Factors associated with quality of life among elderly in Urban Vietnam

Nội dung text: Factors associated with quality of life among elderly in Urban Vietnam

1. JOURNAL OF MEDICAL RESEARCH FACTORS ASSOCIATED WITH QUALITY OF LIFE AMONG ELDERLY IN URBAN VIETNAM Dao Thi Minh An¹, Vu Toan Thinh¹, Dunne P Michael² ¹Institute for Preventive Medicine and Public Health, Hanoi Medical University ²School of Public Health, Queensland University of Technology, Australia. Quality of life (QoL) among the elderly is a big problem in Vietnam due to a growing proportion of the elderly in Vietnam while many conditions, including policies, social facilities, culture and other fac- tors are not ready to support for QoL among elderly. This cross-sectional study was conducted to explore QoL and factors associated with QoL among the elderly in Trung Tu ward, Ha Noi, Viet Nam. The findings showed that the four domains of QoL the among elderly fluctuated around 50.Mean scores of social and psychological QoL were higher than those in the physical and environmental domains. A statistically significant difference in mean scores of QoL by socio-demographics was re- corded (age profile, educational attainment, and occupation). All four domains of QoL were positively correlated with each other. Furthermore, age, psychological, social and environmental domains col- lectively contributed to 47.59% of the physical domain; while the physical, social, and environmental domains accounted for 56.13% of the psychological domain. We also found that occupation (work- er), as well as physical, psychological, and environmental metrics, accounted for 34.19% of the so- cial domain. Moreover, physical, psychological, social domains and occupation (home-wife) collec- tively accounted for 45.92% of the transformation of environmental domain. Our study suggests that it is essential to evaluate overall QoL to have a comprehensive view of its effects in the long run. Keywords: Quality of Life, Elderly, Hanoi, WHO QoL-Bref I. INTRODUCTION Vietnam’s population structure is in a pe- elderly population. Statistics from the Living riod of dramatic change, presenting a num- Standard Survey of Households in Vietnam ber of public health benefits as well as chal- showed that the number of elderly people lenges. Today, one of the most prominent (defined as men and women aged 60 years issues is how to address a rapidly growing plus) grew from 3.71 million people in 1979 (6.9% of the total population) to 7.72 million Corresponding author: Vu Toan Thinh, Institute for in 2009 (9% of the total population). At this Preventive Medicine and Public Health, Hanoi Medical rate, by 2020, it is estimated Vietnam’s el- University derly population will be greater than 12 mil- Email: vutoanthinhdhy@gmail.com lion [1]. Received: 05 June 2017 With this in mind, quality of life (QoL) Accepted: 16 November 2017 among the elderly is the most pressing is- 114 JMR 111 E2 (2) - 2018
2. JOURNAL OF MEDICAL RESEARCH sue. QoL is a multi-dimensional, highly relating to health problems, but most of all, subjective concept and, as recommended less integrated neighborhood relationships by the World Health Organization (WHO), compared to those in rural areas. Further, is measured using four major domains, in- after retirement, may confront psychological cluding physical, psychological, social, and loneliness, emptiness, and even abandon- environmental [2]. Within these categories, ment by their children and neighbors, this QoL has its own characteristics according to would put the elderly in isolated situations. different economic and socio-cultural levels, Hanoi is the capital of Viet Nam where producing trend where an overall negatively there is a rapidly developing economy and asociates with age QoL [3]. growing population in which many Vietnam- Within Vietnam’s cultural context of ese households have 2 to 3 generations live multiple generations living together in the together [7]. QoL of the elderly in Hanoi af- same household, as well as the impact of ter retirement is often influenced by many urbanization on a rapidly aging population, factors such as home economics, relation- QoL and mental disorders among the el- ship with their spouse and children, social derly need to be paid more attention. A re- issue, physical and mental health, and the cent study conducted in 8 provinces on the medical system [8 - 10]. However, few stud- health status of Vietnam’s elderly population ies have specifically analyzed the extent showed that about 95% of the participants that these factors impact QoL among the were infected with at least one disease. On elderly, especially among those living in ur- average an elderly person suffers from 2.6 ban wards in Hanoi. In Vietnam, there were diseases. With this in mind, about 23% of some studies conducted on QoL among the the elderly people have difficulties in their elderly [11]; however, none focused on the daily life, of which more than 90% need sup- population living in major cities. ports from other people [4]. According to the Therefore, this study aims to analyze the statistic of the National Institute of Geron- quality of life based on the four main do- tology, 9.2% of the Vietnamese population mains among the elderly population living in suffer from depression, one third of which Hanoi’s Trung Tu ward. were elderly and largely retired populations II. SUBJECTS AND METHODS in major cities [5; 6]. This is an important point to understand 1. Subjects in an age of rapid urbanization. The propor- Target population is the elderly living in tion of elderly in urban areas is quickly rising urban areas in Hanoi city. Particularly, the and becoming a far more difficult problem study population is defined as the elderly to properly address. Compared to the elder- living in Trung Tu ward, Hanoi. Participants ly living in rural areas, the elderly in urban who were recruited into this study if they met zones have distinct lifestyles such as exten- the following criteria 1) People who living in sive free time, more available information Trung Tu ward, Hanoi for at least 1 year; 2) JMR 111 E2 (2) - 2018 115
3. JOURNAL OF MEDICAL RESEARCH Aged ≥ 60 years old (according to the or- toll-free number for registration. After being dinance of the elderly, issued by the Presi- contacted by potential subjects, the sec- dent of the National Assembly on 28th April ond step was to screen them for eligibility 2000, the elderly are defined as citizens of using a questionnaire that assessed each the Socialist Republic of Vietnam from 60 participant’s recruiting criteria. They were years old or more [12]); and 3) Willing to then recruited into the study based on these participate in this study after giving informed criteria until the target sample size of 299 consent. Individuals were excluded if they elderly people was met. In the last step of were living in Hanoi temporarily, refused to sampling, collaborators contacted regis- participate, or had difficulties in understand- tered participants at home and provided ing or completing the questionnaire. them with consent forms. After reading the 2. Methods consent form, if the elderly agree to partici- pate in the study, they would then receive a Research site self-administered questionnaire from collab- This cross-sectional study was conduct- orators. They then allowed at least 2 weeks ed in Trung Tu ward, Hanoi, which is locat- for participants to complete their question- ed in Northern Viet Nam. This ward has one naires and return them to health collabora- of the densest populations in Hanoi and is tors in Trung Tu ward, either by themselves mainly comprised of government officers or their relatives. If their relatives delivered that live in 62 dormitories and 2 residential the questionnaire, it would be sealed in districts with convenient transportation and an envelope to ensure confidentiality. The close proximity to entertainment venues, self-administered questionnaires were im- national hospitals, and schools. Until 2012, mediately screened to check for missing there were 1,593 elderly people in Trung Tu, information to ensure participants could cir- accounting for 11.78% of the total popula- cle responses they missed. If their relatives tion of the ward. delivered their questionnaires, we used the Sample size and data collection telephone number which was recorded on This is a pilot study, so we decided on that questionnaire to call the elderly. After a convenience sample of 2% (or 299) of that, the participants' phone number was Trung Tu ward’s total elderly population, deleted to secure their personal informa- who volunteered for the study. The first step tion. If the elderly refused to answer, that of recruiting participants was effectively questionnaire was considered as ineligible. announcing the study. Ten health collabo- Measures rators of Trung Tu’s health center wrote an Demographics: Includes 7 questions introduction about the study and announced about participants’ age, marital status (mar- the recruitment on the boards at dwelling ried vs. unmarried), education level, living areas that they are in charge of. The an- arrangements, and occupation before re- nouncement ordered those who wanted to tirement. voluntarily participate in the study to call a 116 JMR 111 E2 (2) - 2018
4. JOURNAL OF MEDICAL RESEARCH Quality of Life: WHO QoL-Bref ques- points; and environmental domain are from tionnaire is self-assessment that antains 8 to 40 points. The raw scores of each do- 24 items, each presenting one facet of QoL main were then converted to a scale of 0 to and two “benchmark” items in an individ- 100 to compare with other populations, with ual’s overall QoL and general health. The lower scores indicating poor QoL. A domain facets are defined as those aspects of life was treated as missing when over 20% of that are considered to contribute to a per- its items were missing. With regard to QoL son’s QoL. QoL comprises of four main do- scores, they are on a positive scale (high- mains – physical health (7 items relating to er scores represent better QoL) and there pain and discomfort, dependence on medi- is no cut-off point to determine a specific cal treatment, energy and fatigue, mobility, score by which the QoL could be assessed sleep and rest, activities of daily living, and as “good” or “bad” [13]. working capacity), psychological health (6 Data analysis items relating to positive feelings, spiritual- Data had been cleaned by checking ity, religion and personal beliefs, thinking, missing data before it was entered into the learning, memory and concentration, body database. Data was entered and cleaned image, self-esteem, negative feelings), so- for outlier and illogical data using Epidata cial relationship (3 items relating to personal software, then converted into file.data to be relations, sex life, practical social support), analyzed in Stata version 10. and environment (8 items relating to phys- The results were initially analyzed using ical safety and security, physical environ- means, standard deviations, and frequen- ment, financial resources, information and cies. Mean and standard deviation were skills, recreation and leisure, home environ- used to assess normal distribution. Subse- ment, access to health and social care, and quently, Man-Whitney tests were employed transportation). These facets were scored to compare means between the four do- on a Likert scale from 1 to 5 with 1 = Very mains of QoL by socio-demographics. poor, 2 = Poor, 3 = Neither poor or good, 4 = The relationships between each domain Good, and 5 = Very good; 1 = Very satisfied, of QoL were identified by conducting Spear- 2 = Dissatisfied, 3 = Neither dissatisfied or man tests, since domains of QoL were not satisfied, 4 = Satisfied, and 5 = Very satis- normally distributed. To analyze the influ- fied; 1 = Not at all, 2 = A little, 3 = A mod- ence of independent variables of each do- erate amount, 4 = Very much, and 5 = Ex- main of QoL, bivariate and multiple linear tremely; or 1 = Never, 2 = Seldom, 3 = Quite regression analysis were used, in which often, 4 = Very often, and 5 = Always. The dependent variables were transformed into raw score from each domain of QoL include ranks because of the absence of normal varying scales; for instance, the physical distribution (physical and social variable domain ranges from 7 to 35 points; psycho- was squared to meet this condition). Some logical domain ranges from 6 to 30 points; socio-demographic factors (age, marital social domain scores ranges from 3 to 15 status, gender, occupation, education lev- JMR 111 E2 (2) - 2018 117
5. JOURNAL OF MEDICAL RESEARCH els, and living arrangement) and signifi- consent form with participants’ agreement cant factors in bivariate linear regression or to participate in the study and their admin- in literature documents were then put into istration group numbers was detached from multiple linear regression for the full model. the main body of the questionnaire and sent The final model was selected by performing to the principle investigator (PI) to be se- stepwise linear regression. The significance curely stored. Therefore, all individual infor- level adopted for statistical test was 5%. mation will be separate throughout the data Co-efficient, constant, p value, confidence collection procedure. Our approach was to interval and R-square for each model were ensure that participants feel that they have calculated and presented. control over the proceedings of the survey. The final model was tested for its fitness They were clearly advised that all informa- by 1) checking its linear predicted value tion is anonymous and will only be analyzed (_hat) and linear predicted value squared at the group level. In the consent form, the (_hatsq); 2) check goodness of fit ("predict PI’s contact number was printed and par- resid, r"; 3), by checking for multi-collinear- ticipants were instructed to if they have any ity. questions. If participants do become dis- 3. Ethics tressed during or after filling out the ques- tionnaire, they could also contact the PI for The risk of discomfort to participants and further counseling. risk of confidentiality loss were marginal. All survey questionnaires were anon- There were some questions about individ- ymous (no name and individual address ual feelings among the elderly about their identified) and securely stored. This study happiness with their life, family members, was submitted and approved by the Ethical sex life, and surrounding physical environ- Committee of the School of Public Health ment, as well as their social connectedness. and accepted in May, 2012. To reduce these risks, in the consent form, participants were advised that they can with- III. RESULTS draw at any time and that they can refuse Among the 299 participants, the propor- to answer any question which made them tion of males to females was balanced at uncomfortable. They were also advised that 48.8% and 51.2%, respectively. The mean all their refusal or withdrawal will not have age of study participants was 70.6 years, any effect on them in any way. Moreover, while the mean age of males was higher an anonymous self-administered question- than females (p < 0.05). The proportion of naire was developed and used, in which the elderly in the group under 70 years was can complete by participants without the 45.5% compared to these age 70 years and survey privately. Additionally, participants older 54.5%. The majority of participants were asked to return their completed ques- (40.6%) were post-graduation, working tionnaire by themselves to the field workers, as government officers (80.3%), married who are outside the participants’ wards. The (84.6%) and living primarily with their hus- 118 JMR 111 E2 (2) - 2018
6. JOURNAL OF MEDICAL RESEARCH band or wife and children (47.8%). Table 1. Mean scores of four domains of quality of life by socio-demographics Mean of scores (Mean ± SD)a Socio - demographic Social characteristics Physical Psychological Environment relationship Mean ± SD 53.4 ± 12.1 57.4 ± 11.3 60.4 ± 14.1 54.3 ± 11.3 Gender Male 52.9 ± 13.4 57.0 ± 11.8 60.2 ± 14.1 54.3 ± 11.7 Female 53.8 ± 10.9 57.7 ± 10.9 60.7 ± 14.2 54.2 ± 10.9 p value* 0.88 0.76 0.81 0.82 Age group < 70 56.7 ± 10.2 59.4 ± 10.6 62.7 ± 13.8 55.2 ± 11.3 >= 70 50.6 ± 13.0 55.7 ± 11.6 58.5 ± 14.1 53.6 ± 11.2 p value 0.0001 0.0042 0.0095 0.396 Marital status Single 53.6 ± 12.1 57.6 ± 11.4 60.6 ± 13.9 54.6 ± 11.1 Married 51.8 ± 12.5 56.2 ± 11.1 59.6 ± 15.1 52.7 ± 12.1 p value 0.26 0.57 0.79 0.60 Occupation Government officers 53.4 ± 12.2 58.0 ± 11.7 61.0 ± 14.1 55.1 ± 11.3 Others 53.1 ± 12.2 54.9 ± 9.5 59.3 ± 14.1 50.8 ± 10.7 p value 0.83 0.04 0.35 0.0138 Education College/Intermedi- 52.5 ± 12.1 55.7 ± 11.1 58.6 ± 14.7 52.5 ± 11.2 ate school and less Post-graduation 54.6 ± 12.1 59.9 ± 11.3 63.2 ± 12.9 56.9 ± 10.9 P value 0.13 0.0006 0.011 0.0012 Living arrangement Alone 51.5 ± 15.0 53.1 ± 13.3 59.7 ± 14.1 53.1 ± 12.8 Family 53.4 ± 12.0 57.5 ± 11.2 60.5 ± 14.1 54.3 ± 11.2 p value 0.47 0.18 0.77 0.89 JMR 111 E2 (2) - 2018 119
7. JOURNAL OF MEDICAL RESEARCH aScore in range from 0 - 100; *Man-Whit- ment officers and others (57.9 vs. 54.9; and ney test 55.1 vs. 50.8 with p < 0.05, respectively). The mean scores of four domains of QoL These differences were not seen in the fluctuated around 50 (table 1). Meanwhile, physical and social domains. The more the social domain had the highest score highly educated participants were, the bet- (60.4), followed by the psychological, phys- ter their QoL in psychological, social and ical, and environmental domain (57.4; 53.4; environmental domains (55.7 vs. 59.9; 58.6 and 54.3, respectively). We found that par- vs. 63.2; 52.5 vs. 56.9 with p < 0.05, respec- ticipants under 70 years had higher QoL in tively), however this was not the case in the physical, psychological and social domains physical domain. We did not find statistical- than those aged at or over 70 years of age ly significant differences in mean scores on (56.7 vs. 50.6; 59.4 vs. 55.7; and 62.7 vs. all four domains based on gender, marital 58.5 with p < 0.01, respectively). However, status and living arrangement (whom living this trend was not observed in the environ- with) (p > 0.05). mental domain. Additionally, statistically Interestingly, all domains of QoL were significant differences were found in the correlated positively with each other (p < psychological and environmental domains 0.001) (Figure1). Specifically, high correla- among the elderly who worked as govern- tions were identified between the physical (0.6), environmental (0.5), social (0.5), and psychological domains. Table 2. Factors associated with physical domain Number of obs 299 Model summary Prob > F 0.0000 R-squared 0.4759 Physical_QoL Coef. P > t [95% Conf. Interval] Age - 29.04 0.00 - 42.93 - 15.14 Psychological_QoL 48.80 0.00 36.66 60.95 Environmental_QoL 16.64 0.03 4.87 28.41 Social_QoL 9.44 0.01 1.82 18.06 Cons 771.47 0.22 - 46.09 20.03 For the physical domain (table 2), R-square equal 0.4759 (p < 0.001), meaning that age, psychological, social, and environmental domains contribute 47.59% to this facet of partici- pants’ QoL. All determinants were positively correlated except for age, which was inversely correlated. For every one unit increase in psychological, environmental and social domains, we would expect a 48.80; 16.64; and a 9.44 unit increase in the physical domain, respectively. The coefficient for age was 29.04, meaning that for a one unit increases with age; a 29.04 unit 120 JMR 111 E2 (2) - 2018
8. JOURNAL OF MEDICAL RESEARCH decreases in physical domain. Table 3. Factors associated with psychological domain Number of obs 299 Model summary Prob > F 0.0000 R-squared 0.5613 Psychological_QoL Coef. P > t [95% Conf. Interval] Physical_QoL 0.36 0.00 0.27 0.44 Social_QoL 0.14 0.00 0.06 0.21 Environmental_QoL 0.36 0.00 0.26 0.45 Cons 10.57 0.00 5.71 15.45 For psychological domain (Table 3), we found that physical, social, and environmental domains were positively correlated with psychological domain, which collectively accounted for 56.13% (p < 0.001). The domain that contributed the most to psychological domain were physical and environmental (whose coefficient was 0.36, meaning that the psychological do- main increases 0.36 ranked units, p < 0.001), followed by social (whose coefficient was 0.14, meaning that the psychological domain increases 0.14 ranked units, p < 0.001). Table 4. Factors associated with social domain Number of obs 299 Model summary Prob > F 0.0000 R-squared 0.3419 Social_QoL Coef. P > t [95% Conf. Interval] Physical_QoL 19.81 0.02 3.17 36.45 Psychological_QoL 41.30 0.00 21.51 61.09 Environmental_QoL 30.91 0.00 13.11 48.71 Workers 539.23 0.04 32.01 106.44 Business man - 61.64 0.23 - 160.91 385.63 Freelance worker - 38.70 0.30 - 119.02 363.61 Home wife 63.43 0.91 - 100.85 113.73 Others - 50.91 0.51 - 207.25 109.43 Cons - 124.25 0.01 - 211.56 - 36.94 For the social domain (Table 4), occupation (worker), physical, psychological, and environ- mental domains were positively correlated and together accounted for 34.19% (p < 0.001). The coefficient for occupation was 539.23, meaning that the elderly individulas who worked JMR 111 E2 (2) - 2018 121
9. JOURNAL OF MEDICAL RESEARCH as workers have a score of 539.23 ranked units greater than those who were government offi- cers; the coefficient for physical, psychological and environmental domain was, in turn, 19.81; 41.30 and 30.91, meaning that a one unit increases in physical, psychological or environmen- tal domain produces a 19.81; 41.30 and 30.91unit increase in the social domain, respectively. Table 5. Factors associated with environmental domain Number of obs 299 Model summary Prob > F 0.0000 R-squared 0.4592 Environmental_QoL Coef. P > t [95% Conf. Interval] Social_QoL 0.14 0.00 0.06 0.22 Psychological_QoL 0.44 0.00 0.32 0.56 Physical_QoL 0.14 0.01 0.03 0.25 Worker - 3.11 0.06 - 6.34 0.13 Businessman 1.34 0.68 - 5.02 7.71 Freelance worker - 0.37 0.88 - 5.13 4.40 Home wife - 10.22 0.00 - 17.11 - 3.33 Others - 4.14 0.40 - 13.81 5.52 Cons 13.38 0.00 7.85 18.90 Data from Table 5 shows factors associ- IV. DISCUSSION ated with the environmental domain. Physi- We found that QoL scores of the elder- cal, psychological, social, and occupational ly living in Trung Tu ward fluctuated around (home-wife) determinants together account- 50 and compared to the maximum score in ed for 45.92%. The physical, psychological, the 0-100 scale, they presented a moderate and social domains were positively correlat- QoL level for the four domains of WHO QoL- ed with the environmental domain and the Bref (table 1). These results are very similar correlation coefficient of these domains to other studies on QoL among the elderly were 0.14, 0.44, and 0.14, respective- in Brazil [13] and two studies conducted in ly, meaning that for a one unit increase in Can Tho and Ho Chi Minh city, Viet Nam, physical, or psychological, or social domain, which indicated that the QoL of people aged we would expect that a 0.14, 0.44, and 0.14 at 18 and over stayed at moderate level [14]. unit increase in the environmental domain. These similarities in QoL between these lo- Working as a homemaker was inversely re- cations can be explained by rapid economic lated and its coefficient was 10.22, meaning development and urbanization. However, that elderly with working as homemakers the average scores of all four domains of have a score of 10.22 ranked units lower QoL in this study were lower than findings than those with government officers. detected in other developing countries, 122 JMR 111 E2 (2) - 2018
10. JOURNAL OF MEDICAL RESEARCH such as among the elderly living in South on QoL of the elderly [18; 19]. The higher Jakarta (Indonesia), in Taiwan (2010), and age was, the lower QoL on physical, psy- in adults with sickle cell disease in Jamaica chological and social domains (Table 1). , as well as in France among people age 80 These results are similar to the findings by and patients after intensive care unit [15]. Barua et al. in 2007, Abhay Mudey et al. in In this study, the mean scores of physical 2011 [2], Abdul Rashid in 2013 [20], Phung and environmental domain were lower com- Duc Nhat et al. in 2011 [14]. As seen in pared to the psychological and social do- García et al., old age was associated with main (Table 1). These findings were similar the worst levels of health-related to QoL. to the results of other studies and indicated Likewise, Laxmikant Lokare’s study in 2011 that social domain had the highest mean indicated that the mean score in the age score when compared to other domains [8; group of under 70 years old and above 70 16]. In a study conducted on 240 partici- years old were significantly differences in pants, Sanghee Chun et al. also indicated the psychological domain (p < 0.05) [21]. that environmental and psychological do- We found that those with higher educa- mains had higher mean scores compared tion level attained better QoL. This finding to physical and social domains (78.9; 74.2 supports a study conducted in Can Tho vs. 73.4; 65.6, respectively) [17]. Likewise, city, Vietnam, which indicated that people a Vietnamese study performed by Phung aged 18 years or over with the highest lev- Duc Nhat et al. also showed this trend [14]. . el of education had better QoL on all four We suggest that the elderly in Trung Tu domains compared to the lower educated ward have a lower perception of their QoL participants [14]. In a study by Ping Xia et in the physical domain. This was indicated al., participants who had a degree, voca- by their self-reported pain and discomfort, tional training or above had mean scores in medicine dependence, energy and fatigue, all domains higher than those without (p < issues related to mobility capability, as well 0.001) [16]. A study conducted on 205 el- as sleeping and rest, activities of day-to- derly in Malaysia indicated that the elderly day life, and working abilities. This was also who had secondary school level education the case of it environmental QoL, which in- had higher QoL as compared to those with cludes a diversity of physical security; sup- primary level or no education (26.7% vs. ports for finance; information sources and 21.5% and 2.2%, with p < 0.01, respec- skills; entertainment; housing environment; tively). Likewise, the elderly who worked as accessibility to health services and social government officers had better QoL than care; and transportation as well. This high- other participants. This result supports pre- lights the importance of improving elderly’s vious studies indicating that the elderly who physical and environmental QoL via urging were employed had 22.6% of higher level of them to participate in clubs and recreational QoL when comparing to those who were not activities while accessing to health services. (13.4%) [22]. Additionally, a study conduct- Several studies showed the effect of age ed in Nonthaburi, Thailand revealed that the JMR 111 E2 (2) - 2018 123