Exploring direct and indirect predictors of heart disease information seeking

ABSTRACT

Background

Based on the integrative model of behavioral prediction, we examined predictors of heart disease information seeking. We also examined demographic and individual factors associated with seeking-related perceived norms, attitudes, and perceived behavioral control.

Methods

Non-Hispanic White and Black participants, aged 45 and older, completed a cross sectional online survey (N = 383). Stepwise logistic and multiple linear regression models were tested to assess study hypotheses, as well as tests of indirect effects.

Results

Perceived norms, attitudes and perceived behavioral control were positively associated with heart disease information seeking, but when controlling for distal variables only the perceived norm-behavior association remained significant (p <.05). Indirect effects of distal variables (race, heart disease risk, perceived heart disease susceptibility and information engagement orientation) on information seeking were also detected via perceived norms.

Conclusions

Our results provide support for the integrative model as a framework for predicting information seeking, but further highlight the important role of distal predictors and perceived norms on heart disease seeking intentions. When communicating to promote heart disease information acquisition, communicators should pay particular attention to promoting information seeking as a normative behavior, particularly among those who perceive a lower risk of heart disease and who may be less engaged with health information more generally.

KEYWORDS:

• Information seeking behavior
• attitudes
• behavioral control perceived norms
• heart diseases

One of every four deaths in the U.S. is the result of heart disease, making it the leading cause of death for both men and women [1]. Deaths caused by heart disease were highest among non-Hispanic White (23.7%) and Black (23.5%) Americans as compared to other racial and ethnic groups [2]. Close to half of Americans (49%) exhibit at least one heart disease risk factor including high blood pressure, high cholesterol and smoking [3]. Risk factors for heart disease, including hypertension (i.e. having high blood pressure or taking medication to lower blood pressure), obesity and diabetes, are also higher among non-Hispanic Black adults than White adults [4].

Behaviors that may decrease heart disease risk include avoiding smoking, maintaining a healthy weight, engaging in physical activity and disease screening [5]. However, knowledge of heart disease risks and risk factors is still fairly low [6, 7]. Knowledge and awareness of cardiovascular risk factors, in particular, may be associated with preventive actions [8]. Still, only a few studies to date have examined predictors of heart disease information seeking, focusing on individual [9–13] and cognitive [14,15] factors. Understanding the factors that drive seeking decisions could help health communicators develop more effective ways to promote heart health information acquisition, as campaigns may have modest effects on information behavior. In the case of the Heart Truth Campaign, although 61% of women said they recognized the red dress as a symbol for heart disease, only 25% reported searching for more information about heart disease after hearing or seeing information about it [16].

Interventions to increase health information seeking may be a tool for reducing heart disease-related disparities between Black and White Americans. While the two groups are shown to seek information about health at similar rates, research has shown that Black Americans are more likely to use the information sought to change approaches to managing their health and to better understand how to treat an illness or health condition [17]. Studies in other disease contexts also show that seeking information is often associated with disease-related awareness and knowledge and, in some cases, screening and prevention behaviors (Wigfall and Friedman, 2016). We draw on the integrative model of behavioral prediction (IMBP; [18,19] to examine predictors of heart disease information seeking among non-Hispanic Black and White Americans. The IMBP has been used to investigate health-related information seeking [20], although less is known about the fit of the model in this context.

Determinants of behavioral intention

The IMBP proposes that behavioral intentions (and subsequent behaviors) are a function of attitudes (perceived favorability), perceived behavioral control (PBC; perceived capability) and perceived normative (social) pressure to engage in the behavior, which are shaped by distal variables [21]. Attitudes are consistently, positively, and significantly associated with health information seeking and intentions [22–27]. A test of the IMBP also shows attitudes predict information seeking intention [20]. Furthermore, beliefs about heart disease, which are believed to underlie attitudes [21], are shown to predict information seeking [28]. Thus, we expect that those with more positive attitudes towards seeking will report seeking heart disease information (H1).

PBC (i.e. self-efficacy) refers to one’s perceived ability to perform an action [21]. PBC is often a strong predictor of behavior and intention (see [29, 30], although this prediction has had mixed support across studies examining information seeking broadly (e.g [27, 31–35]. However, this prediction was supported in a previous test of the IMBP [20]. PBC also predicted heart disease information-seeking intentions when risk perceptions of heart disease were also high [15]. Thus, we expect PBC will be positively associated with heart disease information seeking (H2).

Perceived norms are conceptualized as expected social pressure regarding the performance of a behavior, comprised of injunctive (perceptions of whether a person’s important social networks will be supportive of the behavior) and descriptive norms (perceptions of what social network members actually do) [21, 36, 37]. Perceived norms are a strong predictor of information-seeking outcomes (See [38, 39] for meta analysis), including tests of the IMBP [40]. Therefore, we propose that perceived norms will be positively associated with heart disease information seeking (H3).

Environmental constraints

The IMBP, unlike its predecessor models, introduced two so-called ‘actual control’ variables that are proposed to influence the intention-behavior relationship [41]. These variables are designed to account for situations when positive attitudes, self-efficacy and normative beliefs are not enough to change behavior [42]. It is hypothesized that people are more likely to act on their intentions when (1) skills needed to perform the behavior are present and (2) environmental factors (e.g. resource-based or logistical) do not constrain their ability to act [19, 21].

Despite the potential influence of skills and environmental constraints, IM studies have typically not included these variables; in fact, one study suggested no IM studies had looked at environmental constraints [42]. This noticeable gap in the empirical literature may be due to the broad and largely undefined nature of the environmental constraint variable, which is typically defined using examples of potential constraints (such as problems related to health insurance or transportation; [21]); thus, the specific factors that may be considered environmental constraints are not entirely clear. While this study does not attempt to account for all the potential environmental constraints that may exist for individuals who desire to seek information, our research provides a starting point for research on this topic by examining information access barriers (i.e. lacking resources to access information; [43]). Information barriers to seeking are shown to affect heart disease information-seeking behaviors [11]. Those who face greater barriers accessing information may have lower expectations regarding their ability to seek, particularly individuals who experience information poverty [44], which may ultimately impact decisions to seek information. Disparities in health information access also have been noted based on demographic and socioeconomic factors [45]. Black Americans are also shown to be less likely to trust health information from health care providers, which may further excerbate disparities [46]. Thus, we propose that information barriers will be negatively associated with information seeking (H4).

Background contributors to behavior

According to the IMBP, normative beliefs, attitudes and PBC are shaped by a variety of distal variables [19]. Tests of the IMBP that account for the indirect effects of distal variables have better model fit than those who simply use them as statistical controls [47], although the indirect effects of distal variables on health outcomes has been tested only a handful of times (e.g. [48–51]. Information seeking differences are noted based on individual and demographic characteristics [52, 53]; this is also true for heart disease information specifically [9–13]. Our study focuses on demographic and individual difference distal variables. As outlined below, and in alignment with the IMBP, we propose that attitudes toward seeking, PBC and perceived norms will mediate the effects of distal variables on information seeking (H5).

Demographic and Health Characteristics

Previous information seeking research in a different context examined the demographic and health-related variables of interest in this study (i.e. race/ethnicity, gender, diagnosed heart disease risk, and socio-economic status), showing them to be significantly associated with attitudes, PBC and perceived norms [54]. Specifically, although past seeking did not differ, their results showed that Black Americans were shown to have more positive attitudes, greater seeking-related subjective norms and perceived behavioral control than White Americans; personal and family disease history also were associated with perceived seeking control and seeking norms [54]. More broadly, socio-economic and cultural factors [11, 13], gender [9] and family health history [12] are associated with heart disease information seeking and intentions. Thus, we might expect demographic and health characteristics to have indirect associations with seeking via cognitive variables identified by the IMBP.

Individual difference variables

We examine three individual difference variables (i.e. perceived heart disease susceptibility, information apprehension, and information engagement) that are prominent within the health information seeking literature. Perceived risk, which incorporates perceived susceptibility, is shown to be a better distal than proximal predictor of intentions and behavior [19, 55]; as perceived susceptibility to risk increases, so might social pressure to engage in behaviors, as well as attitudes and efficacy to manage risk [18]. Perceptions of risk are integral to information seeking theory (e.g. [23, 56]), although it is typically viewed in the information seeking literature as an independent predictor rather than a background variable. Previous tests of the IMBP have also incorporated perceived risk (incorporating severity and susceptibility) as a distal predictor. One study showed perceived risk was negatively associated with attitudes and self-efficacy [57], while another showed perceived invulnerability was positively associated with attitudes, PBC and perceived norms [58].

The other two individual difference variables examined here related to individual information orientation (i.e. one’s health information seeking style and motivations), which reflects individual differences in perceived needs for information [59]. People who seek health information are more likely to be health-conscious or health-oriented [60, 61], which is positively related to health-related attitudes and behaviors [62] and PBC [63]. Black American and Hispanic populations also reported greater levels of health consciousness than non-Hispanic White populations [64]

Individuals who have a stronger health information orientation (i.e. they are not apprehensive about seeking information and like to gather and review information; [59]) may view information seeking more positively and feel greater seeking efficacy. Information orientation has also been found to be indirectly associated with behavioral intention via perceived norms [51].

Methods

Participants were recruited through an online survey research panel coordinated by Qualtrics, Inc in February of 2015. A purposive sampling strategy was used to recruit a sample of non-Hispanic White and Black participants, age 45 or older (N = 383). Although heart disease impacts Americans of all ages, risk increases significantly after age 45 for men and age 55 for women [65]. Deaths in 2015 caused by heart disease were also highest among these two racial/ethnic groups [2]. Eligible panelists were invited to participate via e-mail and provided consent online before taking the survey.

Measures

Information seeking

A single item asked about past seeking. ‘Have you ever looked for information about heart disease from any source?’ (responses included: (1) yes, (2) no, or (3) don’t know). This item was recoded and treated as a binary variable (1 = yes, 0 = no), with ‘don’t know’ responses dropped.

Attitudes towards seeking

Based on existing measures [23, 41], respondents were asked to provide their opinion on heart disease information seeking using seven 10-point differential pairs (e.g. worthless/valuable, bad/good, harmful/beneficial). Confirmatory factor analysis (CFA) confirmed item unidimensionality (loadings from .73 to .90; Cronbach’s α = .94). The items were averaged.

Perceived Behavioral Control (PBC)

PBC items were based on Ajzen [41] and Kahlor [23]. Four items were measured on a 1 (strongly disagree) to 5 (strongly agree) scale (e.g. ‘I know where to look for information about my own heart disease risks’). The mean of the items was used (α = .91, CFA item loadings from .79 to .87).

Perceived norms

Perceived norms measures were based on Parks & Smith [66]. Three items measured injunctive norms, including ‘Most people whose opinions I value would approve of my seeking information about heart disease.’ Two items measured descriptive norms, including ‘most people who are important to me have sought information about heart disease.’ Items were measured on a 1 (strongly disagree) to 5 (strongly agree). The mean of the items was used (α = .87, CFA item loadings from .48 to .92) as a single indicator of perceived norms.

Health information orientation

Six items were used from a scale [59] assessing health information engagement (e.g. ‘I like to gather as much information as I can before making a decision’) and apprehension (e.g. ‘I fear I might find out something I don’t want to know’). Items were measured on a scale from ‘not at all’ (1) to ‘very true’ (5). As in previous research [59], CFA confirmed a two-factor solution with engagement loadings from .73 to .83, and apprehension loadings from .65 to .86. Thus, engagement (α = .82) and apprehension (α = .81) were entered separately in the models tested here.

Information barriers

The barriers to information access scale was used here [43]. Individuals responded to three statements, including ‘I needed health information that I couldn’t afford the time or effort to get.’ Items were measured on a (1) strongly disagree to 5 (strongly agree) scale. The mean of the items was used in the analyses (α = .80, CFA item loadings from .70 to .94).

Perceived susceptibility

A single item was used to measure perceived heart disease susceptibility (e.g. ‘How likely is it that you will get heart disease in your lifetime?’) on a scale from 1 (not likely) to 5 (definitely) scale.

Demographic and health characteristics

Measures were based on the Behavioral Risk Factor Surveillance Survey [67]. Items included age (continuous), sex (1 = female, 0 = male), race/ethnicity (1 = Black or African American, 0 = White), education level (from (1) never attended to (6) 4 years or more of college) and income (from [1] less than $10,000 to [8] more than $200,000).

We also measured the presence of heart disease risk factors, as indicator of objective individual heart disease risk, with 3 items: ‘has a doctor, nurse or other health professional ever told you that you have high blood pressure/high cholesterol/angina or coronary heart disease’ (1 = yes, 2 = no, 3 = don’t know). These three items were combined into a single variable indicating whether a participant had at least one heart disease risk factor (1 = yes, 0 = no); individuals with missing data or don’t know responses for any of the three items were dropped from the analysis.

Data analysis

All analyses were conducted in SPSS version 27.0. Descriptive statistics and model variable correlations were first assessed (Table 1). Stepwise binary logistic regression was used to examine direct and indirect predictors of information seeking with IMBP predictors first, followed by distal predictors [20]. In addition to the distal predictors included in our model, participant body mass index (BMI) was measured (using self-reported height and weight) as an indicator of heart disease risk. Preliminary analyses showed BMI was not a significant predictor of seeking or seeking-related cognition; thus, for clarity, we did not control for BMI in the analyses presented here.

Table 1. Spearman Correlations and Descriptive Statistics for Continuous Model Variables

	M	SD	1.	2.	3.	4.	5.	6.
1. Perceived Susceptibility	2.83	1.12
2. Information Barriers	2.34	0.98	.06
3. Information Apprehension	2.88	1.14	.19**	.31**
4. Information Engagement	4.22	0.82	.04	.04	.03
5. Perceived Norms	3.60	0.75	.24**	-.01	.13*	.31**
6. Attitudes Towards Seeking	8.64	1.69	.18**	-.07	-.08	.33**	.39**
7. Seeking PBC	3.94	0.78	.15**	-.28**	-.11*	.27**	.31**	.28**

Note: N = 383. ** p ≤ .01. * p ≤ .05

Multiple linear regression models were tested to identify associations between distal predictions and cognitive predictors. To assess mediation predictions, indirect effects of distal predictors on information seeking were calculated using the PROCESS macro [68], version 4.0 [69].

Results

Participant characteristics

Of participants, 48% (n = 184) self-identified as White and 52% identified as Black or African American (n = 199). The mean age of participants was 58.45 years (range = 45–87 years) and 64.5% were Female (n = 247). The majority (97.3%) had completed high school, but close to 48% (n = 178) had household incomes of less than $35,000 per year. Many participants (68%; n = 240) self-reported heart disease or a heart disease risk factor (high blood pressure and/or cholesterol), and 40% (n = 152) reported a family history of heart disease.

Predictors of information seeking

Perceived norms, attitudes and PBC were positively associated with information seeking (Table 2; H1-H3 supported), although information barriers were unassociated (H4 not supported). When distal predictors were added to the model, only perceived norms remained significant; perceived susceptibility and education also were associated with seeking. As shown in Table 2, the odds of seeking information were 3.146 times greater with each one-unit increase in perceived norms (compared to odds of 1.061 for attitudes and 1.266 for PBC, respectively, although neither were significant in Step 2). In addition, the odds of seeking information were 1.785 times greater with each one-unit increase in perceived susceptibility and 1.617 times greater with each one-unit increase in education level.

Table 2. Stepwise Binary Logistic Regression Models Predicting Information Seeking

	Step 1			Step 2
	b	SE b	OR (95% CI)	b	SE b	OR (95% CI)
Direct Predictors
Perceived Norms	.944	.202**	2.570 [1.73, 3.82]	1.146	.248**	3.146 [1.94, 5.11]
Attitudes Towards Seeking	.159	.079*	1.172 [1.00, 1.37]	0.059	.095	1.061 [0.88, 1.28]
PBC^a	.398	.179*	1.489 [1.05, 2.12]	0.236	.206	1.266 [0.85, 1.89]
Information Barriers	.025	.122	1.026 [0.81, 1.30]	0.146	.151	1.158 [0.86, 1.56]
Indirect Predictors
Black American^b				−0.364	.285	0.695 [0.40, 1.22]
Female				0.210	.289	1.234 [0.70, 2.17]
Education				0.481	.179*	1.617 [1.14, 2.30]
Heart Disease (HD) Risk ^c				0.200	.303	1.221 [0.68, 2.21]
Perceived HD Susceptibility				0.579	.139**	1.785 [1.36, 2.34]
Information Apprehension				−0.207	.132	0.813 [0.63, 1.05]
Information Engagement				0.311	.202	1.364 [0.92, 2.03]
χ^2 (df)	61.661 (4) **			100.872 (11) **
Pseudo R^2 ^c	.21			.36

Note. N = 383. ^a PBC = Perceived Behavioral Control; ^b 1 = non-Hispanic Black American, 0 = non-Hispanic White American; ^c exhibiting heart disease or at least one risk factor (high blood pressure or high cholesterol); ^c Nagelkerke pseudo-R-Square.

** p ≤ .01. * p ≤ .05

Predictors of perceived norms, attitudes and PBC

Perceived susceptibility and information engagement orientation were the most consistent distal predictors (Table 3); they were both positively associated with attitudes, PBC and perceived norms. Heart disease risk also was positively associated with attitudes and perceived norms, while information apprehension was negatively associated with attitudes and PBC. In terms of demographics those with less education reported more positive attitudes towards seeking, while Black participants reported greater perceived norms.

Table 3. Multiple Linear Regression Models Predicting Information Seeking Norms, Attitudes Towards Seeking and Seeking PBC

	Attitudes		PBC^a		Perceived Norms
	b	SE b	b	SE b	b	SE b
Black American^b	.211	.177	-.063	.084	.259	.078**
Female	.350	.183	.092	.087	.056	.081
Education	-.333	.110*	.029	.052	-.040	.049
Heart Disease (HD) Risk^c	.500	.194*	.190	.091*	.135	.085
Perceived HD Susceptibility	.273	.081**	.079	.038*	.159	.036**
Information Apprehension	-.226	.078*	-.106	.037*	.027	.035
Information Engagement	.469	.111**	.253	.052**	.253	.049**
F (df)	142.952 (7, 319) **		24.967 (7,324) **		36.789 (7, 326) **
R^2 Effect Size	.159		.125		.192

Note. N = 383. ^a PBC = Perceived Behavioral Control; ^b1 = non-Hispanic Black American, 0 = non-Hispanic White American; ^c exhibiting heart disease or at least one risk factor (high blood pressure or high cholesterol). ** p ≤ .01. * p ≤ .05

Finally, tests of indirect effects showed perceived norms to be the only consistent mediator (Table 4; H5 partially supported). Indirect effects of nearly all distal variables on seeking were detected via perceived norms except gender, education and information apprehension.

Table 4. Indirect Effects of Demographic and Individual Difference Variables on Information Seeking via Proposed Mediators

	Proposed Mediators
	Attitudes		PBC^b		Perceived Norms
Independent Variables	Est.	95% BCa	Est.	95% Bca	Est.	95% Bca
Black American ^a	.019	[-.042, .093]	-.015	[-.107, .045]	.249	[ .079, .473] *
Female	.074	[-.004, .186]	.031	[-.034, .128]	.090	[-.073, .258]
Education	-.018	[-.077, .025]	.033	[-.006, .099]	-.012	[-.110, .091]
Heart Disease Risk ^c	.066	[-.031, .224]	.096	[ .006, .249] *	.264	[ .062, .532] *
Perceived HD Susceptibility	.033	[-.014, .103]	.036	[-.005, .099]	.155	[ .064, .290] *
Information Apprehension	-.014	[-.053, .012]	-.029	[-.088, .003]	.062	[-.013, .150]
Information Engagement	.080	[-.006, .186]	.089	[-.001, .207]	.269	[ .135, .447] *

Note. N = 383. ^a 1 = non-Hispanic Black American, 0 = non-Hispanic White American; ^b PBC = Perceived Behavioral Control; ^c exhibiting heart disease or at least one risk factor (high blood pressure or high cholesterol).

* significant indirect effect

Discussion

The goal of the present study was to assess the efficacy of the IMBP framework to predict information seeking. Perceived norms, attitudes and PBC were independently and positively associated with heart disease information seeking. However, when controlling for distal variables, only the perceived norm-behavior relationship remained significant; tests of indirect effects also showed perceived norms was the most consistent cognitive mediator. As have others [38], our results suggest norms are quite influential on seeking decisions (the odds of seeking were 3.146 times greater with each one-unit increase in perceived norms), although the variance explained by seeking-related cognitions on behaviors relative to distal predictors is small.

Unexpectedly, attitudes and PBC played a modest role in predicting past information seeking. Furthermore, these two variables were unassociated with information seeking when distal variables were added to the model; thus, education and perceived heart disease susceptibility appear to be more influential on seeking decisions (although odds ratios indicate they are not as influential as perceived norms). Others have shown that perceived norms are a better predictor of behavior than attitudes and self-efficacy [70]. Furthermore, our results showed perceived norms mediated associations between distal predictors and information seeking. Race indirectly influenced seeking via perceived norms, with Black participants reporting greater perceived norms for information seeking than Whites. This finding may be reflective of the fact that heart disease is the leading cause of death within this racial group [71]. Black Americans also suffer racial disparities in health care due to structural and individual racism (e.g. [72, 73], which is a noted reason they experience differential treatment for coronary heart disease [74] and may lead them to seek information beyond health care providers. Indeed, research shows that Black Americans are more likely to use information obtained outside of a health encouter to maintain their health [17].

Perceived heart disease susceptibility and the presence of actual heart disease risk factors (e.g. exhibiting high blood pressure, high cholesterol and/or angina or coronary heart disease) also indirectly influenced information seeking via perceived norms. Specifically, individuals reporting one or more heart disease risk factors, and also perceived greater heart disease susceptibility, were more likely to view seeking as normative, which was associated with information seeking.

These results support information seeking theories that incorporate variables assessing subjective perceived risk (e.g. [23, 56], as well as actual objective risk, but suggest risk may have a broader influence on cognitions and seeking decisions than previously thought. Future research should continue to examine the influence of risk (actual and perceived) on cognitions. Some have suggested that increased age and the presence of a chronic condition (e.g. heart disease) are more important predictors of health information seeking than race or ethnicity [17], although the relationship between race, ethnicity and health outcomes is complex. Demographic variables often stand-in for complex social relationships and structures such as racism and gender norms [17], thus it is important to include a wide range of distal variables that may be associated with information seeking.

Finally, information orientation appears to influence seeking decisions via cognitive factors. The effect of information engagement orientation on seeking was mediated by perceived norms. Our results provide support for the hypothesis of individual differences in perceived needs for disease-related information [59]. Here, individuals who reported greater information engagement were more likely to view seeking as a normative behavior, which increased the likelihood of seeking. Dissemination of heart disease information is therefore critical to those who may be less engaged with information, as they may be less likely to seek on their own. However, dissemination should involve more than just providing access, as experiencing information barriers was unassociated with information seeking in this study. While this finding may reflect the commonality of heart disease in the U.S. and easy access to information online (85% of adults now own a smartphone; Pew [75]), opportunities to encounter heart disease information more effortlessly through normal media use (e.g. information scanning) or everyday living (e.g. health clinics or blood pressure check kiosks) may help to increase public information acquisition. Communicators should also seek to support and empower individuals to access information [59] and promote seeking as normative.

Limitations

There are several limitations to note. First, we recruited only non-Hispanic White and Black participants; thus, we cannot generalize heart disease information seeking to the population more broadly. We also used a single item to assess information seeking behavior, which may not fully capture the extent that participants sought information. The cross-sectional design is also a limitation of this study; it demonstrates correlations between factors, but not causal relationships. Future research should build on these findings by employing a longitudinal study design (e.g. [20]) or capturing actual information seeking (e.g. [76]). The use of an online panel also limits the generalizability of this study, although our results may still guide future theorizing and research.

Conclusion

The present study extends the IMBP into the context of heart disease information seeking, highlighting the influential role of perceived norms. Future health interventions should consider the role of perceived norms on information seeking, as well as the influence of distal factors such as demographics, disease risk and information orientation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Time Sharing Experiments for the School of Communication, The Ohio State University.

Notes on contributors

S.R. Hovick

Dr S.R. Hovick is an associate professor in the School of Communication at The Ohio State University.

N. Rhodes

Dr N. Rhodes is an associate professor in the Department of Advertising and Public Relations at Michigan State University.

E. Bigsby

Dr E. Bigsby is an assistant professor in the Department of Communication at the University of Illinois.

S. Thomas

Dr S. Thomas is now a Senior Strategist for Egg Strategy.

N. Freiberger

N. Freiberger is a PhD Student in the School of Communication at The Ohio State University.