Social Support Networks and Foreign-Birth Status Associated With Obesity,Hypertension and Diabetes Prevalence Among 21-30 and 50-70 Year Old Adults Living in the San Francisco Bay Area

Abstract

Purpose:

To examine whether social network characteristics of US-and foreign-born individuals are related to hypertension, diabetes and obesity prevalence.

Design:

Cross-sectional.

Setting:

Six San Francisco Bay Area counties.

Participants:

N = 1153 cohorts of young and older adults (21-30 and 50-70 years).

Measures:

Network structure and support measures were calculated using name elicitation and interpreter questions common in egocentric surveys. Hypertension and diabetes were self-reported, and overweight/obesity was determined using body mass index calculations. Foreign-birth status was based on country of birth.

Analysis:

Adjusted and unadjusted logistic regression models were used to examine associations between network characteristics and hypertension, diabetes and overweight/obesity. These relationships were tested for moderation by foreign-birth status, age and gender.

Results:

Higher percentages of family members (AOR = 4.16, CI: 1.61-10.76) and same-sex individuals (AOR = 3.41, CI: 1.25-9.35) in the composition of respondents’ networks were associated with overweight/obesity. Higher composition of family members (AOR = 3.54, CI: 1.09-11.48) was associated with hypertension. Respondents whose networks composed of higher numbers of advice individuals (AOR = 0.88, CI: 0.77-0.99), female respondents (AOR = 0.52, CI: 0.35-0.77) and foreign-born respondents (AOR = 0.54, CI: 0.32-0.92) were less likely to report overweight/obesity. Diabetes was associated with higher composition of individuals living within 5-minutes to respondents (AOR = 5.13, CI: 1.04-25.21).

Conclusion:

Family and network support members such as advice individuals could be potential targets for chronic disease prevention, particularly among older adults and immigrants.

Keywords

chronic diseases older adults foreign-born social networks

Purpose

Chronic diseases or risk factors for developing these diseases cause a heavy health burden in the US and affect approximately 60% of American adults.¹ Chronic diseases and conditions such as obesity, hypertension and diabetes are of particular concern because they are associated with the leading causes of deaths in the US, for example heart disease.² These chronic diseases, however, have different social and behavioral risk factors that complicate their distribution in the US.

Disparities in chronic diseases have been described according to age, gender and immigration status.^1,3-5 For example, obesity prevalence is higher in women and older adults compared to men and younger adults, respectively.⁶ Immigrants experience lower prevalence of morbidity and mortality from obesity, diabetes and hypertension than US-born individuals.⁷ Social networks, access to healthcare, and utilization of screening and other health prevention services may contribute to differences in prevalence of chronic diseases among immigrants compared to US-born individuals. Social networks are especially important because immigrant communities often have a tight knit social structure, which can influence behavioral norms that impact health behaviors and outcomes linked with chronic disease.⁸ Social network refers to the web of social ties among a group of individuals.

These individuals may be influenced by their network through different network characteristics.⁹ These include network structure, which can include the size of a given person’s network, its composition (i.e., the proportion of network members that are friends, family members or coworkers), the strength of social ties, and the quality of network connections. Some researchers use network quality to indicate levels of social support available from the network.¹⁰ While the overall patterns of network characteristics of US-and foreign-born individuals are likely to be different, comparisons between these 2 groups are limited.^11-13

The available network studies suggest that larger network size, increased intimate ties such as close family members and friends, and more memberships in various organizations or network groups are associated with reduced systolic and diastolic blood pressure.^14-17 While examination of networks and diabetes has been more limited, higher levels of network structure (i.e., the count of different network members, e.g., friends and co-workers, that a person has frequent contact with) and perceived social support have been linked with lower odds of diabetes in US Hispanic adults.¹⁸ Obesity has been associated with various types of networks: friend, family and peer networks.^19-21 This effect appears to act through social norms, with individuals imitating obesity risk behaviors of their network, including belief formation, and building social capital.²¹

Foreign-born individuals often migrate to the US for reasons such as preexisting social ties in the US.⁸ The social ties and reasons for immigrating may differ by gender and age. For example, younger men may utilize networks with ties such as friends, coworkers and social groups to obtain employment, while women, who are often older when they migrate, depend more on family networks.^8,22 Changes in network relationships over time complicate reasons for immigration and remaining in the US. These changes, which occur alongside acculturation, have implications on how support networks impact health behaviors. Further examination of the structure and quality of social networks of US-and foreign-born individuals can improve understanding of their differences in development of chronic diseases.⁹ Despite the importance of network characteristics, prior studies have provided limited evidence regarding the role of network characteristics and potential socio-demographic moderators of these effects and have focused on individual chronic diseases.^16,18,19

Therefore, this study examined whether differences in network structure and quality of US-and foreign-born individuals were associated with diagnosis of 3 chronic disease and condition outcomes: hypertension, diabetes and obesity. In terms of network structure, we hypothesized that the outcomes will be positively associated with increases in: proportion of family, friends, neighbors, school and work colleagues, same sex individuals, those that study participants feel close to, and individuals living within 5 minutes to participants. We also hypothesized that the outcomes will be negatively associated with increase in network quality (i.e., network support) measures, including the number of: individuals who offered advice, were social companions, were confidants, or provided practical support to the respondents. To further explore these relationships, we also assessed the moderating effect of foreign-birth status, gender, and age group on these relationships, hypothesizing that observed associations will be stronger for foreign-born individuals, women, and older adults.

Methods

Design

This study used cross-sectional data from the first wave of the University of California Berkeley Social Network and Health Study (UCNets) conducted in 2015. A detailed description of the methods can be found elsewhere.²³ Briefly, UCNets is a 3 wave panel study of change in personal networks overtime among young and older adults (21-30 and 50-70 years) residing across 6 San Francisco Bay Area counties. Participants were surveyed on their social networks, experience of major life events and health.

Sample

The older adults (n = 690) and slightly less than half of the young adults were sampled using stratified address-based sampling. An additional sample of young adults were recruited using a small snowball sampling from participant referrals (n = 32) and a large Facebook advertisement, which resulted in a total of 469 young adult for the first wave of the study. Across the recruitment methods for the young adults, there were relatively few differences in network characteristics.²⁴ Respondents from the stratified random sample and the snowball sample were randomly assigned to complete an online survey or face-to-face survey interviews in their home or a preferred quiet setting. Those recruited via Facebook were directed to an online survey that was virtually identical to the face-to-face survey. Following case-wise deletion of missing data on the relevant health outcomes and network characteristics variables the analytic sample was 1153. Specifically, cases missing information on obesity status, hypertension diabetes, foreign-birth status and total network size (n = 6) were excluded. The institutional review boards of East Carolina University and the University of California, Berkely approved this study.

Measures

Dependent variables

Hypertension and diabetes were dichotomous variables based on whether respondents reported a doctor had ever told them they had either condition. Overweight and obesity status was calculated using body mass index (BMI) with 4 categories, including: underweight (BMI < 18.5), normal weight (BMI 18.5-24.9, or <23 for Asians), overweight (BMI 25.0-29.9, or 23-27.5 for Asians) or Obese (BMI ≥ 30, or >27.5 for Asians).²⁵

Independent variables

UCNets uses a detailed egocentric name-elicitation technique to collect information on the personal networks of study respondents. Network characteristics variables were derived from a series of 9 name elicitation questions.^26,27 For example, participants were asked to name the people they socialize with, confide in, as well as people from whom they receive advice, practical and emergency help. A total list of the names generated by each respondent was compiled, representing the respondent’s total network size.

Network structure

Name-interpreting questions were used to acquire description of the named individuals and their relationship with respondents. Based on study respondents’ answers to questions about their relationship with these individuals, the number of individuals under each of the following relationship categories were identified: family members, friends, neighbors, and school and work individuals. Each relationship category was then divided by the total network size (e.g. number of family members divided by total network size). These calculations resulted in the percent composition of family members (family-network), friends (friend-network), neighbors (neighbor-network) and school and work individuals (school and work-network) that were in respondents’ networks. The name interpretation questions were also used to identify the number of individuals that respondents feel close to, live within 5-minutes to respondents, and of the same sex as respondents. Each of these were also divided by the total network size (e.g., number of individuals respondents feel close to divided by total network size). The calculations resulted in the percent composition of individuals in the respondents’ networks that respondents feel close to (network-respondent feel close to), live within 5-minutes to respondents (network-lives 5-minute to respondent) and of the same sex (network-same sex). These continuous network structure variables ranged from 0 to 1 and Pearson correlation between these variables ranged from r = 0.07 to r = 0.47.

Network support

Network support size was a count of the number of individuals in respondents’ networks that provided them with 4 different types of support: Social companion, Advice individual, Confidant, and Practical support individual. All variables ranged from 0-6, except social companion (0-9). Birthplace. Foreign-birth status was determined based on respondents’ answer about which country they were born in. Responses were dichotomized into US-born, and foreign-born.

Covariates

Participants reported health insurance status and self-rated health status (excellent/very good, good or fair/poor). Sociodemographic characteristics included age group, gender, education, race/ethnicity, employment and personal income. Foreign-born individuals reported total number of years lived in the US, age at migration and citizenship status.

Analysis

All analyses used post-stratification weights, based on separate calculations for each age cohort in order to estimate the population distribution of the Bay Area. Descriptive statistics were used to examine the weighted sample characteristics of all participants and by foreign-birth status. Test of difference between US-and foreign-born individuals were also conducted using t-tests and chi-square tests. Following previous research on networks and chronic diseases,^18,28 separate logistic regression models were developed to test the unadjusted association between each network characteristic and core sociodemographic variable (foreign-birth status, age group and gender) with the chronic diseases and condition. Subsequently, the adjusted association between network characteristics, foreign-birth status, gender and age group with each chronic condition indicator were tested in 3 separate models. The models adjusted for race/ethnicity, marital status, employment, personal income, self-rated health and insurances (Figure 1). The 3 models were run multiple times to separately test the interaction effects of each network characteristic by foreign-birth status, gender, and age group. There were no discernable patterns among the multiple interactions tested that required further stratification by the moderators or inclusion of interaction terms in the final models. The SAS® statistical software version 9.4 (Cary, NC) was used to conduct all analyses.

Figure 1.

Moderation tests of nativity, gender and age group on the association between network characteristics and chronic condition outcomes.

Results

The prevalence of chronic conditions were: 27% hypertension, 12% diabetes and 46% overweight/obese (Table 1). Approximately 15% of respondents were foreign-born. Overweight/obesity prevalence was significantly different between US-and foreign-born respondents (56% vs. 45%, p < 0.05). The majority were non-Hispanic whites and most foreign-born individuals were of Asian ethnicity (46%). As shown in Table 2, the mean percent composition of network members living within 5-minute to a respondent was 20% (95% CI: 0.19-0.22) and was associated with nativity, p < 0.001. The mean number of social companions was 5.39 (95% CI: 5.17-5.61), mean number of confidants was 2.87 (95% CI 2.72-3.02), and mean number of network members that provide practical support was 1.83 (95% CI: 1.69-1.97). These were also associated with nativity, p < 0.05.

Table 1.

Characteristics of Respondents and Their Associations With Foreign-Birth Status.

	Overall (N = 1153) n (%)^a	US-born (n = 975) n (%)	Foreign-born (n = 178) n (%)	P-value^b
Hypertension				0.87
Yes	269 (27.14)	238 (27.42)	32 (26.48)
No	884 (72.86)	737 (72.58)	146 (73.52)
Diabetes				0.30
Yes	115 (12.12)	99 (11.51)	17 (15.30)
No	1038 (87.88)	876 (88.49)	161 (84.70)
Weight				0.04
Overweight/obesity	610 (45.73)	473 (56.34)	70 (45.53)
Normal/Underweight	543 (54.27)	502 (43.66)	108 (54.47)
Age group				0.27
21-30	484 (40.55)	392 (41.63)	92 (36.02)
50-70	669 (59.44)	583 (58.37)	86 63.98)
Gender				0.56
Female	763 (52.59)	646 (51.99)	117 (55.11)
Male	390 (47.41)	329 (48.01)	61 (44.89)
Race/Ethnicity				<.0001
Non-Hispanic White	716 (48.18)	653 (54.06)	63 (23.39)
Black	59 (5.89)	58 (7.24)	1 (0.21)
Asian	163 (18.04)	89 (11.38)	74 (46.15)
Hispanic	121 (18.03)	98 (18.01)	23 (18.12)
Other	94 (9.86)	77 (9.32)	17 (12.13)
Relationship Status				0.01
Married	366 (44.84)	299 (42.03)	67 (56.68)
widowed/divorced/separated	250 (16.95)	225 (18.03)	25 (12.39)
Neve married	537 (38.22)	451 (39.94)	86 (30.94)
Employed				0.40
Yes	677 (55.87)	577 (56.72)	100 (52.25)
No	476 (44.13)	398 (43.28)	78 (47.75)
Income				0.76
< $35,000	514 (50.12)	431 (50.48)	83 (48.59)
$35,000 to < $75,000	322 (25.65)	276 (26.08)	46 (23.82)
$75,000 to < $159,000	216 (17.36)	185 (16.56)	31 (20.81)
≤$150,000	82 (6.86)	70 (6.88)	12 (6.79)
Insurance				0.04
Yes	1122 (97.48)	948 (97.10)	174 (99.10)
No	31 (2.52)	27 (2.90)	4 (0.90)
Overall Health				<.0001
Excellent/ Very good	702 (55.15)	604 (46.59)	98 (8.56)
Good	284 (27.15)	235 (20.59)	49 (6.55)
Fair/Pair	165 (17.70)	134 (13.65)	31 (4.05)
Age at migration
Preteen (under 13 years)	–	–	63 (30.53)	–
Teen (13, 17 years)	–	–	17 (11.10)	–
Adult (18 years +)	–	–	89 (58.37)	–
Years in the US
≤10 years	–	–	39 (21.19)	–
>10 years	–	–	169 (78.81)	–

UCnets, 2015.

^a Unweighted sample size and weighted percentage are presented. Weighted percentage may not add up to 100% because of rounding.

^b p < 0.05 is significant, based on chi-square tests of difference between US-and foreign-born.

^c Weight categories defined using calculated Body Mass Index (kg/m²): underweight/normal weight (< 25, or < 23 for Asians), Overweight/Obese (BMI ≥ 25, or ≥ 23 for Asians).

Table 2.

Social Network Characteristics and Their Associations With Foreign-Birth Status Among Young and Older Adults, UCnets, 2015.

	Overall mean (95% CI)^a	US-born mean (95% CI)	Foreign-born Mean (95% CI)	P value^b
Network Structure (proportion)
Family-network	0.40 (0.38, 0.42)	0.41 (0.39, 0.43)	0.38 (0.33, 0.43)	0.25
Friend-network	0.51 (0.49, 0.54)	0.51 (0.49, 0.54)	0.52 (0.47, 0.57)	0.95
Neighbor-network	0.05 (0.04, 0.05)	0.05 (0.04, 0.06)	0.04 (0.03, 0.06)	0.72
School & work-network	0.13 (0.11, 0.14)	0.13 (0.11, 0.14)	0.13 (0.10, 0.16)	0.78
Network-same-sex	0.60 (0.59, 0.62)	0.60 (0.58, 0.61)	0.64 (0.60, 0.68)	0.05
Network-respondent feel close to	0.46 (0.44, 0.48)	0.47 (0.44, 0.49)	0.43 (0.38, 0.48)	0.25
Network-lives 5-minute to respondent	0.20 (0.19, 0.22)	0.21 (0.20, 0.23)	0.15 (0.11, 0.18)	0.02
Network Support (count)
Social companion	5.39 (5.17, 5.61)	5.52 (5.28, 5.77)	4.84 (4.34, 5.34)	0.01
Advice individual	2.61 (2.48, 2.75)	2.59 (2.43, 2.74)	2.72 (2.41, 3.03)	0.45
Confidant	2.87 (2.72, 3.02)	2.98 (2.82, 3.14)	2.39 (2.03, 2.75)	0.003
Practical support individual	1.83 (1.69, 1.97)	1.90 (1.74, 2.06)	1.55 (1.26, 1.83)	0.03

^a CI: confidence interval 95% CI represents confidence interval of the mean percent composition for the network structure measures and mean count (number) for the social support measures. For network structure, proportion can be interpreted as a percentage, e.g. 0.40 = 40%.

^b p < 0.05 is significant, based on tested mean difference (t-test) between US-and foreign-born individuals were weighted.

Chronic Diseases

Increased odds of hypertension (OR = 5.35, 95% CI: 2.42-11.81) were found among respondents whose networks composed of a higher percentage of family members, relative to the networks of other respondents (Table 3). Conversely, those with networks composed of a higher percentage of friends (OR = 0.40, 95% CI: 0.20-0.80), school and work colleagues (OR = 0.19, 95% CI: 0.07, 0.54), and members of the same sex (OR = 0.25, 95% CI: 0.09-0.69) had lower odds of hypertension. Also, increased numbers of confidants (OR = 0.86, 95% CI: 0.77-0.95) and individuals that provide practical support (OR = 0.86, 95% CI: 0.76-0.98) in respondents’ networks were associated lower odds of hypertension. Females were less likely to be hypertensive compared to males (OR = 0.49, 95% CI: 0.34-0.71). Young adults were less likely to report hypertension compared with older adults (OR = 0.21, 95% CI: 0.13-0.36).

Table 3.

Unadjusted Individual Associations of Network Characteristic and Core Sociodemographic Predictor Variables With Chronic Disease Indicators Among Young and Older Adults, UCnets, 2015.

	Hypertension OR (95% CI)^a	Diabetes OR (95% CI)	Obesity OR (95% CI)
Network Structure
Family-network	5.35 (2.42, 11.81)^b	3.54 (1.36, 9.20)^b	5.31 (2.66, 10.60)^b
Friend-network	0.40 (0.20, 0.80)^b	0.49 (0.20, 1.23)	0.40 (0.23, 0.72)^b
Neighbor-network	1.82 (0.35, 9.40)	4.94 (0.82, 29.88)	1.60 (0.36, 7.02)
School & work-network	0.19 (0.07, 0.54)^b	0.13 (0.03, 0.64)^b	0.21 (0.09, 0.47)^b
Network-same-sex	0.25 (0.09, 0.69)^b	0.58 (0.15, 2.20)	0.96 (0.42, 2.23)
Network-respondent feel close to	1.08 (0.51, 2.29)	0.79 (0.29, 2.15)	1.39 (0.75, 2.60)
Network-lives 5-minute to respondent	0.70 (0.28, 1.75)	1.94 (0.65, 5.81)	1.63 (0.73, 3.61)
Network Support
Social companion	0.96 (0.90, 1.03)	0.94 (0.84, 1.04)	0.97 (0.91, 1.03)
Advice individual	0.92 (0.82, 1.03)	0.85 (0.74, 0.97)^b	0.88 (0.80, 0.96)^b
Confidant	0.86 (0.77, 0.95)^b	0.90 (0.78, 1.05)	0.70 (0.82, 0.97)^b
Practical support individual	0.86 (0.76, 0.98)^b	0.85 (0.71, 1.00)	0.98 (0.90, 1.07)
Individual Characteristics
Birthplace (Foreign vs. U.S.)	0.96 (0.58, 1.59)	1.41 (0.74, 2.70)	0.65 (0.42, 0.99)^b
Gender (Female vs. Male)	0.49 (0.34, 0.71)^b	1.02 (0.61, 1.71)	0.61 (0.44, 0.84)^b
Age group (21-30 vs. 50-70 years)	0.21 (0.13, 0.36)^b	0.13 (0.05, 0.32)^b	0.57 (0.41, 0.80)^b

Each cell represents a separate regression model. All analyses were weighted to account for sampling.

Network structure measures represent percent composition and network support measures represent counts.

^a OR, odds ratio; CI, confidence interval.

^b ORs (95% CI) represent significant findings at the p < 0.05 level.

Respondents with networks composed of a higher percentage of family members (OR = 3.54, 95% CI: 1.36-9.20) relative to other respondents were more likely to have diabetes. Diabetes was inversely associated with age group (OR = 0.13, 95% CI: 0.05-0.32). Overweight/obesity was positively associated with respondents whose networks composed of a higher percentage of family members (OR = 5.31, 95% CI: 2.66-10.60). Those with higher numbers of network members that provide advice (OR = 0.88, 95% CI: 0.80-0.96) and confidants (OR = 0.70, 95% CI: 0.82-0.97) had lower odds of overweight/obesity. Additionally, foreign-born individuals (OR = 0.65, 95% CI: 0.42-0.99), females (OR = 0.61, 95% CI: 0.44-0.84), and young adults (OR = 0.57, 95% CI: 0.41-0.80) were less likely to be overweight/obese.

The relationship between network characteristics and foreign-birth status, gender and age group with each chronic disease was examined, adjusting for sociodemographic factors, (Table 4). Respondents with networks composed of higher percentages of family members (AOR = 3.54, 95% CI: 1.09-11.48) had increased odds of hypertension. Females (AOR = 0.41, 95% CI: 0.25-0.68) and young adults (AOR = 0.23, 95% CI: 0.12-0.44) were less likely to report hypertension in adjusted models. The only network composition variable associated with diabetes was networks composed of a higher number of members living within 5-minute to respondents (AOR = 5.13, 95% CI: 1.04-25.21). Relative to other respondents, individuals with networks composed of higher percentages of family members (AOR = 4.16, 95% CI: 1.61-10.76) and members of the same sex (AOR = 3.41, 95% CI: 1.25-9.35) were more likely to be obese. Respondents with a higher number of advice individuals (AOR = 0.88, 95% CI: 0.77-0.99), foreign-born respondents (AOR = 0.54, 95% CI: 0.32-0.92) and females (AOR = 0.52, 95% CI: 0.35-0.77) were less likely to report obesity.

Table 4.

Adjusted Associations of Network Characteristic and Core Sociodemographic Predictor Variables With Chronic Disease Indicators Among Young and Older Adults, UCnets, 2015.

	Hypertension AOR (95% CI)^a	Diabetes AOR (95% CI)	Obesity AOR (95% CI)
Network Structure
Family-network	3.54 (1.09, 11.48)^b	2.04 (0.46, 9.11)	4.16 (1.61, 10.76)^b
Friend-network	1.39 (0.57, 3.37)	1.55 (0.48, 4.97)	0.82 (0.37, 1.83)
Neighbor-network	2.82 (0.34, 23.52)	1.50 (0.11, 19.99	1.29 (0.21, 7.780)
School & work-network	1.27 (0.38, 4.23)	0.36 (0.06, 2.11)	0.38 (0.14, 1.02)
Network-same-sex	0.55 (0.16, 1.90)	0.53 (0.11, 2.52)	3.41 (1.25, 9.35)^b
Network-respondent feel close to	1.16 (0.50, 2.68)	0.75 (0.23, 2.40)	1.22 (0.59, 2.50)
Network-lives 5-minute to respondent	0.55 (0.17, 1.75)	5.13 (1.04, 25.21)^b	1.64 (0.66, 4.04)
Network Support
Social companion	1.07 (0.98 1.18)	1.00 (0.88, 1.13)	1.03 (0.94, 1.11)
Advice individual	1.07 (0.90, 1.28)	0.91 (0.75, 1.11)	0.88 (0.77, 0.99)^b
Confidant	0.86 (0.73, 1.02)	1.07 (0.87, 1.31)	0.96 (0.84, 1.08)
Practical support individual	0.97 (0.83, 1.13)	0.92 (0.75, 1.14)	1.06 (0.95, 1.19)
Individual Characteristics
Birthplace (Foreign vs. U.S.)	0.69 (0.37, 1.28)	0.74 (0.29, 1.88)	0.54 (0.32, 0.92)^b
Gender (Female vs. Male)	0.41 (0.25, 0.68)^b	1.04 (0.55, 1.96)	0.52 (0.35, 0.77)^b
Age group (21-30 vs. 50-70 years)	0.23 (0.12, 0.44)^b	0.13 (0.05, 0.35)^b	0.78 (0.48, 1.29)

Each column represents a separate model. The models adjusted for race/ethnicity, marital status, employment status, personal income, self-rated health and insurance coverage. All analyses were weighted to account for sampling.

Network structure measures represent percent composition and network support measures represent counts.

^a AOR, adjusted odds ratio; CI, confidence interval.

^b Statistically significant findings at the p < 0.05 level.

Discussion

Principal Findings

In the present study, network structure and support were associated with chronic conditions among US and foreign-born adults. In terms of network structure, those with a higher percentage of family members were more likely to report overweight/obesity and hypertension. Those with higher composition of same-sex individuals were also more likely to be overweight/obese. Respondents with networks composed of a higher percentage of individuals living within 5-minutes to them were more likely to report diabetes. Network quality in the form of having higher numbers of advice individuals was associated with obesity. While foreign-birth status, gender and age group did not modify these relationships, being foreign-born and female were independently associated with overweight/obesity. These findings suggest that network characteristics may have both negative and positive impact on health outcomes among our study participants, which we next discuss across the 3 health outcomes of hypertension, diabetes, and overweight/obesity.

Findings in Context

The increased likelihood of hypertension among respondents with higher family-networks may stem from the constant stress of having negative interactions with family members. Unlike other network relationships, it is likely difficult to avoid familial relationship, which may not always be positive. Negative aspects of network interactions and social support in the form of excessive demands or expectations and criticisms from network members, or lower expectation for support from networks have been linked to hypertension.^29,30 Recent work found that negative social interaction with family, children, and partners was associated with increased incidence of hypertension among female older adults, but not among males.³⁰ In our sample, the positive relationship between a higher composition of family-and same sex individual-network with overweight/obesity may be because the network members were also overweight/obese and or could not provide support for obesity prevention such as healthy diet and exercise information and resources. Earlier evidence suggests that obesity clusters in social networks, based on characteristics of network members including friendship, homophily of gender, and peer groups.²¹ Resources provided by network members in the form of social support might impact obesity, by influencing norms and behaviors that may or may not be favorable to the health outcome.³¹

The positive association between network proximity (i.e., higher network composition of individuals living within 5-minutes to respondent) and diabetes in our sample contradicts the limited evidence of the positive influence of network members in reducing the risk of developing diabetes.^18,32 For example, previous research suggests that better structural and social support is associated with as much as 15% decrease in prevalence of diabetes.¹⁸ Conversely, in the present study network proximity may be important because the network members engage in behaviors that increase their risk of diabetes. Therefore, the study participants also engaged in similar behaviors, since individuals tend to follow the behavioral norms of their networks.³³ Besides the need for further examination of additional network characteristics, e.g. network diversity, associated with diabetes, research is needed to explore the process by which these network characteristics influence chronic disease outcomes.

In terms of the positive impact of network characteristics in our sample, the finding that having more network members that provide advice lowered the odds of overweight/obesity suggests that the quality of support available to respondents may be independent of network structure, such that advice individuals may serve as a resource to mitigate the relationship between network structure and overweight/obesity. Notably, other types of support like having more confidants and practical support individuals lowered the odds of hypertension, but these relationships were mitigated after adjusting for other network characteristics. Previous evidence indicate that the presence of family members, especially among older adults, increases availability of social support which has been linked with reduced risk of high blood pressure.³⁴ Also, greater satisfaction and cohesion with a spouse or partner is associated with decreased blood pressure.³⁵ In the present study, it is likely that network members provide practical support and or serve as confidants by providing information and resources that promote health protective behaviors, e.g. healthy eating and physical activities, among respondents. These behaviors in turn reduce their risk of chronic conditions like hypertension.

As indicated earlier, in our sample foreign birth was independently associated with overweight/obesity in that foreign-born respondents had lower odds of overweight/obesity after adjusting for covariates including network characteristics. This finding contributes to previous studies that reveal that being foreign-born was associated with overweight/obesity.^3,36 Foreign-born respondents were likely to be highly acculturated in this study; over two-thirds were either citizens or had lived in the US for at least 10 years. Yet, due to the small sample size of foreign-born respondents, the impacts of these factors could not be examined. Future research with a more robust sample size of foreign-born individuals is needed to examine the effect of acculturation on the relationship between network characteristics and overweight/obesity among foreign-born individuals.

Strengths and Limitations

A strength of the present study is that the UCNets data contains an extensive egocentric network structure and support measures. Limitations of this study include the data collection approach, which was based on different sampling methods; probability sampling for the older adults and probability and convenience sampling for the young adults. It is possible that this may have caused nonrandom differences across the sample. Since the data is self-reported, it is subject to both recall and social desirability bias. The UCNets data we used are cross-sectional, thus causal inference may not be drawn from significant findings. Also, the findings are not generalizable to adults outside of the counties studied that are demographically comparable to study respondents. The sample size of the foreign-born population was too small for further examination of immigrant-specific factors such as length of residency, age at migration and citizenship.

Conclusion

This study suggests that having a higher composition of family-network in one’s network was significantly associated with hypertension and overweigh/obesity. However, a higher number of advice individuals was associated with reduced risk of overweight/obesity, thereby highlighting the potential effect of the qualitative aspect of social relationships on health outcomes. For each chronic disease or condition, certain network characteristics were related to either increased or reduced odds of the health outcomes, which suggests that network members play critical roles in supporting behaviors that are protective against chronic disease. Future work including a large sample of foreign-born individuals should examine additional network characteristics that may differ between US-and foreign-born respondents, as it may help to further determine which network characteristics could serve as modifiable targets of intervention to prevent and reduce chronic diseases. Moreover, further examination of the mechanism by which network structure and support measures were significantly associated with hypertension and overweight/obesity should be addressed in future studies.

So What? Implications for Health Promotion Practitioners and Research

What is already known on this topic?

The relationship between foreign-birth status and the experience of hypertension, diabetes and obesity in adults have previously been established. Social networks help explain these relationships.

What does this article add?

Higher percent composition of family members (AOR = 4.16, CI: 1.61-10.76) and same-sex individuals (AOR = 3.41, CI: 1.25-9.35) in respondents’ network were associated with overweight/obesity. Higher composition of family members (AOR = 3.54, CI: 1.09-11.48) was also associated with hypertension. Respondents with a higher number of advice individuals (AOR = 0.88, CI: 0.77-0.99), foreign-born respondents and females were less likely to report obesity. Diabetes was associated with higher composition of individuals living within 5-minutes to respondent (AOR = 5.13, CI: 1.04-25.21).

What are the implications for health promotion practice and research?

The findings can inform efforts to develop strategies to reduce the incidence and prevalence of chronic diseases and conditions by considering ways to incorporate network support in health promotion interventions.

Footnotes

Authors’ Note

This study was reviewed and approved by the University of East Carolina Institutional Review Board (UMCIRB 19-002563).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data for this work was supported by the National Institute on Aging (Grant No. R01AG041955). Principal Investigator: Claude Fischer.

ORCID iDs

Leslie E. Cofie, PhD, MPH, MA

Joseph G. L. Lee, PhD, MPH

References

Buttorff

Ruder

Bauman

. Multiple Chronic Conditions in the United States. RAND Corporation; 2017.

Jensen

Ryan

Apovian

, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Obesity Society. Circulation. 2014;129(25 suppl 2):S102–S138. doi:10.1161/01.cir.0000437739.71477.ee

Tsujimoto

Kajio

Sugiyama

Obesity, diabetes, and length of time in the United States: analysis of National Health and Nutrition Examination Survey, 1999 to 2012. Medicine (Baltimore). 2016;95(35):e4578. doi:10.1097/md.0000000000004578

Nwankwo

Yoon

Burt

Hypertension among adults in the United States: National Health and Nutrition Examination Survey, 2011-2012. NCHS Data Brief. 2013;(133):1–8.

Fang

Ayala

Loustalot

. Association of birthplace and self-reported hypertension by racial/ethnic groups among US adults—National Health Interview Survey, 2006-2010. J Hypertens. 2012;30(12):2285–2892. doi:10.1097/HJH.0b013e3283599b9a

Ogden

Carroll

Fryar

Flegal

. Prevalence of obesity among adults and youth: United States, 2011-2014. NCHS Data Brief. 2015;(219):1–8.

Singh

Rodriguez-Lainz

Kogan

. Immigrant health inequalities in the United States: use of eight major national data systems. ScientificWorldJournal. 2013;2013:512313. doi:10.1155/2013/512313

Menjivar

. Immigrant kinship networks: Vietnamese, Salvadorians and Mexicans in comparative perspective. JCFS. 1997;28(1):1–24.

Berkman

Glass

Brissette

Seeman

. From social integration to health: Durkheim in the new millennium. Soc Sci Med. 2000;51(6):843–857.

10.

Seeman

TE.

Social ties and health: the benefits of social integration. Ann Epidemiol. 1996;6(5):442–451. doi:10.1016/s1047-2797(96)00095-6

11.

Landale

Oropesa

. Migration, social support and perinatal health: an origin-destination analysis of Puerto Rican women. J Health Soc Behav. 2001;42(2):166–183.

12.

Vega

Sribney

Miskimen

Escobar

Aguilar-Gaxiola

. Putative psychotic symptoms in the Mexican American population: prevalence and co-occurrence with psychiatric disorders. J Nerv Ment Dis. 2006;194(7):471–477. doi:10.1097/01.nmd.0000228500.01915.ae

13.

Harley

Eskenazi

. Time in the United States, social support and health behaviors during pregnancy among women of Mexican descent. Soc Sci Med. 2006;62(12):3048–3061. doi:10.1016/j.socscimed.2005.11.036

14.

Bland

Krogh

Winkelstein

Trevisan

. Social network and blood pressure: a population study. Psychosom Med. 1991;53(6):598–607.

15.

Horwitz

Prados-Torres

Singer

Bruce

. The influence of psychological and social factors on accuracy of self-reported blood pressure. J Clin Epidemiol. 1997;50(4):411–418.

16.

Vogt

Mullooly

Ernst

Pope

Hollis

. Social networks as predictors of ischemic heart disease, cancer, stroke and hypertension: incidence, survival and mortality. J Clin Epidemiol. 1992;45(6):659–666.

17.

Holt-Lunstad

Uchino

Smith

Olson-Cerny

Nealey-Moore

. Social relationships and ambulatory blood pressure: structural and qualitative predictors of cardiovascular function during everyday social interactions. Health Psychol. 2003;22(4):388–397. doi:10.1037/0278-6133.22.4.388

18.

Gallo

Fortmann

McCurley

, et al. Associations of structural and functional social support with diabetes prevalence in U.S. Hispanics/Latinos: results from the HCHS/SOL Sociocultural Ancillary Study. J Behav Med. 2015;38(1):160–170. doi:10.1007/s10865-014-9588-z

19.

Christakis

Fowler

. The spread of obesity in a large social network over 32 years. N Engl J Med. 2007;357(4):370–379. doi:10.1056/NEJMsa066082

20.

Bahr

Browning

Wyatt

Hill

. Exploiting social networks to mitigate the obesity epidemic. Obesity (Silver Spring). 2009;17(4):723–728. doi:10.1038/oby.2008.615

21.

Powell

Wilcox

Clonan

, et al. The role of social networks in the development of overweight and obesity among adults: a scoping review. BMC Public Health. 2015;15:996. doi:10.1186/s12889-015-2314-0

22.

Menjivar

. Fragmented Ties: Salvadoran Immigrant Networks in America. University of California Press; 2000: xviii, 301 p.

23.

Offer

Fischer

. Difficult people: who is perceived to be demanding in personal networks and why are they there? Am Sociol Rev. 2018;83(1):111–142. doi:10.1177/0003122417737951

24.

Lawton

Wilson

. Comparing Facebook and address-based sampling to recruit young adults. 2018:

25.

World Health Organization Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157–163. doi:10.1016/s0140-6736(03)15268-3

26.

Fischer

. To Dwell Among Friends: Personal Networks in Town and City. University of Chicago Press; 1982: xii, 451 p.

27.

Marsden

. Network data and measurement. Ann Rev Soc. 1990;16(1):435–463. doi:10.1146/annurev.so.16.080190.002251

28.

Hernandez

Carnethon

Giachello

, et al. Structural social support and cardiovascular disease risk factors in Hispanic/Latino adults with diabetes: results from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Ethn Health. 2018;23(7):737–751.

29.

Wickrama

KAS

Lorenz

Lora Ebert

, et al. Family influence on physical health during the middle years: the case of onset of hypertension. J Mar Fam. 2001;63(2):527–539.

30.

Sneed

Cohen

. Negative social interactions and incident hypertension among older adults. Health Psychol. 2014;33(6):554–565. doi:10.1037/hea0000057

31.

Holtgrave

Crosby

. Is social capital a protective factor against obesity and diabetes? Findings from an exploratory study. Ann Epidemiol. 2006;16(5):406–408. doi:10.1016/j.annepidem.2005.04.017

32.

Kumari

Head

Marmot

. Prospective study of social and other risk factors for incidence of type 2 diabetes in the Whitehall II study. Arch Intern Med. 2004;164(17):1873–1880. doi:10.1001/archinte.164.17.1873

33.

Cohen

. Social relationships and health. Am Psychol. 2004;59(8):676–684. doi:10.1037/0003-066x.59.8.676

34.

Strogatz

Croft

James

, et al. Social support, stress, and blood pressure in black adults. Epidemiology. 1997;8(5):482–487.

35.

Gump

Polk

Kamarck

Shiffman

. Partner interactions are associated with reduced blood pressure in the natural environment: ambulatory monitoring evidence from a healthy, multiethnic adult sample. Psychosom Med. 2001;63(3):423–433.

36.

Barrington

Baquero

Borrell

Crawford

. Racial/ethnic disparities in obesity among US-born and foreign-born adults by sex and education. Obesity (Silver Spring). 2010;18(2):422–424. doi:10.1038/oby.2009.212