Scholarlink

<p>Infant mortality remains a critical indicator of population health and health system performance, particularly in low- and middle-income countries. Although South Africa has achieved substantial reductions in child mortality, disparities persist within urban communities, where socio-economic heterogeneity and population mobility may influence maternal and infant outcomes. This study investigated predictors of infant mortality in an urban health and demographic surveillance cohort, with particular emphasis on high-risk pregnancies and rare-event modelling. A population-based analytical study was conducted using baseline data from an urban Health and Demographic Surveillance System covering Hillbrow, Atteridgeville, and Melusi. The cohort comprised 20 119 live births with documented survival outcomes within the first year of life. Infant mortality was defined as death before 12 months of age and further categorised into neonatal and post-neonatal periods. Crude associations were examined using relative risks, and adjusted effects were estimated using Firth penalised logistic regression to account for rare events. Overall survival to one year was high (99.2%), with 157 infant deaths recorded (0.8%). Mortality was heavily concentrated in the neonatal period. Multiple pregnancy was the only variable associated with infant mortality after adjustment (AOR = 34.07; 95% CI: 3.08–377.40; <i>p</i> = 0.004). However, this estimate was characterised by a wide confidence interval, reflecting the small number of deaths (157 among 20,119 live births) and limited precision. In absolute terms, infant mortality occurred in 7 of 522 multiple pregnancies (1.34%) compared with 150 of 19,597 singleton pregnancies (0.77%). Male infants had higher crude mortality than females (95/10,051; 0.95% vs. 62/10,068; 0.62%; RR = 1.54), but this association was not statistically significant after adjustment (AOR = 1.27; 95% CI: 0.15–10.75; <i>p</i> = 0.826). No meaningful temporal trend in infant mortality was observed during the study period (AOR per + 1 year = 0.95; 95% CI: 0.80–1.12; <i>p</i> = 0.515). Infant mortality in this urban cohort was rare and strongly associated with multiple gestation. Among deaths with known timing, most occurred during the neonatal period. These findings underscore the importance of targeted identification and management of multiple pregnancies, while highlighting the potential value of strengthening early neonatal health services to further reduce preventable infant deaths.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictors of infant mortality in a population-based surveillance cohort: the role of high-risk pregnancies and rare-event modelling

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • Phoka C. Rathebe,
  • Masilu D. Masekameni,
  • Khathutshelo Vincent Mphaga,
  • Tintswalo Mercy Hlungwani,
  • Themba Titus Sigudu

摘要

Infant mortality remains a critical indicator of population health and health system performance, particularly in low- and middle-income countries. Although South Africa has achieved substantial reductions in child mortality, disparities persist within urban communities, where socio-economic heterogeneity and population mobility may influence maternal and infant outcomes. This study investigated predictors of infant mortality in an urban health and demographic surveillance cohort, with particular emphasis on high-risk pregnancies and rare-event modelling. A population-based analytical study was conducted using baseline data from an urban Health and Demographic Surveillance System covering Hillbrow, Atteridgeville, and Melusi. The cohort comprised 20 119 live births with documented survival outcomes within the first year of life. Infant mortality was defined as death before 12 months of age and further categorised into neonatal and post-neonatal periods. Crude associations were examined using relative risks, and adjusted effects were estimated using Firth penalised logistic regression to account for rare events. Overall survival to one year was high (99.2%), with 157 infant deaths recorded (0.8%). Mortality was heavily concentrated in the neonatal period. Multiple pregnancy was the only variable associated with infant mortality after adjustment (AOR = 34.07; 95% CI: 3.08–377.40; p = 0.004). However, this estimate was characterised by a wide confidence interval, reflecting the small number of deaths (157 among 20,119 live births) and limited precision. In absolute terms, infant mortality occurred in 7 of 522 multiple pregnancies (1.34%) compared with 150 of 19,597 singleton pregnancies (0.77%). Male infants had higher crude mortality than females (95/10,051; 0.95% vs. 62/10,068; 0.62%; RR = 1.54), but this association was not statistically significant after adjustment (AOR = 1.27; 95% CI: 0.15–10.75; p = 0.826). No meaningful temporal trend in infant mortality was observed during the study period (AOR per + 1 year = 0.95; 95% CI: 0.80–1.12; p = 0.515). Infant mortality in this urban cohort was rare and strongly associated with multiple gestation. Among deaths with known timing, most occurred during the neonatal period. These findings underscore the importance of targeted identification and management of multiple pregnancies, while highlighting the potential value of strengthening early neonatal health services to further reduce preventable infant deaths.