Income Inequality in Quality of Life among Rural Communities in Malaysia: A Case for Immediate Policy Consideration

Quality of life (QOL) is a proxy of health and social well-being. Hence, it is vital to assess QOL as it informs the strategies of policymakers to enhance the living conditions in communities. Rural areas in emerging economies are underserved in terms of modern facilities and technologies, which impact QOL. To address this, this study investigated whether income played a role in the QOL of rural residents within emerging economies using a large survey of Malaysian adults above 18 years old. The study extracted data from a sample of 18,607 respondents of a health and demographic surveillance system survey. A generalized linear model was used to estimate the impact of three income groups, the bottom 40%, middle 40% and top 20%, on perceived QOL, controlling for sociodemographic, chronic disease co-morbidities and mental health status. Results of the study showed a statistically significant association between income and the physical, psychological, social and environmental QOL domains. Using the bottom 40% as a reference category, the middle 40% and top 20% income groups showed a significant and positive association across the four domains of QOL. Hence, intervention programs are necessary to escalate the income levels of rural communities, especially the bottom 40%, to uplift perceived QOL among rural residents.


Introduction
Across the globe, rural populations are overwhelmingly poor [1][2][3][4], which significantly impedes overall life satisfaction. A staggering 580 million (79%) of the world's poor reside in rural areas [2,5]. In developing countries, living in rural households increases the odds of being poor compared to urban counterparts [5,6]. Malaysia successfully attenuated its rural poverty rate to 1.0% in 2016, reduced from 58.7% in 1970 [7]. Although the average household income rose by twenty-fold from MYR 200 in 1970 to MYR 4359 in 2016 in rural areas, it has consistently remained lower than in urban areas over this period [7]. This is evidenced by the high share of the bottom 40% (B40) income group in rural areas (44.0%) [8], where a majority are heavily dependent on agricultural outputs which have low income returns [9].
Quality of life (QOL) often serves as a critical indicator of the government policies implemented and highlights health and psychosocial well-being issues in the population [10][11][12]. It encompasses a wide range of facets within the physical, social, environmental and psychological domains and is perceived as an efficient appraisal of one's life satisfaction, desire, needs and aspirations within the context of one's culture and values systems [13]. Thus, less fortunate rural dwellers, who are not only incapable of fulfilling their basic needs but also reside in less developed remote locations, most likely experience poor perceived QOL.
While the rich place great importance on freedom of choice related to life satisfaction, the poor are largely dependent on survival strategies to make ends meet. Hence, a rise in income levels provides the rural poor with basic necessities that would help them achieve a better QOL. Underpinning this, empirical studies from China [14], India [15], Malaysia [16][17][18] and Russia [19] showed a positive link between increasing income levels and QOL among rural dwellers. However, the findings by [16] were questionable as they used a small sample size that did not show generalizability and did not apply rigorous modeling methods to further substantiate the association between income and QOL. Moreover, Hassan et al. [17] used national level QOL measures and did not capture the individual level QOL that was applied in this study. Furthermore, using qualitative research design [17] only selected high-income female entrepreneurs as key participants, which could have led to a biased conclusion due to the lack of comparison with other income levels. Similarly, the findings by Idris et al. [18] were also inaccurate and not robust, owing to the small sample size and the lack of individual-level measurement of QOL.
On the contrary, Easterlin, in a seminal work, contended the exposition that higher income, on average, engendered a happier and more satisfied population until a certain income threshold, after which it diminished [20]. Echoing Easterlin's findings, several other scholars also documented a curvilinear association between income and QOL across and within countries [21][22][23]. Besides, Sen argued that high income levels did not entirely reflect a good QOL and referred to Kerala and Sri Lanka as examples of achieving the desired QOL outcome despite low income status [24]. In support of Sen's contention, a few other studies also showed the insignificant impact of income on QOL [25][26][27][28][29]. Thus, given the mixed findings and paucity of robust scholarly work on the association between income and perceived QOL in rural settings, this study seeks to investigate this association within the Malaysian context.

Study Site
The sample of this study was collected by the South East Asia Community Observatory (SEACO) team located in the district of Segamat, Johor state, Malaysia. SEACO is a health and demographic surveillance system (HDSS) center, which operates and conducts annual surveys in five out of eleven sub-districts of Segamat. It is a unique research platform in assessing population health and well-being of semi-urban and rural communities, which the majority of the Malaysian population belongs to. Segamat and its five sub-districts were chosen based on the strong pre-existing relationship between the Jeffrey Cheah School of Medicine and Health Sciences (JCSMHS) and the district, as well as state health administration, which was essential to conduct this research. Segamat has a marked ethnic breakdown that closely reflects the national proportions of Malays (60%), Chinese (23%) and Indians (7%), as well as equal gender composition (male: 49%; female: 51%) [30].

Study Design, Sample and Data Collection Process
A total sample of 25,512 respondents enrolled in this study survey. Of this number, 18,607 participants aged 18 years and above were drawn and used in the analysis of this study. The study was only based on a baseline cross-sectional survey collected in 2013. All trained enumerators and staff briefed participants about the objectives of the survey, and only participants who gave written consent were recruited and enrolled. Respondents were approached at their respective residences to gather information on their sociodemographic background (age, gender, education, employment status, income, marital status, ethnicity), health-related conditions (diagnosed chronic illnesses) and quality of life using standardized health data collection tools. This information was recorded directly into Android mobile devices and tablets with survey forms designed in Open Data Kit (ODK). Data recorded on the tablets were then encrypted and uploaded to a secure server. The study was approved by the Monash University Human Research Ethics Committee (2013-3837-3646).

QOL Assessment
Previous studies mostly applied either objective or subjective QOL indicators as an overall measure of QOL [31][32][33]. While early works have focused on evaluating objective QOL using quantifiable indicators, such as health care, education and more [18,34], subjective QOL measures individuals' perception of life experiences [35]. A validated Malay version of the abbreviated World Health Organization Quality of Life (WHOQOL-BREF) self-reported questionnaire, which was tested as reliable and effective among Malaysians, was used to evaluate the perceived QOL [36]. WHOQOL-BREF instruments are valid, reliable and best applied in a diverse cross-cultural setting that is internationally comparable [37,38]. These items assess an individual's life satisfaction and their perception of life from various aspects [38]. Moreover, its high acceptance among individuals worldwide minimizes refusal rates and missing data, which improves the accuracy in decision making processes and policy implementations [39].
The WHOQOL-BREF questionnaire consists of 26 instruments, of which 24 items are differentiated into four domains, namely physical health (seven items), psychological health (six items), social relationships (three items) and environment (eight items). Two other items, self-reported QOL and health satisfaction, are self-explanatory and are not assessed in this study. The items under each domain are listed in the Appendix A. Each item has a five-point Likert-scale response option ranging from 1 (very dissatisfied/very poor) to 5 (very satisfied/very good). The domain scores and their transformed linear scale between 0 and 100 are computed following a scoring guideline [13]. Higher scores suggest favorable perceived QOL environmental, social, psychological and physical domains.

Monthly Household Income
Monthly household income groups were characterized into three categories, namely, bottom 40% (B40), middle 40% (M40) and top 20% (T20), based on the income thresholds provided by the Department of Statistics Malaysia (DOSM) in 2014. The B40 income group was defined as individuals with monthly household income below RM 3860, the M40 income group were those with a monthly household income between RM 3860 and RM 8319 and the T20 income group were classified as those with monthly household income greater than RM 8319.

Control Variables
Past studies have shown that demographic factors, such as age, gender, ethnicity and marital status, socioeconomic variables (education status, employment status), mental health status (mainly depression, stress and anxiety), diagnosed chronic diseases (stroke, heart disease, asthma, arthritis, urinary tract disease, kidney disease), impacts QOL [14,40,41]. Therefore, these variables were included in the analysis to control their influence on the association between income groups and perceived QOL. In doing so, the study achieved a more rigorous understanding of the impact of income groups on perceived QOL.

Statistical Analysis
The frequency of variables was recorded, and a chi-square test was conducted to identify the presence of significant bivariate associations between the predictor variables and income groups. Next, a generalized linear model (GLMz) method was applied to investigate the association between income groups and perceived QOL. GLMz was used because the residual of outcome variables was non-normally distributed. In addition, the analysis was adjusted for the control variables mentioned above to eliminate the effect of potential confounders. A robust Huber-White sandwich estimator was also used to avoid heteroskedasticity issues, further contributing to the rigorousness of the model produced. A multicollinearity test based on the variation inflation factor (VIF) was performed to identify the existence of any collinearity issues among the predictor variables. The VIF values for all variables were less than 5.0, indicating the absence of collinearity. Since the missing values were less than 16.0% and did not pass the 50.0% threshold level, all variables were included in the study [42]. All analyses were performed using SPSS version 20, with a 5% and 10% significance level. Table 1 below presents the prevalence and descriptive statistics of the variables included in the study. The average age of the respondents in the sample is 47.6 years, with a standard deviation of 16.9 (mean ± SD = 47.6 ± 16.9). The sample largely comprises Malays (65.1%), married couples (70.0%) and B40 income groups (68.5%), which closely represents the characteristics of a semi-urban and rural population. Income distributions are often skewed, which is apparent in many past works [43][44][45]. Therefore, the use of mean as a measure of income is not reliable because it loses its power to produce accurate results [44]. Also, previous studies that use mean income often ignore the different impacts of demographic and socioeconomic variables on each income level [43]. Hence, income groups are better and more commonly used to examine income differentials [45]. Table 1 also shows high shares of respondents who do not suffer from self-reported depression (81.9%), stress (90.9%) and anxiety (77.9%) levels, as well as low rates of those diagnosed with heart disease (3.0%), stroke (0.9%), asthma (3.7%), kidney disease (0.8%), urinary tract disease (0.9%) and arthritis (9.2%). In addition, the bivariate associations show highly significant results (0.1% significance level) between the demographic variables (marital status, gender and ethnicity), socioeconomic status (education, employment status), mental health (depression, stress and anxiety), all diagnosed chronic diseases except urinary tract and kidney diseases and the three income groups, B40, M40 and T20 (Table 2). High proportions of those aged 60 years and above, other ethnicities, widowers, females, and those who are unemployed and illiterate belong to the B40 income group ( Table 2). The characteristics of individuals within the M40 income group include those who are less than 20 years old, Chinese, single, male, employed and tertiary educated ( Table 2). The T20 income group consists of individuals between age 20-39, Indians, singles, men and those who are employed and tertiary educated ( Table 2).    Table 3 shows the adjusted effects of income groups on the four different domains of QOL, environmental (Model I), physical (Model II), social (Model III) and psychological (Model IV). The results show that, relative to B40, both M40 and T20 income groups are significantly and positively associated with all four domains of QOL, adjusting for age, gender, ethnicity, marital status, employment status, education level, diagnosed chronic diseases and mental health status. However, those who are unemployed are negatively associated across all the domains compared to housewives/husbands, and this is significant after adjusting for control variables. Similarly, pensioners are also significantly and negatively associated with all domains except the social domain. On the contrary, self-employed individuals are positively linked with all domains except the environmental domain. Against housewives/husbands as the reference category, those with paid employment show a positive but weak association with all domains except for the physical domain. However, this association is only significant at a 10% significance level with the social domain.   In terms of education level, only those with secondary and primary education showed a significant positive association with the physical domain compared to illiterates. Against illiterates as the reference category, those with primary, secondary, tertiary and other education showed a highly significant positive association with the psychological and environmental domain. However, there was a significant association between education level and the social domain. All age groups relative to those 60 years and above showed a significant positive association on all four domains of QOL. Malay and Indian ethnic groups showed a significant positive association across all four domains compared to the Chinese group. Gender, however, did not show any statistically significant association across all four domains of QOL. Those diagnosed with chronic diseases, such as kidney disease, arthritis, stroke and asthma, were all statistically and negatively associated with all domains of QOL compared to those who were undiagnosed. Similarly, those who were severely or extremely severely depressed and with mild or moderate anxiety showed a statistically significant negative association across all four domains controlling for other potential influences of QOL.

Discussion
The evidence produced from this study clearly proved the income inequality of QOL among rural residents. The M40 and T20 income groups had a better QOL in all domains (physical, psychological, social and environmental) compared to the B40 community. This concurred with the findings of previous studies done on rural areas of Russia [19], China [14], India [15], Malaysia [16][17][18] and several other developing countries [46]. However, our outcomes differed from studies that concluded the insignificance of income in explaining QOL [28,29]. The lack of consensus between the two sets of studies is most likely elucidated by the subjective perception of whether income is adequate to satisfy one's need, the relative income position compared to income of other individuals and adaptation to new situations as expectation level changes [47].
Juxtaposed against M40 and T20 income groups, the less advantaged population (B40) experiences a much slower income growth in rural areas [7]. This is believed to be caused by the under-investment in infrastructure and facilities, limited job opportunities, high dependency on declining agricultural output, and relatively low pay in exchange for labor-intensive work [48]. Income scarcity weakens the purchasing power of rural dwellers and their ability to make ends meet, which in turn leads to poor perceived QOL.
The lack of income also inhibits rural people from procuring health-related equipment, products and services, as well as gaining health-related knowledge that would otherwise boost their physical mobility, fitness and health status [49]. This sheds light on the findings of this study, which showed a significant positive association between M40 and T20 income groups and the physical domain of QOL compared to the B40 income group. Likewise, other studies on rural areas in Kerala, India [15] and the Netherlands [50] have also arrived at the same conclusion. However, failure to include other major influences on QOL, such as chronic disease, employment status and ethnicity, questions the robustness of findings from the study in the Netherlands.
Findings on the positive association between the M40 and T20 income groups and environmental domain in this paper are most likely explained by their strong financial capacity that enables them to live in high-quality housing areas, which are quiet, safe, less polluted, with high access to green and open spaces as well as infrastructure [51,52]. In contrast, the outcome of a study on Australia showed that men with high socioeconomic status (SES) were less satisfied with certain aspects of the environmental domain [28] due to the pressure of their surroundings, a circumstance that increased one's income level by comparing their income with that of the better-offs despite having strong financial security [53].
Moreover, the advantages of being exposed to green environments and fresh air, which include the promotion of good emotional health among the rich [51], helps substantiate the results of the significant positive association between M40 and T20 income groups and the psychological domain of QOL in this study. Also, those in the rural B40 group, at the bottom of the social ladder, often experience stress and anxiety problems when their income status is compared to those at higher rungs of the ladder, thus resulting in low scores in the psychological domain [46]. Growing concerns over their income position relative to better-offs often invokes negative feelings and emotions, such as envy, shame, guilt, anger, insecurity, social isolation and more [54][55][56], which negatively impacts the psychological domain of QOL [46]. Although an increase in income has been proven to enhance the emotional well-being of the poor, a study has shown that it does not have the same effect among the rich [57]. This explains the changes in life expectancy among the rich as income level rises [57].
While the results of the study showed a significant negative association between unemployed individuals and all four domains of QOL, paid-employees, the self-employed and others were positively associated across all domains of QOL. This is supported by [58], which concludes that unemployment leads to increased family stress [59] and poorer life satisfaction levels [60] compared to the employed. The lack of job security and scarcity of formal employment in rural areas hinders rural dwellers from the protection of severance payments and unemployment benefits available for those in formal employment, which can lead to severe economic hardship and food insecurity for the worker and those family members supported by this income.
Our results on the significant positive association between M40 and T20 income groups and the social domain relative to the B40 income groups in rural areas could be elucidated by the high rural exodus, especially among low-income earners in the United States [61]. A similar phenomenon was also observed in rural Malaysia [62]. This was accompanied by a declination in rural population growth [48], and engendered the deterioration of social institutions and increasing numbers of single, elderly populations in rural areas [63].
Interestingly, the findings of this study also showed no statistical significance between gender and the four domains of QOL, which was consistent with the findings of several other studies. Corroborating this, a meta-analysis of 146 studies showed negligible effects of gender on well-being, where it only accounted for less than 1% of the well-being scores [58], which also paralleled the findings of a cross-country study [58]. The lack of statistical significance could be a consequence of the increasing awareness of gender equality, which provides equal roles and opportunities to both males and females in the decision making process [64]. However, this negated the results of other studies that concluded a better QOL among men than females in rural areas. The outcome from this study also showed a better perceived QOL across all domains among younger residents (age 18 to 59 years) than the elderly (60 years and above), which was consistent with the results of previous works [41,50,65]. The deteriorating health conditions, including mental health, limited mobility and age-related prejudice, most likely explains the poor psychological, physical, social and environmental domains of QOL among the elderly population [41,64,66].

Conclusions
Overall, the M40 and T20 income groups enjoy a better perceived QOL than the B40 in rural areas. Hence, it is imperative to uplift the QOL of the B40 population in rural areas to help contribute to the country's economic growth and raise its status to a developed nation. A better QOL also ensures equitable opportunities for all segments of the population, including the B40 rural households. With this, the B40 population will not be left behind in participating and benefiting from national development and prosperity. Therefore, the findings support the case for introducing intervention programs, such as entrepreneur development activities, and the provision of infrastructure and services, including roads, broadband, Internet access, e-commerce, telecommunications, education and more in rural areas. These, in turn, will create job opportunities and elevate the income levels of the B40 community, which enhances their QOL in rural areas. Furthermore, findings from this study can also inform policymakers to continuously monitor and implement intervention programs needed to increase the QOL among B40 income groups in rural areas.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses or interpretation of data, in the writing of the manuscript or in the decision to publish the results.