class, wherever it was held, was marketed to multiple local intervention churches. Place was also addressed by positioning messages at intervention churches, and involving church leaders, staff, and ‘messengers’ (trained church members) in message distribution. Marketing messages were communicated through posters, brochures, flyers, coupons, church newsletters and bulletins, pulpit announcements, and person-to-person marketing (promotion), guided by a week-by-week marketing plan implemented during four weeks preceding each scheduled class (marketing ‘wave’). Classes at recreation department facilities were listed in their own course catalogs, flyers and websites, consistent with their standard procedures for disseminating information about class availability. The program marketed N'BalanceTM, which incorporates the key programmatic elements and lesson plans of FallProof!TM (Rose, 2010; Rose et al., 1999, 2001), adapted as a biweekly community-based class limited to 10 participants for individualized attention, delivered over 8 weeks by trained exercise instructors. N'BalanceTM addresses balance, gait, flexibility, and strength, with moderate to high progressive challenge to balance, as recommended (Gillespie et al., 2012; Sherrington et al., 2011). N'BalanceTM was selected because trained

instructors (and new-instructor training) were available locally and local recreation departments were willing to offer N'BalanceTM. The social marketing program was implemented by health department staff, assisted by church messengers, from January 2010 through September 2011. All N'BalanceTM classes, whether offered at church or recreation facilities, were available to the entire community, including people attending control or nonstudy churches or no church, and were subsidized for all participants. The recreation departments registered participants for most classes, whether offered in recreation or church facilities; study staff registered participants for the rest. Study design and methods Randomization Churches were the units of randomization, social marketing program implementation and evaluation. A clustered design was chosen because the marketing program was delivered to both churches and individual congregants. An intra-cluster correlation coefficient (ICC) of 0.04 was used to estimate sample size, conservatively inflating the ICC from another church-based study (Voorhees et al., 1996). Enrollment of 56 churches (estimated mean 84 older adult members/church) was planned, to provide 80% power, two-sided α = 0.05, to detect an increase in class participation from 0.9% (estimated baseline) to 3.5% of older adult church members. A researcher blinded to church enrollment randomly allocated churches to study groups in a 1:1 ratio using minimization (Schulz and Grimes, 2002) with Minim v1.5 software (Evans et al., 1990) to ensure similar distributions of denomination, congregation size, percentage of congregants aged ≥ 60, location, Spanish language services, and church leader's willingness to assist marketing. Data analysts were blinded to study group allocation. The Colorado Multiple IRB approved the trial and consent forms. Participants The two study counties comprised several small- to medium-sized cities (‘urban’ locations) (U.S. Census Bureau, 2009a) and the remaining towns and unincorporated areas (‘rural’ locations), totaling 37,532 residents aged ≥ 60 years, of whom 92.3% are white non-Hispanic (U.S. Census Bureau, 2009b). All study county churches were assessed for eligibility (i.e., ≥15 older adult members, for cost efficiency) by telephone. Study staff recruited church leaders by mail and telephone from May 2009 to July 2010. Participating church leaders gave written consent prior to randomization and received $55 gift cards. At enrollment (pre-randomization), each church leader was surveyed about church and personal demographic characteristics, including minimization data (see Randomization section), and their perceptions of the importance of falls in older adults, fall preventability, and potential benefits from balance classes (Janz et al., 2002), using a 5-point Likert scale (Table 1). Class participants aged ≥60 gave written consent to release attendance records and received $5 gift cards. Church congregants implied consent by returning completed postintervention mail surveys. Outcomes Because participation in balance and strength training classes has been proven to reduce falls and fall injuries (El-Khoury et al., 2013; Gillespie et al., 2012; Sherrington et al., 2011), measures of class participation were chosen as primary outcomes. Primary outcomes included the proportions and numbers of church congregants aged ≥60 enrolled in any N'BalanceTM class during the marketing implementation period. Secondary outcomes included class attendance among older adult class participants; church leaders' perceptions post-intervention about importance of injury from falls in older adults, fall preventability, and benefits from balance classes for reducing risk; and self-reported awareness and recall of marketing messages among older adult congregants. We also assessed program implementation. Data collection To assess class participation, everyone enrolled in any N'BalanceTM class held in either county during the study period was administered a 1-page anonymous form indicating age group, name of church (if any), and if this was their first N'BalanceTM class since the study period start date. Instructors recorded class attendance. Denominators, i.e., numbers of congregants aged ≥ 60 in each church, were obtained from church databases or estimated by church leaders prior to randomization. Secondary outcome data were collected from post-intervention surveys of church leaders and older adult congregants. Surveys were administered a

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81 Table 1 Baseline characteristics of intervention (social marketing program) and control churches and church leaders, western Colorado, U.S.A., 2009–2010. Intervention churches (n = 26) Church characteristics Total membership of congregation (median) Services per week (median) Types of communication resources available (median) Percentage of congregation aged ≥60 (mean %) Protestant denomination Regularly hold Spanish language services Estimated typical household income of congregationa Mostly low to middle Mostly middle Mostly middle to high Rural location Exercise classes currently sponsored by/held at church Church leader characteristics Male White non-Hispanic Religious or spiritual leader (e.g., minister, priest)b Age (mean years) Tenure as church leader (mean years) Willingness to assist marketing (median rating)c Willingness to hold classes at church (median rating) Perceived importance of falls as cause of older adult injury (median rating) Perceived preventability of falls in older adults (median rating) Perceived effectiveness of balance exercise classes for preventing older adult falls (median rating)

Control churches (n = 25)

187 2.5 7.0 44.8% 76.9% 7.7%

220 2.0 8.0 47.8% 88.0% 4.0%

50.0% 30.8% 7.7% 38.5% 3.8%

32.0% 64.0% 4.0% 40.0% 4.0%

80.8% 84.6% 92.3% 51.8 5.9 4.5 5.0 4.0

80.0% 100% 92.0% 54.4 8.7 4.0 5.0 5.0

4.0

4.0

4.0

4.0

a

n = 3 missing responses from intervention church leaders. Other responding church leaders included Bishop, Board President, Church Administrator, Director of Senior Adult Ministries, Elder, and Administrative and other Pastors. c Ratings: 1 = “not at all” to 5 = “extremely”. b

median 12 months after church enrollment (median 9 months after marketing program implementation in intervention churches). Church leader surveys asked the same questions on perceptions about falls and balance classes as were asked at baseline (see Participants section), and also assessed marketing message recall to measure awareness of program implementation among intervention church leaders. Post-intervention congregant surveys assessed recall of information about falls, fall prevention or balance classes (i.e., ‘external cues to action’) (Janz et al., 2002) heard or seen at church or from its members and from other sources (e.g., television, physician); how often they attended church for services or activities; and sociodemographics. The congregant survey also measured fear of falling (7-item Fall Efficacy Scale-International [Short FES-I], with ‘high concern’ defined as score ≥ 11 (Delbaere et al., 2010; Kempen et al., 2008)), fall history (self-report portion of the Elderly Fall Screening Test (Cwikel et al., 1998)), and health status and life satisfaction (from the Behavioral Risk Factor Surveillance System (Link, 2008)). Surveys were mailed to approximately 6% of each study church's members aged ≥ 60 years, randomly sampled from church membership lists. When age information was unavailable, all adult members were randomly sampled, using oversampling to achieve desired numbers of older adults. The planned survey process included pre-notification postcards, followed by questionnaires with $2 incentives and, 2–3 weeks later, second questionnaires (Edwards et al., 2009). This was fully implemented by 58% of churches and partially implemented (e.g., omitting the postcards) by the rest.

Statistical analysis Analyses were performed with SAS software (V9.3; SAS Institute Inc., Cary, NC) based on the original random allocation, regardless of marketing program implementation (intention-to-treat approach). Baseline characteristics of churches and church leaders were compared between groups using χ2 tests, Fisher's Exact Tests, t-tests, or Wilcoxon Two-Sample Tests, as appropriate.

77

Primary outcomes Proportions and mean numbers of older adult congregants from intervention and control churches who enrolled in N'BalanceTM were compared using, respectively, the Wilcoxon Two-Sample Test, and a zero-inflated, negative binomial regression analysis, adjusted for the estimated number of older congregants in each church. The latter analysis accounted for dispersion and the excess number of zero counts over the expected number based on the Poisson distribution. Baseline church and leader characteristics were examined as potential confounders, indicated by ≥ 10% estimate change or significance at α = 0.05. To assess the sensitivity of these results to where classes were held, the two primary outcomes were re-examined after excluding classes held in churches, using the same analytic methods. Secondary outcomes Proportions of classes attended by class participants from intervention versus control churches were compared with the Wilcoxon Two-Sample Test. Church leaders' perceptions post-intervention were compared between study groups using Wilcoxon Two-Sample Tests. Post-intervention message awareness and recall among older adult congregants were compared between study groups using logistic regression models with random effects for church. Respondent characteristics were examined as potential confounders as described above (see Primary outcomes section). Estimates presented at the congregant level are adjusted for clustering by church.

Results Fifty-one churches, comprising 7101 adult congregants aged ≥ 60 (mean 139 older adult congregants/church), were enrolled and randomly allocated to intervention (n = 26) or control (n = 25) group (Fig. 1). Table 1 shows baseline characteristics of enrolled churches and church leaders. There were no statistically significant differences between groups (p ≥ 0.05 for all between-group comparisons). During the 21-month implementation period, study staff partnered with intervention churches to market 34 N'BalanceTM classes, of which 9 classes were held in four recreation facilities and 25 classes in 19 churches. On average, four 4-week marketing waves were implemented in each intervention church. Informal person-to-person marketing was implemented in all intervention churches. Pulpit announcements, bulletin announcements and newsletter articles were each implemented in N75% of intervention churches, and brochures and posters in N90%. Primary outcomes Of the 277 N'BalanceTM participants, 269 (97%) completed anonymous forms. Of these, 195 (72.5%) were first-time participants aged ≥ 60 who attended study churches, including 183 (94%) from intervention churches and 12 (6%) from control churches. Most participants attended classes at churches (80.3% of intervention church participants and 91.7% of control church participants). Compared to control churches, higher proportions of intervention church congregants aged ≥ 60 enrolled in N'BalanceTM (Fig. 2; mean 9.8% vs. 0.3%; z = −4.47, p b 0.001). Intervention churches also had significantly greater numbers of older adult congregants who took classes (mean 7.0 vs. 0.5 per church; IRR = 11.2 [95%CI: 7.5, 16.8]). This estimate was unconfounded by differences between study groups. In sensitivity analysis limited to recreation facility classes, proportions of congregants aged ≥ 60 who enrolled remained significantly higher in intervention versus control churches (mean 1.5% vs. 0.2%; Z = −2.7, p b 0.007). Similarly, mean number of older adult congregants who enrolled in recreation facility classes remained significantly higher in intervention versus control churches (mean 1.8 vs. 0.1; IRR = 36.4 (95% CI: 1.5, 897.2)). Secondary outcomes Of the 195 first-time N'BalanceTM participants from study churches, 174 (89%) released their attendance records. Intervention church

78

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81

Fig. 1. Enrollment and retention of churches into trial of social marketing to promote balance classes, western Colorado, U.S.A., 2009–2010.

participants attended, on average, 3.7 more class sessions than control church participants, although the difference in attendance was not statistically significant (average 79.5% vs. 56.5% of 16 sessions attended; p = 0.20).

Fig. 2. Balance and strength class enrollment among congregants of intervention (social marketing program) versus control churches, western Colorado, U.S.A., 2010–2011.

In post-intervention surveys (response rate 98%), intervention church leaders recognized the two key marketing messages: preventing falls by taking balance classes (80%) and availability of balance classes in their community (76%). Compared to control church leaders, intervention church leaders perceived older adult falls as more preventable (median rating 4 vs. 3 [1 = not at all; 5 = extremely], p = 0.03), but had similar perceptions of the importance of falls for causing older adult injuries (5 vs. 5, p = 0.30) and the effectiveness of balance classes for preventing falls (4 vs. 4, p = 0.31). Older adult congregants completed 280 mail surveys (response rate ≈ 62%). Intervention church respondents were somewhat less likely to attend church at least weekly and more likely to be Hispanic but did not differ statistically from control church respondents (Table 2). Compared to control church respondents, intervention church respondents were substantially more likely to have received information about balance classes, falls or exercise to prevent falls at their church or from its members, and to recall the two key marketing messages (Table 3), after adjustment for frequency of church attendance. There was no evidence of confounding by other respondent characteristics. Receipt of information from non-church sources (not part of the marketing program) and recall of information about the seriousness of falls and fall injuries (not a marketing message) were similar between study groups.

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81

79

Table 2 Characteristics of survey respondents aged ≥60 from intervention (social marketing program) and control churches, western Colorado, U.S.A., 2010–2011.

Attends church for services, meetings, or other activities at least weekly High concern about falling (Short FES-I ≥11)a Fell at least once in past 12 months Very good or excellent health Satisfied or very satisfied with life Sex (female) Ethnicity (Hispanic) Race (White) Age (mean years) Education (mean years) a

Attended intervention church (n = 149)

Attended control church (n = 131)

p-Value

70.1% 35.2% 46.4% 56.1% 93.7% 61.5% 6.7% 96.3% 72.3 14.6

81.6% 37.7% 37.5% 52.9% 92.5% 66.9% 2.0% 95.8% 72.7 15.0

0.07 0.66 0.20 0.63 0.69 0.36 0.07 0.83 0.67 0.27

Short FES-I = 7-Item Falls Efficacy Scale-International.

Discussion Older adults were successfully motivated to join balance and strength classes for fall prevention by a social marketing program implemented through their churches. Compared to control churches, the average proportion of older adult congregants who enrolled in a class was nearly 10% higher in intervention churches and the mean number of participants was more than ten times higher. By increasing participation in balance and strength classes, church-based social marketing has the potential to substantially reduce risk of falls and fall injuries among older adults (El-Khoury et al., 2013; Gillespie et al., 2012; Sherrington et al., 2011). Although the absolute increase in number of older adults who participated in balance classes was small, this in part reflects the small sizes of congregations enrolled, with most having b100 older adult congregants, and the relatively small number of classes offered (with each limited to 10 participants). Many countries have high levels of attendance at religious services (McCleary and Barro, 2006), so even a small increase in participation among religious congregants has the potential to benefit a sizable number of older adults. In the United States, for example, where 54% of the population aged ≥ 65 attends services at least weekly (Pew Forum on Religion and Public Life, 2008b), an increase in regular church-goers attending balance classes of 10% translates to more than 3 million older adults. The high levels of message awareness and recall among intervention church members likely explain, at least partly, the marketing program's efficacy. The marketing program matched materials and messages to observable cultural, social and behavioral characteristics of the target population, e.g., using photos of older adults, incorporating messages about fellowship, and holding classes in familiar locations, ensuring its ‘fit’ within the culture of older adult churchgoers, which likely also contributed to its efficacy (Resnicow et al., 2001). ‘Place’ appears to play an important role in motivating participation, since enrollment by both intervention and control church members was much higher in church

than recreation facility classes. Nevertheless, participation was significantly higher among intervention compared to control church members in recreation facilities as well, indicating that class location was not the sole influence on participation. Classes were offered at subsidized rates for all participants, but only intervention church members were given information about discounted rates, which may have favorably influenced participation. Research on the contribution of subsidies to program effectiveness would be helpful, since program costs influence its wider implementation and sustainability. Churches were chosen for marketing program implementation primarily to access the target audience. Additional potential benefits of church-based marketing are the existing communication channels and activities and the influence of religious leaders. Intervention church members most often recalled receiving program messages through existing channels, both formal (church bulletins, newsletters, and pulpit announcements) and informal (conversations with other church members), confirming their value for effective communication. Programgenerated brochures and posters, implemented in N 90% of intervention churches, were less commonly cited as sources of program messages. Congregants often cited pulpit announcements as their source of balance class information, suggesting that religious leaders may have influenced target audience decision-making by communicating marketing messages. Themes of building on the existing social support shared among congregants and between pastors and congregants to motivate enrollment, as well as using “relationships and support from others in the class” to maintain participation, were used to create the study social marketing program (Clark et al., 2013). Churches' existing social networks may make them particularly effective venues for targeting physical activity, since community-based interventions that focus on building, strengthening, and maintaining social networks that support behavior change have been proven to increase physical activity (Task Force on Community Preventive Services, 2001). The social support provided

Table 3 Marketing message recall among survey respondents aged ≥60 from intervention (social marketing program) versus control churches, western Colorado, U.S.A., 2010–2011.

Heard or seen any information about balance classes, falls or exercise to prevent falls at church or from church members? What information … did you hear or see at church or from church members? Preventing falls by taking a balance exercise classc Availability of balance exercise classes in my communityc Seriousness of falls and fall injuries in older adults Heard or seen any information about balance classes, falls or exercise to prevent falls from any other sources?

Attend intervention church (n = 149)

Attend control church (n = 131)

Crude ORa (95% CI)

Adj. ORb (95% CI)

52.2%

15.0%

6.2 (2.7, 14.1)

7.7 (3.1, 18.7)

32.4%

8.1%

31.0%

7.2%

10.1%

9.8%

40.2%

46.0%

5.4 (2.4, 12.4) 5.8 (2.1, 15.5) 1.0 (0.4, 2.7) 0.8 (0.5, 1.3)

6.2 (2.6, 14.8) 7.7 (2.6, 22.9) 1.2 (0.4, 3.3) 0.8 (0.5, 1.4)

Bold font indicates statistically significant results. a Adjusted for clustering by church. b Adjusted for clustering by church and frequency of church attendance for services, meetings, or other activities. c Key message of the social marketing program.

80

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81

by church leaders and members may have contributed to both the higher class enrollment and attendance of intervention compared to control church members. This study's findings are consistent with several earlier studies of social marketing approaches to motivate older adults to exercise for fall prevention (John-Leader et al., 2008; Li et al., 2008a, 2008b; York et al., 2011). These programs, like ours, identified themes of independence, accessibility, safety, and fun in formative research that directed message development. Both the Tai Chi and “Stay Active and Independent for Life” (SAIL) marketing programs (Li et al, 2008b; York et al., 2011), like ours, achieved good class participation. However, no other study used a comparison group to rule out secular trends or unmeasured confounding. This cluster RCT therefore adds substantially to the existing body of evidence. Using an RCT design to examine marketing program effects reduced the likelihood that unmeasured differences between study groups accounted for the findings. Selection bias is also unlikely to explain group differences for the primary outcomes, given a nearly 100% response rate for the anonymous form used to collect primary outcome data. The response rate for the congregant mail survey, which assessed secondary outcomes related to message recall, was only 62%. However, the two respondent groups had similar sociodemographic characteristics, awareness of fall-related information from sources other than churches, and recall of information that was not targeted in the marketing messages. Further, there was no evidence of confounding of recall estimates by sociodemographic characteristics, fear of falling, or fall history. The study counties had no large cities, no churches with substantial minority congregants, and few non-Christian congregations, potentially affecting generalizability. Nevertheless, the intervention churches varied widely in denomination, size, rurality, and income level, suggesting that the marketing program is likely to be effective in a wide range of churches and communities. While regular church attendance, especially among older adults, is common in many middle and high income countries (McCleary and Barro, 2006; Pew Forum on Religion and Public Life, 2008b), church-based social marketing programs will not reach older adults that do not attend religious services. Whether the marketing program would be equally effective among non-adherent older adults in other settings (e.g., senior centers) is an important future research question. Conclusion Testing the social marketing program in other locales, populations, and faith-based congregations (e.g., synagogues, mosques), as well as with other types of balance classes (e.g., Tai Chi), would help determine its broad applicability for promoting exercise classes to prevent older adult falls. It would be useful to discover if social marketing approaches can increase participation in interventions to prevent other injury problems affecting older adults, for example, safe driving courses. Comparisons of church-based social marketing programs to social marketing in other venues (e.g., senior centers or social agencies), and to targeted motivational counseling by health care providers, should also be explored. The positive results of this RCT indicate that church-based social marketing that addresses product, price, place and promotion with more convenient, lower-cost classes and messages about staying independent and building social relationships, can successfully motivate older adults to enroll in balance and strength classes for fall prevention. These messages can be effectively disseminated through existing church communication channels. The involvement of church leaders and informal member-to-member contacts, rather than reliance on brochures and posters, appears important to marketing program success. Since many older adults throughout the world regularly attend religious

services, church-based social marketing programs have the potential to influence millions of older adults. Church leaders' receptivity to delivering these messages indicates that larger scale replication is feasible. With wider adoption of this approach, increased participation in fall prevention programs is likely, thereby reducing fall risk for many older adults. Further, capitalizing on existing church-based social networks to boost class enrollment, and building stronger church-based social relationships through class participation, may have important beneficial effects on other causes of mortality (Holt-Lunstad et al., 2010). Conflict of interest statement The authors declare that there are no conflicts of interest.

Acknowledgments Research assistants Faye Boss, Joyce Chynoweth, Sarah Morehouse, and Vedavani Tiruveedhula made substantive contributions to this study. This work was supported by Grant Number R49/CCR811509 to the Colorado Injury Control Research Center from the Centers for Disease Control and Prevention (CDC). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC. References Aaker, J., Brumbaugh, A., Grier, S., 2000. Nontarget markets and viewer distinctiveness: the impact of target marketing on advertising attitudes. J. Consum. Psychol. 9 (3), 127–140. Alexander, B., Rivara, F., Wolf, M., 1992. The cost and frequency of hospitalization for fallrelated injuries in older adults. Am. J. Public Health 82 (7), 1020–1023. Clark, L., Thoreson, S., Goss, C., Zimmer, L., Marosits, M., DiGuiseppi, C., 2013. Understanding fall meaning and context in marketing balance classes to older adults. J. Appl. Gerontol. 32 (1), 96–119. Cwikel, J.G., Fried, A.V., Biderman, A., Galinsky, D., 1998. Validation of a fall-risk screening test, the Elderly Fall Screening Test (EFST), for community-dwelling elderly. Disabil. Rehabil. 20 (5), 161–167. DeHaven, M., Hunter, I., Wilder, L., Walton, J., Berry, J., 2004. Health programs in faithbased organizations: are they effective? Am. J. Public Health 94 (6), 1030. Delbaere, K., Close, J., Mikolaizak, A., Sachdev, P., Brodaty, H., Lord, S., 2010. The Falls Efficacy Scale International (FES-I). A comprehensive longitudinal validation study. Age Ageing 39 (2), 210–216. Edwards, P., Roberts, I., Clarke, M., et al., 2009. Methods to increase response to postal and electronic questionnaires (Review). Cochrane Database Syst. Rev. 3, MR000008. El-Khoury, F., Cassou, B., Charles, M., Dargent-Molina, P., 2013. The effect of fall prevention exercise programmes on fall induced injuries in community dwelling older adults: systematic review and meta-analysis of randomised controlled trials. BMJ 347, f6234. Evans, S., Day, S., Royston, P., 1990. Minimisation programme for allocating patients to treatments in clinical trials(V. 1.5) Department of Clinical Epidemiology, London Hospital Medical College, London. Gillespie, L., Robertson, M., Gillespie, W., et al., 2012. Interventions for preventing falls in older people living in the community. Cochrane Database Syst. Rev. 9 (11), CD007146. Grier, S., Bryant, C.A., 2005. Social marketing in public health. Annu. Rev. Public Health 26, 319–329. Griffiths, J., Blair-Stevens, C., Thorpe, A., 2008. Social marketing for health and specialized health promotion: stronger together — weaker apart. A paper for debate(online) Rsph.org.uk (Available at: http://www.rsph.org.uk/en/Utilities/site-search/index. cfm?q=social+marketing, Accessed 15 Mar. 2014). Hausdorff, J., Rios, D., Edelberg, H., 2001. Gait variability and fall risk in community-living older adults: a 1-year prospective study. Arch. Phys. Med. Rehabil. 82 (8), 1050–1056. Holt-Lunstad, J., Smith, T., Layton, J., 2010. Social relationships and mortality risk: a metaanalytic review. PLoS Med. 7 (7), 1000316. Hornbrook, M., Stevens, V., Wingfield, D., Hollis, J., Greenlick, M., Ory, M., 1994. Preventing falls among community-dwelling older persons: results from a randomized trial. Gerontologist 34 (1), 16–23. Hughes, S., Williams, B., Molina, L., et al., 2005. Characteristics of physical activity programs for older adults: results of a multisite survey. Gerontologist 45 (5), 667–675. Janz, N., Champion, V., Strecher, V., 2002. The health belief model, In: Glanz, K., Rimer, B., Lewis, F. (Eds.), Health Behavior and Health Education: Theory, Research, and Practice, 3rd ed.Jossey-Bass, San Francisco, pp. 45–66. John-Leader, F., Van Beurden, E., Barnett, L., et al., 2008. Multimedia campaign on a shoestring: promoting ‘Stay Active-Stay Independent’ among seniors. Health Promot. J. Aust. 19 (1), 22–28. Kempen, G.I.J.M., Yardley, L., Van Haastregt, J.C.M., et al., 2008. The Short FES-I: a shortened version of the falls efficacy scale-international to assess fear of falling. Age Ageing 37, 45–50.

C.G. DiGuiseppi et al. / Preventive Medicine 67 (2014) 75–81 Kotler, P., Roberto, N., 1989. Social Marketing, 1st ed. Free Press, New York. Li, F., Harmer, P., Mack, K., et al., 2008a. Tai Chi: moving for better balance — development of a community-based falls prevention program. J. Phys. Act. Health 5 (3), 445–455. Li, F., Harmer, P., Glasgow, R., et al., 2008b. Translation of an effective Tai Chi intervention into a community-based falls-prevention program. Am. J. Public Health 98 (7), 1195–1198. Link, M., 2008. Behavioral Risk Factor Surveillance System (BRFSS), Encyclopedia of Survey Research Methods: A-M, 1st ed.. , vol. 1. Sage Publications, Thousand Oaks, CA, pp. 52–53. Lipkus, I., Green, L., Marcus, A., 2003. Manipulating perceptions of colorectal cancer threat: implications for screening intentions and behaviors. J. Health Commun. 8 (3), 213–228. Magaziner, J., Hawkes, W., Hebel, J., et al., 2000. Recovery from hip fracture in eight areas of function. J. Gerontol. A Biol. Sci. Med. Sci. 55 (9), 498–507. Maibach, E., Rothschild, M., Novelli, W., 2002. Social marketing, In: Glanz, K., Rimer, B., Lewis, F. (Eds.), Health Behavior and Health Education: Theory, Research, and Practice, 3rd ed.Jossey-Bass, San Francisco, pp. 437–461. McCaul, K., Johnson, R., Rothman, A., 2002. The effects of framing and action instructions on whether older adults obtain flu shots. Health Psychol. 21 (6), 624–628. McCleary, R., Barro, R., 2006. Religion and economy. J. Econ. Perspect. 20 (2), 49–72. Muse, T., 2005. Motivation and adherence to exercise for older adults. Top. Geriatr. Rehabil. 21 (2), 107–115. National Center for Injury Prevention and Control, 2014a. Web-based Injury Statistics Query and Reporting System (WISQARS)(online) Cdc.gov (Available at: http:// www.cdc.gov/NCIPC/wisqars/, Accessed 15 Mar. 2014). National Center for Injury Prevention and Control, 2014b. CDC – older adult falls – falls among older adults: an overview – home and recreational safety – injury center(online) Cdc.gov (Available at: http://www.cdc.gov/HomeandRecreationalSafety/Falls/adultfalls. html, Accessed 15 Mar. 2014). O'Keefe, D., Jensen, J., 2007. The relative persuasiveness of gain-framed loss-framed messages for encouraging disease prevention behaviors: a meta-analytic review. J. Health Commun. 12 (7), 623–644. Peterson, J., Yates, B., Atwood, J., Hertzog, M., 2005. Effects of a physical activity intervention for women. West. J. Nurs. Res. 27 (1), 93–110. Pew Forum on Religion and Public Life, 2008a. U.S. Religious Landscape Survey. Religious Beliefs and Practices: Diverse and Politically Relevant(online) Pew Research Center (Available at: religions.pewforum.org/pdf/report2-religious-landscape-study-full. pdf, Accessed 15 Mar. 2014). Pew Forum on Religion and Public Life, 2008b. U.S. Religious Landscape Survey. Religious Affiliations: Diverse and Dynamic(online) Pew Research Center (Available at: religions.pewforum.org/pdf/report2-religious-landscape-study-full.pdf, Accessed 15 Mar. 2014). Reger, B., Cooper, L., Booth-Butterfield, S., et al., 2002. Wheeling Walks: a community campaign using paid media to encourage walking among sedentary older adults. Prev. Med. 35 (3), 285–292. Resnicow, K., Braithwaite, R., Ahluwalia, J., DiIorio, C., 2001. Cultural sensitivity in public health, In: Braithwaite, R., Taylor, S. (Eds.), Health Issues in the Black Community, 2nd ed.Jossey-Bass, San Francisco, pp. 516–542.

81

Resnicow, K., Jackson, A., Blissett, D., et al., 2005. Results of the Healthy Body Healthy Spirit trial. Health Psychol. 24 (4), 339–348. Rose, D., 2010. FallProof! A Comprehensive Balance and Mobility Program, 2nd ed. Human Kinetics, Champaign, IL. Rose, D., Jones, C., Dickin, C., Lemon, N., Bories, T., 1999. The effect of a community-based balance and mobility training program on functional performance and balancerelated self-confidence in older adults with a history of falls(abstract) J. Aging Phys. Act. 793, 265. Rose, D., Jones, C., Lemon, N., 2001. Effectiveness of a fall risk reduction program for older adult women with arthritis. Presented at The First Joint Conference of the American Society on Aging and the National Council on Aging, New Orleans, LA. Scheffer, A., Schuurmans, M., Van Dijk, N., Van der Hooft, T., De Rooij, S., 2008. Fear of falling: measurement strategy, prevalence, risk factors and consequences among older persons. Age Ageing 37 (1), 19–24. Schulz, K., Grimes, D., 2002. Generation of allocation sequences in randomised trials: chance, not choice. Lancet 359 (9305), 515–519. Sherrington, C., Tiedemann, A., Fairhall, N., Close, J., Lord, S., 2011. Exercise to prevent falls in older adults: an updated meta-analysis and best practice recommendations. N. S. W. Public Health Bull. 22 (4), 78–83. Sterling, D., O'Connor, J., Bonadies, J., 2001. Geriatric falls: injury severity is high and disproportionate to mechanism. J. Trauma 50 (1), 116–119. Stevens, J., 2005. Falls among older adults—risk factors and prevention strategies. J. Saf. Res. 36 (4), 409–411. Stevens, J., Noonan, R., Rubenstein, L., 2010. Older adult fall prevention: perceptions, beliefs, and behaviors. Am. J. Lifestyle Med. 4 (1), 16–20. Task Force on Community Preventive Services, 2001. The community guide – summary – physical activity, behavioral: social support interventions in community settings(online) Thecommunityguide.org (Available at: http://www.thecommunityguide.org/pa/behavioral-social/community.html, Accessed 10 Jun. 2014). U.S. Census Bureau, 2009a. City & towns totals: vintage 2009 — U.S. Census Bureau(online) Census.gov (Available at: http://www.census.gov/popest/data/cities/totals/2009/index.html, Accessed 15 Mar. 2014). U.S. Census Bureau, 2009b. County characteristics: age, sex, race, and Hispanic origin — state datasets — U.S. Census Bureau(online) Census.gov (Available at: http://www. census.gov/popest/data/counties/asrh/2009/CC-EST2009-alldata.html, Accessed 15 Mar. 2014). Van Harrison, R., Janz, N., Wolfe, R., et al., 2003. Personalized targeted mailing increases mammography among long-term noncompliant Medicare beneficiaries: a randomized trial. Med. Care 41 (3), 375–385. Voorhees, C., Stillman, F., Swank, R., Heagerty, P., Levine, D., Becker, D., 1996. Heart, body, and soul: impact of church-based smoking cessation interventions on readiness to quit. Prev. Med. 25 (3), 277–285. York, S., Shumway-Cook, A., Silver, I., Morrison, A., 2011. A translational research evaluation of the Stay Active and Independent for Life (SAIL) community-based fall prevention exercise and education program. Health Promot. Pract. 12 (6), 832–839.