Psychometric Testing of a Cardiovascular Disease Fatalism Instrument Among Adults With Cardiovascular Disease Risks

Abstract

Background:

We modified a general health fatalism instrument to examine cardiovascular disease (CVD) fatalism because there is no specific CVD fatalism instrument (CVD-FI).

Methods:

Adults with two or more CVD risk factors completed a 20-item CVD-FI rated on a 5-point Likert scale. Higher scores indicated higher CVD fatalism. Reliability and construct validity of the CVD-FI were examined using Cronbach’s alpha, factor analysis, and hypothesis testing using correlation respectively.

Results:

Cronbach’s alpha was 0.89 supporting internal consistency. Hypothesis that individuals with lower adherence to healthy lifestyle will have high fatalism score was supported (Pearson’s r = −0.151; p = .001), and factor analysis yielded a 4-factor solution.

Conclusions:

CVD-FI is a reliable and valid measure of CVD fatalism. More research is needed to confirm the emergent 4-factor solution of CVD fatalism.

Cardiovascular disease (CVD), which includes coronary heart disease, stroke, and peripheral artery disease, is the leading cause of morbidity and mortality both in the United States and worldwide (Benjamin et al., 2018). Rates of CVD are expected to increase due to a combination of factors including an aging population and continuing challenges to reduce CVD risk (Heidenreich et al., 2011; Wall et al., 2018). In the United States, total direct medical costs associated with CVD are projected to triple, from $273 billion in 2010 to $818 billion in 2030 (Heidenreich et al., 2011). CVD has become our nation’s costliest chronic disease. Expenses associated with CVD are expected to soar in the coming years and surpass medical cost estimates for other chronic diseases, including diabetes and Alzheimer’s. The aging of the population combined with the growth in per capita medical spending are the primary drivers of increased CVD costs (Heidenreich et al., 2011). Based on prevalence, death rates, disability, and cost, CVD will continue to be the most burdensome disease Americans will face in the next decade (American Heart Association, 2017).

The etiology of CVD is multifactorial, arising from complex interactions between genes, lifestyle behaviors, and environmental factors (Mudd-Martin et al., 2015). In rural communities, high rates of poverty, low levels of education, and limited access to healthcare in combination with unhealthy lifestyle behaviors result in higher burden of CVD morbidity and mortality than in urban communities (Caldwell et al., 2016; Hartley, 2004). While genetic and environmental factors are difficult to modify, behaviors that increase CVD risk, such as smoking, physical inactivity, obesity, unhealthy diet, and medication nonadherence, are modifiable and enhancing engagement in healthy lifestyle behaviors can significantly reduce CVD risk (Havranek et al., 2015; Mozaffarian et al., 2015).

While addressing modifiable risk factors is foundational to improving cardiovascular health among those with high CVD risk, many people are unaware that they are at increased risk and therefore do not make appropriate risk-reducing lifestyle changes (Imes et al., 2014). It would be anticipated therefore that increasing awareness of CVD risk could support lifestyle modification to improve health outcomes. Yet increased awareness can also induce a sense of fatalism or lack of personal control that could undermine strategies to reduce CVD risk (Osokpo and Riegel, 2019).

Fatalism has traditionally been conceived as the perception that events or health issues are out of the control of the individual (Niederdeppe & Levy, 2007; Straughan & Seow, 1998). It has been identified as a doctrine held by individuals who believe that all events are fated to happen and that human beings have no control over their future and are unable to change their outcomes (Abraído-Lanza et al., 2007; Franklin et al., 2007). Fatalism has also been defined as the belief that situations, including illnesses or catastrophic events, are caused by a higher power (such as God) or are just meant to happen and cannot be avoided (Austin et al., 2002).

Fatalistic belief toward health is one of the common reasons cited for nonadherence to lifestyle changes and preventive screening recommendations (Deskins et al., 2006). Fatalistic beliefs are associated with lower odds of engaging in prevention behaviors, including regular exercise, not smoking, and adequate fruit and vegetable consumption (Niederdeppe & Levy, 2007). In a study among Hispanics, Urizar and Sears (Urizar, 2006) found that fatalism was higher in patients with more severe CVD and affected the quality of life with a greater negative impact. In another study among Hispanics, higher fatalism was associated with prevalence of hypertension even after adjusting for socioeconomic status and acculturation (Gutierrez et al., 2017). Also, endorsement of fatalism has been found to be associated with lower rates of engagement in health-screening behaviors, such as reduced utilization of cancer screening services (De Los Monteros & Gallo, 2011).

Because fatalism can be a significant barrier to healthy lifestyle changes and adoption of preventive health behaviors, assessing CVD fatalism can be an important strategy to risk reduction. By identifying people with high CVD fatalism, risk reduction strategies addressing fatalism can be implemented to more effectively support engagement in healthy behaviors. Although there are different fatalism scales, most have been developed to measure either disease-agnostic fatalism or fatalism specific to cancer (Cuéllar et al., 1995; Egede & Bonadonna, 2003; Esparza et al., 2014; Hunt et al., 2000; Shen et al., 2009; Straughan & Seow, 1998). To our knowledge, no instruments have been developed to measure CVD fatalism. To address this gap, we modified a disease-agnostic fatalism scale developed by Shen, Condit, and Wright (2009) to measure CVD fatalism and examined the instrument’s psychometric properties. The specific aims of the study were to (1) evaluate the construct validity of the CVD-FI using exploratory factor analysis and hypothesis testing and (2) assess the reliability and internal consistency of the scale.

MATERIALS AND METHODS

Design and Sample

For this methodological study, baseline data from the HeartHealth in Rural Kentucky study was used. The HeartHealth study was a two-phase community trial conducted from 2009 to 2012. The research was conducted among rural Kentuckians to test effectiveness of an intervention to reduce CVD risk by promoting self-management of risk factors (Mudd-Martin et al., 2015). Purposive sampling was used to recruit community-dwelling members at risk for CVD from rural Kentucky areas. Participants were recruited through flyers distributed by local healthcare and community organizations. Recruitment sites were rural counties as defined by the U. S. Department of Agriculture, Economic Research Service’s Rural-Urban Continuum Codes (United States Department of Agriculture).

Community-dwelling residents 18 years of age or older who had two or more risk factors for CVD were eligible to participate in the HeartHealth study. Risk factors for CVD included sedentary lifestyle; unhealthy diet; overweight or obese; current smoker or tobacco user; greater than 45 years of age for males and 55 years of age for females; self-reported depression, anxiety, or chronic stress; and self-reported history or current treatment of the following: hypertension, abnormal lipids, and diabetes. Individuals who were taking prescription medications that interfered with lipid metabolism, were unable to give consent due to cognitive impairment, were non-English speaking, were chronic drug abusers, or had chronic disease requiring a special diet or prohibiting physical activity were excluded. The design and results have been published previously (Mudd-Martin et al., 2015).

A total of 1,181 participants completed the HeartHealth study. Of these participants (n = 526) without a personal history of a CVD event who completed a questionnaire including the CVD-FI, and a modified Medical Outcomes Study (MOS) scale to measure adherence to 12 health behaviors were included in this analysis. Participants were asked to indicate how often they had engaged in each of the 12 behaviors during the past 4 weeks.

Measures

Demographic Variables.

Demographic data was collected using a self-report questionnaire. Age was assessed as a continuous variable (age in years). Race/ethnicity was based on self-report; most participants identified as Caucasian, hence sample was categorized as Caucasian and others. Education was assessed by response to number of years of education completed; responses were dichotomized as less than 12th education and postsecondary. Financial adequacy was assessed by asking if participants have more than enough to make ends meet, had enough to make ends meet, or did not have enough to make ends meet.

CVD Fatalism.

CVD fatalism was assessed using the CVD-FI, a modified version of the fatalism scale developed by Shen and colleagues in 2009 (Shen et al., 2009). We chose the fatalism scale developed by Shen because it provided a disease-agnostic measure of fatalism and had strong construct validity. Shen and colleagues developed the fatalism scale in three steps. In the first step, a pool of items was developed by adapting items from the Powe Fatalism Inventory (PFI) as well as by creating new items. All items were presented to a multi-cultural Community Advisory Board (CAB) for comments and review for level of clarity (including readability), cultural appropriateness, and cultural inclusion. In the next step, the items were pretested with a sample of college students. Items that were perceived to be awkwardly worded, to have low face validity, or to be difficult to respond to were dropped. In step three, data were collected from a randomly selected nationally representative sample as part of a web-based national survey. Of the 1,145 participants, 49.8% were non-Hispanic White, 22.3% were non-Hispanic Black, 23.7% were Hispanic, and 1.9% were more than one race or ethnicity (Shen et al., 2009).

Shen et al. (2009) conceptualized fatalism as a set of health beliefs that is a combination of predetermination, pessimism, and attribution of one’s health (life events) to luck. Hypothesizing that the fatalism scale encompassed these three factors, construct validity of the final instrument was assessed using confirmatory factor analysis. The alpha reliabilities for the three original subscales were 0.86 for predetermination (10 items), 0.80 for luck (4 items), and 0.82 for pessimism (6 items); reliability was 0.88. Construct validity was supported by examining correlations between subscales and external variables hypothesized to be associated with each subscale. The final instrument consisted of 20 items, with Likert scale response options from 1 (strongly disagree) to 5 (strongly agree). The total score of the scale ranged from 20 to 100, with higher scores indicating more fatalism.

For the purposes of this study, the original disease-agnostic fatalism instrument scale was modified to address CVD fatalism. Whereas multiple items in the original disease-agnostic scale referred to “serious disease,” in the adapted scale, “heart disease” was used instead. As an example, an item on the original scale, “If someone is meant to get a serious disease, they will get it no matter what they do,” was changed to “If someone is meant to get heart disease, they will get it no matter what they do.” Global items from the original scale such as “I will die when I am fated to die” and “Sometimes I feel that I’m being pushed around in life” were retained. As the original scale, Likert scale response options and a summed score were used, with a lower total score indicating lower level of fatalistic beliefs.

Adherence to Healthy Lifestyle Behaviors

A modified version of the Medical Outcome Study Specific Adherence Scale (MOS-SAS) (Hays, 1994) was used to measure adherence to 12 health behaviors: regular exercise; taking medications as prescribed; reducing stress; stopping or cutting down on smoking; following a weight loss diet; eating five or more servings of fruit/vegetable; following a diet low in fat; following a diet low in salt; eating low fat or fat free diary; and eating primarily whole grain rather than processed food (Hays, 1994). Participants were asked to indicate how often they had engaged in each of the health behaviors in the past 4 weeks. Response options range from none of the time (0) to all of the time (5). The total possible scores ranged from 0 to 60, a higher score indicated greater engagement in healthy behavior practices. The MOS-SAS has demonstrated validity and reliability when used in populations with CVD (DiMatteo et al., 1993; Hays et al., 1994). The Cronbach’s alpha for the scale in this study was 0.86.

Procedure

Approval for the study was obtained from the University of Kentucky Institutional Review Board and all participants provided written informed consent. Data from a questionnaire that included demographic variables, the CVD-FI, and the MOS specific adherence scale are included in this secondary analysis. Data were de- identified, examined, and cleaned before psychometrics analysis were performed. All data forms were identified by an assigned number. Data entry was evaluated for accuracy by double entry and corrected to ensure 100% accuracy.

Data Analysis

Data were analyzed using IBM SPSS version 23 for windows (Chicago, United States). Descriptive statistics using means and standard deviations for continuous variables and frequency distributions and proportions for categorical variables were used to summarize the sample characteristics. Construct validity was assessed using exploratory factor analysis and hypothesis testing.

Construct Validity

An exploratory factor analysis using principal components extraction with Varimax rotation was performed to examine the factor structure of the 20-item CVD-FI. Four methods for interpretation were applied; eigenvalues greater than one, the scree plot, total variance explained for factor extraction, and conceptual consideration (Polit & Hungler, 1999). Also, a loading of 0.40 or greater was used to identify items contributing to a given factor (Polit & Hungler, 1999). The Kaiser-Meyer-Olkin (KMO) and Barlett’s test of sphericity were used to assess appropriateness of the sample for the factor analysis.

We performed hypothesis testing to further examine the construct validity of the CVD-FI. Based on the literature we hypothesized that individuals with low adherence to healthy lifestyle behaviors would have high fatalism scores (Niederdeppe & Levy, 2007; Straughan & Seow, 1998; Urizar, 2006). Correlation analyses were conducted to examine associations between total MOS scores and score on the CVD-FI.

Reliability

To examine reliability of the CVD-FI and the subscales, internal consistency was determined using Cronbach alpha reliability coefficients. Reliability of the CVD-FI was also assessed by obtaining the mean inter-item correlations and the item-total correlations; coefficients greater than 0.30 were considered acceptable for item-total correlations. Inter-item correlation coefficients between 0.30 and 0.70 were considered acceptable (Ferketich, 1991).

RESULTS

Sample Characteristics

The average age of participants was 52 ± 13 years. The majority of participants were female (73%), married (73%), and White (95%). The demographic characteristics of the sample are presented in Table 1.

TABLE 1.

Demographics and Clinical Characteristics (N = 526)

Characteristics	Mean (Standard Deviation) or Number (%)
Age in years	52 (±13)
Female	385 (73%)
Marital status
Married/cohabitate	386 (73%)
Single/divorced/widowed	140 (27%)
Education level
<12th grade	183 (35%)
Postsecondary education	343 (65%)
Ethnicity
White	501 (95%)
Other	25 (5%)
Financial status
Have more than enough to make ends meet	208 (40%)
Have enough to make ends meet	276 (52%)
Do not have enough to make ends meet	42 (8%)
Employment status
Employed full time	408 (78%)
Homemaker	14 (3%)
Retired not due to illness	78 (15%)
Others	26 (5%)
Have health insurance	447 (85%)
Fatalism score	43.7 ± 10.9

Construct Validity

Factor Analysis.

The KMO value was 0.90 and the Barlett’s test of sphericity was significant (χ² = 4,399, df =190, p < .001) supporting the use of factor analysis. Results of the factor analysis are presented in Table 2.

TABLE 2.

Factor Analysis with Varimax Rotation of Cardiovascular Disease Fatalism Instrument (CVD-FI) (N = 526)

Item	Factor 1	Factor 2	Factor 3	Factor 4
How long I live is a matter of luck	0.75
My health is a matter of luck	0.73
I will stay healthy if I am lucky	0.72
I will get heart disease if I am unlucky	0.62
If someone is meant to get heart disease, they will get it no matter what they do	0.59
If someone gets heart disease that’s the way they were meant to die	0.58
If someone is meant to get heart disease, it doesn’t matter what they eat, they will get heart disease anyway	0.54
My health is determined by fate	0.53
If someone was meant to have heart disease, it doesn’t matter what doctors and nurses tell them to do, they will get heart disease anyway.	0.49
If someone has heart disease and gets treatment for it, they will probably still die from it.	0.43
How long I live is already determined		0.82
My health is determined by something greater than myself		0.72
I will die when I am supposed to die		0.71
If someone is meant to have heart disease, they will get it		0.61
I often feel helpless in dealing with problems of life			0.83
Everything that can go wrong for me does			0.71
Sometimes I feel I am being pushed around in life			0.64
There is no way I can solve some of the problems I have			0.52
I will suffer a lot from bad health				0.82
I will have a lot of pain from illness.				0.72

The principal components analysis with Varimax rotation revealed four factors with eigenvalues exceeding 1. Factor 1 had the highest value at 7.192, factor 2 was 1.685, factor 3 was 1.408, and factor 4 was 1.292. The cumulative percentage of variance accounted for by the four-factor components was 57.9% with component 1 contributing 20.6%, component 2 contributing 14.3%, component 3 contributing 12.5%, and component 4 contributing 10.5%. All extracted items have factor loadings between 0.43 and 0.83. Items that loaded on factor 1 were conceptually related to predetermination, items that loaded on factor 2 were related to pessimism, while items that loaded on factor 3 and 4 were related to helplessness and suffering, respectively. Three items cross loaded; a “cross loading” item is an item that loads at 0.32 or higher on two or more factors. Item 2 cross loaded on both factors 1 and 2, while item 4 cross loaded on factors 1 and 2. Item 6 cross loaded on factors 1 and factor 4. A four-factor component for the CVD-FI scale was retained following the use of criteria for retaining factors. The extraction of four factors does not replicate the original instrument which had three subscales.

Hypothesis Testing.

There was a negative correlation between total MOS and fatalism scores (Pearson’s r = −0.151; p = .001). Hypothesis that individuals with lower adherence to healthy lifestyle will have high fatalism score was supported.

Reliability

Cronbach’s alpha was 0.89 which supported the internal consistency reliability of the CVD-FI. Cronbach’s alpha was 0.85 for the predetermination subscale, 0.79 for the pessimism subscale, and 0.74 for the helplessness subscale. The suffering subscale had lower reliability with a Cronbach’s alpha of 0.68.

Item scale and subscale means and standard deviations are shown in Table 3. The Cronbach’s alphas if deleted were between 0.889 and 0.898. Each item is important to the scale and none of the items if deleted will improve reliability. The item-total correlations were between 0.36 and 0.68 indicating adequate correlation among all the items. All mean and standard deviations were between 1.79 ± 0.77 and 3.32 ± 1.15. There was no evidence of a ceiling or floor effect.

TABLE 3.

Specific Statistics of the Cardiovascular Disease Fatalism Instrument (CVD-FI) (N = 526)

Items	Mean	Standard Deviation	Item Total Correlation
1: If someone is meant to get heart disease, it doesn’t matter what kinds of food they eat, they will get heart disease anyway.	2.00	0.82	0.65
2: If someone is meant to get heart disease, they will get it no matter what they do.	2.08	0.84	0.68
3: If someone gets heart disease that’s the way they were meant to die.	1.79	0.77	0.52
4: If someone is meant to get heart disease they will get it.	2.25	0.94	0.65
5: If someone has heart disease and gets treatment for it, they will probably still die from it.	2.14	0.83	0.50
6: If someone was meant to have heart disease, it doesn’t matter what doctor and nurses tell them to do, they will get heart disease anyway.	1.98	0.90	0.64
7: How long I will live is already determined.	2.55	1.21	0.48
8: I will die if I am fated to die.	3.32	1.15	0.48
9: My health is determined by fate.	1.84	0.86	0.57
10: My health is determined by something greater than myself	2.70	1.14	0.53
11: I will get heart disease if I am unlucky.	1.96	0.84	0.55
12: My health is a matter of luck.	1.81	0.70	0.62
13: How long I will live is a matter of luck.	1.88	0.77	0.57
14: I will stay healthy if I am lucky.	2.24	0.99	0.52
15: Everything that can go wrong for me does.	1.94	0.85	0.49
16: I will have a lot of pain from illness.	2.34	0.91	0.49
17: I will suffer a lot from bad health.	2.41	1.15	0.36
18: I often feel helpless when dealing with problems of life.	2.13	0.93	0.46
19: I feel I am being pushed around in life.	2.22	1.05	0.46
20: There is really no way I can solve some problems I have.	2.17	0.92	0.44

DISCUSSION

To our knowledge, this is the first study that has been conducted to assess the psychometric properties of an instrument developed to measure CVD fatalism. The results support the reliability and validity of the CVD-FI when used to measure CVD fatalism. Findings provided support for four CVD-FI factors: predetermination, pessimism, helplessness, and suffering. This differed from the original three factors of the general health fatalism instrument reported by the instrument’s developers (Shen et al., 2009). The inability to replicate the three factors structure identified in the original disease-agnostic scale suggests that the modified CVD-FI instrument may be capturing other dimensions of fatalism not previously identified. The variation in the factor structure could be due to several reasons. The original disease-agnostic instrument was tested in a nationally representative sample whereas the CVD-FI was used with a sample of residents from rural Kentucky who were at-risk for CVD. Measuring CVD-specific fatalism in people with high risk for the disease may capture dimensions distinct from a disease-agnostic measure used with people without identified risk. There may also be differences in the complexity of CVD fatalism compared to fatalism associated with general health.

The negative association of CVD-FI with adherence to healthy lifestyle supported the construct validity of the instrument. Results of previous studies have similarly shown that fatalism is associated with the underutilization and failure to adopt such health protective behaviors as engaging in appropriate cancer screenings, timely care seeking, quitting smoking, being physically active, and healthy dietary practices (Niederdeppe & Levy, 2007; Powe & Finnie, 2003; Straughan & Seow, 1998).

The results of this study indicate that the CVD-FI is a reliable measure of CVD related fatalism with the exception of the subscale “suffering” that, while acceptable, resulted in a lower Cronbach’s alpha than the other factors. The Cronbach’s alpha of 0.89 for the CVD-FI was consistent with the results of Shen and colleagues (Shen et al., 2009) who reported a Cronbach’s alpha of 0.88 for the original general health fatalism 20-item scale. This reflects strong internal consistency and evidence for good reliability.

Because high fatalism has been associated with low engagement in health protective behaviors, people with high CVD fatalistic beliefs may need more intensive intervention to support their engagement in healthy lifestyle behaviors. To effectively influence fatalistic beliefs in populations at risk for CVD, it is critical for researchers and health providers to have valid and reliable measures to screen for CVD related fatalistic beliefs. Such an instrument can therefore be an important tool to aid in identifying people who need greater support to effect healthy changes in behaviors.

LIMITATIONS

Because our sample was comprised of predominantly Caucasians, reflecting the demographics of rural Kentucky, the generalizability of the results may be limited. Further research is required to test the psychometrics of the CVD-FI among a more diverse population. This study was also conducted with individuals at risk for CVD, although results may differ with a population without CVD risk, the majority of adults in the United States have at least one risk factor for CVD and it would therefore be anticipated that findings would be similar in a general adult population.

CONCLUSIONS

The CVD-FI is an easy to administer 20-item instrument that can be used to assess CVD fatalism. The findings from this study are promising, providing evidence of reliability and validity of the CVD-FI scale. This instrument can provide a valuable measure for assessing CVD related fatalism to identify individuals with CVD fatalism who may benefit from intensive interventions that address health behaviors and underlying fatalistic beliefs.

Relevance to Nursing Practice, Education, or Research

Given the high rates of CVD in U.S. adults, this instrument provides a clinical tool that can easily be administered by nurses to asses CVD fatalism. Screening for CVD fatalism can support appropriate allocation of intervention resources by healthcare providers especially nurses. A valid and reliable CVD fatalism instrument can also be used by nurse researchers who can examine effectiveness of interventions among people with low compared to high fatalism so that we can begin to understand how to tailor interventions more effectively.