【作者】 王淑康；

【导师】 李士雪；

【作者基本信息】 山东大学 ， 社会医学与卫生事业管理， 2012， 博士

【摘要】 研究背景人口老龄化和慢性病是全球共同面临的严峻挑战,无论是在发达国家还是在发展中国家无疑都会带来巨大的经济负担,为了应对二者对全人类所产生的重大负向作用,健康老龄化被提到重要议事日程上来,即在老龄化社会中,绝大多数老年人处于生理、心理和社会功能的健康状态,使社会发展不受过度人口老龄化的影响。规律体育活动(regular physical activity, RPA)行为对于延缓老年人机能衰退和改善慢性病症状具有极其重要作用,其低成本高效益性更使其具有重大的经济和社会意义,是老年人动态、连续健康管理过程中最重要的干预措施。然而大部分老年人尚未达到推荐的体育活动水平,RPA对于老年人慢性病防治及健康水平的提高所应具有的重大意义尚未充分体现,国内外研究均显示RPA行为干预是健康干预的难题。为了对体育活动行为进行有效的干预,准确锁定干预的切入点,必须明确体育活动行为的影响因素及各因素之间的相互作用关系。健康相关行为研究显示,个体的行为是在其社会文化背景、经济水平、身体心理状况、知识和资源框架内形成和改变的,体育活动行为是健康行为的一个极其重要方面,最终RPA行为的形成亦是错综复杂影响因素相互作用的结果。国外对于RPA行为影响因素的研究虽然融入了社会生态学理论思想,涉及了人口学、社会支持、心理状态、健康状况、健康知识、慢性病等因素,但仍存在一定的缺陷和不足：(1)研究中大多是应用简单方法对单一影响因素进行分析,而且年龄、性别、职业、经济收入等人口学因素的影响作用在各项研究中的结论存在不一致现象；(2)单个研究中考虑因素并不全面,未将尤为重要的社会因素细致分化,未考虑社会因素的各个不同维度的影响作用；(3)尚未研究各因素在影响体育活动行为时的相互作用关系,未说明各因素对于RPA行为除了直接作用外,还存在间接作用及其具体路径关系。国内老年人RPA行为影响因素理论研究部分基本是国外研究的理论综述,人群实证研究多数为老年人体育活动现状的分析和讨论,多数研究样本含量较小,基于社会生态学理论的老年人体育活动行为研究尚属空白。为此,本研究以社会生态学理论为基础,考虑一般因素(年龄、性别、文化程度、就业状况、经济状况等)、个体健康知识、社会支持、环境条件、健康状况、慢病条件、心理状态等多个因素对城市社区老年人体育活动行为的作用,采用偏最小二乘(partial Least Squares, PLS)路径模型分析方法,确定老年人RPA的社会生态学影响因子及各因子之间的路径关系,以期为RPA干预措施的制定提供科学依据。研究目的本研究的总体研究目标是通过理论研究和实证性分析,确定城市社区老年人RPA影响因素,探讨体育活动行为主要影响因素共同作用于体育活动行为时的相互关系及可能存在的间接路径,提出利用可控制因素提高老年人体育活动水平,为社区老年人生活方式健康管理提供政策依据。具体研究目的包括：明确我国城市社区老年人RPA现状及其社会生态学影响因素；将社会因素、健康知识、心理因素、健康状况对体育活动行为的影响细致化,确定在各因素中对体育活动产生影响的具体维度因子；分析一般因素、社会因素、健康知识、心理因素等在影响体育活动行为过程中的相互作用机制；确定各主要影响因子对体育活动作用的直接路径与间接路径。研究方法通过文献复习法和焦点小组访谈法收集了本研究的定性资料,通过社区现场和调查员入户问卷调查收集定量资料。在济南市的青年西路、棋盘街、佛山苑、正大城市花园4个社区对50岁以上老年人进行了问卷调查,共收集了1580人的一般信息、社会因素、健康知识、健康状况、慢性病、心理状况、体育活动等方面的资料。对于该定量资料在统计描述的基础上首先进行了单因素分析,并且应用多因素分析方法明确了体育活动行为的影响因素和各影响因素之间的相互作用关系,最终确定了各因子对于体育活动行为作用的直接路径和间接路径。本研究应用的统计分析方法包括χ2检验、等级相关、t检验,探索性因子分析、PLS路径模型分析。上述分析采用SPSS16.0和SmartPLS2.0统计软件实现。主要研究结果1.RPA行为的现状及单因素分析：社区老年人RPA的参加率为60.8%,尚有近40%的老年人尚无规律的体育活动。单因素分析结果显示：RPA与社区、年龄、就业状况、月收入有关,与性别、文化程度无关。青年西路、棋盘街、佛山苑、正大城市花园社区RPA的参加率分别为46.3%、67.2%、72.2%、90.7%,社区间比较的χ2=98.19,P=0.000；男、女组间RPA参加率进行比较的χ2=1.33,P=0.249；各年龄组间RPA比较的χ2=25.21,P=0.000,≥80岁年龄组RPA参加率最低,低于其他各年龄组,50～59岁年龄组RPA参加次低,低于60～69岁年龄组；文化程度组间RPA比较的χ2=2.12,P=0.832；不同就业状况组间比较的χ2=17.78,P=0.000,多重比较显示在业低于离退休组；不同月收入组间进行比较的χ2=14.89,P=0.005。2.RPA行为影响因素的定性分析：部分老人对RPA的认识还不够,部分老人有活动的意愿,希望能够得到一定的帮助,如有人带头组织体育活动,并且得到家人、朋友共同参加,最好由社区来组织,活动地点也能够在社区内部。对于其他社区活动,老人们也需要有个带头人,居委会也行,其他人也行,能心甘情愿地帮助他们的人就行。老人们都希望能够参加社区组织的活动,原因一是老人对社区的信任和依赖,原因二是多数老人参加活动的障碍是活动地点远不方便,希望能够在社区内参加活动。老年人明确表示支持在聊天中学习健康知识,最好能有机会与专家和朋友一起聊天。目前的社区卫生服务尚不能解决老年人的心理问题、如对老年人进行心理疏导等。3.体育活动行为影响因素的PLS路径分析：潜变量社区、年龄、社区活动、轻度活动、中重度活动与体育活动之间的路径系数分别为0.13、0.15、0.12、0.13、0.13,而且路径系数经检验在α=0.01水准上均具有统计学意义(t>t0.01/2,9991≈2.58)；潜变量非高血压、社区安全感、与直系亲属交往的路径系数分别为-0.07、0.09、0.08,而且路径系数经检验在α=0.05水准上均有统计学意义(t>t0.05/2,999≈1.96)；就业状况、抑郁、非糖尿病、社会活动、高血压特异知识、医学知识的路径系数分别为0.06、-0.07、-0.06、0.06、-0.05、0.06,经检验各路径系数在a=0.20水准上均具有统计学意义(t>t0.20/2,999≈1.28)。4.体育活动行为影响因素之间关系的PLS路径分析：与社区安全相关的潜变量为年龄、社区、抑郁、社区活动,其路径系数分别为0.100、0.182、-0.154、0.190；与社会活动相关的潜变量为社区活动、就业状况,其路径系数分别为0.449、-0.067；直系亲属聚会与其他潜变量均无关,对于体育活动行为的影响具有独立性；与社区活动相关的潜变量为社区、抑郁、医学知识、中重度活动能力、社区安全、社会活动,其路径系数分别为-0.145、-0.083、0.121、0.100、0.169、0.375；与抑郁相关的潜变量为社区、中重度活动能力、社区安全、社区活动,其路径系数分别为-0.150、-0.228、-0.139、-0.107；与非高血压相关的潜变量为中重度活动能力,其路径系数为0.29；与非糖尿病相关的潜变量为高血压特异知识,但路径系数较小为0.100,对糖尿病变异的解释能力较小,故其对于体育活动行为的影响具有独立性；与高血压特异知识相关的潜变量为非糖尿病、社区、医学知识,但对于高血压知识解释能力太低,故其对于体育活动行为的影响具有独立性；与医学知识相关的潜变量为年龄、社区、中重度活动能力、社区活动,其路径系数为-0.178、0.240、0.113、0.131；与中重度活动能力相关的潜变量为非高血压、年龄、抑郁、医学知识、轻度活动能力、社区活动,其路径系数分别为0.168、-0.184、,0.118、0.110、0.491、0.086；与轻度活动能力相关的潜变量为中重度活动能力,其路径系数为0.610,但根据其实际意义,该变量对于体育活动行为的影响具有独立性；与社区相关的潜变量为抑郁、医学知识、社区安全、社区活动,其路径系数为-0.167、0.248、0.191、-0.232；与年龄相关的潜变量为中重度活动能力、医学知识,其路径系数分别为-0.372、-0.169；与就业状况相关的潜变量为中重度活动能力、社会活动,但对其变异解释能力低,故其对于体育活动行为的影响具有独立性。5.各因子对于体育活动行为的直接作用和间接作用：年龄对体育活动的直接作用虽不可变,但2条间接作用路径分别为年龄->社区安全->体育活动、年龄->中重度活动能力->体育活动；除社区对体育活动的直接作用外,还有3条间接作用路径,分别为社区->社区活动->体育活动、社区->社区安全->体育活动、社区->抑郁->体育活动；抑郁对体育活动具有直接作用外,还有2条间接路径,分别为抑郁->社区安全->体育活动、抑郁->社区活动->体育活动；社区活动不仅对体育活动具有重要的直接作用,还有2条间接路径,分别为社区活动->社区安全->体育活动、社区活动->社会活动->体育活动；医学知识对体育活动的直接作用虽不强,但其2条间接路径不容忽视,分别为医学知识->社区活动->体育活动、医学知识->中重度活动能力->体育活动；中重度活动能力对体育活动具有直接作用外,还有2条间接作用路径,分别为中重度活动能力->社区活动->体育活动、中重度活动能力->抑郁->体育活动。结论与建议城市社区老年人RPA行为的影响因素是多维的,包括一般因素(社区、年龄、就业状况)、社会因素(社区安全感、社会活动、直系亲属聚会、社区活动)、心理因素(抑郁)、健康状况(活动能力)、慢性病状态(非高血压、非糖尿病)。多维的影响因子同时作用于老年人体育活动行为时,其作用并不是相互独立的,各因子之间是相互影响的。部分因子既是体育活动行为的直接作用因子又是某些因子间接作用的中间变量,而且某些因子对体育活动行为既具有直接作用又具有间接作用。政策建议：(1)将体育活动行为干预作为老年人健康管理与健康干预的重中之重；(2)将社区作为老年人体育活动行为干预的重要平台；(3)对社区老年人RPA行为实施个性化干预；(4)心理干预应融入社区体育活动行为干预；(5)将慢性病老人体育活动行为干预作为社区体育活动行为干预的突破口；(6)实施以社区为平台的多维体育活动行为干预。创新与不足本研究的创新：1.立题有创新体育活动行为干预是人群健康水平提高的重要措施,也是目前全球卫生工作的重点和难点。本研究在国内首次提出以社会生态学理论为基础,同时全面考虑人口学、社会、心理、健康状态、慢性病、健康知识等因子对于体育活动行为的作用,并首次将社会因素分为社区安全感、社会活动、社会交往、社区活动、社会支持不同维度来分别界定其在老年人RPA行为中所起的作用。本研究在国内外首次在研究体育活动行为影响因素基础上,提出各影响因子作用于体育活动行为时相互作用的假设,并对该假设进行了系统的分析和验证,最终确定了各个因子对体育活动作用的直接路径和间接路径。2.方法应用有创新针对本研究立题中所提出的各因素间复杂关系假设,以及数据的非正态、多维、多元共线性等特点,首次将PLS路径模型分析方法应用于筛选RPA行为的影响因素,并应用该方法理顺了各影响因素作用于体育活动行为时的相互路径关系。3.结论有创新本研究得到RPA行为的影响因子是多维的,而且各因子之间具有复杂的路径关系,某因子在对体育活动行为产生直接作用的同时,还会通过若干间接变量对体育活动行为产生间接作用,实际上构成了一个复杂的影响因素网络,具体人群的干预应根据不同人群的各因子特点找到其突破点,有的放矢地进行干预。本研究的不足：1.因调查条件所致,尚未把社区变量进行细分,细化社区的环境因素。2.本研究仅在济南市4个社区进行了老年人的体育活动行为调查,今后应在此基础上进行多城市多社区的研究比较。更多还原

【Abstract】 BackgroundPopulation aging and chronic diseases have been the severe challenges the whole world have to confront, and will undoubtedly exert a huge economic burden on both developed and developing countries. To cope with the great negative effects resulted from the two, healthy aging has been put high on the agenda, i. e. to ensure that the majority of the elderly remain a state of complete physical, mental, and social well-being, and to prevent social development from being influenced by excessive population aging. For its extremely important role in both delaying the elderly functional decline and improving chronic disease symptoms, and its cost-efficient feature which makes it of greater economic and social significance, to conduct regular physical activities (RPA) is supposed to be the most important intervention in the dynamic and continuous process of health management for the elderly. However, most aged people have not yet reached the recommended physical activity (PA) level, and the great significance in PA for elderly chronic disease prevention and health raise has not been sufficiently reflected. Domestic and foreign studies have shown that to regulate PA behavioral intervention is a problem in health intervention. To effectively intervene PA behavior, and to determine the exact entry point for intervention, factors which influence sports activities and interactions among the various factors should be made clear. Studies on health-related behavior have shown that individual behavior forms and changes within the framework of its social and cultural background, economic level, physical and mental status, knowledge and resources. They have also reveled that PA behavior is an extremely important aspect of health-related behavior, and its final formation is the result of the interaction of complex influencing factors. Foreign researches on these factors involve such theoretical ideas of social-ecology as demography, social support, psychological state, health status, health knowledge, chronic disease factors, but they are still with certain shortcomings and deficiencies:(1) Most of the studies analyze a single impact factor by applying simple methods, and inconsistencies among the effects of such socio-demographic factors as age, gender, profession and income are easily found in these studies.(2) Incomplete consideration for related factors in a single study, no detailed differentiation of the particularly important social and psychological factors, and no account of the effects of social and psychological factors in various dimensions are noted.(3) Studies on the interactions of factors affecting the behavior of sports activities are not yet conducted. Indirect effects and specific path relations, in addition to direct effects, of these factors on PA behavior are also never discussed. The theoretical part of domestic researches on the influence of regular PA-related factors among the elderly is generally based on the theoretical reviews of foreign researches. The majority of domestic population studies, often with a small sample size, are analyzing and discussing the status of the elderly sports activities, and researches based on the social-ecological theory are still blank.Therefore, based on social ecological theory, considering influences on PA behavior of the urban community aged groups by demographic factors (age, gender, education level, employment status, economic status, etc.), individual health knowledge, social support, environmental conditions, health status, chronic disease conditions, psychological state and any other factors, applying Partial Least Squares(PLS) path model analysis method, this study is conducted to determine social ecological factors influencing regular PA of the elderly and specify path relations within these factors, thus to provide interventions for regular PA behavior with scientific basis.Research ObjectivesThe goals of this research are to find out influencing factors on regular PA for the elderly in city community, and search the relationship between the major influencing factors and possible indirect path through theory research and demonstration analysis. It was proposed that we should raise the level of PA of the elderly by use of some factors, so as to provide basis of formulating policy. Specific aims of the research are to:1) Reveal the current situation of regular PA and find out influencing factors on social ecology for the elderly in city community;2) subdivide the influencing effect of social factors or health knowledge or psychology factors or health state on regular PA, and find out specific dimensionality of influencing factors;3) Find out the relationship between the influencing factors such as social factors, health knowledge, psychology factors and so on;4) Find out direct and indirect path of the major influencing factors for regular PA.Study DesignCollect qualitative materials through document analysis and focus group interview. Quantitative materials were collected through field investigation and self-administrated questionnaire survey, we investigate the elder people than50years in Qing Nian Xi Lu community, Qi Pan street community, Fo Shan Yuan community and Zheng Da city garden of Ji Nan city, the data of1580people were collected, the data included individual information, social factors, health knowledge, health state, chronic diseases, psychology state, PA. After statistical description and one-way analysis for the data, determine direct path and indirect path of influencing factors for regular PA. Statistical analysis methods included χ2test and χ2multiple comparisons, rank correlation,t test, factor analysis, PLS path model analysis, using statistical analysis software of SPSS16.0and SmartPLS2.0.Main Results1. current situation and one-way analysis of PA:Participation rate of regular PA for the elderly in community, about40percent have not regular PA. one-way analysis results show:PA is concerned with community, age, employment status, monthly income, not concerned with sex and standard of culture. Participation rate of regular PA in Qing Xi, Qi Pan Jie, Fo Shan Yuan, Zheng Da city parden is46.3percent,67.2percent,72.2percent,90.7percent separately. χ2value amounts to98.19, P value amounts to0.000between the four communities; χ2value amounts to1.33and P value amounts to0.249of Participation rate of regular PA between male group and female group;χ2value amounts to25.21and P value amounts to0.000of Participation rate of regular PA between age group, Participation rate is the lowest for the elderly more than80; χ2value amounts to2.12and P value amounts to0.832of Participation rate of regular PA between standard of culture group; χ2value amounts to17.78and P value amounts to0.000of Participation rate of regular PA between employment status group, in fact the difference is remarkable between working staff and retired people, and it’s P value amounts to0.000; χ2value amounts to14.89and P value amounts to0.005of comparison between monthly income groups, Participation rate of regular PA in less than1000yuan group is less than in from1000to1999yuan group.2. Qualitative analysis of influencing factors of PA:Some old people are not familiar with PA adequately, Other old people want to participate PA and wish for help, for example someone take the lead to organize PA, and wish for the support of family, friends, community. For the other community activity, the old people also wish for organizer such as neighborhood committees or other people that like to help them willingly. The old people all wish to participate community activities, the first reason is trust and dependence of the elderly for community, the second reason is that the further location is not convenient for the old people. The old people wish to learn health knowledge in chat, chanting with specialist and friends. Psychological problems of the elderly have been not resolved such as psychological counseling for the elderly in community. 3. Analyzing influencing factors of PA through PLS path model:The path coefficient between PA and certain latent variable such as community, age, community activity, mild activity, severe activity is0.13,0.15,0.12,0.13,0.13separately, and path coefficient is significant in the0.01level (t>t0012.999≈2.58);The path coefficient between PA and certain latent variable such as hypertension, community safe, contacting with direct relative is-0.07,0.09,0.08, and path coefficient is significant in the0.05level (t>t0.05,999≈1.96);The path coefficient between PA and certain latent variable such as employment status, depression, diabetes, society activity, hypertension knowledge, medical knowledge is0.06,-0.07,-0.06,0.06,-0.05,0.06, and path coefficient is significant in the0.20level (t>t0.202.999≈1.28).4. Analyzing the relationship between influencing factors of PA through PLS path model:Community safe is related to age, community, depression, community activity, and path coefficient is0.100,0.182,-0.154,0.190; Social activity is related to community activity, employment status, and path coefficient is0.449,-0.067; contacting with direct relative is not related to other latent variables, it’s effect is independent on PA; Community activity is relative to community, depression, medical knowledge, severe activity, community safe, social activity, and path coefficient is-0.145,-0.083,0.121,0.100,0.169,0.375; Depression is relative to community, severe activity, community safe, community activity, and path coefficient is-0.150,-0.228,-0.139,-0.107; Hypertension is relative to severe activity, and path coefficient is0.29; Diabetes is relative to hypertension knowledge, and path coefficient is0.100, the path coefficient is less explanation to variation of diabetes, it’s effect is independent on PA; hypertension knowledge is relative to diabetes, community, medical knowledge, but these latent variables have less explanation to hypertension knowledge, it’s effect is independent on PA; Medical knowledge is relative to age, community, severe activity, community activity, and path coefficient is-0.178,0.240,0.113,0.131; Severe activity is relative to hypertension, age, depression, medical knowledge, mild activity, community activity, and path coefficient is0.168,-0.184,-0.118,0.110,0.491,0.086; Mild activity is relative to severe activity, and path coefficient is0.610, but according to reality meaning, it’s effect is independent on PA; community is relative to depression, medical knowledge, community safe, community activity, and path coefficient is-0.167,0.248,0.191,-0.232; age is relative to severe activity, medical knowledge, and path coefficient is-0.372,-0.169; employment status is relative to severe activity, society activity, but these latent variables have less explanation to employment status, it’s effect is independent on PA.5. Some factors direct and indirect effecting on PA:Age has two indirect effecting paths on PA except it’s direct effect on PA, and they are age effecting on PA through community safe, age effecting on PA through severe activity; Community has three indirect effecting paths on PA except it’s direct effect on PA, and they are community effecting on PA through community activity, community effecting on PA through community safe, community effecting on PA through depression; Depression has two indirect effecting paths on PA except it’s direct effect, they are depression effecting on PA through community safe and community activity; Community activity has two indirect effecting paths on PA except it’s direct effect, they are community activity effecting on PA through community safe and society activity; Medical knowledge has two indirect effecting paths on PA except it’s direct effect, they are severe activity effecting on PA through community activity and depression.Conclusions and RecommendationsInfluencing factors are multidimensional for regular PA in city community, included social demography factors (community, age, employment status), social factors (community safe, social activity, contacting with direct relative, community activity), psychological factor (depression), state of health (severe activity), chronic state (hypertension, diabetes). Multidimensional Influencing factors are not independent when they influence PA of the elderly simultaneously, they influence each other. Some factors are not only direct action factors of PA but also intermediate variable of indirect effect of other factors, and some factors have not only direct effect but also indirect effect on PA.Recommendations:(1) PA Interventions should be the most important lesson of all health interventions and health management;(2) Community should be important platform of regular PA interventions for the elderly;(3) Personalized intervention of PA should be done for the elderly in community;(4) should be blended into psychological intervention for the elderly in community;(5) PA interventions for the chronic elderly should be regarded as the breakthrough of PA interventions for the elderly in city community;(6) Multidimensional PA interventions through community platform should be putted into effect.Innovations and LimitationsInnovations in this research:1. Innovations on determining the topicPA interventions is the most important measure to raise general level of the health for the elderly, it is also the most difficult and important health working in the world now. It is proposed that research on the influencing factors of PA should be based on Social Ecology theory for the first time at home, at the same time the influencing factors effect on PA was considered such as demography, society, psychology, health status, chronic diseases, health knowledge, and for the first time divide social factors into community safe, community activity, social communication, social support in order to analyze their effect on regular PA for the elderly. Interaction between the influencing factors was proposed for the first time at home after analyze effect of influencing factors on PA, and analyze and confirm the interaction, and finally determine the direct path and indirect path of these factors to PA.2. Innovations in Methods ApplyingAccording to complex relationship between influencing factors in this research, non-normal, multidimensional, multicollinearity data characteristic, for the first time apply PLS path model to select the influencing factors of PA, and apply the method to determine path relationship of the influencing factors to PA.3. Innovations in conclusionIt is known that the influencing factors of PA are multidimensional in this research, and there are complex path relationship between the factors, some factors have not only direct effect on PA but also indirect effect on PA through indirect variable, in fact form a complex influencing factors network, We should find out the breakthrough point according to the influencing factors characteristic for the different people, and intervene PA according to the breakthrough point.Limitations in this research:1. Because of investigation condition we have not divided environmental factors into different parts in community.2. We only investigate PA for the elderly of four communities in Ji Nan city, afterwards we should investigate PA for the elderly in other communities of other cities, and compare the results each other.更多还原