Unintended pregnancies, defined as pregnancies that are mistimed or unwanted, pose significant public health concern especially in low- and middle-income countries due to their association with adverse health, social, and economic outcomes 1 2 3 4 5 . An obvious consequence of unintended pregnancy is induced abortion, which is often unsafe in countries where the practice is illegal 3 6 . Adverse pregnancy outcomes attributable to unintended pregnancy such as low birth weight, premature birth 1 , maternal depression 7 8 9 , anxiety 10 , poor psychological wellbeing 11 , and poor utilization of antenatal and delivery care 12 13 have been documented. Further documentation of the serious consequences of unintended pregnancy exists from studies within other low- and middle-income countries 12 13 14 15 .

In 2008, an estimated 208 million pregnancies occurred worldwide, out of which 86 million (41%) were unintended. Of these (unintended pregnancies), 33 million (39%) ended in unintended births, 41 million (48%) in abortions, and 11 million (13%) in miscarriages 2 . The 2010 Tanzania Demographic and Health Survey (TDHS) estimated that 26% of births that occurred five years preceding the survey were unintended (22% were mistimed; 4% were unwanted), and these estimates showed no significant change from those observed during the 2004–05 TDHS 16 . Unintended pregnancies contribute significantly to unwanted population growth, which consequently compromises provision of adequate social services 17 . Therefore, elimination of unintended pregnancy is important not only to reduce fertility and the rate of unwanted population growth, but also to enhance the wellbeing of women and their families 18 .

Unintended pregnancies are primarily caused by nonuse and/or failure of contraceptives 17 , implying that correct and consistent use of effective contraceptives can lead to prevention of unintended pregnancies 19 . Results of a paper published in 2013 from a cross-sectional survey conducted in Mwanza, Tanzania, among young adults aged 15–30 years found that older age, lower educational level, unmarried status, lack of knowledge of where to access condoms, increased number of sexual partners, and younger age at sexual debut were significantly associated with unintended pregnancy 19 . In this study, the prevalence of unintended pregnancy was estimated at 30.5% among women who had never been married and 23.7% among women who had ever been married. A community-based survey in Nigeria found that 28.0% of women had experienced an unwanted pregnancy at some point in their lives. Risk factors identified by the study included marital status, parity, place of residence, religion, and socio-economic status 6 . Similarly, a matched case–control study in India reported religion, and socio-economic status (measured by wealth index) as significant predictors of unwanted pregnancy 1 .

Additional predictors of unwanted pregnancy reported from this study were partners’ education of at least high school level and female sex of the most recent child 1 . Unfortunately, behavioral interventions that have attempted to reduce unintended pregnancies have shown no evidence of success.

For example, a cluster randomized trial (CRT) in Tanzania that was conducted to evaluate the MEMA kwa Vijana intervention in Mwanza showed no effect on self-reported unintended pregnancy at any point since its introduction 20 21 . Also an evaluation of the Stepping Stone intervention in South Africa found that women in the intervention arm were 1.45 times more likely than their control counterparts to report unintended pregnancy after 24 months, although the 45% increase was not statistically significant (p = 0.110) 22 . Similarly, the Regai Dzive Shiri project in Zimbabwe showed no effect on unintended pregnancy among young women 23 .

Much is still unknown about the risk factors for unintended or unplanned pregnancy. Also, previous studies have assessed risk factors for unintended pregnancy without unpacking items that actually make up unintended pregnancy and what affects each item.

The literature defines unintended pregnancy within two categories—mistimed (the pregnancy is desired, but at a later time), or unwanted (the pregnancy is not desired at all) 16 24 . While these are often grouped together as unintended pregnancies, there may be differences in the risk factors for mistimed pregnancies versus unwanted pregnancies. Women experiencing mistimed pregnancies likely feel different towards the pregnancy as women experiencing entirely unwanted pregnancies. Therefore, identification of the characteristics associated with mistimed and unwanted pregnancies is vital to understand vulnerable groups and subsequently to inform programs about necessary components for effective interventions. Accordingly, the objectives of this study are (1) to estimate the magnitude of mistimed pregnancy, and unwanted pregnancy, and (2) to identify factors associated with mistimed pregnancy and unwanted pregnancy among women of reproductive age in Rufiji, Kilombero, and Ulanga districts of Tanzania.

Unintended pregnancies pose significant risks to the health and general welfare of individuals and families. As such, measures aimed at reducing mistimed and unwanted pregnancies should be pursued. Implementation of appropriate interventions to reduce unintended pregnancies requires an understanding of the underlying factors that influence such pregnancies. Under this goal, this analysis examined the magnitude and correlates of unintended pregnancy among women of reproductive age in Kilombero, Rufiji, and Ulanga districts of Tanzania.

Results show that close to one half of the women analyzed had not intended their most recent pregnancies (32.5% mistimed and 13.4% unwanted). Although these figures were higher than the national estimates of 22.0% mistimed and 4.0% unwanted from the 2010 Tanzania DHS 16 , in both cases, the proportion of mistimed pregnancies significantly outnumbered the proportion of unwanted pregnancies. The comparatively higher estimates of unintended pregnancies in the study communities over national estimates are partially explained by the increased diversity of contexts accounted for nationally. The 2008–09 Kenya Demographic and Health Survey (KDHS) recorded 43.0% (26.0% mistimed and 17.0% unwanted) of current pregnancies among married women as unintended 27 . This proportion was 45.0% in the 2003 KDHS 28 . These findings from both surveys in Kenya are very similar with those found in our analysis in Kilombero, Rufiji and Ulanga districts of Tanzania. Furthermore, an estimated 46.0% of pregnancies were unintended in East Africa in 2008 (where Tanzania is situated) 2 . This proportion matches the result from our study and is similar to estimates of unintended pregnancies in regions of Asia, Europe, and Latin America and Caribbean 2 .

Multivariate analysis revealed several factors associated with both mistimed and unwanted pregnancies, each independent of the other. Marital status was the strongest factor, with single women showing a proclivity for both mistimed and unwanted pregnancies. This suggests, in part, that single women likely engage in sexual activities for motivations other than childbearing such as pleasure, social status, fiscal gain, or other exchanges. Accordingly, researchers have argued that single women are likely to be young, still in school, and consequently more likely to conceive mistakenly as they are not prepared for childbearing 29 . The observed significance of marital status in predicting unintended pregnancies in the study area is consistent with another study conducted in Tanzania 19 and others from elsewhere 30 31 .

The study results also showed that mistimed pregnancy was most likely to occur in younger aged women, and showed decreasing likelihood with increasing age. This observation supports similar findings from slum settlements of Nairobi, Kenya 31 and findings from Nigeria 6 , but contradicts findings from the Tanzania DHS 16 . As noted previously, many of the younger women are not yet married and are still in school, making them likely to have sexual intercourse for reasons other than childbearing 29 . Therefore, a pregnancy in these situations is likely to be unintended. Additionally, younger women may have lower knowledge and skills surrounding pregnancy control mechanisms such as contraceptive use compared with their older counterparts, thus increasing their likelihood of experiencing unintended pregnancies.

Furthermore, women in relationships where inter-partner discussion about family planning matters was not taking place were significantly more likely than their counterparts to experience unintended pregnancies. Unlike this study, previous studies have ignored inter-partner communication about family planning as a possible predictor of pregnancy intentions, therefore limited comparison of findings. However, extant evidence show that existence of discussions between partners about fertility and family planning-related matters is strongly associated with contraceptive use 32 33 34 35 . This supports the earlier assertion that if effective contraceptives are available and correctly used, unintended pregnancies can be prevented. This underscores the need to empower couples to understand the importance of making family planning decisions jointly 36 since this reduces unplanned pregnancies through increased use of contraception.

Religion was shown to have a significant effect on unintended pregnancy, with traditionalists or those with other unspecified beliefs exhibiting less risk for mistimed pregnancy compared to Christian women, and with Muslim women showing less risk of unwanted pregnancy compared to Christian women. While categorization of religions may be arbitrary depending on contexts, the overall significance of religion in predicting unintended pregnancy shown in this study is consistent with others 1 6 . One study in Mwanza, Tanzania, however, treated religion similar to this study, but yet found no significant relationship between religion and unintended pregnancy 19 . In contrast to our study, other studies have found positive association between Muslim religion and unwanted pregnancy, for example in India, Nepal, and Sri Lanka 37 38 . Further assessment of religion in relation to fertility and population growth exist 39 , and findings are generally mixed, thus a need for further research to reveal the likely underlying mechanisms of the relationship between them.

Finally, the risk of unwanted pregnancy was significantly lower in Ulanga than in Kilombero district. Although specific explanations for this observation are lacking, the variation observed may be attributed to differential access to information and pregnancy control services as well as cultural beliefs, norms, and practices pertaining to pregnancies. Place of residence has been found in other studies as a significant predictor of unwanted pregnancy, where the risk of unwanted pregnancy differs by location 1 6 . In Nigeria for example, Singh and Hussain found a higher risk of unwanted pregnancy in the North than in the South, and also in the rural than in the urban areas 2 . Similarly, Ikamari et al. found in Nairobi that women in non-Slum settlements were more likely than their counterparts in Slum settlements to experience unplanned pregnancy 31 . Therefore place of residence has significant association with pregnancy intentions in the study area.

Close to one half of pregnancies among 15–49 year–old women in Rufiji, Kilombero and Ulanga districts of Tanzania are unintended. Overall, factors affecting mistimed, and unwanted pregnancies are very similar, yet not entirely the same. Women at a higher risk of mistimed pregnancies are younger, single, multi-gravidous (>4 pregnancies) women, and those who lack between–partner communication about family planning. Those at a higher risk of unwanted pregnancies are single, multi–gravidous (>4 pregnancies) women, unlikely to reside in Ulanga district, and have no between–partner communication about family planning. Knowledge of these predictive factors is vital for intervention programs seeking to reduce unintended pregnancy within the study area. Vulnerable groups such as women who are young, unmarried, and the multi–gravidous should be targeted with interventions in order to enable them to avoid unintended pregnancies.

CI: Confidence interval; CRT: Cluster randomized trial; IRB: Institutional review board; KDHS: Kenya demographic and health survey; MDG: Millennium development goals; MRCC: Medical research coordinating committee; NIMR: National Institute for Medical Research; RRR: Relative risk ratio; TDHS: Tanzania demographic and health survey.

The authors declare that they have no competing interests.

AE conceptualized the problem, designed the study, analyzed the data and drafted and revised various drafts of the manuscript. AMK provided assistance with data analysis, study design, and reviewed the manuscript critically. MN, KT, and HV reviewed the manuscript critically. AH and JFP developed a protocol for the Connect Project and reviewed this manuscript critically. All authors read and approved the final version of the manuscript.