Intergenerational Differences in Current Contraceptive Use among Married Women in Uganda
We analyze variations in contraceptive use and age cohort effects of women born from 1957 to 1991 based on the hypothesis that individuals born at different time periods experience different socio-economic circumstances. Differential exposure to socioeconomic circumstances may influence women's attitudes and behavior towards critical life issues such as contraceptive use. We use data from the 2006 Uganda Demographic and Health Survey by limiting the analysis to 5,362 women who were currently married (or cohabiting) at the time of the survey. Logistic regression analyses show a higher likelihood of contraceptive use among older cohorts than younger cohorts. These results call for interventions that target young women to reduce fertility and reproductive health challenges associated with too many or too closely spaced births. Although variations in age cohorts are critical in influencing current contraceptive use, other factors such as wealth status, education level, and place of residence are equally important. Thus, interventions that focused on selected regions (e.g., the North and the West Nile), rural communities, and efforts to increase literacy levels will accelerate uptake of contraception and improve maternal and child health.
Age, period, and cohort effects on sociodemographic outcomes are typically interlinked and generally, researchers cannot study them in isolation. Age is a summary measure of exposure to social influences or norms. Therefore, to understand variations in health and demographic rates, it is important to measure the underlying variables for which age is a proxy . Specifically, age cohort differences refer to variations in effects that people experience by virtue of being born during different time periods . A cohort effect is an outcome of an exterior event that affects individuals born in the same period who eventually share a common history that transforms their exposure to the “treatment” such as education or health policy. In this study, we examine the effect of age cohort differences on contraceptive use among Ugandan women (currently married or cohabiting) interviewed in the 2006 Uganda Demographic and Health Survey (DHS). Our study is motivated by the premise that people born during different periods experience different events and processes that are likely to shape their behavior and attitudes towards critical reproductive health issues such as contraceptive use .
Contraception, defined as methods or devices used to prevent pregnancy, is categorized into two types: modern and traditional methods. Modern methods include clinic and supply methods such as the pill, intrauterine device (IUD), condom, and sterilization whereas traditional methods include periodic abstinence (rhythm), withdrawal, and folk methods . Women in the Uganda DHS were asked if they were using anything to delay or stop child bearing altogether and those who were currently using modern or traditional methods were regarded as current contraceptive users.
Contraceptive use in Uganda still remains low despite the increase from 15.4% in 1995 to 18.6% in 2000/01 and 24.4% in 2006. The increase in contraceptive use accelerated in the five years prior to the 2006 DHS than in the 1990s due to improvements in economic growth, literacy, and education . The increase in contraceptive use may have been influenced by complex interactions between microlevel factors such as age and education as well as macrolevel factors such as urbanization and improvements in health service delivery.
According to the cohort historical model, social changes are mainly influenced by the interaction between individuals’ life experiences and changes in the socioeconomic environment [4–9]. Therefore, new cohort members tend to bring about new rules of behavior and alter the social structure in order to fulfill social and economic needs. Such normative and structural changes, along with changes in the size of the cohorts, and other environmental factors can lead to inter cohort differences. This is possible since one cohort is virtually different form the other due to differences in circumstances associated with birth and growth.
1.1. Recent Socioeconomic Developments in Uganda
The aftermath of independence (gained in 1962) and early 1970s until 1986 was characterized by upheavals and inconsistencies in the economic, political, and social life of Ugandans. Uganda’s economic potential was one of the most vibrant and promising at the time of her independence but was curtailed by 15 years of civil strife between 1966 and 1986. During this period, Uganda’s poverty was rampant, the population grew at an alarming rate and characterized by high dependency rates, and levels of schooling were very low. As a result, social and health systems were dysfunctional and contraceptive use was negligible. These dynamics were followed by an era of HIV/AIDS from 1982 (when the first case was diagnosed) to its peak in the early 1990s. This period was also characterized by an increased proportion of orphans. Despite these challenges, the Ugandan Government implemented a universal primary education policy in 1997 that provided an opportunity for orphans and other Ugandan children to be educated [10–13]. Since then, the economy and selected indicators have shown tremendous improvement .
The aforementioned dynamics, and the civil strife in particular, affected Uganda’s health system and compromised its ability to provide adequate services to people. For instance, although contraceptive prevalence rate increased between 1995 and 2006, it is still very low compared to other sub-Saharan African countries such as Kenya (45.5% in 2008/09), Tanzania (34.4% in 2010), and Zambia (40.8% in 2007) . Further, 41% of the Ugandan married or cohabiting women had unmet need for family planning in 2006. This means that more than a third of the women would desire to use contraception but they do not, due to various factors such as availability of and access to contraceptives. Since low contraceptive use is associated with high fertility, Uganda’s total fertility rate of 6.7, coupled with an annual growth rate of more than 3%, makes the country one of the highest growing population in the world.
Cohort studies account for differences in the attitudes and practices that people exhibit in relation to utilization of services. Hence, cohort studies can be used to demonstrate and document the factors that influence many demographic processes. Majority of age cohort studies have not focused on the effect on contraceptive use in Uganda. Therefore, our aim is to establish the effect of age cohorts on contraceptive use in Uganda and to find out the extent to which contraceptive uptake is affected by other micro- and macrolevel factors.
2. Data and Methodology
2.1. The Sample
Data come from the 2006 Uganda DHS, which was the first survey to cover the entire country compared with previous surveys that were obstructed by insecurity. The 2006 survey was also the fourth in a series of DHS in Uganda. In the survey, a nationally representative sample of 8,531 women aged 15–49 years was selected for interviews. For this study, women living in marital unions and born within the periods 1957–1971, 1972–1981, and 1982–1991 were selected. These periods were translated into age cohorts of 15–24 years, 25–34 years, and 35–49 years, respectively. This cohort allocation was translated into an analytical sample of 5,362 women.
The dependent variable is current contraceptive use with the birth or age cohort as the key independent variable. The dependent variable was coded as “1” when a woman was using contraception at the time of the survey and “0” if otherwise. The age cohorts were grouped into three categories namely, 15–24, 25–34, and 35–49 years. Other independent variables included in the analysis are residence of the woman (urban or rural); parity (because the number of children ever born will determine the need for family planning); education level; region of residence; and wealth index (computed from information on household durable assets such as radio, bicycle, and car ownership).
We use logistic regression analysis because the dependent variable (current contraceptive use) is binary. Analysis was done at three levels and conclusions were drawn at the multivariate level. In the first level of analysis, we include age cohort as a baseline model, whereas the second model adds education level, urban/rural residence, and wealth index. In the final model, we include parity and region of residence to assess the effect of age cohort on current contraceptive use before and after controlling for other factors.
We use the birth cohort since people who were born and grew up during different periods experience different circumstances that may shape their attitudes and practices related to reproductive health. For instance, one would expect that women born between 1982 and 1991 (aged 15–24 years) are more likely to use contraceptives than their counterparts because they were born when Uganda experienced a number of political and socioeconomic changes. As a result, these women were more likely to have benefited from family planning services during that period. Women born during the 1972–1981 period may not report more contraceptive use at the time of the survey than those born much earlier. This may be related to the fact that these women were born during a period of political and socioeconomic turmoil, associated with low levels of schooling.
Table 1 presents the background characteristics of women and results show that 77.6% did not use contraceptives at the time of the survey. Majority (24.3%) of contraceptive users were in the old age cohort (1957–1971) compared with 18.4% of women in the young age cohort who used contraceptives. The mean age of women in the sample was 30.7 years and the mean parity was 4.6 children. More than half (58.5%) of women had only primary schooling while 15.2% reported having secondary or higher education. Close to a fifth (18.6%) of women with secondary or higher education were in the young age cohort, while those without any education had the majority (40.3%) in the old age cohort (1957–1971).
Women from poor households (based on the wealth status index) were a majority (44.8%) and those from middle status households accounted for 18.0%. That more women from poor households were in the young age cohort (1982–1991) may be related to early marriages that are usually associated with low rates of schooling completion and economic challenges. Close to nine out of 10 women (87.4%) were rural residents, and most of them were from the Northern (22.1%) and Eastern (12.4%) regions.
Table 2 presents the results on the association between current contraceptive use and selected characteristics of women. Age cohort was significantly associated with contraceptive use (). More (81.6%) women in the young age cohort were not using contraceptives while 24.3% of women in the old cohort used contraception. There was a slight difference of 1.5% in contraceptive use between the middle age cohort and the old age cohort.
About nine out of 10 women (88.9%) with no schooling did not use contraceptives whereas, 44.8% of those with secondary and higher schooling used contraceptives. Consistent with the literature , we find that more women from rich households (37.5%) used contraception, while only 11.2% of women from poor households did so. Consistent with the literature , Table 2 shows that more women who had between 3–5 children and 6+ children were using contraception. More women in Kampala (48.5%), Central 1 (33.7%), and Central 2 (36.2%) used contraception. Only 8.9% and 13.6% of women in the North and West Nile regions, respectively, used contraceptives. Contraceptive use was also low in Eastern (19.8%) and Western (20.5%) regions of the country.
Table 3 presents regression results of the effect of age cohort on current contraceptive use. Model 1 presents the baseline effect of age cohort on contraceptive use. In Model 1, women in the older age cohort were 1.42 times more likely to use contraceptives than those in the younger age cohort. Women in the middle age cohort were 1.38 times more likely to use contraception that those in the young age cohort. After controlling for education level, wealth status, and residence in Model 2, results show that the odds of using contraceptives slightly increased for all age cohorts (1.44 for 1972–1981 and 1.86 for the 1957–1971 age cohorts, resp.).
In Model 2, the odds of current contraceptive use for women with primary education almost doubled (OR = 1.96) compared with women with no schooling. Women with secondary and higher schooling were 3.52 times more likely to use contraceptives than those without schooling. When residence was included in Model 2, rural residents were 35% less likely to use contraceptives than urban residents. Belonging to households categorized as middle income was associated with 64% likelihood of using contraceptives, whereas those belonging to rich households had odds that almost tripled (OR = 2.94) in using contraceptives.
In Model 3, the effect of age cohort on current contraceptive use was no longer significant after including parity and region of residence. Nevertheless, education, residence, and wealth status were still robust though their effects were attenuated (except for secondary schooling). When compared to women with 0–2 children, the odds for current contraceptive use were higher for women with six or more children (OR = 2.03) than for women with 3–5 children (OR = 1.71). This finding could be explained by the desire to limit births among older women with six or more children while for the young cohorts, they still desire to have children hence were less likely to use contraception. Living in Central 2, Kampala, and South West regions was not associated with contraceptive use. However, women in East Central, Eastern, North, West Nile, and Western regions were 36%, 33%, 59%, 52%, and 30%, respectively, less likely to use contraceptives than women living in Central 1 region.
The importance of age cohort differences in health and demographic outcomes has been highlighted elsewhere [2, 4]. This study assessed the effect of age cohort on current contraceptive use among women in marital union interviewed in the 2006 Uganda DHS. Results showed that older age cohorts were more likely to use contraceptives than younger cohorts aged 15–24 years. These findings complement those reported elsewhere [18, 19]. For example, a study from Indonesia  found that younger women not only exhibit higher discontinuation rates but also tend to use short-term methods than older women. When other factors such as education, wealth status, and residence were controlled for, this study showed that the effect of age cohort on current contraceptive use reduced significantly. This means that the effect of age cohort on current contraceptive use is influenced by other factors thereby confirming the premise that age cohorts are summary indicators of variations in differential exposure to lifetime experiences .
Contraceptive use increases with increased schooling and is highly significant with secondary and higher education than with primary education. This is partly explained by the fact that secondary schooling allows a woman to stay in school longer thereby reducing the risk of exposure to marriage. Educated women are also more likely to earn an income, understand their physiology and social needs, and adopt appropriate reproductive health behavior [20–22].
The results also showed that contraceptive use increased with increased parity. This finding may partly be related to the fact that older women desire to limit their births because they have reached their desired family size. Results from the 2006 Uganda DHS also confirm that current contraceptive use is higher for limiting (12.7%) than for spacing (11.0%) .
Compared with women in Central 1, living in the North and the West Nile regions reduced the likelihood of using contraceptives by more than half. The Northern region was hit hard by the 18-year-old war led by Joseph Kony, which forced people to live in camps with limited family planning services. Schooling opportunities during the civil strife were virtually nonexistent and this negatively impacted the socio-economic life of the people in the region .
Compared with the poor, the rich were associated with odds that more than doubled in using contraception. In addition, most of the contraceptive nonusers were young. Coming from a middle status household was not associated with contraceptive use partly because wealth status is also associated with education. If a woman is from a middle status household but not educated, there may not be significant differences between her and the noneducated women because they are not informed about the benefits of family planning. Rural residence was also associated with lower odds of contraceptive use among women than urban residence, consistent with earlier studies in Uganda .
Generally, the effect of age cohort on current contraceptive use was higher among older women than younger women. This finding is of great concern since young women who are more likely to be exposed to pregnancy and childbirth for longer periods need contraception for spacing their births. It is not surprising that under such circumstances, Uganda’s total fertility rate (TFR) was high at 6.7 in 2006 (7.1 in rural areas, 4.4 in urban areas). The high TFR may be associated with women who are exposed to sexual intercourse frequently (either in marital unions or otherwise) without family planning. These results call for an urgent need to target family planning campaigns to young women because they are more at risk of getting pregnant.
That age cohort was less significant after controlling for other socioeconomic factors such as education calls for efforts that aimed at increasing secondary and tertiary school enrollment and completion rates for women in Uganda in order to increase family planning uptake. According to the United Nations in 2011 , secondary and tertiary school enrolment in Uganda was low at 27.4% and 4.1%, respectively. The implication of low enrolment ratios is the inability of women to achieve tangible reproductive health outcomes. Young girls usually do not make the decision to stay in or out of school because this is primarily in the interest of their caretakers or parents who decide whether they should be in school or not. It is mostly during secondary schooling that children will understand the importance of education, and it is also one of the critical stages in their life when they make pertinent life decisions. Even for cases of child marriages, it is apparent that many of the parents who marry off their young daughters often get them out of school and give them out to rich men .
Urban residence was also associated with increased contraceptive use since health facilities are generally well stocked and within reasonable distance from women’s residencies. Further, urban women are also more likely to be exposed to social and media messages on family planning than their rural counterparts. Efforts to expose women to family planning programs in rural and remote areas are inevitable. Specifically, sensitization campaigns will be relevant because they expose women to family planning messages a great deal .
Efforts to improve wealth status of women will have a long lasting impact on contraceptive use in Uganda. Policies should be aimed at increasing women’s productivity at home and at improving their earnings. This will directly empower them, especially within the family, and boost their ability to make decisions .
The authors would like to thank the United States Agency for International Development, ICF Macro, the Uganda Bureau of Statistics, and other partners for supporting the collection and processing of the Uganda 2006 DHS data as well as making the data available for public use.
J. Hobcraft, J. Menken, and S. Preston, “Age, period, and cohort effects in demography: a review,” Population Index, vol. 48, no. 1, pp. 4–43, 1982.View at: Google Scholar
B. H. Hall, J. Mairesse, and L. Turner, “Identifying age, cohort, and period effects in scientific research productivity: discussion and illustration using simulated and actual data on French physicists,” Economic of Innovation and New Technology, vol. 16, no. 2, pp. 159–177, 2007.View at: Publisher Site | Google Scholar
Uganda Bureau of Statistics (UBOS) and Macro International Inc, Uganda Demographic and Health Survey 2006, UBOS and Macro International, Calverton, MD, USA and UBOS, Kampala Uganda, Bugoloobi, 2007.
L. K. George and S. J. Weiler, “Sexuality in middle and late life. The effects of age, cohort, and gender,” Archives of General Psychiatry, vol. 38, no. 8, pp. 919–923, 1981.View at: Google Scholar
R. Freedman, “Theories of fertility decline: a reappraisal,” Social Forces, vol. 58, no. 1, pp. 1–17, 1979.View at: Google Scholar
C. E. Kaufman, “Contraceptive use in South Africa under apartheid,” Demography, vol. 35, no. 4, pp. 421–434, 1998.View at: Google Scholar
N. B. Ryder, “The cohort as a concept in the study of social change,” American Sociological Review, vol. 30, no. 6, pp. 843–861, 1965.View at: Google Scholar
V. K. Pillai, “The postwar rise and decline of American fertility: the pace of transition to motherhood among 1950-1969 marital cohorts of white women,” Journal of Family History, vol. 12, no. 4, pp. 421–436, 1987.View at: Google Scholar
P. Collier and R. Reinikka, “Reconstruction and liberalization: an overview,” in Uganda's Recovery: the Role of Farms, Firms and Government, R. Reinikka and P. Collier, Eds., Fountain Publishers, Kampala Uganda, Bugoloobi, 2001.View at: Google Scholar
K. Edmonds, “Crisis management: the lessons for Africa from Obote's second term,” in Uganda Now: Between Development and Decay, H. Hansen and M. Twaddle, Eds., James Currey, London, UK, 1988.View at: Google Scholar
G. S. Ibingira, The Forging of an African Nation: The Political and Constitutional Evolution of Uganda from Colonial Rule to Independence, 1894-1962, Viking Press, New York, NY, USA, 1980.
World Bank, Uganda: The Challenge of Growth and Poverty Reduction, World Bank, Washington, DC, USA, 1993.
United Nations Development Programme, “Human development report 2011: sustainability and equity—a better future for all (summary),” http://hdr.undp.org/en/media/HDR_2011_EN_Summary.pdf.View at: Google Scholar
National Committee for Population and Family Planning (NCPFP) [Vietnam]. Vietnam Demographic and Health Survey 1988, NCPFP, Hanoi, Vietnam, 1990.
R. Rahayu, I. Utomo, and P. McDonald, “Contraceptive use pattern among married women in Indonesia,” in International Conference on Family PlannIng: Research and Best Practices, pp. 15–18, 2009.View at: Google Scholar
S. Fathonah, Patterns of Contraceptive Use in Indonesia, State Ministry of Population/National Family Planning Coordinating Board and Macro International, Calverton, Md, USA, 2000.
D. Ojakaa, Trends and Determinants of Unmet Need for Family Planning in Kenya, The DHS Working Papers, Demographic and Health Research, Macro International, Calverton, Md, USA, 2008.
M. S. Ullah and N. Chakraborty, “Factors affecting the use of contraception in Bangladesh: a multivariate analysis,” Asia-Pacific Population Journal, vol. 8, no. 3, pp. 19–30, 1993.View at: Google Scholar
R. E. A. Khan and T. Khan, “How a married woman's characteristics affect her contraceptive behavior?” Journal of Applied Sciences, vol. 7, no. 19, pp. 2782–2787, 2007.View at: Google Scholar
P. Collier, “On the economic consequences of civil war,” Oxford Economic Papers, vol. 51, no. 1, pp. 168–183, 1999.View at: Google Scholar
J. P. M. Ntozi and J. B. Kabera, “Family planning in rural Uganda: knowledge and use of modern and traditional methods in Ankole,” Studies in Family Planning, vol. 22, no. 2, pp. 116–123, 1991.View at: Google Scholar
P. S. Dasgupta, “Population, poverty and the local environment,” Scientific American, vol. 272, no. 2, pp. 26–31, 1995.View at: Google Scholar