The association between social functioning and health literacy among rural Southeastern African Americans with hypertension

Abstract

African Americans living in low socioeconomic circumstances are at high risk of poor health outcomes; this is particularly true for those with hypertension. Many African Americans with chronic hypertension living in the rural impoverished Southeastern USA have low health literacy and are socially isolated. These factors are known to have a negative impact on health outcomes, but it is possible that social support may overcome some of the effect of low health literacy. Since little has been reported about this association, we examined the association between social functioning and health literacy in a rural African American population in the Southeast USA. We used baseline data from participants in the Southeastern Collaboration to Improve Blood Pressure Control, a pragmatic trial that recruited rural African Americans with persistently uncontrolled hypertension and collected survey data. Overall, 33.5% of the 1221 person sample reported social isolation, 26.0% reported low instrumental support, 36.0% reported low emotional support, and 63.4% had inadequate health literacy. All three domains of low social functioning were significantly associated with low health literacy, and this effect was robust to multivariable adjustment for sociodemographics and cognitive functioning for social isolation (adjusted odds ratio 1.62, 95% confidence intervals 1.20–2.20). In conclusion, the majority of this sample living in the NC and AL Black Belt had high social functioning but inadequate health literacy. Tests of interventions to improve social support, especially social isolation, may be warranted to overcome low health literacy in this high-risk rural population.

INTRODUCTION

The rural Black Belt region in the Southeastern United States, known and named for its rich soil, has predominantly African American communities that have the highest cardiovascular disease mortality in the USA (Washington, 1901; Kyalwazi et al., 2022). This region also contains mostly rural communities living in low socioeconomic circumstances (Wimberley, 2010). The high mortality rates may be due in part to the fact that compared to whites, African Americans have higher prevalence of uncontrolled hypertension (Safford et al., 2007; Guo et al., 2012). Americans living in rural Southeastern communities may experience unique stressors that limit their ability to manage chronic health conditions, including lack of social support and poor understanding of their medical conditions due to low health literacy, contributing to worse health outcomes (Dave et al., 2013; Cuffee et al., 2020). This may be particularly critical to evaluate as the extent of disparities experienced by rural communities is not fully reported (Probst et al., 2004).

Low health literacy has been associated with worse health outcomes in African American communities (Mensah et al., 2014). Health literacy refers to ‘the degree to which individuals have the ability to find, understand, and use information and services to inform health-related decisions and actions for themselves and others’ (CDC, 2021). This includes a patient’s ability to navigate interactions with healthcare providers, as well as properly utilize fundamental resources including discharge instructions, educational materials, and medical technology (McDonald and Shenkman, 2018). In rural African American communities, poor health literacy may be reinforced by education disadvantage that limits individuals’ access to resources, including occupations with health insurance and routine health care visits (Probst et al., 2004). Individuals who reside in rural communities have been shown to have lower health literacy, as compared to individuals who reside in urban communities (Zahnd et al., 2009). This is further exemplified by the fact that those with low health literacy tend to have more hospitalizations and higher healthcare expenditures in part due to limited preventative measures, as well as an inability to effectively communicate health concerns and understand health instructions (Davis et al., 2020). This includes being less likely to engage in preventative measures, including mammography screening and influenza vaccines (Berkman et al., 2011). For those with chronic diseases like hypertension, this can be particularly detrimental, leading to worse health outcomes and a 75% increased risk of mortality (Mackey et al., 2016).

Social isolation reflects insufficient social connections or interactions, which has also been shown to have a negative impact on overall health. In the USA, more than a quarter of older adults live alone and more than a third of older adults report chronic loneliness (Vespa et al., 2012). Social isolation has also been demonstrated to cause more deleterious consequences for those with lower socioeconomic status and for non-white ethnic groups, such as the African American population in the Black Belt. Black men and women have been found to have fewer close relationships and participate in less groups, making them more likely to report social isolation (Alcaraz et al., 2019). Individuals may further face isolation as a result of poverty and rurality, as well as experienced racism and medical distrust within their community (Connell et al., 2019). The U.S. census reports a higher percentage of people of color in the southeastern region, almost 58% of African Americans in the USA. The impact of social isolation on health has been amply documented, including health related quality of life and premature mortality (Uchino et al., 1996; Hawton et al., 2011). These factors may contribute to social isolation and loneliness, which have consistently been linked to poorer health and worse health outcomes (Steptoe et al., 2013; Rico-Uribe et al., 2016).

Studies suggest that social relationships may play a significant role in moderating the effect of health literacy on health outcomes (Berkman et al., 2011; Smith et al., 2018). Particularly because people frequently consult friends and family about health concerns, social support can be a critical influence on health management (Verbrugge and Ascione, 1987). High social support may overcome low health literacy by serving as an additional source of medical information. By providing decision-making support and behavioral cues, social relationships may help individuals with low health literacy manage various aspects of their health (Smith et al., 2018). Understanding the various ways in which social functioning relates to health literacy may be critical to helping patients manage and comprehend their health. This may be an especially important goal in rural communities, which have been noted to have lower health literacy than urban populations (Aljassim and Ostini, 2020).

The consequences of social isolation and poor health literacy have been demonstrated to be worse for those with lower socioeconomic status and from non-white ethnic groups, such as the Black Belt population. Patients with low health literacy and more social isolation are significantly more likely to report limiting, chronic disease (Smith et al., 2018). Studies suggest social relationships and health literacy influence one another, but the prevalence and significance of this effect on a rural population living with chronic disease is unknown. If social support can overcome some of the deleterious effects of low health literacy in rural African Americans, interventions designed to improve social support could be implemented in this rural region to improve health outcomes.

Given the significant disparity in health outcomes compared to the rest of the USA and demonstration of low health literacy and social isolation, these factors should be thoroughly evaluated in rural southeastern African Americans. The Southeastern Collaboration (SEC) to Improve Blood Pressure Control is a large trial currently underway in the Black Belt regions of Alabama and North Carolina that aims to eventually enroll 2000 African Americans with persistently uncontrolled hypertension. The trial is testing the effectiveness of two interventions designed to improve blood pressure control among hypertensive patients—peer coaching and practice facilitation (Cherrington, 2016). Trial participants completed a survey at baseline that included validated instruments assessing social support and health literacy, providing an opportunity to examine the relationship between social support and health literacy. In this population of African Americans who live with persistently uncontrolled hypertension in the Black Belt, we expected: (i) participants would have low levels of social functioning (social isolation, instrumental functioning, and emotional functioning) as well as low health literacy; and (ii) greater social support would be associated with higher health literacy in this population.

METHODS

Sample

This study was conducted in the context of the SEC, a pragmatic trial comparing the impact of four approaches to improving blood pressure control in rural African Americans: practice facilitation, peer coaching, both practice facilitation and peer coaching, and enhanced usual care (Cherrington, 2016). The setting is the Black Belt region, a remote rural area of the Southeast with a predominantly African American population and a heavy burden of chronic diseases including hypertension. Poverty rates are more than 40% greater than Southern communities not in the Black Belt (Driskell and Embry, 2007). Medical resources are scarce; the Kaiser Family Foundation estimated that the availability of primary care providers is as low as one tenth of the national average (Kaiser Family Foundation, 2020).

The SEC trial was designed to enroll 2000 participants who self-identify as African Americans aged 19–85 years with persistently uncontrolled hypertension (defined as mean systolic blood pressure 140 mm Hg or greater over the previous year). Participants had to reside in the Black Belt, in a self-reported rural setting, have no advanced illness causing limited life expectancy or advanced chronic kidney disease, and not have plans to move out of the area for the next year. Approximately 50 primary care practices in the Black Belt of Alabama and 30 primary care practices in the Black Belt of North Carolina were recruited and 25 participants were recruited at each practice. Baseline data included the Patient Reported Outcomes Measurement Information system (PROMIS) social functioning scales, Chew’s 3 item health literacy scale, as well as sociodemographics and cognitive functioning.

Measures

Main exposure: social functioning

Social functioning was evaluated using the PROMIS social functioning scales, including perceived social isolation, emotional support, and instrumental support. Each scale contains 4 items with response options of ‘never’, ‘rarely’, ‘sometimes’, ‘usually’, or ‘always’ (see Supplementary materials). A cumulative raw score from 4 to 20 was calculated and converted to a standardized T-score to allow comparison to the U.S. population, which set 50 as the mean with a standard deviation of 10. In order to compare and identify those with low or high social functioning in this naturally skewed data, 50 was used to dichotomize each domain of emotional support, instrumental support, and social isolation (Cella et al., 2010).

Dependent variable: health literacy

Health literacy was evaluated using Chew’s 3 item health literacy scale. The questions are: (i) How often do you have someone help you read instructions, pamphlets, or written materials from your doctor or pharmacy? (ii) How confident are you filling out medical forms by yourself? (iii) How often do you have problems learning about your medical condition because of difficulty understanding written information? (Chew et al., 2004). Items 1 and 3 have four response options: ‘all of the time’, ‘most of the time’, ‘some of the time’, ‘a little of the time’ or ‘none of the time’. Item 2 has four different response options: ‘extremely’, ‘quite a bit’, ‘somewhat’, ‘a little bit’, and ‘not at all’. Items 1 and 3 were analyzed as dichotomous variables with participants who responded ‘none of the time’ considered to have adequate health literacy and all others to have inadequate health literacy. Item 2 was analyzed as a dichotomous variable with participants who responded ‘extremely’ considered to have adequate health literacy and all others to have inadequate health literacy. For purposes of operationalizing modeling, we created a dichotomous variable: Individuals with inadequate health literacy on all three questions were considered to have inadequate health literacy and all others were considered to have adequate health literacy. We selected this dichotomy to screen participants for adequate health literacy if they subjectively have absolutely no difficulty understanding medical information. Those who have difficulties at least some of the time with each question were considered to have inadequate health literacy.

Covariates

We included data on covariates that are known to confound the relationship between health literacy and social functioning. These covariates included age (18–64 and 65+ years), gender, annual household income (<$20 000 and ≥$20 000), and educational attainment (<high school and ≥high school). Cognitive impairment was assessed via the Six-Item Screener; patients with a score ≥3 were defined as being cognitively impaired. This screener assesses participant’s cognition by asking basic questions, including orientation to time and 5-minute word recall in order to understand participants’ ability to apprehend and discern information. It is a validated and sensitive screener of cognition (Callahan et al., 2002).

Statistical analysis

We began by examining the distribution of the outcome (health literacy) and main exposure (social functioning) in the population. We then described characteristics of the sample with low and high social isolation. We used logistic regression to estimate crude and multivariable adjusted odds ratios (OR) and 95% confidence intervals (CI) to investigate the relationship between each social functioning domain and health literacy separately. We then used ordinal logistic regression to estimate crude and fully adjusted OR and 95% CI to examine the relationship between the social functioning domains and health literacy simultaneously.

RESULTS

In total, 1221 subjects were included in our analytic sample, reflecting ongoing recruitment in its final year. The majority (72.0%) of individuals were between 18 and 65 years old, 61.0% were women, 82.0% had more than high school education, 64.0% were unmarried, and 40.0% had annual income of $20 000 or less (Table 1).

Table 1:

Characteristics for those with High vs. Low social isolation among 1221 participants in the SEC

Characteristic	Total	More social isolationa	Less social isolationa	p-value
N	1221	409	812
Age				<0.001
18–65	874 (71.6%)	327 (80.0%)	547 (67.4%)
65+	347 (28.4%)	82 (20.0%)	265 (32.6%)
Female	745 (61.0%)	269 (65.8%)	476 (58.6%)	0.016
< High School Education	213 (17.4%)	81 (19.8%)	132 (16.3%)	0.12
Total annual income				<0.001
≥20 K	486 (39.8%)	133 (32.5%)	353 (43.5%)
<20 K	574 (47.0%)	230 (56.2%)	344 (42.4%)
Refused/Don’t Know	161 (13.2%)	46 (11.2%)	115 (14.2%)
Cognitive impairment (six-item score ≥3)	39 (3.2%)	21 (5.1%)	18 (2.2%)	0.006
Married	445 (36.4%)	124 (30.3%)	321 (39.5%)	0.002
Site				0.90
UAB	560 (45.9%)	191 (46.7%)	369 (45.4%)
ECU	479 (39.2%)	157 (38.4%)	322 (39.7%)
UNC	182 (14.9%)	61 (14.9%)	121 (14.9%)

ECU, East Carolina University; SEC, Southeastern Collaboration to Improve Blood Pressure Control; UAB, University of Alabama Birmingham; UNC, University of North Carolina.

Bold numbers are statistically significant.

^aRaw scores from PROMIS social isolation 4a short item questionnaire translated to standardized t-scores. Score of 50 represents average for the USA general population. Those with social isolation score below 50 defined as having low social isolation, participants with a score above 50 defined has having high social isolation.

Within the sample, 409 (33.5%) had a score >50 on the social isolation PROMIS scale, 317 (26.0%) had a score >50 indicating low instrumental support, and 440 (36.0%) reported low emotional support (Supplementary Tables 1 and 2). Those with more social isolation tended to be between the ages of 18 and 65 (80%) and women (65.8%). The majority (56.2%) of these individuals had an annual income less than $20 000. Only 124 (30.3%) of those reporting more social isolation were married and 5.1% of individuals with more social isolation had cognitive impairment, compared to 2.2% for those with less social isolation (Table 1).

In contrast, 774 (63.4%) reported inadequate health literacy. Of those with inadequate health literacy 68.0% were younger (between 18 and 65), 55.3% were women, 53.5% had annual income greater than $20 000, 36% were married, and 24.9% had more than a high school education (Table 2).

Table 2:

Characteristics for those with adequate vs. Inadequate Health Literacy among 1221 participants in the SEC

Characteristic	Adequate health literacy	Inadequate health literacya	p-value
N	447	774
Age			<0.001
18–65	348 (77.9%)	526 (68.0%)
65+	99 (22.1%)	248 (32.0%)
Female	317 (70.9%)	428 (55.3%)	<0.001
< High school education	20 (4.5%)	193 (24.9%)	<0.001
Total Annual Income			<0.001
≥20 K	230 (51.5%)	256 (33.1%)
<20 K	160 (35.8%)	414 (53.5%)
Refused/Don’t Know	57 (12.8%)	104 (13.4%)
Cognitive Impairment (six-item score ≥3)	2 (0.4%)	37 (4.8%)	<0.001
Married	166 (37.1%)	279 (36.0%)	0.70
Site			0.031
UAB	192 (43.0%)	368 (47.5%)
ECU	173 (38.7%)	306 (39.5%)
UNC	82 (18.3%)	100 (12.9%)

ECU, East Carolina University; SEC, Southeastern Collaboration to Improve Blood Pressure Control; UAB, University of Alabama Birmingham; UNC, University of North Carolina.

Bold numbers are statistically significant.

^aHealth literacy was comprised of 3 questions: (i) How often do you have someone help you read instructions, pamphlets, or written materials from your doctor or pharmacy? (answers range from ‘None of the time’—‘All of the time’), (ii) How confident are you filling out medical forms by yourself? (answers range from ‘Extremely’—‘Not at all’) and (iii) How often do you have problems learning about your medical condition because of difficulty understanding written information? (answers range from ‘None of the time’—‘All of the time’). Participants were considered to less than ideal health literacy if they did not respond ‘None of the time’ or ‘Extremely’ to all 3 questions.

Of the sample, inadequate health literacy was more common in those with less social support across all three domains of social support (Figure 1). Furthermore, 476 (39.0%) responded that they needed help reading instructions, pamphlets, or written materials from their doctor or pharmacy; 600 (49.1%) responded less than extreme confidence in filling out medical forms independently; and 493 (40.4%) responded that they had problems learning about their medical condition because of difficulty understanding written information at least ‘a little of the time’ (Figure 2). When considering each item in the health literacy scale separately for each social support domain, lack of confidence and difficulty learning about medical conditions due to difficulty reading were more common in those with low emotional and instrumental support (Figure 2). All three domains of low health literacy were more common in those reporting social isolation (Figure 2).

Fig. 1:

Percent of Participants with adequate (black) and inadequate (gray) Health Literacy, Among High vs. Low PROMIS Social Functioning^a Groups. a. Raw scores from PROMIS 4a short item questionnaire translated to standardized t-scores. Score of 50 represents average for the United States general population. Those with score below 50 defined as low, participants with a score above 50 defined as high. b. For Emotional and Instrumental Support, low levels indicate poor social functioning. For isolation, more indicates poor social functioning.

Fig. 2:

Percent of Participants with Inadequate Health Literacy For Each Individual Item, Among High vs. Low PROMIS Social Functioning^a Groups. a. Raw scores from PROMIS 4a short item questionnaire translated to standardized t-scores. Score of 50 represents average for the United States general population. Those with score below 50 defined as low, participants with a score above 50 defined as high.

Low emotional support, low instrumental support, and more social isolation were all individually significantly associated with inadequate health literacy in crude analyses (Table 3). After adjusting for sociodemographics and cognition, low emotional support and social isolation remained statistically significantly associated with low health literacy (adjusted OR 1.54 CI [1.17, 2.02] and 1.77 CI [1.34, 2.34], respectively). When all social support domains were modeled simultaneously, emotional support and social isolation but not instrumental support were associated with low health literacy in the crude analysis. In the fully adjusted model, only social isolation remained independently associated (adjusted OR 1.62 CI [1.20, 2.20]). There was no significant interaction by either age or gender for these relationships.

Table 3:

Association between Individual Social Functioning Domains and Less than Ideal Health Literacy^a

Social functioning domains	N	Crude	Fully adjusted
		OR (95% CI)	OR (95%)
Independently modelled domains
Low emotional support	440	1.63 (1.27, 2.09)	1.54 (1.17, 2.02)
Low instrumental support	317	1.33 (1.01, 1.74)	1.26 (0.93, 1.69)
More social isolation	409	1.74 (1.34, 2.25)	1.77 (1.34, 2.34)
Simultaneously modelled domains
Low emotional support	440	1.41 (1.06, 1.87)	1.33 (0.98, 1.79)
Low instrumental support	317	0.99 (0.73, 1.34)	0.96 (0.69, 1.34)
More social isolation	409	1.54 (1.17, 2.04)	1.62 (1.20, 2.20)

Fully adjusted model adjusted for age, gender, education, income, cognitive impairment, marital status, and site.

OR, Odds Ratio; CI, Confidence interval.

Bold numbers are statistically significant.

^aHealth literacy was comprised of 3 questions: (i) How often do you have someone help you read instructions, pamphlets, or written materials from your doctor or pharmacy? (answers range from ‘None of the time’—‘All of the time’), (ii) How confident are you filling out medical forms by yourself? (answers range from ‘Extremely’—‘Not at all’) and (iii) How often do you have problems learning about your medical condition because of difficulty understanding written information? (answers range from ‘None of the time’—‘All of the time’). Participants were considered to less than ideal health literacy if they did not respond ‘None of the time’ or ‘Extremely’ to all 3 questions.

DISCUSSION

In this sample of rural Southeastern African Americans with persistently uncontrolled hypertension enrolled in a community-based randomized trial, a large minority reported low emotional support, low instrumental support, and social isolation, and nearly two thirds reported inadequate health literacy. Low health literacy was significantly associated with low social support, with the strongest effect observed for social isolation. In all domains of health literacy assessed, those with more social isolation were particularly more likely to have low confidence in their ability to fill out medical forms independently. Social isolation may play a more significant role in overall social functioning, than emotional or instrumental support. This may be because social isolation impacts the other domains, as someone with isolation is also likely to have low emotional and instrumental support. Further evaluation of the interactions between emotional and instrumental support and social isolation may be warranted. The effects of poor social support in minority communities may be compounded as they are surrounded by similar minority communities, leading to widespread, institutional disadvantage. This disparity may be improved with increased resources and human capital to create social structures that more effectively reach all residents to prevent social isolation (Probst et al., 2004).

Prior studies have suggested that those who are older, unmarried, and male are more likely to experience social isolation (Smith et al., 2018). We adjusted for sociodemographics and cognition to eliminate the effect factors such as age or mental capacity could have on participants’ ability to engage with others. Individuals in rural areas tend to report significant social isolation and loneliness, which may be more severe for those who cannot access support because of personal or cultural factors. Despite residing in a remote rural region, the majority of this sample did not report high levels of social isolation. Some studies have suggested that this variation may be because some rural communities may be tightly knit, and this may be true in the African American rural communities in which these trial participants lived (Henning-Smith et al., 2019). It may be worthwhile to further investigate how certain demographic factors influence the social isolation experienced by individuals living the Black Belt.

This study suggests that social support may be a strategy to overcome low health literacy, especially in regions with scarce resources. The SEC includes a peer group intervention arm, which could function to improve patients’ understanding of medical conditions and compliance to medical regimens. Furthermore, trained peers can serve as a model to help individuals learn about their health conditions and the role of self-management, improving confidence in the ability to manage health (Lee et al., 2004). In past studies, the most consistent improvement after participating in a coaching intervention is greater confidence in the ability to self-manage chronic diseases, a concept known as self-efficacy. The peer coaching model being tested in the SEC is well-suited to these African American rural communities, where low health literacy is common. Risk factors for low health literacy have traditionally included lower education level, male gender, younger age, and non-White ethnicity, (Christy et al., 2017) which were consistent with the findings in our sample. However, more studies may help evaluate the way in which these demographics intersect and effect low health literacy.

This study is one of the first to evaluate the relationship between health literacy and social functioning in this rural, remote African American population using a validated health literacy scale and PROMIS social functioning scales. Examining this large sample of individuals provides insight to the differences in social functioning and health literacy within a minority group with chronic disease. Though the SEC was still recruiting participants at the time of the study, we felt this large sample size was appropriate to investigate this critical relationship, in order to guide future studies. Our study has limitations worth noting. The sample was composed of individuals enrolled in a randomized trial, possibly limiting generalizability. Our measure of health literacy emphasizes the understanding of written information, without evaluating the specific ability to understand verbal information which may skew results for certain participants. Our dichotomization of the Chew health literacy scale designated those reporting any difficulty understanding health information or filling out forms as having inadequate health literacy, which may not be practical for this subjective scale. Additionally, this population had relatively high social functioning, possibly reflecting the tightly knit communities in the Black Belt but limiting generalizability to other rural communities. Nevertheless, the very high burden of chronic diseases and short life expectancy in the southeastern Black Belt warrants study of this unique population. Additionally, this was an observational study thus we could not determine causality between social functioning and health literacy.

In conclusion, in this sample, most participants reported high social functioning despite their rural environment. Those who reported low social functioning differed from prior studies in that they tended to be younger women, rather than older men. A majority of the sample had inadequate health literacy, with younger men with low educational attainment at significant risk for low health literacy, consistent with past studies. Those with low social functioning were more likely to have inadequate health literacy, suggesting that social support may have a significant role in a patient’s ability to learn about and understand healthcare information. Of the three social functioning domains, social isolation, which may be most easily identified in patients, was most strongly associated with inadequate health literacy, an effect that was similar for both genders and older and younger individuals. These findings support the promise of peer support interventions to assist individuals with chronic diseases living in the remote rural Black Belt region.

Funding

This work was supported by grant NHLBI 5UH3HL130691-05.