Hostname: page-component-78c5997874-fbnjt Total loading time: 0 Render date: 2024-11-04T21:08:35.117Z Has data issue: false hasContentIssue false

Trends in BMI of Indonesian adults between 1993 and 2014: a longitudinal population-based study

Published online by Cambridge University Press:  13 March 2023

Tri Nisa Widyastuti
Affiliation:
Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, 18 Frederick Street, Dunedin 9016, New Zealand Biostatistics Centre, Division of Health Sciences, University of Otago, 18 Frederick Street, Dunedin 9016, New Zealand
Robin Turner
Affiliation:
Biostatistics Centre, Division of Health Sciences, University of Otago, 18 Frederick Street, Dunedin 9016, New Zealand
Helen Harcombe
Affiliation:
Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, 18 Frederick Street, Dunedin 9016, New Zealand
Rachael McLean*
Affiliation:
Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, 18 Frederick Street, Dunedin 9016, New Zealand
*
*Corresponding author: Email [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Objective:

To examine the trajectories of BMI in Indonesian adults from 1993 to 2014, investigating different patterns by sex and birth cohort.

Design:

Longitudinal study: secondary data analysis of the Indonesian Family Life Survey, a large-scale population-based longitudinal study, had their height and weight measured up to five times throughout the 21-year study period (1993–2014). The change in BMI across time was estimated using group-based trajectory models, then differences by sex and birth cohort were investigated using random effect (mixed) models.

Setting:

Thirteen out of twenty-seven provinces in Indonesia.

Participants:

Indonesian adults aged 19 years and older (n 42 537) were included in the analysis.

Results:

Mean BMI in adults increased between 1993 (21·4 kg/m2) and 2014 (23·5 kg/m2). The group-based trajectory model found three distinct groups with mean BMI increasing more rapidly in the most recent time periods. The first group (56·7 % of participants) had a mean BMI entirely within the normal weight range; the second group (34·7 %) started in the normal weight category and were obese, on average by the end of the study period; and the third group (8·6 %) were always in the obese category, on average. The shape of these three trajectories differed by gender (P < 0·001) and birth cohort (P < 0·001).

Conclusions:

The mean BMI among Indonesian adults has increased between 1993 and 2014, driven by those in the most recent birth cohorts. Our findings support the urgent need for targeted overweight and obesity prevention and intervention programmes in Indonesia.

Type
Research Paper
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2023. Published by Cambridge University Press on behalf of The Nutrition Society

The Global Burden of Disease study shows mean BMI has been increasing since the latter half of the 20th century in many countries(Reference Dai, Li and Bragazzi1). In middle-income countries, this increase has happened much more recently than in high-income countries(2,Reference Abarca-Gómez, Abdeen and Hamid3) . This rapid increase in mean BMI has been associated with increases in a number of non-communicable diseases including CVD, diabetes and some cancers(Reference Dai, Li and Bragazzi1,Reference Seidell and Halberstadt4) . Data collected worldwide shows that global adult obesity prevalence has increased from an estimated 29 % to 37 % in men and 30 % to 38 % in women over the last 30 years. Increases have been particularly rapid in many low- and middle-income countries, such as in Pacific Island countries and the Caribbean, Middle East and Central America(Reference Ng, Fleming and Robinson5,Reference Ford, Patel and Narayan6) .

Several repeated cross-sectional studies using the Indonesian Family Life Survey (IFLS) have reported an increasing trend in mean BMI or overweight/obesity prevalence for adults in Indonesia(Reference Roemling and Qaim7Reference Oddo, Maehara and Rah9). The National Report of Indonesia Basic Health Research conducted by the Indonesia Ministry of Health reported a substantial increase in the prevalence of overweight and obesity among the adult population from 8·8 % to 13·6 % and 10·3 % to 21·8 %, respectively, from 2007 to 2018(1014). These studies were based on cross-sectional designs, so they are limited to estimating prevalence and cannot track individual weight change over time. Development of nuanced obesity prevention/intervention programmes and policies requires a profound understanding of under-recognised characteristics of BMI trends(Reference Vaezghasemi, Razak and Ng8).

Crucial questions such as whether the rising trends in mean BMI might be different between males and females and between birth cohorts, which could identify high-risk population groups and lead to more targeted prevention and interventions, have not yet been addressed for the Indonesian population. To fill this evidence gap, we aimed to examine the trajectories of BMI in Indonesian adults from 1993 to 2014, with specific objectives to examine whether the pattern of mean BMI differs by sex and birth cohort.

Methods

Study design and data sources

We used publicly available (https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS.html) (15) data from the Indonesian Family Life Survey (IFLS), a large population-based longitudinal study conducted over a 21-year period (1993–2014). The survey and sampling methods have been described in detail elsewhere(Reference Frankenberg and Thomas16Reference Strauss, Witoelar and Sikoki19). In brief, there are five waves of data available from 1993, 1997, 2000, 2007 to 2014. The initial sampling frame included randomly selected households from thirteen out of twenty-seven provinces in Indonesia, representing around 83 % of the Indonesian population in 1993. The initial households along with their ‘split-off’ (new additional family member/s due to marriage or new birth) were followed in the surveys conducted in waves 2 to 5. The drop-out rate was considered low (less than 10 %) in all follow-up surveys(Reference Strauss, Witoelar and Sikoki19).

We included all adults aged 19 years and older. Individuals with missing values on age or sex, where these were unable to be determined from any wave, were excluded. Pregnant women and individuals with implausible values of height (less than 100 cm or more than 200 cm) and weight (< 25 kg or > 200 kg) were also excluded from the analysis(Reference Vaezghasemi, Razak and Ng8).

The outcome variable for this study was BMI calculated from weight in kilograms (kg) divided by height in meters squared (m2). Height and weight were measured by trained health personnel such as nurses, midwives or medical doctors(Reference Frankenberg and Thomas16). BMI was analysed as a continuous variable. The WHO criteria for Asian populations BMI cut-off points were used as a reference in the graphs. The WHO BMI for Asian population defines BMI < 18·5 kg/m2 as underweight; 18·5–22·9 kg/m2 as normal weight; 23–24·9 kg/m2 as overweight and > 25 kg/m2 as obese(20).

The explanatory variables were sex and birth cohort/age. Sex was treated as dummy variable as 1 if male and 0 if female. Age was measured in years from the difference in years between self-reported date of birth and the date of measurement (or self-reported age if participants did not know their date of birth) and then categorised in either 5-year or 10-year intervals depending on the level of detail required for the analysis. Birth cohort was classified into a decade of birth and further grouped into 1930s and earlier; 1940s; 1950s; 1960s and 1970–1990s.

Statistical analysis

The characteristics of the sample were summarised using number and percentage for sex, age and birth cohort using the first measure per person (regardless of wave). We initially summarised BMI with mean and sd for each wave (by sex) and by plotting the distribution of BMI for each of the five survey waves both overall and by sex, using a kernel density smoother.

To investigate the changing pattern of BMI across time, by age group and birth cohort, we plotted: (a) mean BMI by age group for different periods, observations within each period connected; (b) mean BMI by period for different age groups, observations within each age-group connected; (c) mean BMI by age for different birth cohort, observations within each birth cohort connected and (d) mean BMI by birth cohort for different age groups, observations within each age-group connected.

We used group-based trajectory model (a finite mixture model) to estimate the shape and number of BMI trajectory groups using a censored normal distribution. The number of groups was decided based on having the largest Bayesian information criterion, and the shape was based on the highest-order polynomial parameter where the P value was less than 0·05. We also constrained the groups to have more than 5 % of the sample assigned to them to avoid a non-parsimonious final model(Reference Nagin21). The adequacy of the final model was assessed using diagnostic procedures: the assignment of the average posterior probabilities for each group should be 0·7 or higher; the odds of correct classification for each group should be 0·5 or higher and the proportion of the observed probability for each group should be similar with the proportion of estimated from the model(Reference Nagin and Odgers22).

Mixed models, with random effects to account for the within-person correlation, were used to test if the identified trajectory groups were different by sex and birth cohort(Reference Singer and Willett23,Reference Diggle24) . We conducted a sensitivity analysis where we excluded participants who had one measure only as they could not contribute information about the within-person change. All analysis was conducted in Stata version 17·0(25,Reference Jones and Nagin26) .

Results

There were 105 820 observations from 42 537 individuals included in the final analysis after exclusion criteria were applied. Figure 1 shows the recruitment and exclusions across the study waves. The description of the included participants at first observation is presented in Table 1. The majority of participants (52·8%) were aged 19–29 years when they first participated in the survey, followed by those aged 30–39 years (22·9%), 40–49 years (10%), 50–59 years (7·1%), 60–69 years (4·3%) and > 70 years (2·9%). The proportion of males and females was similar (48·8 % males, 51·2% females). A higher proportion of participants completed two or more surveys (65·5%) than those who only completed one survey (34·5%).

Fig. 1 Flow diagram of exclusion criteria and total sample

Table 1 Characteristics of the Indonesian adult participants at their first observation

Mean BMI increased from 21·0 kg/m2 in 1993 (sd 3·0) to 22·5 kg/m2 (sd 3·9) in males and from 21·7 kg/m2 (sd 3·7) in 1993 to 24·4 kg/m2 (sd 4·8) in 2014 in females (online Supplementary Material).

The distribution of BMI changed over time from 1993 to 2014 both in males and females (Fig. 2), showing that the proportion of participants within the overweight and obese category has increased with time, more so for females than males.

Fig. 2 The distribution of BMI in Indonesian adults, 1993–2014: (a) in males, (b) in females

Figure 3 shows the mean BMI by age, period of survey and birth cohort. Plot (a) shows the mean BMI by age for each survey wave (period). Mean BMI increased with age then declined with the peak in the 40–44-year age group for each survey wave. Mean BMI increased with each survey wave (1993–2014). Plot (b) shows mean BMI by a wave of survey (period) for different age groups. Mean BMI increased over time from 1993 to 2014 among all age groups. The exception is that mean BMI slightly decreased between 1997 and 2007 and substantially increased after 2007 among the 19–29-year age group. Mean BMI was highest among those aged 40–49 years overall. Plot (c) illustrates mean BMI by age for different birth cohorts. Mean BMI changed with age and the birth cohort. Higher mean BMI was present for those born in the recent decades than the mean BMI of those born in the earlier decades. For instance, those born in 1970s had substantially higher mean BMI (24·6 kg/m2) when they were aged 40–44 years than those born in 1940s when they were at the same age (22·3 kg/m2). More recent birth cohorts also have steeper increases in BMI over time, with the earliest birth cohorts showing a decrease in BMI over time. Plot (d) showing mean BMI by birth cohort for different age groups, confirms the illustration shown in plot c that Indonesian adults born in the recent decades had higher mean BMI than those born in earlier decades when they were the same age.

Fig. 3 Mean BMI of adults: (a): Mean BMI of adults by age for each survey year. (b): Mean BMI of adults by year of survey for different age groups. (c): Mean BMI of adults by age group for different birth cohort. (d): Mean BMI of adults by birth cohort for different age groups

Using the group-based trajectory model, we identified three distinct trajectory groups among Indonesian adults (Fig. 4). Each trajectory is quadratic in shape (with an increasing slope in more recent waves). Group 1, the majority of participants (56·7 %), is characterised by a normal weight on average but increasing. Group 2, around one-third (34·7 %), is characterised by being overweight on average and this is increasing with time. Group 3, the trajectory with the smallest proportion of participants (8·6 %), is characterised by the mean BMI being classified as obese throughout the period and the mean BMI is increasing faster than the other groups. There was a good model fit with an average posterior probability of assignment for each group higher than 0·7, the odds of correct classification were higher than 5·0, and the observed v. estimated proportion assigned to each group was similar.

Fig. 4 Estimated BMI trajectories (solid lines) of adults from the group-based trajectory model by survey wave with observed group means at each survey year (dot), with the actual group percentages in the figure legend

The characteristics of participants in each trajectory group is presented in Table 2. Compared with other trajectory groups, participants in group 1 have a higher proportion of males than females. Participants born in 1970s to 1990s comprise the highest proportion in all groups.

Table 2 The characteristics of people in trajectory each group

Using mixed models, we found that these trajectories were different by both sex (P < 0·001) and birth cohort (P < 0·001). Figure 5 displays the interaction effects for BMI by sex (a) and birth cohort (b). Graph (a) (left) shows the differences in mean BMI between males and females. Among groups 1 and 2, in 1993, males had a higher mean BMI than females, but this was reversed by the last survey wave, with females having a higher mean BMI in 2014. Among group 3, females had a higher mean BMI throughout the entire study period. Graph (b) (right) shows the pattern of mean BMI among different birth cohorts for each of the three trajectory groups. The mean BMI of the youngest cohorts (born in 1970–1990s) increased faster over time in all three trajectory groups compared to older birth cohorts (born 1940–1960s). In contrast, the mean BMI in the oldest birth cohort (born in 1930s and earlier) remained stable over time or showed a slight decrease among group 1.

Fig. 5 Mean BMI per trajectory groups, by sex and birth cohort: (a) Mean BMI per trajectory groups by sex. (b) Mean BMI per trajectory groups by birth cohort

Discussion

This study demonstrates mean BMI has increased among Indonesian adults between 1993 and 2014. Longitudinal analysis shows that the increase in mean BMI over the period is being driven by younger cohorts (born in 1970—1990s) who have a higher mean BMI and a greater increase in BMI over time than those from older cohorts. There are also differences by sex, with females having a higher mean BMI than males in the latest survey wave. Using group-based trajectory methods, we identified three groups. All three groups have quadratic increases in mean BMI over time, with mean BMI of group 3 increasing more than groups 1 and 2 (Fig. 4). The greatest increase in mean BMI for all three groups has occurred since 2007. Group 3 (representing 8·6 % of the participants) shows the greatest increase in BMI over the period 1993 to 2014 with mean BMI increasing from 27·1 kg/m2 to 32·6 kg/m2.

Three previous studies examining BMI among Indonesian adults have also found an increased prevalence of overweight and obesity over time; however, these have been based on repeated cross-sectional analysis(Reference Vaezghasemi, Razak and Ng8,Reference Linando and Lee27,28) . Oddo et al. (2019) reported a two-fold increase in overweight prevalence from 17·1 % in 1993 to 33 % in 2014 based on analysis of the IFLS as five repeated cross-sectional surveys rather than longitudinal analysis(Reference Oddo, Maehara and Rah9). Another study using four cross-sectional surveys of the IFLS up until 2007 reported that obesity prevalence increased two-fold both in males (4·1 % in 1993 to 8·9 % in 2007) and females (9·7 % in 1993 to 19·6 % in 2007)(Reference Roemling and Qaim7). Using the same the IFLS data, Vaezghasemi et al. (2016) reported that the mean BMI of adults increased from 21·5 kg/m2 in 1993 to 22·8 kg/m2 in 2007(Reference Vaezghasemi, Razak and Ng8). All of these studies showed that females have higher overweight and obesity prevalence as well as higher mean BMI compared to males.

The findings of the current study are also in line with the current trend of overweight and obesity seen in many low-middle income countries in the Pacific, the Caribbean, the Middle East, Central America and South East Asia(2,Reference Ng, Fleming and Robinson5) . These studies suggest that obesity prevalence is now increasing at a faster rate in low-middle income countries than in high-income countries, although these increases started earlier in high-income countries (Reference Ng, Fleming and Robinson5). This study confirms that this is occurring in Indonesia(Reference Ford, Patel and Narayan6,Reference Jaacks, Vandevijvere and Pan29) .

Studies examining trajectories of BMI using longitudinal study designs in low-middle income countries setting are scarce. A study in China using longitudinal data found four distinct trajectories of weight change with a quadratic effect among the overweight to obese group(Reference Zheng, Du and Zhang30). There are however several studies that have examined BMI trajectories using longitudinal data in high-income countries. Three of these studies from USA, Finland and Australia reported that BMI increased with age from early adulthood into middle age and tended to decrease at older ages(Reference Becnel and Williams31Reference Yang, Lynch and Dugué33). Similarly, the current study also portrayed complex patterns in longitudinal data sets by revealing three distinctive development trajectories of BMI in population. Indonesia is currently experiencing a Double Burden of Malnutrition, indicated by a high prevalence of under-nutrition problems such as underweight, stunting and wasting in conjunction with overweight and obesity issues(Reference Shrimpton and Rokx34). We believe that BMI change in Indonesia cannot be fully described by a single parameter. It is crucial to take the variance in the population into account in order to prioritise obesity interventions targeted to high-risk groups.

Another important finding from this study was that individuals who were already obese at the beginning of the survey tended to increase their BMI much faster over time than those who started from a normal weight. This suggests that the prevalence of overweight and obesity in Indonesia will continue to increase and potentially lead to increases in non-communicable disease prevalence if major prevention and intervention actions are not taken(Reference Wells, Sawaya and Wibaek35).

This is the first study observing trends of BMI in Indonesia adults examining different patterns of trajectories by sex and birth cohort. Strengths of the current study include the use of longitudinal analyses from a large population-based longitudinal survey covering a 21-year period, with more than 90 % re-contact rates at each wave (65 % of participants have two or more measures). Therefore, the dynamic and individual differences in growth rate can be captured through different patterns in BMI trajectory(Reference Ryder, Jacobs and Sinaiko36). The longitudinal nature allows us to investigate how individual people change over time(Reference Fitzmaurice, Laird and Ware37). A further strength is that the anthropometry assessments were taken by trained health workers rather than by self-report. There were some limitations; not all individuals participated in all five waves of data collection, leading to missing data and loss of follow-up(Reference Longhi38). However, the number lost to follow-up was relatively small (less than 10 % per wave). Moreover, the use of mixed model analysis allows us to include all eligible participants regardless of how many times they participated in the study(Reference Singer and Willett23,Reference Diggle24) . BMI is a widely used measure in population-based epidemiological studies; however, there are limitations to BMI as a useful measure in some individuals and population groups(Reference Solorzano, Stevens and Doak39). BMI does not distinguish the origin of body weight (e.g. from muscle mass or fat mass) as it is a measure of body size rather than composition. Accurate assessment of body composition in individuals involves more complex and expensive tools such as dual-energy X-ray absorptiometry(Reference Hu40).

In conclusion, this study provides robust new evidence on the trajectories of BMI among Indonesian adults. The mean BMI patterns confirm that Indonesia is in the process of a rapid and intense nutrition transition(Reference Popkin, Corvalan and Grummer-Strawn41). Both males and females show increased BMI over the period on average. However, the more recent generations show greater increases in mean BMI than the earlier generations. The potential implications of the positive quadratic effect (with greater increases in mean BMI over the period from last three surveys) indicate an accelerated increase in BMI that is alarming in relation to chronic disease risk and the potential impact on public health and health services.

Further research using the same and/or other datasets to examine risk factors of overweight/obesity and different population such as children and adolescents are needed to continue to gain a better understanding of the epidemiology of overweight and obesity in the Indonesian population.

Acknowledgements

Authorship: T.N.W. acquired data, performed data analysis and wrote the initial draft. R.T. guided the statistical methods and contributed to the analysis plan. T.N.W., R.T., H.H. and R.M. developed the paper framework. All authors contributed to the interpretation of results, reviewed the drafts of the manuscript and approved the final version of manuscript. Ethics of human subject participation: The consultation has been made with the Chair of the Preventive and Social Medicine (PSM) Ethics Committee University of Otago regarding the ethics approval requirement and suggested that there is no need to pursue an ethics approval from the department. Given this study is using secondary data that has already been through an ethical review process from relevant institutions.

Financial support:

None.

Conflict of interest:

There are no conflicts of interest.

Supplementary material

For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S1368980023000472

References

Dai, H, Li, F, Bragazzi, NL et al. (2020) Distinct developmental trajectories of Body Mass Index and diabetes risk: a 5-year longitudinal study of Chinese adults. J Diabetes Investig 11, 466474.CrossRefGoogle Scholar
NCD Risk Factor Collaboration (2016) Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants. Lancet 387, 13771396.CrossRefGoogle Scholar
Abarca-Gómez, L, Abdeen, ZA, Hamid, ZA et al. (2017) Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128·9 million children, adolescents, and adults. Lancet 390, 26272642.CrossRefGoogle Scholar
Seidell, JC & Halberstadt, J (2015) The global burden of obesity and the challenges of prevention. Ann Nutr Metab 2, 712.CrossRefGoogle Scholar
Ng, M, Fleming, T, Robinson, M et al. (2014) Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 384, 766781.CrossRefGoogle ScholarPubMed
Ford, ND, Patel, SA & Narayan, KMV (2017) Obesity in low- and middle-income countries: burden, drivers, and emerging challenges. Ann Rev Public Health 38, 145164.CrossRefGoogle ScholarPubMed
Roemling, C & Qaim, M (2012) Obesity trends and determinants in Indonesia. Appetite 58, 10051013.CrossRefGoogle ScholarPubMed
Vaezghasemi, M, Razak, F, Ng, N et al. (2016) Inter-individual inequality in BMI: an analysis of Indonesian Family Life Surveys (1993–2007). SSM – Population Health 2, 876888.CrossRefGoogle ScholarPubMed
Oddo, VM, Maehara, M & Rah, JH (2019) Overweight in Indonesia: an observational study of trends and risk factors among adults and children. BMJ Open 9, e031198.CrossRefGoogle ScholarPubMed
Indonesia Ministry of Health (2007) National Report of Indonesia Basic Health Research 2007–2008. Jakarta: Agency of Health Research and Development – Indonesia Ministry of Health.Google Scholar
Indonesia Ministry of Health (2010) National Report of Indonesia Basic Health Research 2010. Jakarta: Agency of Health Research and Development – Indonesia Ministry of Health.Google Scholar
Indonesia Ministry of Health (2013) National Report of Indonesia Basic Health Research 2013. Jakarta: Agency of Health Research and Development – Indonesia Ministry of Health.Google Scholar
Indonesia Ministry of Health (2016) Report of National Health Indicator Survey 2016. Jakarta: Agency of Health Research and Development – Indonesia Ministry of Health.Google Scholar
Indonesia Ministry of Health (2018) National Report of Indonesia Basic Health Research 2018. Jakarta: Agency of Health Research and Development – Indonesia Ministry of Health.Google Scholar
RAND Corporation (n.d.) The Indonesia Family Life Survey (IFLS). https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS.html (accessed November 2019).Google Scholar
Frankenberg, E & Thomas, D (2000) The Indonesia Family Life Survey (IFLS): Study Design and Results from Waves 1 and 2. Contract No.: DRU-2238/1-NIA/NICHD. http://www.vanneman.umd.edu/socy699j/ifls2design.pdf (accessed November 2019).Google Scholar
Strauss, J, Witoelar, F, Sikoki, B et al. (2004) The Third Wave of the Indonesia Family Life Survey (IFLS3): Overview and Field Report. Contract No.: WR-144/1-NIA/NICHD. https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS/ifls3.html (accessed November 2019).Google Scholar
Strauss, J, Witoelar, F, Sikoki, B et al. (2009) The Fourth Wave of the Indonesia Family Life Survey (IFLS4): overview and Field Report April. Contract No.: WR-675/1-NIA/NICHD. https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS/ifls4.html (accessed November 2019).Google Scholar
Strauss, J, Witoelar, F & Sikoki, B (2016) The Fifth Wave of the Indonesia Family Life Survey (IFLS5): overview and Field Report March. Contract No.: WR-1143/1-NIA/NICHD. https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS/ifls5.html (accessed November 2019).CrossRefGoogle Scholar
World Health Organization (2000) Regional Office for the Western Pacific. The Asia-Pacific Perspective: Redefining Obesity and its Treatment. Sydney: Health Communications Australia.Google Scholar
Nagin, D (1999) Analyzing developmental trajectories: a semi parametric, group-based approach. Psychol Meth 4, 139157.CrossRefGoogle Scholar
Nagin, DS & Odgers, CL (2010) Group-based trajectory modeling in clinical research. Ann Rev Clin Psychol 6, 109138.CrossRefGoogle ScholarPubMed
Singer, JD & Willett, JB (2003) Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. Oxford: Oxford University Press.CrossRefGoogle Scholar
Diggle, P (2002) Analysis of Longitudinal Data, 2nd ed. Oxford: Oxford University Press.Google Scholar
StataCorp (2021) Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC.Google Scholar
Jones, BL & Nagin, DS (2013) A note on a Stata plugin for estimating group-based trajectory models. Sociol Meth Res 42, 608613.CrossRefGoogle Scholar
Linando, AY & Lee, SK (2018) Socioeconomic, dietary, nutrition, body weight and epidemiologic transitions: three nations at different stage of development. Prog Nutr 20, 602615.Google Scholar
Indonesia Ministry of National Planning Development (2019) Nutrition Development in Indonesia. Directorate of Public Health and Nutrition – Deputy for Human Development. Jakarta: Agency of National Planning Development – Indonesia Ministry of National Planning Development.Google Scholar
Jaacks, LM, Vandevijvere, S, Pan, A et al. (2019) The obesity transition: stages of the global epidemic. Lancet Diabetes Endocrinol 7, 231240.CrossRefGoogle ScholarPubMed
Zheng, R, Du, M, Zhang, B et al. (2018) Body mass index (BMI) trajectories and risk of colorectal cancer in the PLCO cohort. Br J Cancer 119, 130132.CrossRefGoogle ScholarPubMed
Becnel, JN & Williams, AL (2019) Using latent class growth modeling to examine longitudinal patterns of body mass index change from adolescence to adulthood. J Acad Nutr Dietetics 119, 18751881.CrossRefGoogle ScholarPubMed
Salmela, J, Lallukka, T, Mauramo, E et al. (2020) Body mass index trajectory–specific changes in economic circumstances: a person-oriented approach among midlife and ageing finns. Int J Environ Res Public Health 17, 3668.CrossRefGoogle ScholarPubMed
Yang, Y, Lynch, BM, Dugué, P-A et al. (2021) Latent class trajectory modeling of adult Body Mass Index and risk of obesity-related cancer: findings from the Melbourne collaborative cohort study. Cancer Epidemiol Biomarkers & Prev 30, 373.CrossRefGoogle ScholarPubMed
Shrimpton, R & Rokx, C (2013) The Double Burden of Malnutrition in Indonesia, World Bank Jakarta, Report 76192-ID. Report No.: 76192-ID. https://documents1.worldbank.org/curated/en/955671468049836790/pdf/761920WP0P12640Box0379884B00PUBLIC0.pdf (accessed September 2021).Google Scholar
Wells, JC, Sawaya, AL, Wibaek, R et al. (2020) The double burden of malnutrition: aetiological pathways and consequences for health. Lancet 395, 7588.CrossRefGoogle ScholarPubMed
Ryder, JR, Jacobs, DR, Sinaiko, AR et al. (2019) Longitudinal changes in weight status from childhood and adolescence to adulthood. J Pediatr 214, 187192.e2.CrossRefGoogle ScholarPubMed
Fitzmaurice, GM, Laird, NM & Ware, JH (2011) Applied Longitudinal Analysis. Chicester: John Wiley & Sons, Incorporated.CrossRefGoogle Scholar
Longhi, S (2015) A Practical Guide to Using Panel Data. London; Thousand Oaks, CA: Sage Publications Ltd.CrossRefGoogle Scholar
Solorzano, AA, Stevens, SM & Doak, CM (2022) Misconceptions in the use of Body Mass Index: a review of the literature. Nutr Today 57, 329335.CrossRefGoogle Scholar
Hu, FB (2008) Obesity Epidemiology. Oxford, NY: Oxford University Press.CrossRefGoogle Scholar
Popkin, BM, Corvalan, C & Grummer-Strawn, LM (2020) Dynamics of the double burden of malnutrition and the changing nutrition reality. Lancet 395, 6574.CrossRefGoogle ScholarPubMed
Figure 0

Fig. 1 Flow diagram of exclusion criteria and total sample

Figure 1

Table 1 Characteristics of the Indonesian adult participants at their first observation

Figure 2

Fig. 2 The distribution of BMI in Indonesian adults, 1993–2014: (a) in males, (b) in females

Figure 3

Fig. 3 Mean BMI of adults: (a): Mean BMI of adults by age for each survey year. (b): Mean BMI of adults by year of survey for different age groups. (c): Mean BMI of adults by age group for different birth cohort. (d): Mean BMI of adults by birth cohort for different age groups

Figure 4

Fig. 4 Estimated BMI trajectories (solid lines) of adults from the group-based trajectory model by survey wave with observed group means at each survey year (dot), with the actual group percentages in the figure legend

Figure 5

Table 2 The characteristics of people in trajectory each group

Figure 6

Fig. 5 Mean BMI per trajectory groups, by sex and birth cohort: (a) Mean BMI per trajectory groups by sex. (b) Mean BMI per trajectory groups by birth cohort

Supplementary material: File

Widyastuti et al. supplementary material

Tables S1-S5

Download Widyastuti et al. supplementary material(File)
File 52 KB