Side effects and the need for secrecy: characterising discontinuation of modern contraception and its causes in Ethiopia using mixed methods

Abstract

Background: Contraceptive discontinuation is a major barrier to reducing global unmet needs for family planning, but the reasons why women discontinue contraception are poorly understood. Here we use data from Ethiopia to investigate (i) the magnitude of contraceptive discontinuation in 2005-2011, (ii) how the risk of discontinuation varies with method type and education level and (iii) the barriers to continuation. Our main hypothesis is that contraceptive discontinuation is driven by the experience of physiological side-effects associated with the use of hormonal contraception, rather than a lack of formal education.

Methods: We used a mixed methods explanatory sequential design to explain the quantitative results in more details through the qualitative data. First, we analysed quantitative data from the 2011 Ethiopian Demographic and Health Survey to study patterns of contraceptive discontinuation and method choice using multilevel multiprocess models. Second, we conducted semi-structured interviews and focus group discussions in the 3 most populated regions of Ethiopia with individuals of reproductive age and health professionals.

Results: The analysis of EDHS data shows that the rate of discontinuation has not reduced in the period 2005-2011 and remains high. Discontinuation mainly takes the form of abandonment, and is a function of method type, age and wealth but not of educational level. Interviews with women and health professionals reveal that the experience of debilitating physiological side effects, the need for secrecy and poverty are important barriers to continuation.

Conclusions: Our findings together suggest that physiological and social side-effects of contraceptive use, not a lack of formal education, are the root causes of contraceptive abandonment in Ethiopia.

Keywords: Contraceptive discontinuation; Education; Ethiopia; Family planning; Mixed methods; Multilevel multiprocess modelling; Reproductive ecology; Semi-structured interviews; Side-effects; Unmet needs.

Fig. 1

Trends in contraceptive adoption, discontinuation and use among Ethiopian women, 2005–2011. This figure is based on the descriptive statistics of the data collected during the 2005 and 2011 Ethiopian demographic and health surveys. The percentage of ever-users of modern contraception has increased by ca. 10% over the period investigated. The percentage of contraceptive discontinuation, which includes failure, switch and abandonment, has not reduced. The percentage of women using contraception at the time of the survey, i.e. contraceptive prevalence, has increased by ca. 5% in the study period

Fig. 2

Predicted odd-ratios & 95% confidence intervals for method choice. The reference category (horizontal line) corresponds to the use of oral contraceptives. At the national level, injectables are ca. 6 times more likely to be used than any other short-acting methods. This figure is based on the results of a multinomial probit model using the calendar data collected in the 2011 Ethiopian Demographic and Health Survey

Fig. 3

Predicted odd-ratios & 95% confidence intervals for contraceptive discontinuation as a function of a woman’s level of education. The reference category (horizontal line) is no education. The risk for a woman to discontinue contraception because of abandonment (a woman stops using contraception) or failure (a woman becomes pregnant) is independent from her educational level. However, the risk of switching between methods is ca. 55% higher for women with the highest level of education. This figure is based on the results of a multilevel multiprocess model using the calendar data collected in the 2011 Ethiopian Demographic and Health survey

Fig. 4

Predicted odd-ratios & 95% confidence intervals for contraceptive discontinuation as a function of method type. The analysis focuses on short-acting contraceptives only and the reference category (horizontal line) corresponds to the use of oral contraceptives. The risk of all three forms of discontinuation is higher for oral contraceptives. This figure is based on the results of a multilevel multiprocess model using the calendar data collected in the 2011 Ethiopian Demographic and Health Survey