Contextual variation and social determinants of health care non-use among adults aged 50+ in low- and middle-income countries: A Study on WHO Global Ageing and Adult Health (WHO-SAGE)
Independent thesis Advanced level (degree of Master (Two Years)), 10 credits / 15 HE creditsStudent thesis
Objectives: To identify predisposing, individual and household enabling, and need factors associated with not getting health care when last needed (health care non-use) among adults aged 50 and over in the six SAGE countries and to identify community-level variation in health care non-use within each SAGE country.
Methods: A secondary data analysis using data obtained from the World Health Organisation’s Study on Global AGEing and adult health (SAGE) was conducted for China, Ghana, India, Mexico, Russian Federation and South Africa, among adults aged 50+ years. SAGE is a longitudinal study on ageing and adult health with nationally representative samples of persons aged 50 years and above, and comparison samples of 18-49 years, from China, Ghana, India, Mexico, Russian Federation and South Africa. Out of SAGE sample of size of 47443, a total of 24854 respondents are included in the analysis comprising 9671 in China, 3809 in Ghana, 5340 in India, 1124 in Mexico, 2685 in Russia and 2225 in South Africa. Outcome variable was defined as either getting healthcare or not getting healthcare last time it was needed. Simple logistic regression between health care non-use and the predisposing, individual and household enabling, and need factors in six SAGE countries, was undertaken to influence inclusion into multilevel analysis. Multilevel logistic regression was undertaken to determine the contextual variation of health care non-use among older adults residing in different communities in China and India.
Results: Adjusting for individual and household level predictor variables, the final multilevel model indicates an ICC value of 45%, indicating extreme variability in not getting health care last time it was needed attributed to unobserved community-level characteristics. The residual heterogeneity of the 64 communities in China generates a median odds ratio of 4.76 indicating that the median risk of an individual not getting health care when needed is almost 5 times higher when moving from an area with low risk to an area with high risk of not getting health care. Meanwhile in India, a higher ICC value (68%) indicates extreme variability of not getting health care between the 372 communities. The residual heterogeneity of these communities in India generates a median odds ratio of 12.04 which indicates that the median risk of an individual not getting health care when needed is 12 times higher when moving from an area with low risk to an area with high risk of not getting health care.
Conclusion: There exists extreme contextual variation in not getting health care among older adults between communities in China and India. By ignoring contextual variation of not getting health care, assumptions have been made that health care non-use is uniform across all older adults, whilst findings from this study have identified that area-level characteristics play a significant role in determining whether older adults will get health care or not, over and above the predisposing, individual and household enabling, and need factors. Future studies must take into account area-level characteristics with an aim to address disparities in health care non-use.
Place, publisher, year, edition, pages
2016. , 60 p.
Centre for Public Health Report Series, ISSN 1651-341x ; 2016:36
health care non-use; contextual variation; low- and middle-income countries; universal health coverage; ageing; social capital; community-level
Public Health, Global Health, Social Medicine and Epidemiology
IdentifiersURN: urn:nbn:se:umu:diva-131692OAI: oai:DiVA.org:umu-131692DiVA: diva2:1075309
SAGE Data - Paul Kowal
Master's Programme in Public Health
2016-05-23, Umeå Universitet, Umeå, 16:48 (English)
Vaezghasemi, Masoud, PhD StudentWilliam Stewart, Jennifer, Senior research engineer
San Sebastian, Miguel, Professor