Multilevel Analysis of Childbearing in Childhood in Tanzania’s Rufiji District Amon Exavery 1 • Almamy Malick Kante ´ 1,2,3 • Sigilbert Mrema 1 • James F. Phillips 2 • Honorati Masanja 1 Published online: 21 November 2015 Ó Springer Science+Business Media New York 2015 Abstract Objectives To examine levels, trends and cor- relates of childbearing in childhood (CiC) in the Rufiji district of Tanzania from 2002 to 2010. Methods Using longitudinal data collected in, and by, the Rufiji health and demographic surveillance system in Tanzania from 2002 to 2010, all women who initiated childbearing in this period (n = 5491) were selected for analysis. CiC was defined as childbearing initiation before age 18. Data analysis involved one–way tabulations of each variable—most of which were socio-demographic—to obtain frequency dis- tributions, cross-tabulations of CiC and each of the inde- pendent variables with a Chi square test for associations, and multivariate analysis using multilevel logistic regres- sion to examine covariates of CiC. Results CiC was 44 % and remained constant over the 2002–2010 period (P = 0.623). The relative odds of CiC was significantly reduced by 83 percent among women with secondary or higher educational attainment relative to CiC among une- ducated women (OR = 0.17, CI 0.12–0.23). Moreover, the odds of CiC significantly declines monotonically as rela- tive household wealth increases by quintile (OR = 0.70, CI 0.57–0.86). CiC also declines significantly with employ- ment and marital status of the respondent. Conclusions CiC represents a challenging social and health problem. Forty- four percent of first time mothers in Rufiji district of Tanzania are of childhood age, and this has not changed over the past 9 years since 2002. Prioritizing girls’ formal education—especially up to secondary level or higher—as well as devising some economic empowerment modalities, may be worthwhile measures towards curbing CiC in the study area. Keywords Childhood childbearing  Childhood fertility  Adolescent fertility trends  Correlates  Rufiji  Tanzania Significance Previous studies have analyzed the problem slightly dif- ferent by defining it as adolescent pregnancy/births, unin- tended or unwanted pregnancy. Most of them have been rather descriptive with colossal reliance on cross–sectional data. They have so far inferred individual–level associa- tions and have identified among others age, school per- formance, education and economic status as key factors. Although the current study is narrow in its geographical focus as it was conducted in only one district, it is based on longitudinal data accrued over a period of nine years, and applies multilevel models to examine the problem. While some individual level factors observed in the current study such as education and employment status have similarly & Amon Exavery aexavery@ihi.or.tz Almamy Malick Kante ´ amk2213@columbia.edu Sigilbert Mrema smrema@ihi.or.tz James F. Phillips jfp2113@columbia.edu Honorati Masanja hmasanja@ihi.or.tz 1 Ifakara Health Institute, P.O. Box 78373, Dar es Salaam, Tanzania 2 Heilbrunn Department of Population and Family Health, Mailman School of Public Health, Columbia University, New York, NY, USA 3 Swiss Tropical and Public Health Institute, University of Basel, Socinstrasse 57, 4002 Basel, Switzerland 123 Matern Child Health J (2016) 20:447–455 DOI 10.1007/s10995-015-1842-7