More than 20 years ago, public health researchers began calling for the application of social marketing principles in the design and implementation of public health education interventions (Lefebvre and Flora 1988; Slater and Flora 1991). These principles include the following:
· The segmentation of a target audience into homogeneous subgroups.
· An examination of knowledge, beliefs, social norms, and behaviors pertaining to the outcome or behavior targeted for change.
· The identification of communication channels relevant to each audience segment.
· The development and targeting of messages and interventions relevant to the particular constellation of knowledge, beliefs, social norms, and behaviors in each audience segment.
· Piloting material or programs with each audience segment to confirm its applicability.
The goal of audience segmentation is "to identify population subgroups that are homogeneous with respect to certain variables associated with a given outcome or behavior" (Boslaugh et al. 2005). The underlying premise of audience segmentation is that product marketing is most effective when messages are meaningfully tailored to certain segments of the population (Slater 1996). In the context of health and social service interventions, audience segmentation research seeks to identify sectors of a target population who, given their unique circumstances, may need tailored outreach and/or service delivery.
Public health researchers posit that the kinds of variables and subgrouping methods used in audience segmentation research are applicable to, and should therefore be used in, designing and targeting public health education programs (Boslaugh et al. 2005; Slater and Flora 1991). They argue that reliance on demographic variables such as age, gender, and marital status for segmenting audiences is insufficient when seeking to describe or predict health behaviors, because meaningful variability with respect to individuals' psychosocial characteristics remains within demographically homogenous subgroups (Kreuter et al. 2003; Slater and Flora 1991). In fact, in their sample of 1,090 adults ages 18 to 65, Boslaugh and colleagues (2005) found that the five subgroups that maximally distinguished levels of physical activity were defined by intrinsic motivation to exercise and current health status. Demographic characteristics (age, race, gender, education, and income) played no role in distinguishing more and less physically active subgroups.
Social marketers consider the following "psychographics," or variables reflecting consumers' interests, activities, and opinions (Grier and Bryant 2005):
· Needs. Consumers who perceive a need for a product or service are more likely to seek out and purchase these products or services. Service providers target services depending on the nature or severity of individuals' needs.
· Wants. Consumers who want a particular product (whether or not they perceive a need for it) are more likely to seek out and purchase these products or services. Service providers use different marketing messages and recruitment strategies to make their service appealing to the target audience.
· Values and lifestyle. Consumers with certain values and lifestyles are likely to want, need, and/or benefit from certain products or services more than others. Service providers seek to accommodate these values and lifestyles-for example, by articulating how program services align with participants' goals, and by making their services available at a time and place of convenience to the target audience.
· Knowledge. Consumers who know what they want and need, and who are well-informed about what products or services can meet these wants and needs, are more likely to seek out and purchase the most appropriate products or services. Public health interventions provide information about healthy and unhealthy behaviors and the consequences of unhealthy behaviors.
· Current behavior. Consumers already engaging in certain behaviors (for example, jogging) may be more likely to purchase a complementary product or service (for example, join a gym), whereas consumers engaging in unhealthful behaviors (for example, smoking) may need to have these behaviors addressed (for example, through a smoking cessation program) before they can effectively participate in and/or benefit from the primary product/service being offered.
· Readiness to change and future intentions. Consumers who are ready and plan to make a purchase in the near future may be more likely to do so, compared to those who are not ready or who have no such plans. Service providers may need to tailor program messages and services based on the target audience's readiness and/or plans for change.
· Social norms. Consumers are more likely to purchase products or services that align with the norms of their family, peers, culture, and community. Service providers may need to take these norms into account when developing and targeting services.
In creating audience segments, public health researchers rely almost exclusively on interactive approaches. For example, in predicting who were more and less likely to receive a flu vaccine, Lemon and colleagues (2003) used classification and regression tree analysis and a handful of key discriminating predictor variables (past pneumonia vaccination, checkup in the past year, and race/ethnicity). Findings revealed that those least likely to have received a flu vaccine had never had a pneumonia vaccine and did not have a checkup in the prior year. The authors noted that this information could help public health officials better target vaccination campaigns.
Also adopting an interactive approach but using an expanded set of variables including psychosocial factors, BeLue et al. (2010) used latent class analysis to explore why some men had low take-up rates of colorectal cancer screenings. In a sample of male African American patients at a health center run by the U.S. Department of Veterans Affairs, the researchers assessed each participant's knowledge (of screening guidelines, prevalence of colon cancer, and typical treatments), perceived susceptibility to cancer, benefits of screening, barriers to screening, and self-efficacy/confidence in following screening procedures, and trust in his or her primary care provider. Four psychosocial risk classes emerged: (1) prepared; (2) unprepared; (3) high perceived barriers/low self-efficacy; and (4) low perceived benefits, barriers, and self-efficacy. Authors concluded that veterans who are nonadherent to colorectal cancer screening recommendations are not a homogenous group and may thus need different outreach and education strategies.
A more detailed example illustrating how the interactive approach of cluster analysis (termed "lifestyle analysis" by audience segmentation researchers) can be informative for targeting messages and services in health interventions can be found in Appendix D.