背景：

我国高血压患者基数庞大，高血压也是多种心血管疾病的主要危险因素，做好高血压患者的规范化管理意义重大。高血压患者遵循专业卫生保健的指导并形成良好的自我保健行为是影响高血压管理效果的关键，而这些行为又受到多种因素的影响。因此，不断寻找重要的、潜在的影响因素及其干预切入点，并阐明有效的新型行为促进策略是极其重要的。过去几十年里，国际上大量研究表明社会网络等社会因素与健康行为呈现高度关联，但国内尚缺乏可以用来指导干预策略的相关证据。

目的：

本研究旨在从社会网络理论视角出发，利用流行病研究设计、统计推断、社会网络分析及定性研究相结合的方法，研究我国社会文化背景下，对高血压患者自我保健行为（聚焦服药依从和休闲身体活动）具有关键影响的社会网络因素，尤其是识别具有关键影响的网络成员，并了解医、患及网络成员（聚焦配偶）三方对开展社会网络干预的意愿或建议，从而为开展创新性的社会网络干预、改进高血压管理策略提供具有指导意义的科学证据。

方法：

使用嵌入式混合研究方法来设计总体研究框架，以定量分析为主，定性分析为定量分析提供启发性和解释性支持。研究现场为湖北省宜昌市。

定量数据收集方法：采用分层多阶段随机抽样法，从宜昌市18家基层医疗卫生机构中抽取到2639名患者，使用“国际身体活动问卷”测量患者的休闲身体活动（LTPA）水平，用“Morisky服药依从性量表”测量患者的服药依从性水平，用“提名法”收集患者的社会网络数据。采用受访者驱动抽样法，抽取到1747名患者的配偶，利用Likert-5级量表测量夫妻双方对关系的满意度及其参与干预的意愿。

定量数据分析：1. 对LTPA、服药依从性的分析分为两个层次：（1）个体层次：以患者为分析单元。使用Logistic回归、限制性立方样条、线性趋势检验分析社会网络结构与患者LTPA水平的关联、社会支持内容与服药依从性水平的关联；（2）关系层次：以患者和网络成员构成的关系/配对为分析单元。采用多水平模型分析关系性质与关系双方LTPA同质性的关联，采用Logistic回归分别在不同关系性质网络中进行亚组回归，分析特定关系性质网络成员LTPA水平与患者LTPA水平的关联、特定关系性质网络成员提供的支持水平与患者服药依从性水平的关联，从而分别识别网络中对患者LTPA具有关键影响的锻炼榜样、对患者服药依从性具有关键影响的支持提供者。2. 对患者及其配偶参与社会网络干预的意愿分析仅为个体层次：使用Logistic回归分别分析患者及其配偶参与干预意愿水平的影响因素。

定性资料收集与分析：采用半结构化访谈收集资料。使用基于扎根理论的主题分析法，提取对LTPA和服药依从性分别具有促进或阻碍作用的社会网络相关因素主题，同时了解医患双方对于开展社会网络干预的看法或建议。

结果：

定量分析主要结果：

1. 高血压患者社会网络对LTPA的影响：高血压患者的LTPA充足率为34.82%，其中男性、中年、来自农村、在职、务农、受教育水平低、低收入、自评健康状况差的患者的LTPA充足率较低。个体层次上，拥有较小、较紧密、LTPA聚集性越高的网络与患者LTPA充足呈正相关；关系层次上，配偶、兄弟姐妹LTPA充足，分别与患者LTPA充足的可能性提高121.9%（95%CI = 79.2%–174.7%）、45.0%（10.1–91.0）相关；而这种关联性在兄弟姐妹中表现出了性别差异，同性兄弟姐妹LTPA充足，与患者LTPA充足的可能性提高48.8%（1.1–118.9）相关，而异性兄弟姐妹未表现出显著关联。

2. 高血压服药患者社会支持对服药依从性的影响：高血压服药患者的服药依从性良好率为56.04%，忘记服药是最常见的不依从服药情况，男性、中年、来自农村、高血压病程长的患者的服药依从性良好率较低。个体层次上，提醒患者服药相关的照护支持得分与患者服药依从性良好的可能性之间存在线性变化趋势（P for trend<0.05）；关系层次上，配偶提供多种支持、兄弟姐妹提供多种支持与患者服药依从性良好呈正相关；配偶提供信息支持、兄弟姐妹提供情感支持、兄弟姐妹提供照护支持，分别与患者服药依从性良好的可能性提高47.6%（0.4–116.9）、53.6%（8.6–117.3）、70.2%（17.1–147.4）相关，而朋友提供经济支持，与患者服药依从性良好的可能性降低49.6%（15.4–70.0）相关。

3. 夫妻双方参与社会网络干预意愿的影响因素：83.34%的患者愿意接受配偶提供的干预支持，女性、自评健康好、对夫妻关系满意的患者更愿意接受配偶的干预支持；87.01%的配偶愿意为患者提供干预支持，对夫妻关系满意的配偶更愿意给患者提供干预支持，而务农职业的配偶给患者提供干预支持的意愿较低。

定性分析主要结果：社会网络成员支持、不支持体育锻炼的观念与行为分别对患者的体育锻炼行为具有促进、阻碍影响；来自社会网络成员的信息支持、情感支持和工具支持对患者的服药依从行为具有促进影响，而患者与网络成员的消极互动对患者的服药依从行为具有阻碍影响。

结论：

由基层医疗卫生机构管理的宜昌高血压患者中，仍有约三分之二的患者休闲身体活动不足，男性、中年、来自农村、在职、务农、受教育程度低、低收入、自评健康状况差的患者应成为未来干预的重点人群。仍有约一半的患者服药依从性不佳，忘记服药是最常见的不依从服药情形，男性、中年、来自农村、高血压病程长的患者成为未来干预的重点人群。

身边人普遍支持休闲身体活动的社会规范、拥有很小且紧密的核心社交网络、身边拥有较多支持休闲身体活动的人员是促进高血压患者休闲身体活动的积极因素；配偶、同性兄弟姐妹是网络中的关键锻炼榜样。身边人提供与服药相关的多样化支持或帮助、尤其是提供提醒吃药相关的照护支持是促进高血压患者服药依从性的积极因素；配偶、兄弟姐妹是网络中的关键支持提供者。对夫妻关系满意是夫妻意愿参与社会网络干预的主要促进因素。因此，社会网络中的行为规范与社会支持成为高血压患者自我保健行为促进干预设计中值得考虑的重要内容，夫妻关系良好的配偶、同性兄弟姐妹可能成为网络中潜在的正向干预资源。

Background:

Hypertension is highly prevalent in China and is a major risk factor for various cardiovascular diseases, resulting in heavy health and social burden in China. Standardized management of hypertension is crucial to mitigate these burdens. The self-care behaviors of hypertensive patients, guided by healthcare professionals, are pivotal to effective hypertension management and are influenced by various factors. Therefore, it is essential to continuously identify underlying influencing factors and develop effective behavior promotion strategies. While international studies have shown a strong correlation between social networks and health behaviors, there is a lack of evidence to guide interventions for hypertension management in China.

Purposes:

Focusing on medication adherence and leisure-time physical activity (LTPA), this study aims to investigate the social network factors that have critical influences on the self-care behaviors of hypertensive patients in the social context in China, especially to focus on the identification of key influential network members, from the perspective of social network theory and utilizing various methodologies such as epidemiology, statistics, social network analysis and qualitative research. The intention or suggestions of doctors, patients and network members (focusing on spouses) on social network interventions were also understood. The ultimate goal is to provide scientifically guided evidence for innovative social network interventions and improvement of hypertension management strategies.

Methods:

The embedded mixed research method was utilized to design the overall research framework. Quantitative analysis served as the primary approach, complemented by qualitative analysis which provided valuable illumination and explanatory support. The study took place in Yichang City, located in Hubei Province.

Quantitative data collection: A total of 2639 patients were selected from 18 primary healthcare institutions in Yichang City by hierarchical multi-stage random sampling method. LTPA was measured by the International Physical Activity Questionnaire (IPAQ), and medication adherence was assessed by Morisky Medication Adherence Scale (MMAS). Information regarding patients’ social networks was gathered through the name generator method. 1747 spouses of patients were sampled using respondent-driven sampling, and their relationship satisfaction and willingness to participate in the future social network interventions were measured using the Likert-5 scales.

Quantitative data analysis:

1. The analysis of LTPA or medication adherence can be approached in two levels:

(1) Individual-level: In this level, the patient is considered as the unit of analysis. Logistic regression, restricted cubic spline, and linear trend tests were employed to examine the relationship between social network structures and patients’ LTPA, as well as the association between social support content and medication adherence.

(2) Tie-level: In this level, the focus shifts to the ties/relationships/pairs between patients and their network members. The multi-level model was used to investigate the link between these relationships and the homogeneity of LTPA. Logistic regression was performed among different relationship subgroups. This allowed for the examination of the association of the LTPA levels of different types of network members with patients’ LTPA levels, and the connection of the support provided by different types of network members with patients’ medication adherence, so as to identify the exercise leaders and the key support providers within patients’ social networks, respectively.

2. The assessment of patients’ and their spouses’ inclination to engage in social network interventions was solely conducted at an individual level: Logistic regression was employed to examine the factors influencing the willingness of patients and their spouses to partake in the interventions.

Qualitative data collection and analysis: Semi-structured interviews were utilized to gather qualitative data. Thematic induction analysis based on the grounded theory was used to extract themes related to facilitators and barriers within social networks for LTPA and medication adherence. Additionally, this approach helped comprehend the opinions or suggestions of doctors and patients regarding social network interventions.

Results:

The main findings of quantitative analysis were as follows:

1. The influence of social networks on LTPA of hypertensive patients: the prevalence of sufficient LTPA in hypertensive patients was 34.82%, with lower rates observed among males, middle-aged patients, those residing in rural areas, working as farmers or in paid employment, having lower education level, low income, or poor self-rated health. The individual-level analysis results indicated that smaller, more closely-knit networks or networks with a higher concentration of LTPA were positively linked to engaging in patients’ sufficient LTPA. In terms of tie-level analysis, it was found that the presence of spouses and siblings who engaged in sufficient LTPA was associated with a 121.9% (95%CI = 79.2%–174.7%) and 45.0% (10.1–91.0) increase in the likelihood of sufficient LTPA of patients, respectively. However, there appeared to be a gender difference among siblings. Having same-sex siblings who engaged in sufficient LTPA increased the likelihood of patients also achieving sufficient LTPA by 48.8% (1.1–118.9), whereas no significant association was observed among opposite-sex siblings.

2. The influence of social support on medication adherence of hypertensive patients: 56.04% of patients exhibited optimal medication adherence. Among the various situations of non-adherence, forgetting to take medication was identified as the most prevalent. Certain demographic factors were found to be associated with lower levels of adherence, such as being male, young, residing in rural areas, or having a longer duration of hypertension. The results of the individual-level analysis suggested a clear link between caregiving support in terms of reminding patients to take their medication and the probability of achieving optimal medication adherence (P for trend<0.05). Regarding relationships, spouses providing multiple supports and siblings providing multiple supports were positively correlated with optimal medication adherence. Specifically, informational support provided from spouses, emotional support provided from siblings and caregiving support provided from siblings showed associations with 47.6% (0.4–116.9), 53.6% (8.6–117.3) and 70.2% (17.1–147.4) increase in the likelihood of optimal medication adherence of patients, respectively. However, financial support from friends was associated with a 49.6% (15.4–70.0) decrease in the likelihood of optimal medication adherence.

3. Influencing factors of couples’ willingness to participate in social network interventions: 83.34% of patients expressed their willingness to accept intervention support from their spouses. Among these patients, female patients, those with good self-rated health, and those satisfied with their marital relationship were more likely to accept such support. 87.01% of spouses indicated their willingness to provide intervention support for the patients. It was found that spouses who were satisfied with their marital relationship were more inclined to offer support. However, spouses who worked as farmers showed lower levels of willingness to provide intervention support.

The qualitative analysis yielded several key findings: The norms and behaviors of social network members who support and do not support LTPA promote and hamper the LTPA behavior of patients, respectively; Informational support, emotional support and instrumental support from social network members had positive influences on patients’ medication adherence, while negative interaction between patients and network members had a negative influence on patients’ medication adherence.

Conclusion:

Among the hypertensive patients managed by primary healthcare institutions in Yichang, approximately two-thirds still have insufficient LTPA. Males, middle-aged patients, those from rural areas, working in paid employment or as farmer, with low education levels, low income, or poor self-rated health should be the target population for future interventions. About half of the patients exhibit poor medication adherence, with forgetting to take medication being the most common non-adherence behavior. Males, middle-aged patients, those from rural areas, or having a longer duration of hypertension are important target groups for future interventions.

The social norms that generally support LTPA, having small and close-knit core social networks, and having more people who support LTPA in social networks are positive factors that promote LTPA among hypertensive patients. Spouses and same-sex siblings serve as key exercise role models within the network. The diverse support related to medication from social networks, particularly caregiving support in terms of reminding patients to take their medication, is a positive factor in promoting medication adherence among hypertensive patients. Spouses and siblings also play a crucial role as key providers of support within the network. Satisfaction with marital relationships is a major facilitating factor for couples’ willingness to participate in social network interventions. Therefore, behavioral norms and social support within the social networks emerge as important considerations in the design of interventions promoting self-care behaviors among hypertensive patients, and spouses with good relationships and same-sex siblings may serve as potential positive intervention resources.