Abstract
The age, period and cohort (APC) associations of behavioural and biological NCD risk factors are essential in unveiling underlying population-level changes over time for informed decision-making in health-related policies and interventions. However, there has not been an APC study in Malaysia. Therefore, the present study aimed to determine the trajectories of five behavioural (number of cigarettes smoked per day, total AUDIT score, total METs-min/week, total sitting hours and total serving of fruit/vegetable per day) and six biological (BMI, waist circumference, total cholesterol, blood glucose,
systolic and diastolic blood pressure) NCD risk factors across age, period and cohort (APC) and examine the differences in such trajectories by sex and ethnicity. Subsequently, the mediating roles of these NCD risk factors in the associations between sex and ethnicity with all-cause mortality were also determined. We pooled data from four National Health and Morbidity Surveys (NHMS) conducted in 1996, 2006, 2011, and 2015. The extended Hierarchical APC (HAPC) analysis modelled the APC trajectories using the mixed command in STATA 14.0. Then, the mediation roles of these NCD risk factors were explored using causal mediation analysis in STATA 14.0 via the med4way command. The APC analysis demonstrated that Malaysian adults were getting less healthy as they aged (positive age effect), and younger cohorts replaced older cohorts (positive cohort effect). Generally, the trajectories of behavioural risk factors across age and birth cohorts did not vary by sex. On the other hand, postmenopausal women and
women of more recent cohorts had more pronounced increasing biological NCD risk factor trajectories. Despite increasing sedentarism and low fruit/vegetable intake trajectories, the Chinese had the least pronounced increasing trajectories of biological NCD risk factors over age and cohort than other ethnicities. Mediation analysis revealed that current smoking, physical inactivity and sedentarism are significant mediators that explain about 10%, 2% and 0.5% of the male-excess mortality risk. Furthermore, current smoking alone accounted for almost 20% of the Malay-excess all-cause mortality risk (compared to Chinese). Abdominal obesity accounted for about 9% of male-excess mortality risk. In conclusion, the present study identified important underlying population-level changes in behavioural and biological NCD risk factor trajectories across the life course and birth cohorts, and most of these changes varied by sex and ethnicity. Future studies should consider the sex- and ethnic-specific age and cohort effects in health data surveillance, formulation of health policies and preventive health
measures to tackle the rising prevalence of behavioural and biological NCD risk factors in Malaysia, with a particular emphasis on smoking and abdominal obesity.
random effect, inter-cohort differences, mediation
Links
BibTeX (Download)
@phdthesis{phd-Teh-Chien-Huey-2024,
title = {Sociodemographic Patterns in Non-Communicable Disease (NCD) Risk Factors and Their Impact on Mortality Risk in Malaysia, 1996-2015},
author = {Teh Chien Huey},
year = {2024},
date = {2024-08-22},
abstract = {The age, period and cohort (APC) associations of behavioural and biological NCD risk factors are essential in unveiling underlying population-level changes over time for informed decision-making in health-related policies and interventions. However, there has not been an APC study in Malaysia. Therefore, the present study aimed to determine the trajectories of five behavioural (number of cigarettes smoked per day, total AUDIT score, total METs-min/week, total sitting hours and total serving of fruit/vegetable per day) and six biological (BMI, waist circumference, total cholesterol, blood glucose,
systolic and diastolic blood pressure) NCD risk factors across age, period and cohort (APC) and examine the differences in such trajectories by sex and ethnicity. Subsequently, the mediating roles of these NCD risk factors in the associations between sex and ethnicity with all-cause mortality were also determined. We pooled data from four National Health and Morbidity Surveys (NHMS) conducted in 1996, 2006, 2011, and 2015. The extended Hierarchical APC (HAPC) analysis modelled the APC trajectories using the mixed command in STATA 14.0. Then, the mediation roles of these NCD risk factors were explored using causal mediation analysis in STATA 14.0 via the med4way command. The APC analysis demonstrated that Malaysian adults were getting less healthy as they aged (positive age effect), and younger cohorts replaced older cohorts (positive cohort effect). Generally, the trajectories of behavioural risk factors across age and birth cohorts did not vary by sex. On the other hand, postmenopausal women and
women of more recent cohorts had more pronounced increasing biological NCD risk factor trajectories. Despite increasing sedentarism and low fruit/vegetable intake trajectories, the Chinese had the least pronounced increasing trajectories of biological NCD risk factors over age and cohort than other ethnicities. Mediation analysis revealed that current smoking, physical inactivity and sedentarism are significant mediators that explain about 10%, 2% and 0.5% of the male-excess mortality risk. Furthermore, current smoking alone accounted for almost 20% of the Malay-excess all-cause mortality risk (compared to Chinese). Abdominal obesity accounted for about 9% of male-excess mortality risk. In conclusion, the present study identified important underlying population-level changes in behavioural and biological NCD risk factor trajectories across the life course and birth cohorts, and most of these changes varied by sex and ethnicity. Future studies should consider the sex- and ethnic-specific age and cohort effects in health data surveillance, formulation of health policies and preventive health
measures to tackle the rising prevalence of behavioural and biological NCD risk factors in Malaysia, with a particular emphasis on smoking and abdominal obesity.
random effect, inter-cohort differences, mediation},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
