Internal and Emergency Medicine https://doi.org/10.1007/s11739-019-02075-7 IM - ORIGINAL Is social support associated with hypertension control among Ghanaian migrants in Europe and non‑migrants in Ghana? The RODAM study Gertrude Nsorma Nyaaba1,2  · Karien Stronks1 · Karlijn Meeks1 · Erik Beune1 · Ellis Owusu‑Dabo3 · Juliet Addo4 · Ama de‑Graft Aikins5 · Frank Mockenhaupt6 · Silver Bahendeka7 · Kerstin Klipstein‑Grobusch8,9 · Liam Smeeth4 · Charles Agyemang1 Received: 31 August 2018 / Accepted: 14 March 2019 © The Author(s) 2019 Abstract Hypertension (HTN) control is crucial in preventing HTN-related complications such as stroke and coronary heart disease. Yet, HTN control remains suboptimal particularly among sub-Saharan African (SSA) populations partly due to poor self- management. Self-management of HTN is influenced by social support, but the evidence on the role of social support on HTN control particularly among SSA populations is limited. This study assessed the association between multiple proxies for social support and HTN control among Ghanaians resident in Ghana and Europe. The Research on Obesity and Diabetes among African Migrants (RODAM) study participants with HTN and who self-reported HTN (n = 1327) were included in this analysis. Logistic regression was used to assess the association between proxies of social support and HTN control (SBP < 140 mmHg and DBP < 90 mmHg) with adjustments for age and socioeconomic status (SES). Among Ghanaian males in both Europe and Ghana, cohabiting with more than two persons was associated with increased odds of having HTN controlled. Male hypertensive patients cohabiting with ≥ 5 persons had the highest odds of having HTN controlled after adjustment for age and SES (OR 0.30; 95% CI 0.16–0.57; 0.60; 0.34–1.04, respectively). This association was not observed among females. Relationship status, frequency of religious activity attendance and satisfaction with social support did not show any significant association with HTN control. Our study shows that cohabitation is significantly associated with HTN control but in males only. The other proxies for social support appeared not to be associated with HTN control. Involving persons living with Ghanaian men with HTN in the treatment process may help to improve adherence to HTN treatment. Further research is needed to explore in-depth, how these social support proxies could contribute to improved HTN control among SSA populations. Keywords Hypertension control · Social support · Sub-Saharan Africa · Migrant health · Ethnic minority groups · Ghana · Europe Introduction Migrants (RODAM) study show a 56.8% and a 34.4% HTN prevalence among Ghanaian migrants living in Europe and Evidence shows a high prevalence of hypertension (HTN) Ghanaians in Ghana with suboptimal HTN control rates of but suboptimal control rates for HTN among Sub-Saharan 20% and 52.5% among Ghanaian migrant men in Europe African (SSA) origin groups in Europe [1–6]. Recent stud- and non-migrant rural Ghanaian men, respectively [5, 7]. ies on the Research on Obesity and Diabetes among African Adequate blood pressure (BP) control (i.e. systolic and dias- tolic BP (SBP/DBP) less than 140/90 mmHg [8]) consider- ably lessens the rate of cardiovascular events such as stroke, renal insufficiency, coronary heart disease, peripheral vas- Gertrude Nsorma Nyaaba cular disease, congestive heart failure and premature deaths * g.n.nyaaba@amc.uva.nl [3, 9, 10]. Extended author information available on the last page of the article Vol.:(012 3456789) Internal and Emergency Medicine The reasons for the poor control of HTN among SSA ori- homogenous SSA populations living in different loca- gin groups remain unclear although untreated severe HTN, tions in Europe and in rural and urban Ghana [23]. Using differences in the prescription practices, fear of side effects data from the RODAM study, we assessed the association and poor patient self-management (patient’s daily activities between social support and HTN control among Ghanaians surrounding adherence to medication and behavioural rec- resident in Europe and Ghana and whether the associations ommendations) are thought to be potential underlying fac- varied in the different contexts. We hypothesised that living tors [5, 11]. While it is generally accepted that behavioural with more people, been in a relationship, regular religious factors and low socioeconomic status (SES) contribute to activity attendance and high satisfaction with social sup- individual’s risks for HTN [12, 13], the extent to which port received would positively influence HTN control among social support mechanisms could influence HTN control SSA migrants and non-migrants. has been relatively unexplored especially among SSA ori- gin populations. Evidence shows that social support contrib- utes significantly to improving health outcomes [14–16]. A Methods review on the influence of social support on chronic illness self-management shows that social support has a protective Data source effect on self-management particularly in supporting dietary changes [17]. The bulk of evidence on social support has Data from the RODAM study, which was carried out from been on the characteristics of family support such as per- 2012 to 2015, were used for this analysis. The study received ceived family support [17, 18], family solidity, articulateness ethical approval in all participating countries and included and struggles [17, 19], which enable adherence to treatment Ghanaians aged 25–70 years residing in three European for chronic conditions. Data on other social support indica- countries (The Netherlands, Germany and UK) and Ghana- tors such as cohabitation, relationships, patient satisfaction ians living in rural and urban Ghana; and all participants with social support and social networks such as religious gave written consent before data collection. The rationale, attendance are lacking. This suggests the need for further conceptual framework, methodology, response rates and research on the role of other key indicators of social support other details have been published elsewhere [5, 23]. on chronic illness self-management. Within African societies where members live together with shared responsibilities and possessions, studies show Description of measurements that during times of illness, members play a “brokerage function” between patients and healers and are regarded as The RODAM study used a structured questionnaire to col- the “therapy managing group” [20]. A recent study on HTN lect information on sociodemographic, medical history, self-management in Nigeria found religion and marital status treatment and lifestyle, and psychosocial stress. Validated as significant determinants of social support, which contrib- devices were used to conduct physical examination accord- uted to adherence to HTN treatment [16]. Another study on ing to standardised operational procedures across all study community perceptions of HTN in rural Ghana found that sites. A validated semi-automated device (the Microlife social support could contribute to improvement in adherence WatchBP home Widnau, Switzerland) with appropriate to treatment particularly for adherence to dietary regimes cuffs was used to measure BP three times in a sitting posi- and HTN medication [21]. Evidence indicates that social tion after at least, a 5-min rest. HTN was defined as SBP support is one of the main determinants of adherence to ≥ 140 mmHg or DBP ≥ 90 mmHg and or being on antihy- HTN treatment among SSA migrants living in high income pertensive medication. HTN awareness was defined as the countries (HICs) [22]. While these studies [16, 21, 22] have proportion of individuals with HTN who reported HTN prior provided evidence on the role of social support to adherence to BP measurements by the study team [5]. HTN treatment to HTN treatment, the specific indicators of social support, was defined as the proportion of persons with HTN who which could influence HTN control among SSA migrants, had been prescribed antihypertensive medication for high have not yet been clearly elucidated. Identifying specific BP management, while HTN was defined as controlled if indicators of social support, which could enhance patient a participant was on antihypertensive medication with BP self-management of HTN is key to identifying pathways for < 140/90 mmHg [5]. implementing targeted interventions to improve HTN control among these populations. Study design and measurements The RODAM study provides an opportunity to assess the association between social support and HTN control as it We selected Ghanaian participants from the RODAM study used a highly standardised approach to collect data among with HTN. Because social support requires a degree of 1 3 Internal and Emergency Medicine patient awareness of HTN status, we excluded participants the responses, we categorised responses into low, medium who were unaware of their HTN status. We therefore con- and high levels of satisfaction with social support received. ducted our analysis on 1327 participants who were hyperten- sive and were aware of their HTN status (Fig. 1). Data analysis A person’s social network often evolves social contacts [24] around the place of abode, work, relationship and reli- Characteristics of study participants were expressed as gion. Four proxies for social support were used in this anal- means and standard deviation (SD) for continuous vari- ysis: number of cohabitants, relationship status, religious ables and percentages for categorical variables. Logistic attendance frequency and satisfaction with social support regression was used to assess the association between the received. Participants recorded the number of persons living four proxies for social support and HTN control. Logistic with them in their household (cohabitants) and responded regression models were adjusted for age, sex and SES. to their relationship status. Based on the distribution of the Level of education, occupation, income or a composite of relationship status responses, participant responses were these factors are most commonly used to determine SES re-categorised into a binary variable (yes/no) where “yes” [25]. For this study, SES refers to the level of education represents married/registered partnership and cohabiting, and employment status of a participant as they are key while “no” represents unmarried, divorced or separated and socially derived economic factors that influence the status widow/widower. In addition, frequency of religious activity held by participants within their societies. Based on the attendance was categorised into two groups (once a week distribution of the employment status responses, partici- or more and less than once a week) based on the distribu- pants were categorised into employed and unemployed tion of participant responses. Furthermore, because we where employed referred to participants who were actively were interested in patient satisfaction with social support engaged in income generating activities. We found signifi- received, the Social Support Scale for Satisfaction with sup- cant statistical interactions between sex and some of the portive transactions (SSQS) was used to measure five items proxies for social support in relation to HTN control and (warmth, friendliness, willingness to lend a friendly ear, therefore stratified the analyses by sex. All analyses were are not problem-oriented and affection of people in the par- performed using IBM SPSS Statistics 24. ticipant’s environment) on a 1–4 point Likert scale. Answer categories were ordinal and included much less than I like (1), less than I like (2), just as much as I like (3) and more than I like (4). Higher scores indicated more satisfaction Results with social support received. Based on the distribution of Sociodemographic characteristics Table 1 presents a detailed description of study partici- 5898 RODAM study participants pants’ sociodemographic characteristics. The mean age of all participants was 53.2 (± 9.0), and the average num- ber of years that migrants had lived in Europe was 21.2 Excluded 3471 participants with SBP (± 9.5). HTN was uncontrolled in 68% of all participants <140mmHg or DBP <90mmHg and/or most of whom had lower vocational or secondary educa- not on antihypertensive medication tion (35.8%). While 26.9% were non-migrant Ghanaians, 73.1% were migrant Ghanaians living in Europe. Among 2427 participants with SBP at least men, HTN was uncontrolled in 41.6% of study partici- 140mmHg or DBP <90mmHg and/or on pants. Of all the participants, 36% were fully employed, antihypertensive medication 73.7% were using antihypertensive medication and males accounted for 36.3% of the total population. 3.8% of migrants reported current smoking compared with 0.8% Excluded 861 participants for not been aware of hypertensive status and 239 for of non-migrants. 61.7% of non-migrants and 29.8% of missing data migrants reported being on a special diet because of HTN. More than half of non-migrant participants (55.2%) and less than half of migrants (40.6%) were married or were 1327 participants included in the analysis in registered partnerships. The average number of cohabit- ants was 5.2 (± 2.6) for non-migrants and 3.3 (± 1.6) for migrants. The majority of the participants attended reli- Fig. 1 Sample selection gious activities at least once a week (85.4%), and 58.7% 1 3 Internal and Emergency Medicine Table 1 Characteristics of participants selected for analysis (n = 1327) All Ghana Europe Total Hyper- Hypertension Total Hyper- Hypertension Total Hyper- Hypertension tension uncontrolled tension uncontrolled tension uncontrolled controlled controlled controlled n = 1327 n = 424 n = 903 357 (26.9) 105 (29.4) 252 (70.6) 970 (73.1) 319 (32.9) 651 (67.1) Demographics Age, years (SD) 53.1 (9.0) 53.9 (8.3) 52.7 (9.3) 55.2 (10.5) 55.8 (8.9) 54.9 (11.1) 52.3 (8.3) 53.3 (7.9) 51.9 (8.4) Site Ghana 357 (26.9) 105 (24.8) 252(27.9) 357 (100) 105 (29.4) 252 (70.6) NA NA NA  Rural Ghana 235 (65.8) 35 (33.3) 87 (34.5) 235 (65.8) 35 (33.3) 87 (34.5) NA NA NA  Urban Ghana 122 (34.2) 70 (66.7) 165 (65.5) 122 (34.2) 70 (66.7) 165 (65.5) NA NA NA Europe 970 (73.1) 319(75.2) 651(72.1) NA NA NA 970 (100) 319 (32.9) 651 (67.1)  Amsterdam 468 (48.2) 168 (52.7) 300 (46.1) NA NA NA 468 (48.2) 168 (52.7) 300 (46.1)  Berlin 194 (20.0) 47 (14.7) 147 (22.5) NA NA NA 194 (20.0) 47 (14.7) 147 (22.5)  London 308 (31.8) 104 (32.6) 204 (31.3) NA NA NA 308 (31.8) 104 (32.6) 204 (31.3) Sex Male 482 (36.3) 106 (25.0) 376 (41.6) 88 (24.6) 21 (20.0) 67 (26.6) 394 (40.6) 85 (26.6) 309 (47.5) Female 845 (63.7) 318 (37.6) 527 (62.3) 269 (75.4) 84 (31.2) 185 (68.8) 576 (59.4) 234 (40.6% 342 (59.4) Length of stay in Europe (years) Mean (SD) NA NA NA NA NA NA 21.2 (9.5) 21.4 (9.3) 21.0 (9.5) 1–18 NA NA NA NA NA NA 338 (36.7) 111 (34.8) 227 (34.9) 19–25 NA NA NA NA NA NA 319 (34.6) 113 (35.4) 206 (31.6) 26–64 NA NA NA NA NA NA 264 (28.7) 83 (26.0) 181 (27.8) Education completed Elementary or less 436 (33.3) 133 (31.8) 303 (34.0) 200(56.0) 52 (49.5) 148 (58.7) 236 (24.8) 81 (25.9) 155 (24.3) Lower vocation or 481 (36.7) 162 (38.8) 319 (35.8) 115 (32.2) 43 (41.0) 72 (28.6) 366 (38.4) 119 (38.0) 247 (38.7) lower secondary Intermediate voca- 269 (20.6) 93 (22.2) 176 (19.8) 31 (3.1) 8 (7.6) 23 (9.1) 238 (25.0) 85 (27.2) 153 (23.9) tional or interme- diate secondary Higher vocation or 123 (9.4) 30 (7.2) 93 (10.4) 11 (3.1) 2 (1.9) 9 (3.6) 112 (11.8) 28 (8.9) 84 (13.1) university Employment status Employed full-time 470 (36.0) 147 (35.1) 323 (36.4) 73 (20.4) 22 (21.0) 51 (20.2) 397 (41.8) 125 (39.8) 272 (42.8) Employed part- 388 (29.7) 119 (28.4) 269 (30.3) 204 (57.1) 55 (52.4) 149 (59.1) 184 (19.3) 64 (20.4) 120 (18.9) time Retired 88 (6.7) 33 (7.9) 55 (6.2) 22 (6.2) 7 (6.7) 15 (6.0) 66 (6.9) 26 (8.3) 40 (6.2) Unemployed 75 (5.7) 20 (4.8) 55 (6.2) 4 (1.1) 2 (1.9) 2 (0.8) 71 (7.5) 18 (5.7) 53 (8.3) Unable to work 131 (10.0) 47 (11.2) 84 (9.5) 49 (13.7) 19 (18.1) 30 (11.9) 82 (8.6) 28 (8.9) 54 (8.5) On social benefits 118 (9.0) 43 (10.3) 75 (8.5) 2 (0.6) – 2 (0.8) 116 (12.2) 43 (13.6) 73 (11.5) Full-time home- 30 (2.3) 8 (1.9) 22 (2.5) 3 (0.8) – 3 (1.2) 27 (2.8) 8 (2.5) 19 (2.9) maker Student 7 (0.5) 2 (0.5) 5 (0.6) – – – 7 (0.7) 2 (0.6) 5 (0.8) Hypertension management characteristics Use of antihyper- 977 (73.6) 424 (100) 553 (61.2) 209 (58.5) 105 (100) 104 (41.3) 768 (79.2) 319 (100) 449 (69.0) tensive medica- tion Smoking, % Yes 40 (3.0) 8 (1.9) 32 (3.6) 3 (0.8) – 3 (1.2) 37 (3.8) 8 (2.5) 29 (4.5) Special diet for 508 (40.4) 182 (45.3) 326 (38.2) 219 (61.7) 75 (72.1) 144 (57.4) 289 (29.8) 107 (35.9) 182 (30.2) HTN, % Yes Alcohol, % No 818 (61.6) 270 (63.7) 548 (60.7) 251 (70.3) 77 (73.3) 174 (69.0) 567 (58.5) 193 (60.5) 374 (57.5) 1 3 Internal and Emergency Medicine Table 1 (continued) All Ghana Europe Total Hyper- Hypertension Total Hyper- Hypertension Total Hyper- Hypertension tension uncontrolled tension uncontrolled tension uncontrolled controlled controlled controlled n = 1327 n = 424 n = 903 357 (26.9) 105 (29.4) 252 (70.6) 970 (73.1) 319 (32.9) 651 (67.1) Physical activity Low 420 (37.3) 132 (31.1) 288 (27.8) 146 (41.0) 45 (42.9) 101 (40.2) 274 (35.5) 87 (33.3) 187 (36.7) Moderate 214 (19.0) 69 (16.3) 145 (19.1) 53 (14.9) 14 (13.3) 39 (15.5) 161 (20.9) 55 (21.1) 106 (20.8) High 493 (37.2) 165 (38.9) 328 (42.1) 157 (44.0) 46 (43.8) 111 (44.2) 336 (43.6) 119 (45.6) 217 (42.5) Types of cohabitants Living with partner 631 (60.5) 201 (59.0) 430 (60.8) 204 (60.0) 62 (62.6) 142 (58.9) 427 (60.7) 139 (58.6) 288 (61.8) Living with 184 (19.7) 61 (19.5) 123 (18.3) 77 (22.8) 23 (23.7) 54 (22.5) 107 (16.5) 38 (17.6) 69 (16.0) children 3 and younger Living with chil- 781 (75.6) 253 (76.2) 528 (75.3) 309 (91.4) 90 (90.9) 219 (91.6) 472 (67.9) 163 (70.0) 309 (66.9) dren 4 years and older Living with other 186 (18.6) 53 (16.8) 133 (19.4) 54 (15.9) 19 (19.4) 35 (14.5) 132 (19.9) 34 (15.7) 98 (22.0) adults Social support-related characteristics Number of cohabitants Mean (SD) 3.9 (2.1) 3.9 (2.1) 3.9 (2.1) 5.2 (2.6) 5.3 (2.6) 5.2 (2.5) 3.3 (1..6) 3.4 (1.5) 3.3 (1.7) 1–2 cohabitants 369 (30.3) 111 (28.7) 258 (31.0) 56 (15.7) 15 (14.3) 41 (16.3) 313 (36.3) 96 (34.0) 217 (37.4) 3–4 cohabitants 443 (36.3) 147 (38.0) 296 (35.6) 90 (25.2) 23 (21.9) 67 (26.6) 353 (41.0) 124 (44.0) 229 (39.5) 5 or more cohabit- 407 (33.4) 129 (33.3) 278 (33.4) 211 (59.1) 67 (63.8) 144 (57.1) 196 (22.7) 62 (22.0) 134 (23.1) ants Frequency of religious attendance Once a week or 910 (85.4) 302(86.3) 608 (84.9) 261 (95.3) 74 (96.1) 187 (94.9) 649 (81.9) 228 (83.5) 421 (81.1) more Once every 2 51 (4.8) 18 (5.1) 33 (4.6) 3 (1.1) 2 (2.6) 1 (0.5) 48 (6.1) 16 (5.9) 32 (6.2) weeks Once a month 41( 3.8) 8 (2.3) 33 (4.6) 3 (1.1) – 3 (1.5) 38 (4.8) 8 (2.9) 30 (5.8) Less than once a 36 (3.4) 13 (3.7) 23 (3.2) 4 (1.5) – 4 (2.0) 32 (4.0) 13 (4.8) 19 (3.7) month Never 28 (2.6) 9 (2.6) 19 (2.7) 3 (0.8) 1 (1.3) 2 (1.0) 25 (3.2) 8 (2.9) 17 (3.3) Relationship status Married/registered 580 (44.6) 168 (40.8) 412 (46.3) 197 (55.2) 57 (54.3) 140 (55.6) 383 (40.6) 111 (36.2) 272 (42.7) partnership Cohabiting 133 (10.2) 48(11.7) 85 (9.6) 25 (7.0) 9 (8.6) 16 (6.3) 108 (11.4) 39 (12.7) 69 (10.8) Unmarried 154 (11.8) 49 (11.9) 105 (11.8) 6 (1.7) 1 (1.0) 5 (2.0) 148 (15.7) 48 (15.6) 100 (15.7) Divorced or sepa- 331(25.4) 110 (26.7) 221(24.9) 51 (14.3 15 (14.3) 36 (14.3) 280 (29.7) 95 (30.9) 185 (29.0) rated Widow/widower 103 (7.8) 37 (9.0) 66 (7.4) 78 (21.8) 23 (21.9 55 (21.8) 25 (2.6) 14 (4.6) 11 (1.7) Satisfaction with social support received Low 329 (25.7) 107 (26.0) 222 (25.6) 86 (24.4) 30 (28.8) 56 (22.5) 243 (26.2) 77 (25.1) 166 (26.8) Medium 751 (58.7) 240 (58.4) 511 (58.9) 232 (65.7) 64 (61.5) 168 (67.5) 519 (56.0) 176 (57.3) 343 (55.4) High 199 (15.6) 64 (15.6) 135 (15.6) 35 (9.9) 10 (9.6) 25 (10.0) 164 (17.7) 54 (17.6) 110 (17.8) Data are n (%) SD standard deviation, NA not applicable No cases = – 1 3 Internal and Emergency Medicine of patients expressed a medium satisfaction with social religious activities in all males or all females and across support received. sites. In addition, satisfaction with social support received was not associated with HTN control in both sexes. In both Social support and hypertension control Europe and Ghana, neither medium nor high satisfaction with social support received was associated with HTN The association between number of cohabitants as a proxy control in both sexes. The models that were additionally for social support and HTN control is presented in Table 2a. adjusted for SES showed similar results. Among all males, cohabiting with two persons or less or with 3–4 persons was associated with lower odds of having con- trolled HTN compared to those cohabiting with five or more Discussion persons after adjustment for age and SES [odds ratio (OR) 0.30; 95% CI 0.16–0.57; 0.60; 0.34–1.04, respectively]. Key findings The odds were similar among male migrants in Europe and male non-migrants in Ghana when the analysis was strati- Our findings show that living with more than two other per- fied by site. No significant association was observed among sons increases the odds of having HTN controlled among migrant and non-migrant females even after adjusting for migrant and non-migrant Ghanaian men but not among age and SES (OR 1.26; 95% CI 0.86–1.85 for cohabiting migrant and non-migrant Ghanaian women. There was no with 3–4 persons; and 1.27; 0.89–1.81 for cohabiting with significant association between the other social support five persons or more, respectively). The association between proxies (relationship status, frequency of religious activity cohabiting and HTN in females did not vary by geographical attendance and satisfaction with social support received) and locations (Europe and Ghana). HTN control in both sexes. There was no association between relationship status and HTN control in both sexes and across sites as shown Discussion of the key findings in Table 2b. Among all males, the OR was 0.94 (95% CI 0.58–1.51), while in all females the OR was 0.98 (95% CI Migrant and non-migrant Ghanaian males were more likely 0.72–1.32). to have HTN controlled if they lived with more than two Table 3a, b shows the sex stratified associations between people, but this association was not observed among female frequency of attendance of religious activities and sat- migrants and non-migrants. Our findings are consistent isfaction with social support received and HTN control. with the findings of a study conducted in England, where After adjusting for age, there was no significant associa- living alone was negatively associated with HTN control tion between HTN control and frequency of attendance of among English males but not among English females [26]. Table 2 The association between number of cohabitants, relationship status and hypertension control stratified by sex Model Sex Location 2a. Cohabitation 2b. Relationship status 1–2 cohabitants 3–4 cohabitants 5 and more cohab- Yes OR (95% CI) No OR (95% CI) OR (95% CI) OR (95% CI) itants OR (95% CI) Model 1: Crude + age Male All 0.31(0.16–0.59) 0.60 (0.35–1.04) 1.00 (Ref) 1.00 (Ref) 0.99 (0.62–1.58) Europe 0.34 (0.16–0.70) 0.67 (0.35–1.29) 1.06 (0.64–1.76) Ghana – 0.21 (0.04–1.03) 0.46 (0.05–3.98) Model 2: Model All 0.30 (0.16–0.57) 0.60 (0.34–1.04) 1.00 (Ref) 1.00 (Ref) 0.94 (0.58–1.51) 1 + Educational Europe 0.32 (0.15–0.68) 0.66 (0.34–1.29) 1.05 (0.62–1.78) status and Employ- ment Ghana – 0.28 (0.5–1.48) 0.40 (0.04–3. 3.65) Model 1: Crude + age Female All 1.28(0.88–1.86) 1.29 (0.91–1.82) 1.00 (Ref) 1.00 (Ref) 0.97 (0.73–1.29) Europe 1.13(0.68–1.87) 1.28 (0.79–2.06) 1.00 (0.71–1.41) Ghana 0.99(0.49–2.05) 0.88 (0.46–1.64) 0.82 (0.47–1.43) Model 2: Model All 1.26(0.86–1.85) 1.27 (0.89–1.81) 1.00 (Ref) 1.00 (Ref) 0.98 (0.72–1.32) 1 + Educational Europe 1.12 (0.67–1.88) 1.13 (0.75–1.71) 1.05 (0.72–1.52) status and Employ- ment Ghana 0.99 (0.48–2.07) 0.87 (0.46–1.65) 0.82 (0.46–1.44) – No cases 1 3 Internal and Emergency Medicine Table 3 The association between hypertension control and frequency of attendance of religious activities and perceived social support received stratified by sex Model Sex Location 3a. Religious activities attendance 3b. Satisfaction with social support received Once a week or Less than once a Low OR (95% CI) Medium OR (95% High OR (95% CI) more OR (95% week OR (95% CI) CI) CI) Model 1: Male All 1.00 (Ref) 1.10 (0.58–2.07) 1.00 (Ref) 0.79 (0.48–1.28) 0.58 (0.28–1.21) Crude + age Europe 1.09 (0.57–2.10) 0.82 (0.48–1.40) 0.57 (0.26–1.27) Ghana – 0.67 (0.21–2.17) 0.63 (0.09–4.11) Model 2: Model All 1.00 (Ref) 1.06 (0.55–2.03) 1.00 (Ref) 0.79 (0.49–1.31) 0.59 (0.28–1.23) 1 + educational Europe 1.06 (0.54–2.08) 0.82 (0.47–1.43) 0.57 (0.25–1.18) status and employ- ment Ghana – 0.59 (0.17–2.15) 0.34 (0.04–2.56) Model 1: Female All 1.00 (Ref) 1.06 (0.61–1.83) 1.00 (Ref) 0.99 (0.70–1.39) 1.13 (0.71–1.78) Crude + age Europe 1.03 (0.57–1.87) 1.22 (0.80–1.86) 1.25 (0.73–2.12) Ghana 0.79 (0.15–4.08) 0.67 (0.37–1.23) 0.71 (0.27–1.90) Model 2: Model All 1.00 (Ref) 0.98 (0.56–1.742 1.00 (Ref) 0.99 (0.71–1.42) 1.15 (0.72–1.82) 1 + educational Europe 0.94 (0.52–1.73) 1.19 (0.78–1.83) 1.27 (0.74–2.17) status and employ- ment Ghana 0.76 (0.15–4.02) 0.71 (0.38–1.33) 0.68 (0.25–1.86) – No cases On average, non-migrants live with five other persons in are primarily caregivers during periods of ill health in Afri- a household, which is in line with the documented aver- can societies. Another possible explanation could be that age Ghanaian national household size of 4.0 [27]. Migrants females may have broader social networks and interactions, living in Europe on the other hand, live with an average of which provide them with alternative social support mecha- three other persons in a household, which is slightly higher nisms beyond family or other extended kin living with them, than the 2.3 persons per European Union (EU) household which could potentially explain why cohabitation was not size in 2016 reported by Eurostat. Within African societies, significantly associated with HTN control among females. cohabitants potentially comprise members of the same line- Existing evidence shows that females have wider sources of age including relatives such as parents, siblings, spouses, emotional support and may not necessarily consider their children, uncles and aunts who form the extended family. partners as their closest support [30]. Evidence shows that treatment and health care evolve around Contrary to our hypothesis, we did not observe a sig- kinship (common lineage) systems within African societies nificant association between relationship status and HTN with relatives brokering and facilitating health care for mem- control for both sexes and in migrants and non-migrants. bers during times of ill health [20, 28]. Male patients might Several studies have shown the association between marital therefore feel more comfortable discussing their illness and status and HTN control particularly for males [24, 26, 31], or health problems with their cohabitants than their female and thus, we anticipated similar associations in our study counterparts. Among all study participants, over 60% lived population. It is plausible that given the extended nature with their partners, over 70% lived with children 4 years of the African family system, social support maybe offered and above, and over 18% lived with other adults who could directly and/or indirectly by other relations or adult children provide both direct and indirect social support in facilitat- and not necessarily from a partner. As such, even among ing patients adherence to medication through reminders to migrants, while being in a relationship was not associated take medication and adherence to dietary recommendations. with HTN control, living with more than two other persons For instance, nearly half of the participants with controlled was associated with HTN control in men. HTN were on special HTN diet, which may be facilitated by Evidence shows that the SSA migrant populations in living with partner or older child or other adults. Indeed, a Europe are close knitted and organised around social struc- study found that African American men recognise the role tures evolving around community and faith-based or reli- of having supportive family relations, particularly females, gious organisations, which provide social support systems in controlling HTN as they support them to integrate treat- for their members [32]. Studies have shown that social net- ment recommendations into their daily lives [29]. Female works have an influence on cardiovascular health outcomes members, however, may not receive such support during [33–36]. As such, we hypothesised that hypertensive patients their HTN treatment possibly, because traditionally, females who attended religious actives at least weekly, would have 1 3 Internal and Emergency Medicine increased odds of having their HTN controlled compared These activities may not be perceived as forms of social with those that did not attend religious activities weekly. support by patients within African societies because they Contrary to our hypothesis, we did not find any associa- are expected and considered societal norms and duties of tion between regular attendance at religious activities and family members during times of ill health as they play the HTN control. While the lack of association observed in our “brokerage function” of the “therapy managing group” iden- study is similar to the results of an earlier study conducted tified by Janzen et al. In addition, sociocultural norms and in southern Africa [37], other studies have shown a nega- practices within Ghanaian communities frown on smoking tive association between HTN control and social networks habits, which potential explains the low (3%) prevalence of among males in Spain and African American males in North current smokers recorded in this study and thus facilitating Carolina, USA [24, 38]. It may well be that such gatherings adherence to smoking recommendations. at religious activities may not present the right context or Our results show that of the four proxies for social sup- environment for patients to discuss their HTN status. While port, only cohabitation was associated with HTN control participation in religious activities provides opportunities for among men while none of the social support proxies was social interactions, which are particularly common within associated with HTN control among females. The lack of the Ghanaian culture, they may not necessarily build suf- association between social support and HTN control among ficient trust networks for hypertensive patients to discuss females has also been reported among older women in Spain their health condition, concerns or share experiences. This [24]; and a study conducted in England showed that low is particularly more evident among males who may be appre- perceived social support was associated with low HTN con- hensive of disclosing or discussing their HTN status given trol rates [26]. While the key social indicators that could the connotation between antihypertensive medication and contribute to improving HTN control among men are unrav- sexual weakness, which has been shown to be a barrier to elling, the underlying mechanisms which could potentially medication adherence among males in this population [11]. improve HTN control among females, particularly among In addition, the fear of stigma due to HTN has been found female African migrants, remain unclear. Perhaps, it is to be a key reason for non-disclosure of hypertension sta- reasonable to assume that given the health carer role that tus to community members [11] and this could potentially females, in particular African females play within the fam- explain the lack of association between this kind of social ily and community system, they do not perceive themselves support and HTN control as patients may not have disclosed as having similar access to socially supportive resources or their HTN status and hence religious members may not even spousal support as their male counterparts. For instance, be aware of a person’s status. A recent qualitative study on if a woman is diagnosed with HTN, adherence to dietary community perceptions of HTN found that non-disclosure of changes such as reduction of salt content in meals may be HTN status was a key barrier to members ability to support hindered by the inability of relatives to accept such dietary members with HTN [21]. The frequency of religious attend- changes because of relatives’ poor recognition of their own ance may not fully capture the social processes and interac- risk to developing HTN [21, 41]. Given that only 37.6% of tions surrounding regular attendance at religious activities. all female participants had HTN controlled and the health For instance, some religious groups such as the Pentecostal caretaker roles that females assume in this context, if efforts Council of Churches (PCC) in the Netherlands have been are not made to address HTN control rates among females, shown to have more health-centred activities [32] and an the rates of uncontrolled HTN among African females may individual’s personal involvement may influence the kind supersede the rates of uncontrolled HTN among males in of social support religious groups or members provide [39]. coming years. Additional studies are therefore needed to The association between satisfaction with social sup- identify and explore the specific social support mechanisms port and BP control has been reported in other studies [24, that could contribute to improved adherence and HTN con- 40], but in this study, the odds were similar for high and trol among SSA females. medium patient satisfaction when compared with low patient satisfaction with social support received. However, given Limitations the sociocultural and communal relationships predominant within African communities, the perception of what consti- The main limitation to this study was that the definition of tutes social support might differ between our study popu- HTN was based on three blood pressure (BP) measurements lation and other populations. For instance, younger family on a single occasion as in most epidemiological studies. members and children may directly facilitate adherence to However, it is unlikely that this differs between the groups medication by reminding patients to take medication while under study and thus this limitation would not affect the females may facilitate adherence to dietary recommenda- associations observed in this study. Moreover, despite a tions by ensuring that meals are salt free and serving meals standardised research protocol, because of differences in reg- together with HTN medication for male HTN members. istration systems in the various sites, recruitment had to be 1 3 Internal and Emergency Medicine modified to suit the settings, which could account for lower distribution, and reproduction in any medium, provided you give appro- sample size recorded in some sites. In addition, we were priate credit to the original author(s) and the source, provide a link to unable to further stratify the analysis by religious groups the Creative Commons license, and indicate if changes were made. due to small sample size for HTN control. Furthermore, we lack data on other factors relating to cohabitation such as the numbers and gender of children living with hyperten- References sive study participants, which may influence HTN control. Lastly, the quantitative nature of the questionnaire hinders 1. Addo J, Smeeth L, Leon DA (2007) Hypertension in Sub-Saharan Africa a systematic review. Hypertension 50(6):1012–1018 extrapolations about the association between these social 2. Agyemang C et al (2005) Prevalence, awareness, treatment, and support proxies and HTN control as the underlying concepts control of hypertension among Black Surinamese, South Asian of these proxies may not be fully captured, which could Surinamese and White Dutch in Amsterdam, The Netherlands: account for the associations or lack of associations observed. the SUNSET study. J Hypertens 23(11):1971–1977 3. Agyemang C et al (2015) Hypertension control in a large multi- ethnic cohort in Amsterdam, The Netherlands: the HELIUS study. Int J Cardiol 183:180–189 Conclusion 4. Agyemang C et al (2010) A cross-national comparative study of blood pressure and hypertension between English and Dutch South-Asian- and African-Origin Populations: the role of national Our study findings show that living with more than two context. Am J Hypertens 23(6):639–648 people is positively associated with HTN control among 5. Agyemang C et  al (2018) Variations in hypertension aware- migrant and non-migrant Ghanaian males. It provides evi- ness, treatment, and control among Ghanaian migrants living in Amsterdam, Berlin, London, and nonmigrant Ghanaians living dence of the role of cohabitation as a key determinant for in rural and urban Ghana–the RODAM study. J Hypertension social support, which enhances HTN management among 36(1):169–177 male patients. Further research is needed to explore in-depth, 6. Glynn LG et al (2010) Interventions used to improve control of how SSA patients conceptualise social support in order to blood pressure in patients with hypertension. Cochrane Database Syst Rev. https ://doi.org/10.1002/14651 858.CD0051 82.pub4 identify modifiable determinants and the specific social sup- 7. Awuah R, Agyemang C, de-Graft Aikins A (2018) 5.10-P11 psy- port mechanisms that could contribute to adherence to HTN chosocial factors and hypertension among Ghanaians living in treatment and HTN control among SSA populations. different geographic locations: the RODAM study. Eur J Public Health 28(suppl_1):cky048–197 8. Weber MA et al (2014) Clinical practice guidelines for the man- Acknowledgements The authors are very grateful to the Advisory agement of hypertension in the community: a statement by the Board members of the RODAM study for their valuable support in American Society of Hypertension and the International Society shaping the methods, the research assistants, interviewers and other of Hypertension. J Hypertens 32(1):3–15 staff of the five research locations who have taken part in gathering the 9. Agyemang C et al (2012) Stroke in Ashanti region of Ghana. data and to the Ghanaian volunteers participating in this project. The Ghana Med J 46(2):12–17 RODAM study was supported by the European Commission under the 10. Stamler J (1991) Blood pressure and high blood pressure. Aspects Framework Programme (Grant Number: 278901). Liam Smeeth was of risk. Hypertension 18(3 Suppl):I95 supported by the Wellcome Trust (grant number WT082178). 1 1. Beune EJAJ et al (2008) How Ghanaian, African-Surinamese and Dutch patients perceive and manage antihypertensive drug treat- Funding This work is part of Ms. Nyaaba’s doctoral studies funded by ment: a qualitative study. J Hypertens 26(4):648–656 the Erasmus Mundus Joint Doctorate Program of the European Union 1 2. Grotto I, Huerta M, Sharabi Y (2008) Hypertension and socioeco- Specific Grant Agreement 2015–1595. nomic status. Curr Opin Cardiol 23(4):335–339 13. Leng B et al (2015) Socioeconomic status and hypertension: a Compliance with ethical standards meta-analysis. J Hypertens 33(2):221–229 14. Berkman LF (2000) Social support, social networks, social cohe- sion and health. Soc Work Health Care 31(2):3–14 Conflicts of interest The authors declare that they have no competing 15. Lavis J, Stoddart G (1999) Social cohesion and health. In: Work- interests. ing Paper No. 99–09 of centre for health economics and policy analysis. McMaster University, Hamilton, Ontario, Canada Statement on human and animal rights Data from the RODAM study, 16. Osamor PE (2015) Social support and management of hyperten- which was carried out from 2012 to 2015 was used for this analysis. sion in south-west Nigeria: cardiovascular topic. Cardiovasc J Afr The RODAM study received ethical approval from the respective ethics 26(1):29–33 committees in Ghana, and the three European countries prior to data 1 7. Gallant MP (2003) The influence of social support on chronic collection in each country. illness self-management: a review and directions for research. Health Educ Behav 30(2):170–195 Informed consent Each study participant provided informed written 18. Lo R (1999) Correlates of expected success at adherence to health consent prior to study enrolment. regimen of people with IDDM. J Adv Nurs 30(2):418–424 19. Christensen AJ et al (1992) Family support, physical impairment, Open Access This article is distributed under the terms of the Crea- and adherence in hemodialysis: an investigation of main and buff- tive Commons Attribution 4.0 International License (http://creat ering effects. J Behav Med 15(4):313–325 iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, 1 3 Internal and Emergency Medicine 20. Lemba JJM (1982) 1650–1930: a drum of affliction in Africa and 3 2. Agyemang C et al (2018) Your health is your wealth: faith-based the New World. Garland Publishing Inc., London community action on the health of African migrant communities 21. Nyaaba GN et al (2018) Lay community perceptions and treatment in Amsterdam. J Epidemiol Community Health 72(5):409–412 options for hypertension in rural northern Ghana: a qualitative 3 3. Vogt TM et al (1992) Social networks as predictors of ischemic analysis. BMJ Open 8(11):e023451 heart disease, cancer, stroke and hypertension: incidence, survival 22. Meinema JG et al (2015) Determinants of adherence to treatment and mortality. J Clin Epidemiol 45(6):659–666 in hypertensive patients of African descent and the role of cultur- 3 4. Mookadam F, Arthur HM (2004) Social support and its relation- ally appropriate education. PLoS One 10(8):e0133560 ship to morbidity and mortality after acute myocardial infarction: 23. Agyemang C et al (2015) Rationale and cross-sectional study systematic overview. Arch Intern Med 164(14):1514–1518 design of the Research on Obesity and type 2 Diabetes among 35. García EL et al (2005) Social network and health-related quality African Migrants: the RODAM study. BMJ Open 4(3):e004877 of life in older adults: a population-based study in Spain. Qual 24. Redondo-Sendino Á et al (2005) Relationship between social net- Life Res 14(2):511–520 work and hypertension in older people in Spain. Revista Española 3 6. Berkman LF, Glass T (2000) Social integration, social networks, de Cardiología (English Edition) 58(11):1294–1301 social support, and health. Social Epidemiol 1:137–173 25. Winkleby MA et al (1992) Socioeconomic status and health: how 37. Daniels A et al (1999) Blood pressure and social support obser- education, income, and occupation contribute to risk factors for vations from Mamre, South Africa, during social and political cardiovascular disease. Am J Public Health 82(6):816–820 transition. J Hum Hypertens 13(10):689 2 6. Shah S, Cook DG (2001) Inequalities in the treatment and control 38. Wagner EH et al (1984) The edgecombe county high blood pres- of hypertension: age, social isolation and lifestyle are more impor- sure control program: I. Correlates of uncontrolled hypertension tant than economic circumstances. J Hypertens 19(7):1333–1340 at baseline. Am J Public Health 74(3):237–242 2 7. Service, G.S. (2014) Ghana living standards survey round 6 3 9. Abanilla PKA et al (2011) Cardiovascular disease prevention in (GLSS 6): labour force report. Ghana Statistical Service Accra, Ghana: feasibility of a faith-based organizational approach. Bull Ghana World Health Organ 89(9):648–656 2 8. Ankrah EM (1993) The impact of HIV/AIDS on the family and 40. Baker B et al (2003) Marital support, spousal contact and the other significant relationships: the African clan revisited. AIDS course of mild hypertension. J Psychosom Res 55(3):229–233 Care 5(1):5–22 4 1. de Graft Aikins A et al (2012) Lay representations of chronic 2 9. Rose LE et al (2000) The contexts of adherence for African Amer- diseases in Ghana: implications for primary prevention. Ghana icans with high blood pressure. J Adv Nurs 32(3):587–594 Med J 46(2 Suppl):59–68 30. Fuhrer R, Stansfeld SA (2002) How gender affects patterns of social relations and their impact on health: a comparison of one Publisher’s Note Springer Nature remains neutral with regard to or multiple sources of support from “close persons”. Soc Sci Med jurisdictional claims in published maps and institutional affiliations. 54(5):811–825 31. Glynn LM, Christenfeld N, Gerin W (1999) Gender, social sup- port, and cardiovascular responses to stress. Psychosom Med 61(2):234–242 Affiliations Gertrude Nsorma Nyaaba1,2  · Karien Stronks1 · Karlijn Meeks1 · Erik Beune1 · Ellis Owusu‑Dabo3 · Juliet Addo4 · Ama de‑Graft Aikins5 · Frank Mockenhaupt6 · Silver Bahendeka7 · Kerstin Klipstein‑Grobusch8,9 · Liam Smeeth4 · Charles Agyemang1 1 Department of Public Health, Amsterdam University 6 Charité—Universitaetsmedizin Berlin and Institute Medical Centres, Amsterdam Public Health (APH) Research of Tropical Medicine and International Health, Berlin, Institute, University of Amsterdam, Meibergdreef 9, Germany 1105 AZ Amsterdam, The Netherlands 7 MKPGMS—Uganda Martyrs University, Kampala, Uganda 2 Barcelona Institute for Global Health (ISGlobal), University 8 of Barcelona, Barcelona, Spain Julius Global Health, Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, 3 School of Public Health, Kwame Nkrumah University Utrecht, The Netherlands of Science and Technology, Kumasi, Ghana 9 Division of Epidemiology and Biostatistics, School 4 Department of Non-communicable Disease Epidemiology, of Public Health, Faculty of Health Sciences, University London School of Hygiene and Tropical Medicine, London, of the Witwatersrand, Johannesburg, South Africa UK 5 Regional Institute for Population Studies, University of Ghana, Legon, Ghana 1 3