Subjective Happiness , Health and Quality of Life and Their Sociocultural Correlates among Younger Population in Malawi

Public health research and policy in sub-Saharan Africa are generally disease-oriented, with the focus being largely confined within the biological determinants of health. So far, little attention has been given to developing a more health-oriented approach by emphasising the psychosocial dimensions of health, especially among the younger population. To this regard, we conducted the present study to assess the prevalence and sociocultural correlates of perceived happiness, health, and life satisfaction among the adolescent and young (15–24 years) population in Malawi. We analysed cross-sectional data on 12,610 men and women based on a Malawi multiple indicator cluster survey conducted in 2013–2014. Data were analysed using descriptive and multivariable regression methods. According to the findings, more than 80% of the men and women reported being satisfied about happiness, health, and life. Multivariate analysis showed an inverse relationship between being currently or formerly married and perceived happiness. Ethnic disparities in perceived health and happiness were more pronounced in men, whereas that of life satisfaction was more pronounced in women. Living in households of the highest wealth quintile was positively associated with health and life satisfaction, but not with happiness. These findings highlight the need for prioritising the psychosocial needs of the adolescent and youth populations in designing health and social policy in Malawi. The findings need to be interpreted in light of the factors specific to the sociocultural environment in Malawi.


Introduction
The concepts of the social determinants of health (SDH) and life outcomes have been gaining widespread attention in various disciplines, including sociology, environment, public health, and international development (Ngamaba and Soni 2018;Amit et al. 2012;Becker et al. 2009;Benjamin et al. 2012;Mail Online 2010).Led by the growing understanding of the nonbiological determinants of health, researchers and clinicians are placing higher emphasis on the factors embedded within the environment in which individuals live and function (Braveman and Gottlieb 2014).The importance of socioeconomic, environmental, and cultural factors on health outcomes are better understood in terms of their role in providing the necessary conditions required for maintaining a healthy lifestyle, avoiding risk factors, and being able to seek proper care when facing health issues (Zarini et al. 2014;Bishwajit et al. 2017;Gate et al. 2016).These factors are broadly classified to encompass the social, cultural, behavioural, and environmental aspects of health, quality of life, and well-being, especially among the adult population.However, unlike clinical diagnosis, the psychosocial aspects of health are rather subjective and not possible to capture by objective measurement (Bishwajit et al. 2017).This difficulty was greatly dealt with by the development of the single-item question by the World Health Organization (WHO).Self-rated health (SRH) is now a widely used measure to capture the overall sociocultural condition of individual health and has been found to be a strong predictor of health condition and mortality among individuals living with and without disability and chronic conditions, such as cardiovascular diseases and functional ability (Zarini et al. 2014;Idler and Kasl 1995;Idler and Benyamini 1997;DeSalvo et al. 2006).
Apart from that, SRH is also strongly associated with quality of life (QoL), which is a measure of general well-being that reflects an individual's satisfaction with health, education, employment, living and social environment, relationships, and safety (Zarini et al. 2014;Idler and Kasl 1995;Idler and Benyamini 1997;DeSalvo et al. 2006;Hays et al. 2015;Balogun et al. 2017).Over the course of the last few decades, both SRH and QoL have been growing in popularity, especially among public health researchers in high-income countries.However, the two constructs have remained largely underappreciated in the low-middle-income settings.The relatively higher disease burden, shorter life expectancy, and lower living standards in developing countries can be interpreted in terms of sociocultural and environmental determinants, such as age structure, gender disparities, low literacy rates, living environment, occupation and economic freedom, and the values and norms that guide peoples' beliefs and behaviour (Agorastos et al. 2014;Galavotti and Schnell 1994;Tsai et al. 2017;Development Impact Guest Blogger 2016;Decker and Constantine 2011;Gooneratne and Vitiello 2014).For instance, individuals with a higher socioeconomic status are more likely to be able to afford better housing, lead a healthy lifestyle, and enjoy higher quality of health and life compared to those in the lower socioeconomic strata (Zarini et al. 2014;Bishwajit et al. 2017;Amo-Adjei and Kumi-Kyereme 2015).So far, only one longitudinal study conducted in a rural setting in Malawi from the perspective of subjective well-being found that higher income was associated with better general health status and subjective well-being of the participants (Chin 2010).
The underlying sociocultural disparities in health, diseases, and quality of life are well-documented in many high-income countries.By contrast, health research and care delivery in developing countries have been concerned mainly with clinical outcomes, with little attention on the psychosocial aspects of health, such as life satisfaction and perceived sense of happiness.There are certain countries, such as Bhutan (Sithey et al. 2015) and Costa Rica (Rosero-Bixby and Dow 2016), where public policy is oriented towards developing social capital and promoting general happiness.However, research evidence on these constructs from African countries is still scarce.To this end, we carried out the present study on Malawi, one of the least developed of all African countries (Yaya et al. 2016), to investigate the situation of perceived health, life satisfaction, and happiness.Our secondary objective was to measure whether these three constructs vary across gender, demographic, and sociocultural factors.Due to lack of first-hand data, we relied on secondary extracted data from the most recent multiple indicator cluster survey conducted in Malawi (MICS 2013(MICS -2014)), which included a special section on subjective well-being among men and women aged between 15 and 24 years.As a poverty-and internal conflict-ridden country, Malawi is facing significantly challenges to invest on human development programmes and improving people's quality of life.

Data Source
Data for this study were sourced from the fifth round of the Malawi multiple indicator cluster survey conducted in 2013-2014 (MICS 5) (MICS 5).Field work for the survey was started in November 2013 and lasted until April 2014.The MICS programme operates in about 108 countries and over 280 surveys have been carried out since its inception in 1995.While known as being the most comprehensive and reliable source of data on maternal and child health issues in developing countries, in recent years, the surveys have been collecting data on issues such as living conditions and perceived and quality of life among men and women aged between 15-24 years.
Sampling technique: As indicated by the name, MICS surveys employ multistage sampling techniques that involve systematic selection of primary sampling unites (PSUs) and then selections of certain number of households from the PSUs.Participants are approached in their households and interviewed face-to-face by trained surveyors.In 2013-2014 MICS, 24,230 women aged between 15 and 49 years were interviewed in total, with a response rate of 95.3%.However, data on the outcome variables used in this study were collected among participants aged up to 24 years only; therefore, the rest were not included in the analysis.Detailed methods of the survey, such as questionnaire building and field work, were published elsewhere (Yaya et al. 2016).

Outcome Variables
The dependent variables included in this study were subjective: (1) Estimation of happiness, (2) satisfaction with health, and (3) life satisfaction.These were assessed based on the following questions: Answering the questions on estimation of happiness and life satisfaction was facilitated by displaying a card to the respondents with smiling faces corresponding to the degree of response categories: 'very satisfied', 'somewhat satisfied', 'neither satisfied nor unsatisfied', 'somewhat unsatisfied', and 'very unsatisfied'.For happiness estimation, similar questions have been used in previous surveys, e.g., the World Values Survey, Euro-Barometer Survey, and U.S. General Social Survey (Benjamin et al. 2012):

•
All things considered, how satisfied are you with your life as a whole these days?

•
Taken all together, how would you say things are these days-would you say that you are very happy, pretty happy, or not too happy?
Similar approaches have been used for assessing life satisfaction as well (Cheung and Lucas 2014): • All things considered, how satisfied are you with your life as a whole these days?

•
In general, how satisfied are you with your life?• Overall, how satisfied are you with your life nowadays?
Empirical studies have shown that these single-item life satisfaction questions perform very similarly compared with the Satisfaction with Life Scale (SWLS), as they produce virtually identical answers to substantive questions regardless of the measure used (Cheung and Lucas 2014).
Regarding health status, the personal rating of general health during the survey or the recent time is more commonly used in the public health literature (Hirve 2014;Hays et al. 2015):

•
In general, how would you rate your health, today?

•
In general, how would you rate your physical health?
For this study, all three dependent variables were dichotomised in the following way: (1) Satisfied-Very satisfied, somewhat satisfied, and (2) Not satisfied-Neither satisfied nor unsatisfied, somewhat unsatisfied, Very unsatisfied (Abdulrahim and Asmar 2012;He et al. 2018).Happiness was categorised as: Happy-Very happy, somewhat happy, and Not happy-neither happy nor unhappy, somewhat unhappy, very unhappy.

Predictor Variables
Selection of explanatory variables was guided by a thorough literature review on subjective well-being in PubMed.The review suggested the involvement of factors across several domains of health and wellbeing: Demographic (e.g., age, sex), sociocultural (e.g., religion, ethnicity), financial (e.g., income sufficiency), behavioural (e.g., dietary factors, smoking, drinking), and relational (e.g., living arrangement, quality of marriage).For this study, it was not possible to include variables from all these domains, as subjective well-being was not the main focus of the survey.Based on the availability of the variables on the dataset, the following were included in the analysis: [20][21][22][23][24] (Ngamaba and Soni 2018;Bishwajit et al. 2017;DeSalvo et al. 2006;Balogun et al. 2017;Chin 2010;Cheung and Lucas 2014;He et al. 2018;Son et al. 2016;Ghose 2017).
Satisfaction with living condition was dichotomised in the following way: Not satisfied (Dissatisfied/Very dissatisfied/Neutral) and Satisfied (Very satisfied/satisfied).As there was no direct information on income, we used a household wealth index, which is a composite indicator of household wealth status.The index is constructed by principal components analysis of scores assigned on individual household possessions, e.g., consumer goods, dwelling characteristics.Households are then ranked based on individual scores to range between poorest, poorer, middle, richer, and richest (Son et al. 2016).Media use was measured by self-reported use of TV and radio.

Data Analysis
Data analyses were performed using Stata 14.The main inclusion criterion for this study was availability of information on the outcome variables (subjective health, happiness, and quality of life).Datasets were first checked for missing values and outliers.Survey weight provided on the dataset was used before conducting the analyses to ensure data are representative of the entire population.Background information about the participants is presented using descriptive statistics.The five categories of happiness, health, and life satisfaction were merged into two groups before analysis.Percentages of perceived happiness, health, and life satisfaction were presented along with 95% CI for all the predictor variables.Ungrouped proportions of perceived happiness, health, and life satisfaction were also calculated and presented as bar charts.These analyses were stratified by age and sex to provide a better understanding of the demographic patterns in the distribution of the outcome variables.Chi-square bivariate tests were used to assess the statistical significance between the three outcomes with the predictor variables.Variables that showed association at 0.25 at bivariate level were retained for multivariable analysis (Ghose 2017).
The relationship between the three outcome and explanatory variables were measured by multivariable regression methods.Given the dichotomous nature of the outcome variables, a binary logistic regression model was used to generate the odds ratios and their 95% confidence intervals.The variance inflation factor (VIF) was used as a measure of collinearity to ensure that none of the predictor variables in the final model was highly associated with each other.All statistical tests were two tailed and p values below 0.05 were considered statistically significant (Except for the cross-tabs).

Ethical Approval
All participants gave informed consent prior to taking part in the interviews.Data were open-access and available online in anonymised form; therefore, no additional approval was necessary.

Descriptive Results
The survey population comprised 12,610 men and women aged between 15 and 24 years.The majority of the participants were female (77.4%) and aged below 20 years (54.8%).The basic sociodemographic characteristics are presented in Table 1.
Table 1 also shows that percentages of being positive/satisfied with happiness, health, and life were higher among those aged 15 to 19 years, female, who never married, who attended school, and who had children.Those who expressed satisfaction with life were also more likely to be rural residents, never married, residents of the Southern region, satisfied with living condition, followers of Christianity, have a higher wealth status (except for health), of Chewa ethnicity, with higher education (primary and secondary), no children, have never smoked tobacco, never drunk alcohol, and be non-users of electronic media.

Prevalence of Satisfaction with Happiness, Health and Life
Figure 1 illustrates the overall proportion of satisfaction with happiness, health, and life.Nearly 70% of the participants expressed high satisfaction (very satisfied), with less than 5% reporting dissatisfaction about all three.

Prevalence of Satisfaction with Happiness, Health and Life
Figure 1 illustrates the overall proportion of satisfaction with happiness, health, and life.Nearly 70% of the participants expressed high satisfaction (very satisfied), with less than 5% reporting dissatisfaction about all three.Figure 2 presents the prevalence of perceived happiness stratified by age and sex.The percentage of reporting "Very happy" was higher among women of both age groups.Percentages of reporting "very unhappy" were higher among women aged 15-19 years, whereas the percentage was higher among men in the 20-24 age group.These differences were statistically significant (p < 0.05).
Figure 3 presents the prevalence of health satisfaction stratified by age and sex.Unlike for happiness, the percentage of reporting "Very satisfied" was higher among men in both age groups.However, the percentages of reporting "Somewhat unsatisfied" and "Very unsatisfied" were somewhat similar higher among men and women in both age groups.These differences were not statistically significant (p > 0.05).
Figure 4 shows the prevalence of life satisfaction stratified by age and sex.Similar to satisfaction about health, men had slightly higher percentages of reporting "Very satisfied" compared to women in both age groups.Women also had higher percentages of reporting "Very unsatisfied".However, these differences were not statistically significant (p > 0.05).Figure 2 presents the prevalence of perceived happiness stratified by age and sex.The percentage of reporting "Very happy" was higher among women of both age groups.Percentages of reporting "very unhappy" were higher among women aged 15-19 years, whereas the percentage was higher among men in the 20-24 age group.These differences were statistically significant (p < 0.05).
Figure 3 presents the prevalence of health satisfaction stratified by age and sex.Unlike for happiness, the percentage of reporting "Very satisfied" was higher among men in both age groups.However, the percentages of reporting "Somewhat unsatisfied" and "Very unsatisfied" were somewhat similar higher among men and women in both age groups.These differences were not statistically significant (p > 0.05).
Figure 4 shows the prevalence of life satisfaction stratified by age and sex.Similar to satisfaction about health, men had slightly higher percentages of reporting "Very satisfied" compared to women in both age groups.Women also had higher percentages of reporting "Very unsatisfied".However, these differences were not statistically significant (p > 0.05).

Multivariate Regression Analysis
Results of multivariate regression analysis for happiness, health, and quality of life are presented in Tables 2-4, respectively.From Table 2, the odds of reporting positive estimation of happiness were lower among ever married men and women compared to those who were never married.Women in the Southern region had higher odds of reporting satisfaction with happiness compared to those in the Northern region (OR = 1.419, 95% CI = 1.107, 1.819).Men belonging to Ngoni ethnicity had 1.624 times (95% CI = 1.070, 2.464) higher odds of satisfaction with happiness compared to those belonging to Chewa.Not drinking alcohol (OR = 1.140, 95% CI = 1.020, 1.761) and media use (OR = 1.273, 95% CI = 1.082, 2.285) showed higher odds satisfaction with happiness among men, but not among women.

Multivariate Regression Analysis
Results of multivariate regression analysis for happiness, health, and quality of life are presented in Tables 2-4, respectively.From Table 2, the odds of reporting positive estimation of happiness were lower among ever married men and women compared to those who were never married.Women in the Southern region had higher odds of reporting satisfaction with happiness compared to those in the Northern region (OR = 1.419, 95% CI = 1.107, 1.819).Men belonging to Ngoni ethnicity had 1.624 times (95% CI = 1.070, 2.464) higher odds of satisfaction with happiness compared to those belonging to Chewa.Not drinking alcohol (OR = 1.140, 95% CI = 1.020, 1.761) and media use (OR = 1.273, 95% CI = 1.082, 2.285) showed higher odds satisfaction with happiness among men, but not among women.
Table 3 shows women who were formerly married had lower odds (OR = 0.687, 95% CI = 0.537, 0.879) of reporting health satisfaction compared to those who were currently married.Compared to the Northern region, being located in the Central (OR = 1.620, 95% CI = 1.246, 2.105) and Southern (OR = 1.904, 95% CI = 1.483, 2.444) region showed higher odds of health satisfaction among women.Living in the highest wealth quintile (Q5) households was associated with higher odds of health satisfaction among both men (OR = 1.960, 95% CI = 1.261, 3.044) and women (OR = 1.523, 95% CI = 1.188, 1.952).Compared to men in Chewa, those in other ethnic groups had higher odds of health satisfaction (except for Tumbuka).Women who had higher education (OR = 4.627, 95% CI = 1.520, 14.08) and had no children (OR = 1.382, 95% CI = 1.098, 1.740) also had higher odds of reporting health satisfaction.
Table 4 shows that being in the 20-24 year group was associated with lower odds of life satisfaction and among men only (OR = 0.731, 95% CI = 0.535, 0.999).Compared to men, women had lower odds of reporting life satisfaction (OR = 0.696, 95% CI = 0.592, 0.818).Compared to the Northern region, being located in the Southern (OR = 0.707, 95% CI = 0.558, 0.897) region showed lower odds of life satisfaction among women.Women living in the highest wealth quintile households had higher odds of reporting life satisfaction (OR = 1.563, 95% CI = 1.037, 2.999).Compared to Chewa women, those in the Lonwe and Ngoni ethnic groups were also more likely to report life satisfaction.

Discussion
The present analysis generated several interesting findings.Firstly, the descriptive analysis showed that more than four-fifths of all participants expressed satisfaction with their happiness, health, and life.The relatively higher proportions of health and life satisfaction are quite remarkable given the level of human development in the country.As of 2017, the per capita GDP was estimated to be 338.48USD, which is one the lowest globally.The country also performs low in terms of human development index, ranking 170 out of 188 countries and territories in 2017 (AfricaNews 2017).
The relationship between the level of national wealth and happiness has been a topic of widespread attention among researchers.In the Western world, the continued economic prosperity and rise of strikingly unequal societies has renewed the debate on whether or not material wealth is actually translating to better health and well-being.In high-income countries, human development policies have been more focused towards enhancing economic prosperity, population health, quality of life, and overall well-being.In resource-poor countries, on the other hand, the psychosocial aspects of health have so far failed to attract adequate funding and policy attention, especially in African countries.However, in order to foster a more equitable and people-centric development approach, greater investment in psychosocial research is necessary to understand the underlying determinants of health, happiness, and quality of life among the burgeoning population in the African continent.
Our second main result is the disparate association between happiness, health, and life satisfaction with socioeconomic factors.Quite contrary to expectation, satisfaction with happiness did not vary significantly across the wealth quintiles.Higher wealth status also did not consistently result in health and life satisfaction, meaning that the relationship between material well-being and these constructs is not necessarily a straightforward one.Surprisingly, education also did not show any noticeable association with the outcomes, except for health satisfaction.With regard to health satisfaction, the protective effect of wealth status was observed for the highest wealth quintile only.Findings from previous studies reflect a cyclical relationship between health and economic prosperity, as lack of wealth can lead to poorer health, which in return causes a decline in financial well-being through low productivity and increased medical expenditures (Guttmann 2001;Murray 2006;Wagstaff 2002;Cheng et al. 2016).
Intuitively, being financially well-off is a good predictor of higher living standards and providing the preconditions for better health, happiness, and well-being among individuals (Wagstaff 2002).Contrary to some studies, the present study found no clear relationship between media use, smoking or drinking status with happiness, health and life satisfaction (Staccini and Douali 2014;Lew et al. 2018;Welch et al. 2016).One possible explanation behind this may be the lower prevalence of smoking and drinking and decreasing popularity of TV and radio as modes of entertainment among youth.Regarding sociocultural factors, religion did not show any significant association with any of the outcome measures.However, we observed that ethnic disparities in perceived health and happiness were more pronounced in men, whereas that of life satisfaction was more pronounced in women.These contrasting findings are suggestive of the fact that perceived physical and psychosocial well-being at population level can vary greatly depending on the cross-cultural contexts and hence should be given special priority in meeting the health and social needs of the adolescent and youth populations.
Another apparently counterintuitive finding was the inverse relationship between marital status and happiness and health.Having children also did not show any protective effect on the outcomes.In both sexes, the odds of reporting happiness were significantly lower among the currently unmarried participants compared to those who are married.Literature on health and happiness from a marital perspective suggests that currently married individuals report better health status and happiness and share a higher mortality risk than their unmarried counterparts (Robards et al. 2012;Ngamaba and Soni 2018;Bookwala 2011;Chung and Kim 2014).However, the findings may be confounded by the fact that we had no information regarding the quality of marriage.Unhealthy conjugal relationships have been shown to have adverse effects on various aspects of health (Lapate et al. 2014;Choi and Marks 2008;Chen et al. 2015).
Another possible reason might be that the participants were predominantly young, with the majority aged below 20 years.Thus, it is assumable that marital status among the younger population may not result in better health and happiness.The diminishing role on better health outcomes in the context of progressive industrialisation and modernisation have been reported in several Asian countries as well (Fu and Noguchi 2016).As countries in Africa continue to embrace modern lifestyles and culture characteristics of Western countries, exploring the dynamics of family, marriage, and fertility can be rewarding in terms of promoting evidence for better psychosocial health of the population.
Previous studies have shown subjective health and quality of life outcomes among older population and people living with disease conditions (Bishwajit et al. 2017;Klotz et al. 2018;Matsuguma et al. 2018;Calys-Tagoe et al. 2014;Ryan et al. 2007).Currently, there are no other studies showing the predictors of subjective health, happiness, and life satisfaction among healthy adolescent populations in an African context.A cross-sectional study involving elderly people in Ghana reported that age, sex, educational level, income, and ethnic background were significantly associated with subjective well-being (Calys-Tagoe et al. 2014).A similar study on the elderly population in Nigeria found that socioeconomic status was the biggest indicator of quality of life (Gureje et al. 2008).In South Africa, the strongest predictor of a poor QoL was reported to be psychiatric morbidities, such as depression and anxiety (Mapatwana et al. 2018).
In the traditional model of care, health and illness are generally assessed based on clinical measures that fail to account for the psychosocial dimensions of health (Tuchtenhagen et al. 2015).This is particularly the case in the resource-poor countries like Malawi, where public health remains a largely underappreciated issue on the national development agenda.With the growing understanding of the social determinants of health, however, psychosocial constructs such as quality of life and well-being are getting increasing attention from health economists and policy makers.For instance, individual perceptions of health-related quality of life (HRQoL) are used as a multidimensional concept characterising the overall environmental, social, physical, emotional, and cognitive aspects of life.From this perspective, the findings of our study can be of special importance among health and social researchers, especially in the low-resource settings in Africa.As mentioned earlier, the data were not particularly suitable to explore the predictors of complex constructs, such as happiness and life satisfaction.Nonetheless, the present study is an important contribution to the current literature and also sets the avenue for more in-depth and cross-cultural studies in the areas of subjective well-being in Africa.
Study limitations: Apart from the contribution, we have several important limitations to declare as well.Firstly, the data were not primary, and as such, we were unable to have any control over the selection and measurement of the variables.Lack of data on some key indicators, such as parental background, satisfaction about marriage and sex life, social relationships, leisure-time, safety, and occurrence of health conditions, is perhaps the most important limitation of the study.Secondly, the data were self-reported, which means that the answers are subject to reporting bias.Young people are generally more sensitive about maintaining their self-image and end up giving answers that they find more socially appropriate.Secondly, the relationship between health and life satisfaction with health behaviours such as smoking and drinking is likely to vary by the level of consumption.However, we are unable to measure such nuances of smoking and drinking habits and therefore cannot ensure the robustness of these associations.It is therefore suggested that the findings be interpreted in light of the local contexts that influence these behavioural factors.Further, the sample consisted only of members of the population aged up to 24 years, and hence, the results are not generalisable for the adult population in the country.Lastly, the cross-sectional nature of the data precludes making any causal inference between the explanatory and response variables.

Conclusions
Our findings suggest that more than four-fifths of Malawian adolescents and youth reported being satisfied with their happiness, health, and (quality of) life.We also observed the absence of any progressive and positive relationship between happiness, health, and life with some of the most commonly entrusted predictors of general well-being, such as marital status, education, and household wealth status.These findings imply that improving socioeconomic status alone may not result in better subjective well-being and should be accompanied with more conducive environmental and sociocultural conditions congenial to higher living standards in psychosocial health among the younger populations.More in-depth surveys are required to examine the underlying meaning and to test the robustness of the present findings.

Figure 1 .
Figure 1.Proportion of self-reported happiness, health, and life satisfaction.

Figure 1 .
Figure 1.Proportion of self-reported happiness, health, and life satisfaction.

Figure 2 .
Figure 2. Proportion of men and women by self-reported happiness.

Figure 3 .
Figure 3. Proportion of men and women by self-reported health status.Figure 3. Proportion of men and women by self-reported health status.

Figure 3 .
Proportion of men and women by self-reported health status.

Figure 4 .
Figure 4. Proportion of men and women by self-reported life satisfaction status.

Figure 4 .
Figure 4. Proportion of men and women by self-reported life satisfaction status.

Table 2 .
Predictors of perceived happiness among men and women.Malawi multiple indicator cluster survey (MICS) 2013-2014.