Determinants of Under-five Mortality in Cameroon (1991–2018): A Survey-weighted Comparison of Cox and Discrete-time Hazard Models
Abstract:
Identifying independent determinants of under-five mortality (U5M) in Cameroon requires survival methods that properly account for the complex sampling design of DHS data. This study compares survey-weighted Cox proportional hazards regression and a discrete-time hazard model, applied to the same pooled dataset of five DHS waves (1991–2018): 35,257 live births, 2,968 under-five deaths, and 1,821 primary sampling units. Cox models were built sequentially across four covariate blocks, with model discrimination tracked at each step. As a robustness check, a discrete-time logit model with a natural cubic spline for time was fitted to 922,995 person-months. The proportional hazards assumption held globally and for every covariate. In the final Cox model (concordance=0.655), female sex (HR = 0.915, p = 0.033), maternal education (HR = 0.867 per level, p < 0.001), maternal age (HR = 0.956 per year, p < 0.001), and residence in the West region (HR = 0.675, p = 0.001) were significantly protective, while parity was the strongest risk factor (HR = 1.167 per child, p < 0.001), two other regions carried significant excess hazard, and the 2004 survey wave showed an independently elevated hazard (HR = 1.341, p < 0.001) unattenuated by covariate adjustment. The discrete-time model reproduced every Cox finding within 1%, with full concordance across all eighteen coefficients, and five sensitivity analyses left every finding unchanged (concordance 94–100%). These results demonstrate that Cox and discrete-time hazard models yield equivalent, highly robust conclusions for Cameroonian U5M determinants, and that parity reduction, maternal education, and later maternal age represent the most consistent protective targets for child survival policy, alongside the persistently protective profile of the West region.References:
[1].
Rutstein SO, Rojas G. Guide to DHS Statistics.
Calverton: ORC Macro; 2006.
[2].
Croft TN, Marshall AMJ, Allen CK, et al. (Eds). Guide
to DHS Statistics. DHS-7. Rockville: ICF; 2018. Available from: https://dhsprogram.com/pubs/pdf/DHSG1/Guide_to_DHS_Statistics_DHS-7.pdf
[3].
Cox DR. Regression models and life-tables. J R Stat Soc B.
1972;34(2):187-220.
[4].
Allison PD. Discrete-time methods for the analysis of
event histories. Sociol Methodol. 1982;13:61-98.
[5].
Adedini SA, Odimegwu C, Bamiwuye O, et al. Regional variations
in infant and child mortality in Nigeria: a multilevel analysis. J Biosoc Sci.
2015;47(2):165-187.
[6].
Kayode GA, Adekanmbi VT, Uthman OA. Household and
community determinants of childhood mortality in Nigeria. J Trop Med.
2012;2012:601489. https://doi.org/10.1155/2012/601489
[7].
Gwatkin DR, Rutstein S, Johnson K, et al.
Socio-Economic Differences in Health, Nutrition, and Population within
Developing Countries. Washington DC: World Bank; 2007.
[8].
Singer JD, Willett JB. It's about time: using
discrete-time survival analysis to study duration and the timing of events. J
Educ Stat. 1993;18(2):155-195.
[9].
Cleves MA, Gould WW, Marchenko YV. An Introduction to
Survival Analysis Using Stata. 3rd ed. College Station: Stata Press; 2010.
[10].
Conde-Agudelo A, Rosas-Bermudez A, Kafury-Goeta AC.
Birth spacing and risk of adverse perinatal outcomes: a meta-analysis. JAMA.
2006;295(15):1809-23.
[11].
Caldwell JC. Education as a factor in mortality
decline: an examination of Nigerian data. Popul Stud. 1979;33(3):395-413.
[12]. Mosley WH, Chen LC.
An analytical framework for the study of child survival in developing
countries. Popul Dev Rev.
1984;10(Suppl):25-45.
[13].
Institut National
de la Statistique du Cameroun. Cartographie des Indicateurs de Santé au
Cameroun. Yaoundé: INS; 2019.
[14].
World Health Organization Regional Office for Africa.
Meningitis Epidemiological Report, Cameroon 2004. Brazzaville: WHO-AFRO; 2005.
[15].
Munafo MR, Davey Smith G. Robust research needs many
lines of evidence. Nature. 2018;553(7689):399-401. https://doi.org/10.1038/d41586-018-01023-3
[16].
Black RE, Cousens S, Johnson HL, et al. Global,
regional, and national causes of child mortality in 2008: a systematic
analysis. Lancet. 2010;375(9730):1969-87.
[17].
Binder DA. Fitting Cox’s proportional hazards models
from survey data. Biometrika. 1992;79(1):139-47.
[18].
Rutstein SO. Effects of preceding birth intervals on
neonatal, infant and under-five years mortality and nutritional status in
developing countries: evidence from the demographic and health surveys. Int J
Gynaecol Obstet. 2005;89:S7-24.
[19].
Amouzou A, Ziqi M, Carvajal L, Quinley J. Skilled
attendant at birth and newborn survival in Sub-Saharan Africa. J Glob Health.
2017;7(2):020504.
[20].
Yaya S, Bishwajit G, Okonofua F, Uthman OA. Under five
mortality patterns and associated maternal risk factors in sub-Saharan Africa:
A multi-country analysis. PLoS One. 2018;13(10):e0205977.
[21]. Noori N, Proctor JL, Efevbera Y, Oron AP. The effect of adolescent pregnancy on child mortality in 46 low- and middle-income countries. BMJ Glob Health. 2022;7(5):e007681.
