Main Organiser

Julius Centre University of Malaya


Department of Social and Preventive Medicine, Faculty of Medicine, University of Malay

Supported by

University of Malaya



Chiu-Wan Ng 1 , Nagaraj Shyamala 2 , Chinna Karuthan 1 , Nai-Peng Tey 3


1 Julius Centre, Department of Social and Preventive Medicine, Faculty of Medicine, University of Malaya
2 Department of Statistics, University of Michigan- Ann Arbor, MI, USA
3 Department of Economics, Faculty of Economics and Administration, University of Malaya


Objectives: Universal health coverage has been advocated as one of the solutions to reduce health disparities especially in developing economies. Malaysia, an upper middle income country, has long achieved universal coverage mainly via a public health delivery system financed through general taxation. The existence of this system which provides comprehensive health benefits has led to complacency and the issue of disparities in health is one which has received little attention in policy and research circles. This study is aimed at examining the extent of a social gradient in premature mortality in the country for all-cause mortality as well as maternal and accidental deaths.

Methods: This analysis used a near complete electronic mortality database in Peninsular Malaysia for the period 1998 to 2002 comprising a total of 457,000 deaths. Deaths in Sabah and Sarawak had been excluded as it was not possible to quantify underreporting of deaths in these states. The mortality data were pooled to estimate the age and sex standardized mortality using the 2000 Malaysian population as the reference population. Premature mortality was assessed using the measure of potential years of life lost (PYLL), defined as the difference between contemporaneous life expectancy and age at death. We used a 5% sample of the Malaysian 2000 national population census to construct an index of socio-economic status (SES) for small areas (82 districts) using principal component analysis of selected household level assets. We estimated the concentration index (CI) as the summary measure of socio-economic disparities in mortality. We examined the social gradient for all- cause mortality as well as maternal and accidental deaths. Data issues preclude examination of other causes of deaths. We also present visual depiction of the mortality distribution using the Geographical Information System (GIS).

Results: Our results show that mortality rates decrease with increasing SES for males and females separately and combined. This social gradient was also observed for maternal and accidental deaths. Premature mortality as measured by PYLL were higher in poorer compared to richer districts and displayed a steeper social gradient among males compared to females. Each year, accidental deaths made up about a fifth of the total PYLL for the year and of these, about three quarters were due to premature deaths among males. In contrast, only about 1% of PYLL each year were attributable to maternal deaths. Premature mortality due to these two causes was again higher in poorer compared to richer districts.

Conclusion: This study provides evidence of a social gradient in premature mortality in a country with universal health coverage lending support to the notion that access to health care per se is not the only factor contributing to good health status. The country needs to also focus attention on other social determinants of health such as education and household economic wellbeing.