Effects of Culturally Targeted Message Framing on Colorectal Cancer Screening Among African Americans

Abstract

Objective:

This study examined how standard and culturally targeted versions of gain and loss-framed messaging affect African Americans’ colorectal cancer (CRC) screening receptivity and behavior, as well as their anticipation of experiencing racism in undertaking CRC screening.

Methods:

Screening-deficient African Americans (N=457) viewed an informational video about CRC risks, prevention, and screening and were randomized to receive a gain or loss-framed message about screening. Half of participants viewed an additional culturally targeted message about overcoming racial disparities in CRC by obtaining screening. Using the Theory of Planned Behavior, we measured general receptivity to CRC screening. We also measured arousal of anticipatory racism in response to messaging. Finally, we offered participants a no-cost fecal immunochemical testing kit (FIT Kit) and measured uptake and use.

Results:

Message framing interacted with culturally targeted messaging to affect CRC screening receptivity and behavior. Participants were no more receptive to CRC screening when standard loss-framing was used, but were more favorable if loss-framing was culturally targeted. Targeted loss-framing also reduced anticipatory racism, which partially mediated effects on screening receptivity. Finally, although participants least often accepted a FIT Kit with standard loss-framing, effects of messaging on FIT Kit uptake and use were not significant.

Conclusion:

This study adds to growing recognition of important cultural nuance in effective use of message framing. Current finding also suggest that targeted and framed messaging could synergistically impact the extent to which African Americans engage in CRC screening, although specific impacts on FIT Kit screening are less certain.

Worldwide, colorectal cancer (CRC) is the third most diagnosed cancer and fourth highest cause of cancer mortality (Mattiuzzi, Sanchis-Gomar, & Lippi, 2019). In the United States, CRC is especially burdensome for African Americans, who have higher CRC incidence and mortality than any racial group (ACS, 2020; Alcaraz et al., 2016). Although lifestyle and prevention are contributing factors (Carethers & Doubeni, 2020), this disparity also reflects that African Americans continue to be screened for CRC less often than other racial groups (Alcaraz et al., 2016; Burnett-Hartman et al., 2016). One route to reducing CRC disparity is thus to facilitate better uptake of available CRC screening among African Americans. In turn, and given obstacles that African Americans can face in obtaining CRC screening (May, Almario, Ponce, & Spiegel, 2015), communicating effectively to encourage CRC screening is crucial.

Extending recent research (Lucas, Hayman, Blessman, Asabigi, & Novak, 2016; Lucas, Manning, Hayman, & Blessman, 2018), the current study examined how gain and loss-framed messaging compel African Americans to participate in recommended CRC screening (Rothman & Salovey, 1997), and whether culturally targeted messaging can be added to message framing to further enhance receptivity to CRC screening (Kreuter & McClure, 2004). The current research also assessed whether loss-framed messaging arouses racism-related angst among African Americans, and the subsequent potential of thinking about racism to reduce CRC screening receptivity (Lucas, Hayman, et al., 2016; Lucas, Lumley, et al., 2016). Finally, the current research examined whether and how message framing and culturally targeted messaging affect willingness to engage in at-home fecal immunochemical (i.e., FIT Kit) testing as an option for obtaining recommended CRC screening.

Brief Overview of Message framing

Attempts to improve uptake of cancer screening often rely on persuasive communication, and gain/loss message framing strategies derived from prospect theory have been used extensively. Stemming from both psychological science and behavioral economics (for recent review, Thaler & Ganser, 2015), prospect theory underscores that individuals often make subjective as opposed to rational choices between risky alternatives (Tversky & Kahneman, 1979). In turn, risk tolerance can be influenced by selectively presenting risk information through gain and loss framings (Kahneman, 2003). Gain-frames emphasize benefits to be had from taking action, motivate risk-aversion, and better persuade individuals to accept small risks to realize probable gain, whereas loss-frames highlight costs of inaction, motivate risk-taking, and compel individuals to assume greater risk to avert potential loss. Rothman and Salovey (1997) developed health communication strategies from insights provided by prospect theory. Prevention health behaviors like exercise typically convey small and certain benefits. As such, individuals may be better persuaded to perform them when a gain-framed presentation is used. Alternatively, and since they can reveal illness, detection behaviors like cancer screening can be psychologically risky. In turn, individuals may be better encouraged to use screening tools when the possible consequences of non-use are emphasized. Although theoretically compelling, evidence supporting selective use of loss-framing to encourage illness detection has been mixed. On one hand, meta analytic review highlights that loss-framed messaging is indeed associated with a discernible overall increase in detection behavior, including cancer screening (O’Keefe & Jensen, 2009). Yet, other reviews did not find compelling support for use of loss-framed messaging to promote screening (Gallagher & Updegraff, 2012; Van ‘t Riet et al., 2014).

Cultural Differences and Culturally Targeted Message Framing

Overlaying generally mixed support, evidence that loss-framing increases screening behavior among racial minorities has lagged, including among African Americans (for reivew, Schneider, 2006). Moreover, relatively few studies have focused on message framing and CRC screening (Ferrer, Klein, Zajac, Land, & Ling, 2012; Lipkus, Johnson, Amarasekara, Pan, & Updegraff, 2019; Lucas, Hayman, et al., 2016; Lucas, Manning et al., 2018). These gaps require attention not only given current CRC disparities, but also considering continued wide use of message framing strategies. Of current interest, recent preliminary research suggests that loss-framed messaging may adversely influence CRC screening among African Americans (Lucas, Hayman, et al., 2016). Consistent with much prior research, White Americans in this study reported greater receptivity to obtaining CRC screening when exposed to a loss-framed message that emphasized consequences of not being screened. However, African Americans were more receptive when exposed to a gain-framed message that emphasized benefits of obtaining CRC screening. The adverse effect of loss-framed messaging on African Americans was mediated by arousal of perceived racism (see also, Nabi et al., 2019; Nicholson et al., 2008), suggesting that loss-framed messaging could be interpreted as a negative social cue among racial minorities that produces thoughts about racial discrimination (for review, Major, Quinton & McCoy, 2002). Thus, Lucas and colleagues (2016) demonstrate not only the potential for a profound cultural difference in predicted message framing effects among African Americans, but also that loss-framed messaging can trigger racism-related emotion that reduces message persuasiveness.

In tandem to revealing a crucial cultural difference, Lucas and colleagues (2016) also showed that effects of message framing can be altered by parallel use of culturally targeted messaging. Whereas a standard loss-framed message increased perceived racism and reduced receptivity to CRC screening among African Americans, including a culturally targeted message – in which the potential to overcome racial disparities in CRC through ensuring that one is personally screened (c.f. Neighbors, Braithwaite, & Thompson, 1995; Neighbors, Hudson, & Bullard, 2012) – mitigated these effects. These additional findings align with a vast literature supporting the use of culturally targeted messaging (e.g., Kreuter et al., 2003; Sherman, Uskul, & Updegraff, 2011; though see Langford, Larkin, Resnicow, Zikmund-Fisher & Fagerlin, 2017), including to encourage health behavior among African-Americans (Kreuter & Haughton, 2006). Such findings also suggest that culturally aligned versions of more general message framing strategies may be vital to effective use with racial minorities.

Messaging Framing and Stool-Based Colorectal Cancer Screening

Relative to many cancer screenings, CRC screening is unique in that individuals may select from among multiple recommended options, if screening options are presented to them (McQueen et al., 2009). Currently, men and women between the ages of 45 and 75 who are at average risk for CRC can choose from among screening options that include 1) yearly high sensitivity stool-based testing; 2) flexible sigmoidoscopy every 5 years; 3) colonoscopy every 10 years (Wolf et al., 2018). Colonoscopy is considered the gold standard and appeals to many because it can detect and remove precancerous polyps, in addition to detecting cancer. However, uptake of colonoscopy may be hindered by factors such as time away from work and risk of complication (Vanderpool, Nichols, Hoffler, & Swanberg, 2017), and also by psychological factors including potential embarrassment and discomfort (Beydoun & Beydoun, 2008). For these reasons, many individuals opt for annual stool-based testing, which can broaden overall participation in CRC screening (Quintero et al., 2012). Although providing no ability to detect or remove polyps, stool-based testing appeals to many because this form of CRC screening may be conducted at home, and one may avoid ever needing a colonoscopy if screening is maintained and results are always negative. Widely available options for stool-based testing now include annual fecal immunochemical (i.e., FIT Kit) testing, which achieves superior behavioral compliance and has a higher positive predictive value for CRC relative to traditional occult blood tests (Smith, Young, Cole, & Bampton, 2006). Although FIT Kit screening is generally accessible, many CRC screening-eligible individuals remain unaware of this alternative. Of current interest, whether message framing and culturally targeted communication strategies impact FIT Kit testing uptake among racial minorities also remains unknown. Better understanding links between health communication and FIT Kit screening is critically important in that CRC screening outreach programs, including those that target African American communities, often only provide FIT Kit screening as an initial option (McDonald et al., 2019; Potter et al., 2011).

The Present Research

In the present research, we examine how standard and culturally targeted versions of gain and loss-framed messaging affect African Americans’ colorectal cancer (CRC) screening receptivity and behavior. We also examine how these health messaging approaches affect the extent to which African Americans expect to experience racism if they were to obtain CRC screening (anticipatory racism). Finally, we evaluate effects of messaging on participants’ willingness to accept and use a no-cost take-home FIT Kit screening. Based on preliminary research (Lucas, Hayman, et al., 2016), we expected that African Americans would be most receptive to CRC screening when loss-framing was used, but only if this message was culturally targeted. In addition, we expected that standard loss-framing would arouse anticipatory racism, but that including culturally targeted messaging would mitigate this effect. Finally, we expected that effects on anticipatory racism would mediate links between targeted loss-framed messaging and CRC screening receptivity and performance.

Method

Participants

Following institutional review board (IRB) approval by Wayne State University, participants were recruited through face-to-face outreach in the metropolitan Detroit (Michigan) community, and through advertisement and online recruitment. Participants were invited to enroll if they were African American, between the ages of 50 and 75 years and had never been screened for CRC, or were currently off schedule with recommended screening (e.g., more than ten years since last colonoscopy). Although CRC screening guidelines are transitioning to recommend regular screening beginning at forty-five for average risk individuals (Wolf et al., 2018), fifty was selected as the lower age limit based on CRC guidelines in place at the time (Bibbins-Domingo et al., 2016). All participants were enrolled via an online website, where a brief prescreen questionnaire confirmed eligibility. Five-hundred and thirty-nine eligible participants completed an IRB-approved electronic consent form online and were initially recruited. Using two quality control procedures, eighty-one of these participants were eliminated due to inattentiveness, which was identified using timestamps (N = 59) and a subsequently described knowledge retention assessment (N = 22). For timestamps, and based on a priori pilot testing, we excluded participants who took less than 45 minutes or greater than 3 hours to complete the survey. Quality control procedures resulted in a final effective sample size of four hundred and fifty-seven African Americans. Table 1 presents sample sociodemographic characteristics. All participants received a modest financial remuneration in exchange for participating in an online session that lasted approximately ninety minutes (M = 91.04, SD = 29.65).

Table 1.

Sample Characteristics (N=457).

Gender
Male	117 (25.6)
Female	340 (74.4)
Age
50–60	313 (67.4)
61–70	131 (29.2)
71 & above	13 (3.4)
Income
Less than $25,000	190 (41.6)
$25,000-$49,999	139 (30.4)
$50,000-$99,999	99 (21.7)
$100,000 and above	29 (6.3)
Education
High School/GED	126 (27.6)
Some College or Trade School	213 (46.6)
College Graduate	75 (16.4)
Professional/Advanced Degree	43 (9.4)
Insurance
Medicare/Medicaid	228 (49.9)
Private	145 (31.7)
Combination of Govt/Private	41 (9.0)
No Insurance	43 (9.4)

Design and Procedure

This study was conducted as a 2 (gain-framed versus loss-framed) × 2 (standard versus culturally targeted) fully-crossed between-participants experimental design. Experimental message manipulations were embedded in an online video module and survey that was presented using Qualtrics software. Animated video segments were professionally created using GoAnimate (Vyond) software, and were narrated using an animated and racially ambiguous male character.

Following prescreen, the video module and survey were completed as a single session that was administered in four phases. In the first phase, all eligible participants began by consenting to participate and completing an initial set of sociodemographic and individual differences measures. In the second phase, all participants viewed an immersive online informational video module about CRC that was created for this study. This module was designed both to educate participants about CRC etiology, risk factors and prevention, and to describe several screening options and recommendations. This included an even-handed review of colonoscopy, sigmoidoscopy, and FIT Kit as generally available recommended options for engaging in CRC screening, in which tradeoffs provided by each screening modality were reviewed. CRC informational content was developed using publicly available resources provided by the Centers for Disease Control (https://www.cdc.gov/cancer/colorectal/basic_info/). To promote attentive viewing, the informational module was broken into eight video segments that ranged in length from approximately two to four minutes. Following each segment, participants were required to answer a series of low difficulty true-false knowledge retention questions that reviewed information presented moments prior in each video segment. Participants were required to correctly answer 16 out of 22 true-false items (> 70%) for their survey response to be considered attentive (M = 21.10, SD = 1.71).

In the third phase, participants completed an identical set of subsequently described outcome measures to gauge receptivity to CRC screening. In the fourth and final phase, all participants viewed an additional four-minute video segment that described an offer to receive a FIT Kit through the mail. The video offer conveyed that a FIT Kit would be made available at no-cost, could be completed at home and returned through the mail, and that results would be directly communicated to the participant by the research team, which included individuals with a medical background who were available to advise on next steps in the event of a positive screen.

Gain/Loss and Culturally targeted Messaging Manipulations.

The message framing and culturally targeted experimental manipulations were embedded at the end of the second phase of the video survey, just prior to collection of outcome measures. The essential content of both message manipulations is presented in Table 2. All participants were randomly assigned to receive an approximately one-minute video about the importance of CRC screening that was presented either as a gain-framed or loss-framed message. Message frames were adapted from prior research on CRC screening and have been shown to effectively induce gain and loss-framings in both African American and non-African American viewers (Lucas et al., 2016). As a completion check, and to further reinforce messaging, all participants were required to check three items indicating that they had viewed and understood the assigned gain/loss-framed health message (Table 2).

Table 2.

Gain/Loss Framing and Cultural-Targeted Video Message Manipulations.

Gain-Framed Message	Loss-Framed Message	Culturally-Targeted Message
Timely action can save your life!	Delaying action can cost you your life!	Colorectal Cancer - Overcoming Adversity
Colorectal cancer screening effectively promotes a life free from cancer. By participating in recommended colorectal screening, you could remain free from cancer and easily add years to your life. Being screened also may make you feel relaxed and safe about gaining a future that is free of colorectal cancer, which is also a benefit worth having.	Colorectal cancer screening effectively reduces loss of life from cancer. By not participating in recommended colorectal screening, you could neglect a treatable cancer, and lose years off your life. Not being screened also may make you feel anxious and unsafe about losing a future free from colorectal cancer, which is also a cost worth avoiding.	We have invited you here today because the benefits of screening (costs of not being screened) may be highly important to you as an African American. Colorectal cancer rates in America are higher among African Americans than among any other racial group. This difference is due in part to controllable personal choices about colorectal cancer screening. Some research suggests that African Americans do not get screened for colorectal cancer as soon or as often as members of other groups. Thus, the injustice of colorectal cancer could be (will not be) overcome if (unless) more African Americans take personal control and obtain screening
Colorectal cancer screening may add years to my life.	Not performing colorectal cancer screening could cost me years off my life.	Colorectal cancer screening may be particularly important to me as an African American.
Colorectal cancer screening may help me feel relaxed and safe about my health.	Not obtaining colorectal cancer screening make me feel anxious and unsafe about my health.	The injustice of colorectal cancer can be overcome if more African Americans take control and obtain screening.
Being screened for colorectal cancer provides benefits that are worth having.	Not being screened for colorectal cancer has many drawbacks that are worth avoiding.

Note. Gain/loss-frames were either presented as stand-alone messages or paired with culturally targeted personal prevention message.

Half of participants were randomly assigned to receive an additional culturally targeted personal prevention message that emphasized overcoming adversity through personal will – a culturally enshrined coping approach among African Americans (Neighbors et al., 1995). The targeted message also lasted approximately one minute and contained three key elements. To begin, participants viewed a title slide labeled “Colorectal Cancer: Overcoming Adversity.” This title was accompanied by an animated image of a balance scale that was labeled “African Americans” on one side and “Other ethnicities” on the other. Initially shown in imbalance, the scale slowly adjusted to show “African Americans” and “Other Ethnicities” in equilibrium over an approximately five second duration. Inclusion of this visual was based on prior research that has shown balance scales and equilibrium are an effective visual means to induce thinking about justice and injustice (Lucas, Strelan, et al., 2018). Second, participants were broadly informed of current CRC disparities among African Americans in the United States, and that disparities were attributable in part to personal choices about CRC screening. Third, African Americans were told that the current injustice of CRC could be overcome if more African Americans took personal control and obtained screening. As with message framing, participants were required to check two additional items indicating that they had viewed and understood the culturally targeted health message (Table 2).

FIT Kit characteristics and logistics.

Participants who accepted the offer to receive a no-cost FIT Kit and provided a mailing address were sent the Insure^™ FIT Kit (Clinical Genomics), which was supplied by Quest Diagnostics (Chicago, IL). In brief, the Insure^™ FIT is a brush-based fecal immunochemical screening kit that requires collection of two samples from two separate bowel movements. After labeling and brushing a sample collection card, the kit is returned in a self-addressed prepaid envelop to a designated Quest laboratory for processing. Samples collected via this FIT Kit are viable for up to 14 days, ensuring adequate time for mailing and processing. The Insure ^™ FIT Kit utilized presently has demonstrated 87% sensitivity and 98% specificity for detecting colorectal cancer (Cole et al., 2003). Completed FIT Kits were sent by mail from participants to Quest Diagnostics laboratories for processing. Screening results were relayed to the research team by a Quest Diagnostics maintained and privacy protected software program, who in turn relayed results to participants.

Measures

Following message manipulations, we collected Likert measures of general receptivity to CRC screening and anticipatory racism. Table 3 presents means and standard deviations, as well as bivariate correlations and internal consistency coefficients. We also collected subsequently described behavioral measures of FIT Kit uptake and use.

Table 3.

Means, Standard Deviations, Internal Consistency Coefficients, and Bivariate Associations (N=457).

	Mean	SD	1.	2.	3.	4.	5.
1.Attitudes	6.27	0.95	.77
2. Norms	5.33	1.07	.57***	.90
3. Perceived Control	6.31	0.99	.46***	.33***	.90
4. Intentions	6.07	1.34	.65***	.55***	.47***	.95
5. Anticipatory Racism	2.05	1.42	−.24***	−.06	−.21***	−.18***	.86

Notes. Cronbach’s alpha for multi-item measures reported on diagonal.

^*p < .05,

^***p < .001.

CRC Screening Receptivity.

Overall receptivity to CRC screening was measured using a Theory of Planned Behavior (TPB) framework that assessed CRC screening attitudes, normative beliefs, perceived behavioral control, and intentions to be screened. Measures of TPB variables were adapted from recently published applications of TPB constructs to CRC screening (Lucas et al, 2016). These items were constructed following recommended procedures to ensure construct validity and adequate behavioral specificity (Fishbein & Ajzen, 2011). All items used Likert-type scales that ranged from 1 (Strongly Agree) to 7 (Strongly Disagree). Attitudes were measured using four items (e.g., “obtaining CRC screening would be good for me”), normative beliefs were measured using five items (e.g., “most people who are important to me think I should obtain CRC screening”), and perceived behavioral control was measured using three items (e.g. “I am confident that I can obtain CRC screening if I wanted to”). Intentions to be screened were also measured with three items (e.g., “I expect to obtain CRC screening”). Four separate TPB measures were calculated by averaging appropriate subscale items.

Anticipatory Racism.

All participants read instructions that asked them to consider whether they believed racism would impact the benefits and burdens they could experience from obtaining CRC screening. Participants then responded to three questions that asked, “In some way, my obtaining CRC screening would be impacted by racism,” “Racism would undermine the value of CRC screening for me,” and “Racism would negatively impact the accuracy of my CRC screening.” Responses were collected using a seven-point Likert-type scale that ranged from 1 (Strongly Agree) to 7 (Strongly Disagree). A composite was computed as the three-item response average (see Supplemental Materials for available validity information).

FIT Kit Behavior.

Two aspects of FIT Kit screening behavior were collected. First, we assessed whether participants opted to receive the no-cost FIT Kit offered to them (yes-no). Second, we assessed whether participants who opted to receive FIT Kits completed and returned them (yes-no). All participants who opted to receive a FIT Kit received written instructions for completing that were designed by the manufacturer, which were supplemented by the research team with a link to an instructional FIT Kit completion video. Participants who opted to receive but did not return a FIT Kit after 14 days were mailed a reminder card that matched the assigned messaging condition.

Overview of Statistical Analysis

Effects of messaging on TPB outcomes and anticipatory racism were evaluated using a series of 2 (framing: gain versus loss) × 2 (targeting: standard versus targeted) analyses of variance (ANOVAs). For TPB outcomes, we performed a mixed ANOVA with LSD comparison, in which TPB outcomes were specified as a four-level within-participant factor (for univariate consideration of TPB measures, see Supplemental Material). We also conducted 2 × 2 univariate ANOVA with LSD comparison on anticipatory racism, and we then performed a mediation analysis to examine whether anticipatory racism linked message framing to TPB-indicated CRC screening receptivity. Preliminary analyses revealed that income was significantly correlated with TPB outcomes (r’s = .10 – .17, p’s < .05), anticipatory racism (r = −.13, p = .005). Thus, we covaried for income in both ANOVAs, although results were consistent with and without including (see Supplemental Material). Effects of messaging on FIT Kit behavior were assessed by computing yes-no percentages across the four messaging conditions, which were then compared to the overall mean percentage of each outcome using a series of one-sample test of proportions (Fleiss, Levin & Paik, 2003).

Results

CRC Screening Receptivity and Anticipatory Racism

The 2 (framing) × 2 (culturally targeted) mixed ANOVA revealed no significant within-participants main effects or interactions involving TPB outcomes (p’s ≥ .497). At the between-participants level, the main effect of message framing was significant (F(1, 452) = 4.39, p = .037, η²_partial = .010). Participants were more receptive to CRC screening across TPB measures in the loss-framed condition than in the gain-fram0065d condition (M_Loss = 6.12, SD = 0.838 versus M_gain = 5.96, SD = 0.900, d = .184). There was also a notable main effect of culturally targeted messaging (F(1, 452) = 3.29, p = .070, η²_partial = .007). Participants were more receptive to CRC screening in the culturally targeted condition than in the untargeted condition (M_Targeted = 6.11, SD = 0.818 versus M_Untargeted = 5.97, SD = 0.923, d = .161). Of greater interest, main effects were qualified by the hypothesized framing × targeted interaction (F(1, 452) = 5.07, p = .025, η²_partial = .011). As seen in Figure 1 (see also Supplemental Material Table S1), there were no differences between standard and targeted gain-framed messages (p = .756; d = .011). However, targeted loss-framing produced greater CRC receptivity than standard loss-framing (p = .004; d = .374). Planned contrasts comparing each of standard and targeted loss-framing to all other conditions revealed that standard loss-framing did not significantly differ across TPB outcomes (t (453) = −1.039, p = .299, d = .080), whereas targeted loss-framing was significantly higher (t (453) = 3.394, p < .001, d = .319).

Figure 1.

Standard and culturally targeted message framing predicting CRC screening planned behavior and anticipatory racism.

The hypothesized framing × culturally targeted interaction was also significant for anticipatory racism (F(1, 456) = 5.32, p = .022, η²_partial = .012). As seen in Figure 1, there were no differences between standard and targeted gain-framed messages (p = .330; d = .092). However, targeted loss-framing produced lower anticipatory racism than standard loss-framing (p = .023; d = .299). Planned contrasts comparing standard and targeted loss-framing to all other conditions revealed that standard loss-framing increased anticipatory racism (t (453) = 1.83, p = .035, d = .172), whereas targeted loss-framing reduced anticipatory racism (t (453) = −1.82, p = .036, d = .171).

Mediation Analysis

A moderated-mediation analysis was conducted to examine whether anticipatory racism differentially linked culturally targeted messaging to CRC screening receptivity for gain versus loss-framed messaging. Following results of the mixed ANOVA, which indicated no significant TPB outcome main effects or interactions, mediation analysis was conducted on the overall (i.e. average) TPB construct. Furthermore, we positioned anticipatory racism as the mediating variable based on a priori experimental manipulation of messaging (Spencer & Zanna & Fong, 2005) and on its predicted mechanistic functions (Lucas et al., 2016). As seen in Figure 2, Baron and Kenny (1986) multiple regressions suggested a partially-mediated effect of culturally targeted messaging for loss-framing only. Moderated-mediation was corroborated using SPSS PROCESS macro 3.4 (Hayes, 2017). We selected PROCESS model number 8 and positioned culturally targeted messaging as the focal predictor and message framing as the moderator. The conditional indirect effect of culturally targeted messaging on CRC screening receptivity was significant for loss-framed messaging (B_{indirect effect} = 10.0248, SE_{indirect effect} = .013; 95% CI Upper = .053; Lower = .003), but not for gain-framed messaging (B_{indirect effect} = −0.010, SE_{indirect effect} = .012; 95% CI Upper = .011; Lower = −.037). Additionally, the difference between conditional indirect effects on CRC screening intentions was significant (Index = −0.035, SE = .019; 95% CI Upper = −.004; Lower = −.077), further supporting that loss-framed messaging moderated the pathway linking culturally targeted messaging to higher CRC receptivity through decreasing anticipatory racism. Finally, the conditional direct effect of culturally targeted messaging remained significant for loss-framed messaging, corroborating that results were best characterized as partial rather than full mediation (B_{direct effect} = 0.134, SE_{direct effect} = .056, p = .018; 95% CI Upper = .245; Lower = .024).

Figure 2.

Moderated-mediation: Mediated effects of culturally targeted messaging on CRC screening planned behavior for gain and loss-framed messaging. *p<.05, **p<.01, ***p<.001.

FIT Kit Screening Behavior

Overall, 292 participants opted to receive the FIT Kit (M = 63.9%; see also Supplemental Material Figure S1-left). Compared to the overall mean, participants opted to receive the FIT Kit no more often in the standard gain (67.3%; z = 0.74, p = .458) or targeted loss conditions (64.6%; z = 0.15, p = .878). Likewise, participants opted to receive the FIT Kit no less often in the standard loss (60.0%; z = −0.91, p = .364) or targeted gain conditions (63.6%; z = −0.07, p = .945). Across the four messaging conditions, FIT Kits were returned by 56 participants (M = 19.18%; see also Supplemental Material Figure S1-right). Among returned kits, there were 43 negative results, 7 positive results, and 6 incomplete results due to expired samples or improper collection. Of specific interest, participants returned completed FIT Kits no more often in the standard gain (20.27%; z = 0.29, p = .772) and targeted loss conditions (23.61%; z = 1.19, p = .236). Likewise, participants returned completed FIT Kits no less often in the targeted gain (14.29%; z = −1.37, p = .172) or standard loss conditions (18.84%; z = −0.93, p = .926).

Discussion

This study corroborates and extends recent preliminary research showing that message framing can be overlaid with culturally targeted messaging to better promote CRC screening among African Americans (Lucas, Hayman, et al., 2016). Contrary to predictions of prospect theory, standard loss-framed messaging did not enhance African Americans’ receptivity to CRC screening, and this was partly attributable to loss-framed messaging increasing anticipatory racism among African Americans when considering CRC screening. Crucially, however, combining loss-framing with a culturally targeted message reduced anticipatory racism and augmented CRC screening receptivity. Additionally, when presented with an opportunity to actively engage in CRC screening via a no-cost FIT Kit, participants were least receptive when standard loss-framed messaging was used, but were more receptive if standard gain-framed or culturally targeted loss-framed messaging was used, although effects of messaging on FIT Kit uptake and use were not statistically significant.

Aligned with cultural explorations of message framing effects, and with momentum towards better attending to moderating characteristics and contexts in message framing theory and research (Rothman, Desmarais & Lenne, 2020), the present findings support that effective use of message framing is attainable, but may be nuanced for African Americans (Lucas, Hayman, et al., 2016). Specifically, our findings illuminate that overlaying loss-framed messaging with a culturally targeted message may be vital to achieving theoretically predicted message framing effects. Although more corroboration and clinical trial evaluation are needed, health communication practitioners working with African American individuals and communities might one day deploy multiple messaging strategies, including culturally targeted messaging, to marshal more effective uses of generally proffered communication strategies, such as message framing. However, such approaches will require precision in both developing and deploying appropriate messaging, and thus should be undertaken cautiously. Notably, whereas culturally targeting loss-framed messaging was generally compelling, targeting gain-framed messaging was somewhat counter-persuasive. Thus, imprecise formulations of multiple messaging strategies could be counterproductive and result in unintended consequences.

Our findings also corroborate that racism-related thoughts play a central role in receptivity to CRC screening among African Americans, and that racism-related angst is differentially activated by gain and loss-framed messaging (Lucas, Lumley, et al., 2016; Lucas, Manning, et al., 2018). This finding is novel to the extent that racism-related constructs are generally formulated as trait-like individual difference in the available health literature, including when assessed for their potential to act as moderating variables (e.g., Benkert, Peters, Clark, & Keves-Foster, 2006; Drolet & Lucas, 2020). Two insights can be gleaned from the present mechanistic finding. First, health communication theory and research must further explore how and why persuasive health messages can trigger activation of racism-related cognition and emotion, as resulting insights may ultimately lead to more effective communication strategies for use with racial minorities. With respect to the present findings, one possibility is that loss-framed messages could be interpreted as a negative social cue that signals threat to members of socially disadvantaged groups, and may produce a need to engage in psychological self-protection. In turn, activation of racism-related feelings in response to thinking about loss could act as a defense mechanism. Indeed, some research suggests that members of stigmatized or disadvantaged groups may preserve personal well-being by attributing negative social cues to the prejudiced attitudes of others (for review, Major et al., 2002).

Second, activation of racism-related cognition and emotion must be recognized for its capacity to impede receptivity to recommended preventive screenings, including CRC screening. Future studies must further consider that racism-related angst may act as a barrier to cancer screening among African Americans and other racial minorities. With an eye towards clinical trial evaluation, future research should generally evaluate the potential to focus on activation of racism-related cognition and emotion through a wide array of health communication strategies. Presently, we demonstrate that culturally targeting loss-framed messaging may be one viable approach for lessening activation of anticipatory racism. However, other singular or combined messaging strategies may also be effective in this respect. We also note that anticipatory racism only partially mediated effects of culturally targeted message framing, suggesting that future research might also explore additional mechanistic pathways.

Beyond effects of the currently evaluated messaging strategies on CRC screening receptivity, the present results provide two additional insights that may one day aid in developing effective screening programs for addressing extant CRC disparities. First, although we observed significant messaging effects across a broad range of outcomes, African Americans in this study were generally receptive to FIT Kit screening in all messaging conditions, with no less than 60 percent of participants accepting the offer to receive a no-cost FIT Kit. Thus, an especially potent intervention may be comprised of effectively and efficiently presenting information about CRC risks, prevention, and screening options, and providing a ready opportunity to engage in CRC screening. From this perspective, persuasive communication strategies, including culturally targeted message framing, might be regarded as opportunities for supplemental intervention that can work in tandem with broader health behavior education interventions and programs. Second, although most participants in this study accepted the offer to receive a no-cost FIT Kit, only approximately 20 percent of these kits were completed and returned. Available literature highlights that FIT Kit completion rates vary widely by context (Crosby, Stradtman, Collins, & Vanderpool, 2017), and that completion may be especially challenging in low-income and predominantly African American communities (Gupta et al., 2014). Beyond promoting initial uptake of stool-based CRC screening, effective CRC screening programs may require use of additional intervention strategies to encourage the volitional act of completing a take-home CRC screening kit (Weiner, Lewis, Clauser, & Stitzenberg, 2012).

Although we provided an opportunity to engage in no-cost FIT Kit screening, a primary limitation is that we were not able to provide or track either colonoscopy or sigmoidoscopy use among study participants. In turn, the current study did not assess impacts of message framing and culturally targeted messaging on completion of CRC screening through other screening modalities. Effects of the presently evaluated and other health communication strategies on preferences for specific CRC screening modalities have not been well articulated. A ready possibility overlaying the present findings is that, in addition to altering general receptivity to CRC screening, message framing might also alter preferences screening modalities. For example, although culturally targeted loss-framed messaging most effectively compelled CRC screening receptivity, it did not enhance actual FIT Kit screening behavior, perhaps due to promoting a greater desire for more thorough screening than FIT Kit screening provides, such as colonoscopy. Future research must consider that the availability or absence of other CRC screening modalities in a given context could alter the impacts of a persuasive health messaging strategy.

Beyond failing to provide and track colonoscopy and sigmoidoscopy use, several additional limitations suggest a cautious interpretation of the present research, while also pointing towards future research. First, the current findings are characterized by some instances of modest effects sizes, especially for differences across messaging conditions in FIT Kit uptake and use, which were not statistically significant (see Supplemental Material for required sample size calculations). Message framing research has been critiqued for producing discernable, but diminutive effects (O’Keefe & Jensen, 2009), especially when used to compel illness detection behavior among racial minorities (Schneider, 2006). This critique is offset to some extent in that take-home CRC screening programs are frequently conducted through broad outreach, where completion rates are often low, particularly in underserved communities. In this context, seemingly small effects can produce notable differences in cancer detection, and may be ultimately highly valued by public health practitioners (Greenwald, Banaji & Nosek, 2015). Nonetheless, future research will be needed to demonstrate how the proposed messaging approaches can be implemented on a larger scale to reduce racial disparities. Second, only African Americans were included. Although initially focusing on African Americans yielded important insights and is appropriate given extant CRC disparities in the United States, future research will also be needed to develop and evaluate use of culturally targeted message framing to augment CRC screening in other underserved groups. Third, although the current research demonstrates an effective use of culturally targeted messaging, tailored messaging strategies were not considered. Persuasive communication that is crafted to match individual-level as opposed to group-level characteristics is also effective (e.g., Resnicow et al., 2009), including in promoting cancer screening (Kreuter, Lukwago, Bucholtz, Clark, & Sanders-Thompson, 2003). Moreover, some research suggests that tailored messaging can overlay targeted messaging, including among racial minorities using culturally-relevant individual differences (Lucas, Manning, et al., 2018). A vital future direction is thus to decipher whether CRC screening can be further promoted through a combined targeting and tailoring messaging approach, especially given some evidence that culturally-targeted messaging may not always be persuasive when used in isolation (Langford et al., 2017). Fourth, our research only considered a single CRC screening occasion and only provided one screening option (i.e., FIT Kit). Future research must consider effects of the current messaging strategies on maintenance of CRC screening across multiple screening options and occasions, including effects on alternate CRC screening modality preferences and completion. Longitudinal research designs may also be used bolster evidence of mediation through anticipatory racism that was presently observed (Mitchell & Maxwell, 2013)

Limitations notwithstanding, the current research provides a critical advance in further demonstrating that culture is an important consideration when deploying message framing strategies to promote illness detection behaviors, such as CRC screening. Although loss-framed messaging may indeed promote illness detection among African Americans, such messages may require a simultaneous delivery of supplemental culturally directed messaging. Moreover, messaging in the context of CRC screening should be recognized for its capacity to affect racism-related cognition and emotion, which can ultimately alter receptivity to recommended screening. Finally, health messaging may alter preferences for specific CRC screening modalities that are routinely available both to racial minorities and others, highlighting a critical need to better understand these influences, and to decipher whether messaging strategies can be matched to suit particular CRC screening programs and contexts.