https://orcid.org/0000-0001-7963-2139
Figures
Abstract
Background
In Southeast Asia, the prevalence of multimorbidity is gradually increasing. This paper aimed to investigate the association between educational level and multimorbidity among over 15-years old adults in Southeast Asia.
Methods
We conducted a systematic review of published observational studies. Studies were selected according to eligibility criteria of addressing definition and prevalence of multimorbidity and associations between level of education and multimorbidity in Southeast Asia. The Newcastle-Ottawa Scale (NOS) was used to measure the quality and risk of bias. The methodology has been published in PROSPERO with registered number ID: CRD42021259311.
Results
Eighteen studies were included in the data synthesis. The results are presented using narrative synthesis due to the heterogeneity of differences in exposures, outcomes, and methodology. The prevalence of multimorbidity ranged from 1.7% to 72.6% among over 18 years-old adults and from 1.5% to 51.5% among older people (≥ 60 years). There were three association patterns linking between multimorbidity and education in these studies: (1) higher education reducing odds of multimorbidity, (2) higher education increasing odds of multimorbidity and (3) education having no association with multimorbidity. The association between educational attainment and multimorbidity also varies widely across countries. In Singapore, three cross-sectional studies showed that education had no association with multimorbidity among adults. However, in Indonesia, four cross-sectional studies found higher educated persons to have higher odds of multimorbidity among over 40-years-old persons.
Conclusions
Published studies have shown inconsistent associations between education and multimorbidity because of different national contexts and the lack of relevant research in the region concerned. Enhancing objective data collection such as physical examinations would be necessary for studies of the connection between multimorbidity and education. It can be hypothesised that more empirical research would reveal that a sound educational system can help people prevent multimorbidity.
Citation: Feng X, Kelly M, Sarma H (2021) The association between educational level and multimorbidity among adults in Southeast Asia: A systematic review. PLoS ONE 16(12): e0261584. https://doi.org/10.1371/journal.pone.0261584
Editor: Masaki Mogi, Ehime University Graduate School of Medicine, JAPAN
Received: September 14, 2021; Accepted: December 3, 2021; Published: December 20, 2021
Copyright: © 2021 Feng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: As this is a systematic review, data are available in the published articles included in the review.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors declare they have no competing interests
Abbreviations: NOS, Newcastle-Ottawa Scale; LMICs, Low- and middle-income countries; SES, Socioeconomic status; PRISMA, Preferred Reporting Items for Systematic reviews and Meta-Analyses; ASEAN, the Association of Southeast Asian Nations; CI, Cumulative incidence; IR, Incidence Rate; CI, Confidence interval; aOR, Adjusted odd ratio; SMHS, Singapore Mental Health Study; TB, Tuberculosis; WiSE, Well-being of the Singapore Elderly; IFLS, Indonesian Family Life Survey
Introduction
Multimorbidity, also known as multiple chronic conditions, is generally defined as the presence of two or more chronic health conditions in a person at the same time [1]. Multimorbidity is more common in older adults (more than 60 years old), such as suffering from hypertension, diabetes and chronic kidney disease concurrently [2,3]. Due to global aging, multimorbidity is also becoming a global public health issue [2]. The magnitude of multimorbidity in low- and middle-income countries (LMICs) was estimated at 10% to 11% recently, and it is predicted to increase in the coming years [4]. Due to the complexity of symptoms and higher mortality rates, the treatment requirements for multimorbidity are more complicated than those for single diseases, and patients with multimorbidity often do not receive cost-effective treatment [5]. This situation may increase the economic and medical burden of these patients and lead to a reduced quality of life and greater damage to their physical and mental health [5,6].
In Southeast Asia, the prevalence of multimorbidity is also gradually increasing, from a prevalence of about 4.5% at the beginning of the 21st century to about 10% in recent years [4,7]. There are many factors contributing to the rising prevalence of multimorbidity in Southeast Asia, among which is the rapid socio-economic development of region and the concurrent growth in socioeconomic inequality [8–10]. Socio-economic development is connected to epidemiological transition and the growth of the prevalence of chronic diseases [9]. Moreover, inequalities in socioeconomic status (SES) may be reflected in unequal access to health care, participation in health activities, and life stressors, which would contribute to an increased burden of multiple chronic conditions [9,10]. Furthermore, lower socioeconomic groups who suffer from multimorbidity will suffer more because they have limited access to diagnosis and the burden of expensive treatment [8–10].
Education level generally refers to the highest attained level of education by individuals and is often classified into these levels: no education; elementary school; middle school, junior high school, senior high school, university or higher [9,10]. Moreover, education level, which is a key indicator of SES, may have a greater impact on the prevalence of chronic diseases compared with the other two factors (income and occupation) of SES. This is because in health studies on multimorbidity, measures of SES include income, occupation, and education, but education usually logically determines subsequent occupational and income development [9,10]. Moreover, education may also influence health literacy, leading to a potential role in reducing the prevalence of multimorbidity. Furthermore, educational inequalities are evident in Southeast Asian and are possibly influencing patterns of multimorbidity. In Southeast Asian nations, disparity in access to educational resources is common, often also along gender lines [11–13]. Thus, in this article, education level will be used as an exposure factor to measure its effect on multimorbidity in Southeast Asia.
Studies to date in Southeast Asia which have assessed associations between the education level and multimorbidity have had mixed outcomes. For example, some studies have concluded that lower education levels cause an increase in the prevalence of multimorbidity [14,15]. But other studies have established that the level of education was not connected with the prevalence of multimorbidity [16]. It is hypothesis that in high-income countries, education may prevent the occurrence of multimorbidity, but in LMICs, education may be a risk factor for multimorbidity [8,9]. Nevertheless, to date, no article has conducted a comprehensive systematic review of the association between multimorbidity and education level among populations living in Southeast Asia.
Therefore, this paper is the first attempt to provide an overview of studies on multimorbidity and educational attainment in Southeast Asia and to systematically evaluate published observational studies. The aim of this paper was to better understand the association between educational attainment and multimorbidity in Southeast Asia, which may help to identify potential causes of multimorbidity in these places and to design appropriate interventions to prevent or reduce the occurrence of multimorbidity.
Methods
A systematic review of published articles reporting multimorbidity and educational level among adults in Southeast Asia was conducted using the terms of the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement (S3 Table) [17]. The methodology has been published in PROSPERO with registered number ID: CRD42021259311 (S1 File).
Inclusion and exclusion criteria
Articles were selected based on the following inclusion criteria, (1) quantitatively designed observational studies; (2) studies reporting multimorbidity being defined as a person having two or more chronic conditions at the same time; (3) studies reporting or having available detailed data on associations between level of education and multimorbidity; (4) studies in which the study sites including either individual Southeast Asian countries or the Southeast Asian region, with the definition according to countries belonging to the Association of Southeast Asian Nations (ASEAN), which included Brunei, Cambodia, East Timor, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand and Vietnam; (5) the associations between education level and multimorbidity in the Southeast Asian region or Southeast Asian countries being reported in studies examining the regional level, such as at the level of LMICs or at the global level; (6) participants in studies being over 15 years of age; and (7) English-language studies being published between 1990 and 2021 due to the term of “multimorbidity” first being coined in the early 1990s [18].
The exclusion criteria were [18,19]: (1) book series and conferences; (2) qualitative studies; (3) non-observational studies; (4) studies reporting co-morbidity (studies with an index disease), such as multimorbidity among patients with diabetes, HIV or hypertension; (5) studies where detailed data was not available regarding the association between educational level and multimorbidity; (6) study sites not in Southeast Asian region or in Southeast Asian countries; (7) studies of examining the regional level such as LIMCs level and global level not reporting the associations between education level and multimorbidity in Southeast Asia region or Southeast Asian countries; (8) participants in studies was younger than 15 years; and (9) studies not published in English, and not published between 1990 and 2021.
Search strategy and the selection of literature
The databases of Scopus, PubMed and ProQuest were used to search for relevant articles. We classified the search terms according to exposure, outcomes and location (S1 Table) [18,19]: (1) Exposure: ‘education, literacy, educational status, educational level, educational attainment’. (2) Outcome: ‘multimorbidity, multimorbidity, multimorbid, multiple morbidities, multiple morbidity, multiple conditions, multiple diseases, multiple chronic diseases, multiple chronic conditions, multiple illnesses, multiple diagnoses, multi-pathology’. (3) Location: ‘Southeast Asia, Association of Southeast Asian Nation, ASEAN, Brunei, Cambodia, East Timor, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand, Vietnam”. The ‘AND’ was used to the combination of search terms across the categories and ‘OR’ was combined within the categories. In addition, a term similar to the definition of multimorbidity is comorbidity, but in 2018, a distinction has been made between the definition of comorbidity and multimorbidity, and while both terms emphasize the co-existence of multiple chronic conditions in the same individual, the term of “comorbidity” means the combined effects of additional conditions with reference to the index chronic condition such as the comorbidity of diabetes, stroke, or depression [1]. Namely, comorbidity is the presence of one or more additional diseases, as a result of the presence of the index condition [19]. Although, these two terms have different definition, multimorbidity and comorbidity are commonly used interchangeably [1,8]. Thus, after the initial search, the addition of ‘comorbidity’ was added into our search to test if any articles had been missed through the exclusion of the term of ‘comorbidity’ and linguistic changes of the term of ‘comorbidity’ [18].
Furthermore, the studies in this paper only involved human participants, were published from 1990/01/01 to 2021/06/15, and had abstracts available. We used the hand search in the references of retrieved studies to identify additional relevant papers.
The first reviewer (XF) performed an initial screening of titles and abstracts for all keywords searched. The second reviewer (MK) conducted a 20% random sample of all references to ensure that eligible studies were not omitted [18]. Studies that met all of the above eligibility criteria were retained for full-text screening. Full-text screening was done independently by two reviewers (XF and MK). When disagreements arose, XF and MK resolved them through discussion. When agreement could not be reached, a third reviewer (HS) was consulted. Disagreements were eventually resolved by consensus.
The assessment of quality
Two authors (XF and MK) independently assessed the risk of bias and study quality in cohort studies and case-control studies using the Newcastle Ottawa Scale (NOS), one of the risk of bias assessment tools recommended by the Cochrane Collaboration for use in observational studies [20], and cross-sectional studies using adaptations of the NOS (S2 File and S2 Table) [19,21]. For observational studies, the checklist focused on three aspects (Selection, Comparability and Exposure/Outcome).
The examined selection of observational studies according to NOS was shown in the table below (Table 1) [19,21]. Observational studies included cross-sectional studies, cohort studies and case-control studies. The selection table was based on a summary of the scored items for each of the observational studies in NOS (S2 File), with reference to other systematic reviews of multimorbidity that have applied NOS as an assessment tool to evaluate and examine the literature [19,21].
There were three levels of quality to assess the scores of the individual study in the table below (Table 2) [19,21]. The scores of NOS corresponding to the three levels mentioned in the table were summarized from similar levels and scores that appeared in other systematic reviews of multimorbidity [19,21].
Two reviewers (XF and MK) independently decided on the summary measurement of the relevant articles. The risk of bias was assessed as the sum of the scores for each item. Each reviewer independently determined an overall quality score for each article. It should be noticed that if one of the three aspects (Selection, Comparability and Exposure/Outcome) was given a zero score, the level was Poor, regardless of the total scores of all three aspects. The final article selection was based on the scores of the three aspects (Selection, Comparability and Exposure/Outcome). To be retained in our systematic review, articles should have had a quality score of four scores or above (S2 Table) [19,21].
Data extraction
For each included study, we extracted the following information [18,19]: (1) the author(s) and publication year; (2) study country/location; (3) study design and study population (4) sample size: total number; (5) sample characteristics: the percentage of male (%); (6) sample characteristics: mean age and/or the range of age (years); (7) data collection; (8) the definition of multimorbidity and the number of conditions or diseases; (9) the subdivision of educational attainment; (10) the prevalence/incidence of multimorbidity (11) the prevalence/incidence of multimorbidity in terms of educational level; and (12) the main results of the association between multimorbidity and education (including analytical adjustments made).
Data synthesis
Because of the different exposures, outcomes, and study methods, the included studies were judged to be heterogeneous in distinct ways such as populations, the definition of variables and different confounders. Taking this heterogeneity into account, we were unable to perform a meta-analysis of the findings [18,22].
The findings regarding methods, exposures, and outcomes were described using narrative synthesis [18,22]. The prevalence of multimorbidity was extracted from individual studies. The extraction of the prevalence of multimorbidity was divided into three cases in these studies. (1) The prevalence of multimorbidity was reported directly in the study, thus we could extract data on the prevalence of multimorbidity directly from the article. (2) If the prevalence of multimorbidity was not directly reported in the study but it reported the total number of participants and the number of people with multimorbidity, we calculated this indicator using the number of participants as the denominator and the number of people with multimorbidity as the numerator. (3) If studies did not report the prevalence of multimorbidity, the total number of participants and the number of multimorbidity cases, we categorised those studies “not available” (N/A).
The available data regarding education level in the survey area were pooled in the same way to extract and calculate the prevalence (%) of multimorbidity at different education attainments. The prevalence of multimorbidity (%) at different education levels in the study was all obtained by calculation in selected studies. The prevalence of multimorbidity for a particular education level group was calculated by dividing the number of participants experiencing multimorbidity at that educational attainment level by the total sample number of participants. For studies that we could not calculate the prevalence of multimorbidity (%) at different educational levels because of lack of reporting the number of patients with multimorbidity (numerator) at different educational levels, we would put “N/A” to show the prevalence.
All studies reported the ratios of multimorbidity at different educational levels. Therefore, the ratios in this paper were used directly from the ratios given in each study.
Results
Yield of search strategy
The electronic and manual searches yielded 7,558 articles. After eliminating duplicate articles, 6,786 articles were selected for the title and abstract screening. After carefully screening the titles and abstracts, 36 articles were found to be available for full-text review. A further 18 articles were then removed due to their addressing comorbidity (17 articles) and being unable to obtain detailed data of multimorbidity in terms of educational level in Southeast Asia (1 article). Finally, 18 articles were included in this systematic review (Fig 1) [14–16,23–37]. Seventeen papers were cross-sectional studies [14–16,23–28,30–37], and one paper was a longitudinal study [29]. No studies were case-control studies.
Description of article selection process.
Study selection and characteristics
Table 3 detailed the key characteristics of the included studies. Two multi-country studies implemented in different Southeast Asian countries were included in this review [14,16]. Five studies were from Singapore [23,27,34–36], four studies were from Indonesia [24,28,30,31], three studies were from Vietnam (all from different parts of Vietnam) [15,26,32], two studies were from Thailand [33,37], one was from Myanmar [25] and one was from Malaysia [29]. Sample sizes of selected studies ranged from 729 [24] to 13,798 [29]. Just one article did not calculate the prevalence or incidence of multimorbidity [30]. The only longitudinal study showed the incidence of half year cumulative incidence and the incidence rate (IR) for multimorbidity [29]. Nine cross-sectional articles directly showed the number of participants experiencing multimorbidity at that educational attainment level and the total sample number to get the prevalence of multimorbidity at different educational levels [15,23–26,31–34]. Six studies included only participants aged 60 years and older [15,25,29,32,34,36].
On the NOS, six studies were scored Good level of quality [15,24,28,30,31,36] nine studies (including the only longitudinal study [29]) were at Fair level [14,16,23,25,26,29,33,34,37], and three studies were at Poor level [27,32,35] (Table 3).
Definition and measure of multimorbidity
Twelve studies defined multimorbidity as two or more chronic conditions (Table 3) [14–16,24–26,28–31,33,34]. Two studies defined only multiple diseases but did not emphasize a specific number of diseases [32,35]. Four studies defined multimorbidity as the coexistence of two chronic conditions [23,27,36,37], but not the combined effects of additional conditions with reference to the index chronic condition, unlike the definition of co-morbidity. The number of chronic conditions measured were from two [23,36,37] to twenty-one [14] in these studies. Multimorbidity in twelve studies included psychological disorders in addition to physical disorders [14,16,23–28,31,33,34,36], and four studies’ multimorbidity included tuberculosis (TB) (an infectious disease) [15,28–30].
To determine the outcome of the condition, seven studies used self-report to collect data [16,26,29,32–35] and four studies used a combination of medical or professional examination and self-reports [24,28,31,36]. But one study used only health examination [37].
The prevalence of multimorbidity and multimorbidity in terms of education
The prevalence of multimorbidity ranged from 3.2% to 72.6% among over 15 years-old participants. Moreover, the prevalence ranged from 1.5% to 51.5% among older people (more than 60 years old) (Table 3).
The prevalence of multimorbidity was from 20.84% (≥ 40 years old) [24] to 35.7% (≥ 40 years old) [28] in Indonesia [24,28,30,31], and in all four studies [24,28,30,31] higher education related to higher odds of multimorbidity among over 40-years-old persons. The prevalence was from 1.5% [36] to 51.5% among people over 60 years old [34] and from 5.8% [23] to 45.2% [33] among people (≥ 18 years old) in Singapore, and three [23,35,36] in all five studies [23,27,34–36] showed that education had no association with multimorbidity. The prevalence of multimorbidity was from 12.7% among older people (≥ 60 years old) [32] to 39.2% among over 60-years-old people [15] in different parts of Vietnam [15,26,32], and two studies [15,26] suggested that education may reduce people’s odds of developing multimorbidity but the age of the study population (≥ 18 years vs. ≥ 60 years) for these two articles were different (Tables 3 and 4).
Association between educational level and multimorbidity
There were three outcomes of these studies. First, higher education reduced odds of suffering from multimorbidity [14,15,26,27,30,33,34], second was that lower education reduced likelihood of having multimorbidity [24,35,28,30,31,33,37] and third was that educational attainment was not related to multimorbidity [16,23,28,29,32,33,35,36] (Table 4).
Higher education reducing odds of multimorbidity
Seven cross-sectional studies [14,15,26,27,30,33,34] found education has been associated with reducing odds of multimorbidity. The study of Liew [30], found a higher level of education was beneficial for the health of over 40-year-old Indonesian women to fight against multimorbidity (all p-value < 0.05) after controlling for mobility problems, age, marital status, and smoking. Similar results were found in studies in other countries such as Vietnam [15,26]. For example, in accordance with the study of Ba et al in Central Highlands Region of Vietnam [26], after controlling for gender, age, education, and occupation, over 15-year-old participants with a high school education had significantly lower odds of suffering from multimorbidity compared to those with secondary or less school education (adjusted odd ratio (aOR) 0.8, 95% Confidence interval (CI) 0.59–0.98). Furthermore, in the other study focused on the population older than 60 years in southern Vietnam [15], literate individuals (aOR 0.68, 95% CI 0.54–0.85) were associated with lower odds of multimorbidity in comparison with illiterate individuals after controlling for the related confounders (Table 4).
Higher education increasing odds of multimorbidity
Seven cross-sectional studies found education has been associated with increasing odds of multimorbidity [24,25,28,30,31,33,37] and four of these studies [24,28,30,31] were located in Indonesia. Anindya et al [24] showed that higher educated people (more than 40-year-old) had more odds of suffering from multimorbidity after controlling for socio-demographic variables (all p-value > 0.05). Moreover, in another study in Indonesia [31], comparing lower levels of education to high level of education, 50-year-old and older participants would be more likely to suffer from multimorbidity. Furthermore, another study in Thailand [37] showed, after adjusting for potential confounders for socio-demographic variables, those 15 years old and older with less than 6 years of education (aOR 1.83, 95% CI 1.03–3.38) were more likely to have multimorbidity compared to those with no formal education (Table 4).
Education having no association with multimorbidity
There were eight studies (including the only longitudinal study) which showed no association among educational attainment and multimorbidity [16,23,28,29,32,33,35,36]. Three cross-sectional studies [23,35,36] were located in Singapore. For instance, one study [35] showed that the odds of having multiple chronic medical conditions were not associated with educational attainment, but this study did not control for relevant confounders. Furthermore, in other countries like Laos, Malaysia, Myanmar, and Philippines [16], the results showed that educational levels were not associated with multimorbidity in these four countries after controlling for gender and age (all p-value > 0.05) (Table 4).
Discussion
This was the first study to systematically review and assess the available literature on the prevalence of multimorbidity and the relationship between education level and multimorbidity in Southeast Asia.
We identified a small number of relevant publications and found heterogeneity between these studies. For example, the estimation of sample size, the grouping of education levels, the age groups of study participants, and the inclusion and exclusion criteria varied considerably among the studies we included, making comparability difficult [18] that did not allow us to clarify the association between education level and multimorbidity using meta-analysis.
There were two obvious associations between education and multimorbidity in Singapore and Indonesia, respectively. Firstly, educational level was not associated with multimorbidity in Singapore, alternately higher levels of education were associated with higher odds of developing multimorbidity in Indonesia. The reasons for these results may be related to differences in the purpose and methods of these studies, such as different patterns of multimorbidity, classification of education levels, and the confounding factors, and may also be related to the databases used in these studies.
Most studies in Singapore [23,35,36] mentioned that education level was not associated with multimorbidity. According to the authors, this is likely to be related to the fact that these studies [23,27,34–36] from Singapore used only two databases, the Singapore Mental Health Study (SMHS) [23,27,35] and the Wellbeing of the Elderly in Singapore (WiSE) [34,36]. It may be that the association between education level and multimorbidity was not significant in these two databases [23,27,34–36]. In addition, it is worth noting that the results would be influenced by many factors, even when a common data source was used for the analysis, the definition of multimorbidity, adjustment factors, and differences in education level, may impact on the results. For example, the study of Chong et al [27] and the study of Subramaniam et al [35] used the SMHS-2010, but the classification of education level (the study of Subramaniam et al [35] combining the level of Pre-U/Junior College/Diploma together) and the pattern of multimorbidity differed (the pattern of multimorbidity including mental disorders in the study of Chong et al [27]), thus the results of the connection of multimorbidity and education of these two studies were different. Similarly, the study of Picco et al [34] and the study of Subramaniam et al [36] used WiSE, but their patterns of multimorbidity and adjustment factors were very different, leading to distinct results of the association between educational level and multimorbidity.
While all four Indonesia studies found [24,28,30,31] that higher education levels were associated with higher odds of developing multimorbidity, all of them used the Indonesian Family Life Survey (IFLS). It was likely that the association between higher education level and higher odds of developing multimorbidity was evident in the IFLS database. However, there were still many differences between the results of these studies. The reasons for this include factors such as the purpose of the study, the selection of the population, and the method of analysis affected the association between multimorbidity and education level. Also relevant was that the ILFS is a longitudinal study collecting data every 4–5 years. The various studies using ILFS data conducted cross sectional analysis on various waves of this study, which may also help explain the variation. A typical example was the study by Marthias et al [31]. This study consisted of two cross-sectional sub-studies whose purpose and methods were consistent. However, the difference in the pattern of multimorbidity and participants between these two sub-studies made the results of the association between multimorbidity and education level different [31].
High levels of education have been reported to reduce the odds of developing multimorbidity in western countries such as the United States and Canada [8]. However, in some studies of LMICs, the relationship between education level and multimorbidity was more complex than in developed countries. For example, in some studies from Bangladesh [38], India [39–41] and China [42,43], there was a positive [38,39,42], negative [40,43] or no association [41] between high education level and higher odds of multimorbidity. This was the same the results of our study.
The improvement of educational level was recognized as reducing risk of multimorbidity [10,44]. The potential reason would be that as the level of education increased, people’s SES would also increase, making them more aware of healthy living and gaining health literacy and decreasing the risk [8,10,44]. In developed countries such as the United States and Canada, people with low SES were more likely to suffer from multimorbidity due to their lack of health knowledge, higher stress levels, and inability to afford healthy and adequate diet, which made them less healthy [8,10]. At the same time, their inability to afford the high cost of medical care after suffering from multiple diseases caused their health level to further worsen, which would be a vicious cycle [8,10]. The same problem was faced in developing countries. However, few studies have investigated multimorbidity in depth in developing countries, and the current focus in developing countries was still on single diseases [9,18]. This was because although aging has been also increasing in LMICs, the proportion of the elderly population is relatively low compared to that in developed areas, and aging is positively correlated with the incidence of multimorbidity [2,3,5]. These have led to under-research of multimorbidity in LMICs and a great variation from study to study, resulting in a diversity of results between multimorbidity and factors such as education level.
Several studies have shown that higher levels of education increase the chances of developing multimorbidity, and the reasons for this phenomenon include the following. Firstly, although people of lower educational level in LMICs were more likely to consume a poor diet quality–low in fruit and vegetables and high in red met and processed food–higher educated people tended to favour sedentary lifestyle habits, thus increasing the odds of developing multimorbidity [38]. In addition, people with higher levels of education had better access to medical care and better health knowledge, especially in developing countries. Since people with low SES in developing countries had more difficulty accessing effective medical care with relevant health knowledge compared to their counterpart in high-income countries [24,38]. These highly educated people were more likely to receive a diagnosis of their condition and to be reported as having chronic conditions [24]. As for the results of no association between education and multimorbidity, it may be due to information bias in the study itself, selection bias, and confounding factors that have an impact on the results [35,41].
Most studies in this review did not describe the criteria for selecting chronic conditions [18,22]. Most common was selection of common or high prevalence diseases for inclusion in the study. For example, some studies included infectious diseases such as tuberculosis (TB) [15,28–30] and mental diseases [14,16,23–28,31,33,34,36], and the definition of multimorbidity in four [23,27,36,37] studies was two comorbid conditions. Because the number and type of chronic diseases determined the estimates of multimorbidity, prevalence rates and associations with education differed [18,22]. What is more, differences in study location, size and characteristics of the study population, data collection methods, and educational attainment classification may lead to selection bias and thus to different results [18,22]. Different data collection methods could also result in bias, and the recall bias of self-report could be more pronounced compared to medical examination for the identification of multimorbidity. In addition, different confounders had a significant impact on the results. For example, education level can be confounded by other socio-demographic factors including age, gender, and income [18,19,22]. Generally, the older age of the studied population implied a higher prevalence of multimorbidity, as we found in this paper [2,3,5,45]. As for the classification of educational attainment, the same data using different classifications of educational attainment could also lead to differences in the results that emerged, such as in the two studies [23,27] in Singapore.
Strength and limitation
The main strength of this paper included the systematic listing of education level separately in relation to multimorbidity, rather than including it in the SES, which provided a more precise understanding of the association between education level and multimorbidity.
The results of most studies [16,32,33,35] obtained with self-report data showed no association between education level and multimorbidity, and the sample size and age range of these studies varied widely, suggesting that such results may be related to self-reported having recall bias. The studies of multimorbidity tended to favour objective test of multimorbidity such as physical examination, or a combination of physical examination (major) and self-report (minor), because this could reduce the bias in the outcomes. In addition, in studies with participants older than 40 years, most of the results [24,25,28,30,31] showed that high education level was associated with high odds of having multimorbidity. This may be due to the same set of datasets used in number of these studies [24,28,30,31]. We did not find a pattern in the effect of sample size on the results. However, in general, the larger the sample size, the more accurate the results should be, controlling for other variables such as age and data collection [8,19,22].
The differences in inclusion and exclusion criteria for chronic diseases, sample size estimates, differences in study areas, different age groups of study participants, differences in data collection methods, distinction in educational level groupings, differences in confounding factors may all lead to large biases across studies and, to some extent, explain the large differences in observed outcomes. The inherent bias in the estimates of the original studies prevented the assessment of the quantification of prevalence of multimorbidity and the estimation of the association between education level and multimorbidity. The similar heterogeneity was seen in the systematic review of studies on the prevalence of multimorbidity in South Asia [18], with different methodologies and research settings contributing to this phenomenon.
Additionally, there are seventeen cross-sectional studies but only one longitudinal study. The lack of longitudinal studies made it impossible to definitively state whether there was a causal relationship between education level and multimorbidity [22,45].
Another limitation was that it was difficult to ensure that all relevant literature was included. Since multimorbidity was not well indexed in literature databases and was often used interchangeably with the term “comorbidity” [18,22]. Although the term “comorbidity” was also searched separately in the database after the initial search to compensate for this search omission, inadvertent omissions could not be excluded [22]. In addition, other keywords such as “education level” and “Southeast Asia” may have other names that we did not mention, and some literature may have been omitted. An inherent limitation of any systematic review was the limitation of the search period, which in our case was from January 1, 1990, to June 15, 2021, implying the exclusion of new studies after the end date, which may have led to the omission of more recent studies [18]. We also restricted the search to English publications, resulting in relevant articles in other languages that we could not retrieve [18,22]. Furthermore, another major limitation was the large amount of statistical and methodological heterogeneity, causing the inability to combine studies to obtain overall estimates of prevalence in multimorbidity and the overall estimates of association of education level with multimorbidity [18].
Implications for policy, practice, and future research
The Association between multimorbidity and education remains an underappreciated area of research in Southeast Asia. As mentioned above, the small number of relevant studies included in this article and the large differences of statistics and methodology between individual studies. The results of the available global systematic review [8] on SES and multimorbidity suggest that the higher the level of education is, the lower the odds of developing multimorbidity people have. Although this result is inconsistent with the association between multimorbidity and education in Southeast Asia derived in this paper, the salient heterogeneity of this study imposes limitations on the overall estimation of the connection between education level and multimorbidity to calculate a uniform association. Therefore, in the next studies in Southeast Asia, the panel of chronic conditions should be prepared with a standardized definition of each disease and a uniform operational definition of multimorbidity, which could reduce the selection bias of chronic conditions and lead to a more reliable and comparable estimate of multimorbidity [18,22]. In addition, there should be uniform and objective criteria for the classification of education level, such as no education; elementary school; middle school, junior high school, high school, university, or higher, and reducing self-reported data collection methods and strengthening objective tests, such as physical examinations, would also bring accurate results.
Conclusion
This study is a comprehensive mapping of research related to the association between education and multimorbidity in the Southeast Asian region and reveals the neglect and lack of studies on the connection between multimorbidity and educational level in this region. The heterogeneity of the findings did not allow us to reach a definitive conclusion about the association between educational level and multimorbidity. The prevalence of multimorbidity ranged widely in this study, and the associations between educational attainment and multimorbidity were inconsistent. Reasons for this result include the different national contexts and the lack of relevant studies on the relationship between educational status and multimorbidity in Southeast Asian countries. This study may indicate the need to reduce the use of subjective data collection methods, such as self-report, to improve credibility and accuracy when studying the relationship between multimorbidity and education. Finally, there were several different connections between educational attainment and multimorbidity in Southeast Asia in this paper, however, it is predicted that a sound and completed educational system could help people to raise health awareness and thus effectively prevent multimorbidity.
Supporting information
S2 Table. NOS checklist for selected studies (cross-sectional study).
https://doi.org/10.1371/journal.pone.0261584.s002
(DOCX)
S1 File. The PROSPERO-registered number-CRD42021259311.
https://doi.org/10.1371/journal.pone.0261584.s004
(PDF)
S2 File. The Newcastle-Ottawa Scale (NOS) checklist.
https://doi.org/10.1371/journal.pone.0261584.s005
(PDF)
Acknowledgments
Authors thank Dr. Nasser Bagheri for his help with the content and structure of the article.
References
- 1. Harrison C, Fortin M, van den Akker M, Mair F, Calderon-Larranaga A, Boland F, et al. Comorbidity versus multimorbidity: Why it matters. J Comorb. 2021 Mar 2;11:2633556521993993. pmid:33718251
- 2.
Sheikh A, Donaldson LJ, Dhingra-Kumar N, Bates DW. Multimorbidity-Technical series on safer primary care. WHO [Internet]. 2016 Dec [cited 2021 Sep 8];[about 1 p.]. Available from: https://apps.who.int/iris/bitstream/handle/10665/252275/9789241511650-eng.pdf?sequence=1.
- 3.
United Nations, Department of Economic and Social Affairs, Population Division. World Population ageing 2019: Highlights (ST/ESA/SER.A/430). UN [Internet]. 2019 [cited 2021 Sep 8];[about 1 p.]. Available from: https://www.un.org/en/development/desa/population/publications/pdf/ageing/WorldPopulationAgeing2019-Highlights.pdf.
- 4. Abebe F, Schneider M, Asrat B, Ambaw F. Multimorbidity of chronic non-communicable diseases in low- and middle-income countries: A scoping review. J Comorb. 2020 Oct 16;10:2235042X20961919. pmid:33117722
- 5. Wallace E, Salisbury C, Guthrie B, Lewis C, Fahey T, Smith SM. Managing patients with multimorbidity in primary care. BMJ. 2015 Jan 20;350:h176–85. pmid:25646760
- 6. Amuna P, Zotor FB. Epidemiological and nutrition transition in developing countries: impact on human health and development. Proc Nutr Soc. 2008 Feb;67(1):82–90. pmid:18234135
- 7. Hajat C, Stein E. The global burden of multiple chronic conditions: A narrative review. Prev Med Rep. 2018 Oct 19;12:284–93. pmid:30406006
- 8. Pathirana TI, Jackson CA. Socioeconomic status and multimorbidity: a systematic review and meta-analysis. Aust N Z J Public Health. 2018 Feb 14;42(2):186–94. pmid:29442409
- 9. Hosseinpoor AR, Bergen N, Mendis S, Harper S, Verdes E, Kunst A, et al. Socioeconomic inequality in the prevalence of noncommunicable diseases in low- and middle-income countries: results from the World Health Survey. BMC Public Health. 2012 Jun 22;12:474–86. pmid:22726343
- 10. Johnson-Lawrence V, Zajacova A, Sneed R. Education, race/ethnicity, and multimorbidity among adults aged 30–64 in the National Health Interview Survey. SSM Popul Health. 2017 Mar 28;3:366–72. pmid:29349230
- 11. Hallinger P. Using faculty evaluation to improve teaching quality: A longitudinal case study of higher education in Southeast Asia. Educ Asse Eval Acc. 2010 Sep 22;22:253–74.
- 12. McGuinness S, Kelly E, Pham TTP, Ha TTT, Whelan A. Returns to education in Vietnam: A changing landscape. World Dev, 2020 Nov 11;13:1–19.
- 13. Sadiman AS. Challenges in Education in Southeast Asia. SEAMEO Secretariat [Internet]. 2004 Nov [cited 2021 Aug 30];[about 1 p.]. Available from: https://www.seameo.org/VL/library/dlwelcome/publications/paper/india04.htm.
- 14. Pengpid S, Peltzer K. Multimorbidity in Chronic Conditions: Public Primary Care Patients in Four Greater Mekong Countries. Int J Environ Res Public Health. 2017 Sep 6;14(9):1019–27. pmid:28878150
- 15. Ha NT, Le NH, Khanal V, Moorin R. Multimorbidity and its social determinants among older people in southern provinces, Vietnam. Int J Equity Health. 2015 May 30;14:50–6. pmid:26024877
- 16. Afshar S, Roderick PJ, Kowal P, Dimitrov BD, Hill AG. Multimorbidity and the inequalities of global ageing: a cross-sectional study of 28 countries using the World Health Surveys. BMC Public Health. 2015 Aug 13;15:776–85. pmid:26268536
- 17. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ. 2021 Mar 29;134:178–89. pmid:33782057
- 18. Pati S, Swain S, Hussain MA, van den Akker M, Metsemakers J, Knottnerus JA, et al. Prevalence and outcomes of multimorbidity in South Asia: a systematic review. BMJ Open. 2015 Oct 7;5(10):e007235. pmid:26446164
- 19. Nguyen H, Manolova G, Daskalopoulou C, Vitoratou S, Prince M, Prina AM. Prevalence of multimorbidity in community settings: A systematic review and meta-analysis of observational studies. J Comorb. 2019 Aug 22;9:2235042X19870934. pmid:31489279
- 20.
The Cochrane Collaboration. Method in Cochrane Handbook-ROBINS-I tool. Cochrane [Internet]. 2021 [cited 2021 Sep 8];[about 1 p.]. Available from: https://methods.cochrane.org/methods-cochrane/robins-i-tool.
- 21. Tadeu ACR, E Silva Caetano IRC, de Figueiredo IJ, Santiago LM. Multimorbidity and consultation time: a systematic review. BMC Fam Pract. 2020 Jul 28;21(1):152–9. pmid:32723303
- 22. Ingram E, Ledden S, Beardon S, Gomes M, Hogarth S, McDonald H, et al. Household and area-level social determinants of multimorbidity: a systematic review. J Epidemiol Community Health. 2020 Nov 06;75(3):232–41. pmid:33158940
- 23. Abdin E, Chong SA, Vaingankar JA, Shafie S, Seah D, Chan CT, et al. Changes in the prevalence of comorbidity of mental and physical disorders in Singapore between 2010 and 2016. Singapore Med J. 2020 Aug 17;1–27. pmid:32043156
- 24. Anindya K, Ng N, Atun R, Marthias T, Zhao Y, McPake B, et al. Effect of multimorbidity on utilisation and out-of-pocket expenditure in Indonesia: quantile regression analysis. BMC Health Serv Res. 2021 May 5;21(1):427–38. pmid:33952273
- 25. Aye SKK, Hlaing HH, Htay SS, Cumming R. Multimorbidity and health seeking behaviours among older people in Myanmar: A community survey. PLoS One. 2019 Jul 11;14(7):e0219543. pmid:31295287
- 26. Ba NV, Minh HV, Quang LB, Chuyen NV, Ha BTT, Dai TQ, et al. Prevalence and correlates of multimorbidity among adults in border areas of the Central Highland Region of Vietnam, 2017. J Comorb. 2019 May 29;9:2235042X19853382. pmid:31192142
- 27. Chong SA, Abdin E, Nan L, Vaingankar JA, Subramaniam M. Prevalence and Impact of Mental and Physical Comorbidity in the Adult Singapore Population. Ann Acad Med Singap. 2012 Mar;41(3):105–14. pmid:22538737
- 28. Hussain MA, Huxley RR, Al Mamun A. Multimorbidity prevalence and pattern in Indonesian adults: an exploratory study using national survey data. Bmj Open. 2015 Nov 3;5:e009810. pmid:26656028
- 29. Hussin NM, Shahar S, Din NC, Singh DKA, Chin AV, Razali R, et al. Incidence and predictors of multimorbidity among a multiethnic population in Malaysia: a community-based longitudinal study. Aging Clin Exp Res. 2019 Feb; 31(2):215–24. pmid:30062670
- 30. Liew HP. Depression and Chronic Illness: A Test of Competing Hypotheses. J Health Psychol. 2012 Jan; 17(1):100–9. pmid:21712338
- 31. Marthias T, Anindya K, Ng N, McPake B, Atun R, Arfyanto H, et al. Impact of non-communicable disease multimorbidity on health service use, catastrophic health expenditure and productivity loss in Indonesia: a population-based panel data analysis study. Bmj Open. 2021 Feb 17;11(2):e041870. pmid:33597135
- 32. Mwangi J, Kulane A, Van Hoi L. Chronic diseases among the elderly in a rural Vietnam: prevalence, associated socio-demographic factors and healthcare expenditures. Int J Equity Health. 2015 Nov 17;14:134–41. pmid:26578189
- 33. Pengpid S, Peltzer K. Chronic conditions, multimorbidity, and quality of life among patients attending monk healers and primary care clinics in Thailand. Health Qual Life Out. 2021 Feb 23;19(1):61–9. pmid:33622328
- 34. Picco L, Achilla E, Abdin E, Chong SA, Vaingankar JA, McCrone P, et al. Economic burden of multimorbidity among older adults: impact on healthcare and societal costs. Bmc Health Services Research. 2016 May 10;16:173–84. pmid:27160080
- 35. Subramaniam M, Abdin E, Picco L, Vaingankar JA, Chong SA. Multiple chronic medical conditions: prevalence and risk factors—results from the Singapore Mental Health Study. Gen Hosp Psychiat. 2014 Jul-Aug; 36(4):375–81. pmid:24713327
- 36. Subramaniam M, Abdin E, Vaingankar JA, Picco L, Seow E, Chua BY, et al. Comorbid Diabetes and Depression among Older Adults—Prevalence, Correlates, Disability and Healthcare Utilisation. Ann Acad Med Singap. 2017 Mar;46(3):91–101. pmid:28417133
- 37. Tiptaradol S, Aekplakorn W. Prevalence, awareness, treatment and control of coexistence of diabetes and hypertension in thai population. Int J Hypertens. 2012 Jul 19;2012:386453. pmid:22888406
- 38. Khan N, Rahman M, Mitra D, Afsana K. Prevalence of multimorbidity among Bangladeshi adult population: A nationwide cross-sectional study. BMJ Open. 2019 Nov 28;9(11):e030886. pmid:31784434
- 39. Mini GK, Thankappan KR. Pattern, correlates and implications of non-communicable disease multimorbidity among older adults in selected Indian states: A cross-sectional study. BMJ Open. 2017 Mar 8;7(3):e013529. pmid:28274966
- 40. Garin N, Koyanagi A, Chatterji S, Tyrovolas S, Olaya B, Leonardi M, et al. Global multimorbidity patterns: A cross-sectional, population-based, multi-country study. J Gerontol A Biol. Sci Med Sci. 2015 Jun;71(2):205–14. pmid:26419978
- 41. Pati S, Agrawal S, Swain S, Lee JT, Vellakkal S, Hussain MA, et al. Non communicable disease multimorbidity and associated health care utilization and expenditures in India: Cross-sectional study. BMC Health Serv Res. 2014 Oct 2;14(1):451–9. pmid:25274447
- 42. Zou S,Wang Z, Bhura M, Zhang G, Tang K. Prevalence and associated socioeconomic factors of multimorbidity in 10 regions of China: An analysis of 0.5 million adults. J Public Health. 2020 Dec 10;396:s12–26. pmid:33300571
- 43. Zhang Y, Zhou L, Liu S, Qiao Y, Wu Y, Ke C, et al. Prevalence, correlates and outcomes of multimorbidity among the middle-aged and elderly: Findings from the China health and retirement longitudinal study. Arch Gerontol Geriatr. 2020 Jun 04;90:104135. pmid:32554217
- 44. Nagel G, Peter R, Braig S, Hermann S, Rohrmann S, Linseisen J. The impact of education on risk factors and the occurrence of multimorbidity in the EPIC-Heidelberg cohort. BMC Public Health. 2008 Nov 11;8:384–93. pmid:19014444
- 45. Marengoni A, Angleman S, Melis R, Mangialasche F, Karp A, Garmen A, et al. Aging with multimorbidity: A systematic review of the literature. Ageing Res Rev. 2011 Mar 23;10(4):430–9. pmid:21402176