Introduction

In 2016, over 1.9 billion adults over the age of 18 were overweight or obese. This number has tripled since estimates from 1975 [1]. Overweight and obesity are leading metabolic risk factors for many non-communicable diseases (NCDs), including cardiovascular and chronic respiratory diseases, diabetes, and cancer among others. Altogether NCDs account for 71% of all global deaths, over 85% of premature deaths due to NCDs occur in low- and middle-income countries [2].

Decades ago overweight was considered a problem exclusive to high-income countries. However, in the current global scenario prevalence and distribution of overweight are not confined within high-income countries anymore. According to the 2018 Global Nutrition Report, undernutrition has slightly declined whereas anemia has risen to 32.8% among women [3]. According to the 2017 United Nations International Children’s Emergency Fund (UNICEF)-led State of Food Security and Nutrition report, global undernourishment prevalence has decreased since the early 2000s. However, the decline was less than 20% and has begun to reverse since 2015 [4]. Due to the continuous global increase in overweight prevalence, it now exceeds that of underweight in all regions. The ineffective tackling of the problem of underweight, combined with the encroaching problem of overweight, has left many low- and middle-income nations caught under the weight of the double-burden of malnutrition (DBM) [5]. Countries affected by this double-burden must focus on the health-related consequence of being underweight as well as the increasing prevalence of NCDs associated with being overweight [5]. The co-occurrence of multiple forms of malnutrition across individual life histories, socioeconomic levels, as well as within households and countries remain an area of importance [6, 7].

There are a variety of short-term and long-term factors that lead to malnutrition. The most common driver of undernutrition is poverty [5, 8]. This association results in a feedback loop as poverty is associated with lower education. Especially among women this also leads to chronic malnutrition throughout childhood [8]. A similar feedback loop between childhood and adulthood overweight has also been observed [8]. A major reason causing the weight gain is the imbalance between caloric intake and energy expenditure. Two of the leading causes for this imbalance is the increased availability of energy-dense foods and the physical inactivity associated with sedentary lifestyles [9]. It is interesting to note that the average portion sizes of many packaged and restaurant foods have increased while their costs have decreased. Conversely, the cost of fresh food products has increased [10].

Timor-Leste is an island country in Southeast Asia, which occupies the eastern half of the island of Timor. Timor-Leste’s relatively short history as an independent nation (officially recognized in 2002) includes significant bloodshed and political instability. Recent stability has allowed the nation to refocus on achieving economic solvency and rebuilding and strengthening nationwide infrastructure [11]. Timor-Leste has been classified as a Least Developed Country (LDC) since its inception. However, it is on track to graduate from this status in the coming years [12]. In terms of healthcare, this may be a double-edged sword, as the country heads toward an epidemiological transition in terms of malnutrition. As of 2018, 41.3% of Timorese women of reproductive age were anemic, which makes Timor-Leste a significant outlier at its relative economic level [13]. Compared to global averages this is relatively low. However, this number indicates a five-fold increase among adult women specifically in the last decade [13]. Consequently, despite making some progress in Global Nutrition Targets related to under-five weight, progress towards all other targets including adult obesity and diabetes has been lacking significantly [14].

Despite significant research regarding nutrition, there is a lack of nationally representative data on the prevalence and socio-demographic determinants of underweight, overweight, and obesity in the adult population of Timor-Leste. An updated data on the prevalence and risk factors of underweight, overweight, and obesity will help the policymakers and public health managers design strategies and public health interventions effectively. In 2016, the Timor-Leste Demographic Health Survey (TLDHS) was conducted. It collected nationally representative anthropometric data on the 15–49 years population in Timor-Leste. This study aimed to find out the prevalence and factors associated with underweight, overweight and obesity among 15-49-year-old men and women in Timor-Leste using the TLDHS 2016 data.

Methods

Data analysis

At first, descriptive analyses were conducted, and the findings were reported in frequencies and percentages. Then bivariate analyses were performed between the selected covariates and the BMI status. Finally, bivariate logistics regression analyses were used to identify the factors associated with underweight and overweight/obesity, as compared to normal BMI. The variables which yielded a p-value of <0.2 (which was considered enough to control residual confounding) were included in the final multivariable model [20]. Both crude odds ratio (COR) and adjusted odds ratio (AOR) were reported, along with 95% confidence interval (CI). A p-value <0.05 was considered to be statistically significant. All analyses were done using Stata version 16.0.

Ethical consideration

TLDHS 2016 study protocol was approved by the ICF Institutional Review Board (ICF IRB FWA00000845). Before data collection, informed consent was collected from the study participants [15]. Permission for using TLDHS 2016 for this study was obtained from DHS program in March 2021. Due to utilization of publicly available and deidentified data in this study, it was deemed exempted from the by the institutions’ Institutional Review Board (IRB).

Findings

This study analyzed the data from a weighted sample of 16,488 Timorese (aged 15–49 years) to estimate the prevalence of malnutrition and the associated factors among them. The prevalence of normal weight, underweight, and overweight or obesity were found to be 55.2% (95% CI: 54.2%-56.2%), 25.5% (95% CI: 24.4%-26.7%), and 19.3% (95% CI: 18.3%-20.3%), respectively. The description of background characteristics of the respondents as well as the distribution of BMI across different categories of background characteristics are outlined in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Distribution of background characteristics of respondents according to BMI status, (N = 16,488), TLDHS 2016.

https://doi.org/10.1371/journal.pone.0262999.t001

The majority of respondents were female (75.8%), aged 15–24 years (40.9%), hailed from rural areas (68.8%), and belonged to the Dili (23.6%), the capital city of Timor-Leste. More than half of the respondents were married at the time of the survey, whereas more than one-third were not (38.7%).

The distribution of BMI revealed statistically significant variation (P < .05) across different categories of background characteristics. Among the background characteristics, age, sex, type of settlement (urban/rural), township, and wealth, marital, and educational status were found to have statistically significant relationship (p < .05) with BMI (Table 1).

In the logistic regression models (Tables 2 & 3), the associated factors of underweight and overweight or obesity were identified after adjusting for the potential confounders. After adjusting for the covariates, apart from the type of settlement (urban/rural) and educational status, the rest of the background characteristics (i.e. age, sex, township, and wealth and marital status), were found to be statistically significant correlates (p < .05) of underweight (Table 2). The older age groups were found to have lower adjusted odds of being underweight (aged 25–34 years: AOR: 0.7, 95% CI: 0.6–0.7; aged 35–49 years: AOR: 0.6, 95% CI: 0.5–0.7, p < .001) compared to the 15–24-year-old age-group. Females were found to have 20% higher odds of being underweight than males (AOR: 1.2, 95% CI: 1.1–1.3, p < .001). In terms of wealth status, compared to the ‘poorest’ quintile, those who belonged to the higher quintiles had lower odds of being underweight. Also, a statistically significant association was found for all the quintiles except in the case of the ‘poorer’ group. The Timorese who were married at the time of the survey, were found to have 30% lower odds of being underweight (AOR: 0.7, 95% CI: 0.6–0.8, p < .001) than the ones who were not married. The same pattern echoed for the ones who were separated, divorced, or widowed, although a statistically significant association was not found in that case (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Crude and adjusted odds ratio (95% CI) estimates of underweight compared to normal weight by respondent background characteristics, TLDHS 2016.

https://doi.org/10.1371/journal.pone.0262999.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Crude and adjusted odds ratio (95% CI) estimates of overweight/obesity compared to normal weight by respondent background characteristics, TLDHS 2016.

https://doi.org/10.1371/journal.pone.0262999.t003

On the other hand, in the matter of overweight and obesity, all the background characteristics included in this study were found to be statistically significant correlates, after adjustment for the covariates (Table 3). In comparison with the 15–24-year-old age group, the older age groups were found to have around two times higher odds of being overweight or obese (aged 25–34-years: AOR: 1.9, 95% CI: 1.7–2.2, p < .001; aged 35–49 years: AOR: 2.3, 95% CI: 2.0–2.7, p < .001). Females were found to have 40% higher odds of being overweight or obese than males (AOR: 1.4, 95% CI: 1.3–1.6, P < .001). The urban Timorese had 20% lower odds of being overweight or obese than the rural compatriots (AOR: 0.8, 95% CI: 0.7–0.9, p < .001. In terms of educational status, compared to the Timorese having no formal education, those with higher educational attainment had significantly higher odds of being overweight or obese. As to wealth status, compared to the ‘poorest’ quintile, those who belonged to the higher quintiles had higher odds of being overweight or obese; the odds in the ‘richest’ quintile were more than twice as high as that of the ‘poorest’ (AOR: 2.3, 95% CI: 1.9–2.8, p < .001). The Timorese who were married at the time of the survey were found to have greater than twice the odds of being overweight or obese (AOR: 2.2, 95% CI: 1.9–2.5, p < .001) compared to the ones who were not married. The same pattern echoed for the ones who separated, divorced, or widowed, although a statistically significant association was not found (Table 3).

Discussion

To the best of our knowledge, this study is the first one to provide evidence regarding the nationwide prevalence and correlates of underweight, overweight and obesity among 15-49-year-old men and women in Timor-Leste. To ensure optimum generalizability of our findings, we have used nationally representative data from the Timor-Leste Demographic Health Survey (TLDHS) 2016. Specifically, this study has precisely delineated the prevalence of normal weight, underweight, and overweight or obesity to be 55.2% (95% CI: 54.2%-56.2%), 25.5% (95% CI: 24.4%-26.7%), and 19.3% (95% CI: 18.3%-20.3%) among the respondents. Likewise, it revealed that age, sex, township, and wealth and marital status are significant covariates of underweight. Furthermore, it reported age, sex, township, type of settlement, education, wealth and marital status to be significant covariates of overweight or obesity. We expect this study to provide the policymakers with timely evidence regarding the burden of overweight and obesity in the age group of 15–49 years.

The prevalence of underweight, 25.5% (95% CI: 24.4%-26.7%), was alarmingly high compared to Indonesia (11.2%)–which is also an island country that occupied Timor-Leste for more than two decades [21]. However, compared to another Asian country Bangladesh, the prevalence was lower(30.4%). On the other hand, the prevalence of overweight or obesity,19.3% (95% CI: 18.3%-20.3%), was lower compared to that of Indonesia (30.4%), Malaysia (50.2%), Thailand (40.9%), and Bangladesh (23.5%) [19, 21–23]. That said, one important aspect that might explain the discrepancies in prevalence is that in our study the age range of the study participants was from 15–59 years, whereas the studies conducted in other countries considered a different age range. Although all of the aforementioned countries are in Asia, they differ in many socio-geographical aspects that includes differences in health policies and behaviors. This, too, might explain the difference in prevalence rates of underweight, overweight or obesity.

Our study revealed that 21.8% of the respondents received no formal education. As the majority (68.8%) of the respondents hailed from rural (68.8%) areas, it is not possible to ascertain if the proportion of people with no formal education is comparatively lower among the urban population. However, it is probable that there has been an under-ascertainment, as the true proportion of people with no formal education could be slightly higher. Because the majority of the respondents in this survey was people aged 15–24 years (40.9%) and this age group is likely to have higher proportion of formally educated people compared to their older counterparts. Consequently, it could slightly inflate the overall prevalence of people with no formal education.

In our study, Females were found to have 20% higher odds of being underweight than males. This finding is congruent with two similar studies carried out using nationally representative data in the context of Indonesia and Bangladesh [19, 21]. We have also found that those who belonged to the higher wealth quintiles had lower odds of being underweight. This finding is supported by a previous study carried out in the context of Bangladesh where they reported that children from households in the highest wealth index quintile had lower odds of being underweight (OR  =  0.44, 95% CI: 0.37, 0.53) compared to children from households in the lowest quintile [24]. Furthermore, another study from Indonesia buttressed our findings [21].

Interestingly, our study revealed that Timorese who were married at the time of the survey were found to have 30% lower odds of being underweight. A previous study in the context of Bangladesh echoed our findings as they reported, through a pooled analysis, that not being married was positively associated with being underweight [25]. Two other previous studies in the context of Ethiopia and Iran also supported our findings [26, 27]. In many developing countries, being married provides women with greater financial stability which in turn could work as a protective factor from being underweight. Some other factors such as usage of contraceptive pills, weight gain in the postpartum phase etc. are more likely to be prevalent among married women in many countries’ contexts [28, 29]. Although these are plausible hypotheses that might explain the phenomena, a proper factual explanation warrants a deeper exploration of the socio-cultural dynamics of Timor Leste. Further study to explore this would also either help establish or refute any plausible causal connection between being married and having lower odds of underweight.

A previous study from Bangladesh mimicked our finding regarding females having 40% higher odds of being overweight or obese [19]. Our study also noted that urban Timorese had 20% lower odds of being overweight or obese than rural compatriots. However, a study by Hashan et al reported in the context of Malaysia that urban or rural residence was not correlated with being overweight/obese. That said, they found a slightly higher prevalence of overweight/obesity among rural women compared to urban women [17]. This contradiction of findings can be explained by the fact that the demarcation between rural and urban places is not distinct or constant across countries. For example, in some countries, rural and urban women have almost equal access to outdoor physical activities, nutritious foods, and healthcare facilities. However, in some developing countries, rural women are socially expected to work hard in order to complete household chores [30]. They also face more barriers to outdoor physical activities or local gymnasium compared to their urban counterparts [31]. Therefore, the heterogeneity in the difference between rural and urban places, the social expectations from women, and access to physical activities might be some of the issues that explain the contradiction in findings. Our study also found higher odds of obesity/overweight among those who belonged to higher wealth quantile and had higher educational attainment. Templin et al reported, through using 182 Demographic and Health Surveys and World Health Surveys, that low-income countries have a higher prevalence of overweight/obesity among wealthier individuals [32]. Cohen et al systematically analyzed 289 articles from 91 countries and concluded that a positive correlation between higher educational attainment and obesity was more common in lower-income countries [33]. Lastly, we reported that Timorese who were married had greater than twice the odds of being overweight or obese. This evidence is in line with a previous study from Poland which reported similar findings [34].

Government and stakeholders need to take multilayer obesity preventive initiatives including legislative approaches. Additionally, multisectoral initiatives containing pragmatic yet culturally appropriate strategies need to be formulated in order to achieve the objective. Similarly, policy initiatives and effective interventions in healthcare settings are important in order to effectively tackle the issue of malnutrition.

Strengths and limitations

This remains, to date, the first study that leveraged a nationally representative dataset to explore the correlates of underweight, and overweight or obesity in the context of Timor-Leste. We have used an Asia specific BMI index to categorize the BMI which helped us increase the precision of our findings. Given the lack of existing scientific evidence in Timor-Leste, this study will build the base of further research which would in-depth explore these associated factors.

The dataset we used did not have several important socio-behavioral and clinical covariates which are known to be associated with underweight and overweight or obesity. As such, we were not able to explore their association with our outcome variables of interest. Furthermore, the absence of including these variables in our logistic regression model could fail to block the confounding pathways and backdoor paths which in turn could confound our effect estimate. Finally, although the male-female ratio in Timor-Leste is 1.03:1.00, the male-female ratio in our study was 1:3 [35]. This is probably due to the primary target population of TLDHS was the women of the reproductive age group [15]. As a result, the estimated prevalence may not represent the nationally-representative prevalence.

Conclusion

Findings from this study depict a comprehensive picture of the status quo of the burden, distribution, and determinants of underweight, overweight or obesity. We suggest further studies should be carried out to: explore the germane covariates, casually investigate the explanatory-outcome variable relationships, and inspect the socio-cultural and clinical context-specific issues that might affect the prevalence. We further urge that further in-depth studies be carried out using novel approaches building on the evidence reported by our study.

Acknowledgments

We are grateful to the DHS program for providing access to the dataset.