Socio-ecological model explaining contraceptive behavior among abortion seekers

Previous studies on the determinants of contraceptive use after abortion reveal varying results due to contextual differences, study designs and outcomes of interests [5,23–26]. Research also shows that contraceptive uptake is highest when immediately offered on-site, or within less than six months or up to a year [23,26–29]. For instance, Adelman, Free and Smith [29] examined the predictors of post-abortion contraception at four and 12 months among Cambodian women. They found that previous contraceptive use, intention to use contraception, and number of living children were strongly associated with post-abortion contraception use. Occupation was significantly associated at four months whilst abortion method predicted post-abortion contraception at 12 months.

In an intervention study conducted by Johnson and colleagues [23], their results showed that women who received post-abortion family planning services during an intervention programme had significantly lower unplanned pregnancies and few repeat abortions during a 12 month follow-up period compared to post-abortion women who were not included in the intervention programme. In addition, Delvaux and colleagues [30] found that post-abortion contraceptors were more likely to be older, married, have at least one child and use contraception before the abortion.

Similarly, other studies have reported greater acceptance and/or use of modern contraceptives among women who received PAFP counselling [31] and services if they attended health centers or maternity homes compared to hospitals [32,33], and if women attended urban hospitals compared to rural hospitals or attended Protestant hospitals compared with Catholic hospitals [34]. Women were also more likely to use modern contraception after abortion if the family planning counselling and services was provided by gynaecological ward staff compared to other models of provision [3].

The socio-ecological model can best explain factors associated with contraceptive uptake among women with abortion experiences. Studies have explored factors associated with contraceptive use at the individual (intrapersonal), partner (interpersonal) and institutional or structural levels [35]. Age, education, religion, marital status, ethnicity, number of living children, number of abortions and personal motivation for abortion are individual-centred factors that can determine use among women. While some studies have found that older women are more likely than younger women to use contraception, especially long-acting methods, because of prior experience and the desire to limit childbearing [13], others have found that younger women are more likely to adopt contraception post-abortion [8,30,36]. Also, women with some level of education may have a greater likelihood of contraceptive adoption than uneducated women while women belonging to all other religious affiliations other than Catholics and all other ethnic group other than Mole-Dagbanis are predisposed to use a contraceptive method [12,13].

Important interpersonal factors have included those associated with the male-partner. The few studies on this have discussed that financial, communicative, and emotional support are predictors of contraception uptake immediately post-abortion [37]. Men’s participation and support during the abortion process may be linked to women’s acceptance and use of contraception post abortion [12,37,38]. In the literature, institutional level factors are characterized by facility-level factors and these are also key to understanding contraceptive use. The abortion procedure, facility type and type of provider may determine use through these mechanisms [13]. Structural level factors may also determine access to facilities for contraceptive use [28].

The two aims of this study are to examine the key factors associated with immediate post-abortion contraceptive use, as well as those associated with subsequent or future use, suggesting sustained use. We argue that abortion seekers in Ghana may use post-abortion contraception when partner support and health provider or facility-level factors are available.

Study design and setting

The 2017 Ghana Maternal Health Survey dataset is a publicly accessible document which was obtained from www.DHSprogram.com [11]. The survey was jointly implemented by the Ghana Statistical Service (GSS) and the Ghana Health Service (GHS) with technical assistance from the ICF (originally Inner City Fund) through the DHS program. Participants in the survey provided written informed consent and participation in the survey was voluntary. The dataset was completely anonymized, de-identified and aggregated before we were provided with access to it for analysis. The identities of participants were therefore, not linked to the data. The 2017 GMHS is the second nationally representative household two-stage cluster survey which comprises comprehensive information on maternal and reproductive health issues, maternal mortality, and specific causes of death among women in Ghana [11]. The survey protocols and biomarkers were reviewed and approved by the ICF Institutional Review Board. The data that were collected comprised events within five years preceding the survey. The women’s individual dataset used for this study was limited to a weighted sample of 1,880 women (both married and unmarried) who had ever had an abortion within the five years preceding the survey.

Measures

The primary outcome variables were immediate and subsequent or later post-abortion family planning or contraceptive use. In the dataset, immediate post-abortion contraceptive use was captured by two questions. The first question asked respondents, ‘After this abortion, did the doctor or health worker give you a method of contraception, prescribed a method of contraception, or refer you to a family planning clinic?’ with responses as either ‘Yes’ or ‘No’. a subsequent question for those responding in the affirmative was “Did they give you the method of contraception, give you a prescription, or give you a referral?” and the corresponding options were ‘Gave method’, ‘Prescribed a method’, ‘Gave a referral’ or ‘Don’t know’. Respondents who chose the option ‘gave method’ were categorised as immediate users and were coded as ‘Yes’ while all the other options were categorised as ‘No’. Given the data limitation of questions on the immediacy of actual contraceptive use post-abortion, we assume that those ‘given a method’ suggests acceptance of and actual uptake of a contraceptive method after the abortion. Subsequent post-abortion uptake of contraception was measured by women’s current use of contraception at the time of the survey. In this study, modern contraceptive users should be using implants, intrauterine device (IUD), oral contraceptive pills, injectables, male and female condoms. All others were coded as ‘No’.

The predictor variables in the study included socio-demographic, reproductive, and abortion-related service characteristics, partner factors, and health facility indicators. The socio-demographic characteristics examined include age, highest educational status, religious affiliation, ethnicity, marital status, and place of residence. The reproductive factors were number of abortions (1, 2, 3+) and contraceptive use at index pregnancy (No/Yes). Abortion-related service delivery characteristics consisted of pre-or post-abortion health provider counselling and services (No/Yes) and abortion method (Safe/Unsafe). Respondents’ abortion method was categorised as safe or unsafe based on the WHO classification (place abortion was performed, category of health provider who performed the abortion and equipment used). Partner-related factors were measured based on three variables: attitude towards the abortion (Favourable/Opposed), provision of financial support for the abortion (Yes/No), and partner-related reasons for the abortion (Health-related, Partner-related and Other-related).

Data analysis

The data were analysed in three stages. First, we employed univariate techniques to explore frequencies of the categorical variables. Second, for the bivariate analyses, descriptive statistics were reported for all categorical and continuous predictor variables with their differences assessed by selected characteristics using cross-tabulations and chi-square tests. Third, multivariate analyses were conducted using two binary logistic regression models to ascertain the odds of all the factors significantly associated with immediate and subsequent post-abortion contraceptive use. All analyses were performed using STATA version 13.

There were few limitations which pertained to restrictions with the dataset in measuring some variables. First, we were unable to measure immediate post-abortion contraceptive use due to the limited questions on actual contraceptive use immediately post-abortion; hence, the assumption that ‘given method’ is actual immediate contraceptive use. Second, another limitation was our inability to measure some individual characteristics such as occupation and number of surviving children. In addition, the type of health facility—primary, secondary, or tertiary—where women obtained induced abortion could not be ascertained more specifically for and considered in the analysis since it was not captured in the data.

Individual level factors of post-abortion contraceptive use

At the individual/intrapersonal level, age, marital status, place of residence, and contraceptive use before pregnancy were associated with subsequent/future post-abortion contraceptive use. Age as a determinant of post-abortion contraception use has been found in other studies [8,30,36] and these findings are consistent with our study which showed that older women (>35 years) were less likely to use contraception subsequently after abortion compared to adolescents and young women. Attitudinal resistance and prior experience with contraceptive use may account for older women’s unwillingness to initiate contraception following an abortion. Fertility intentions and reproductive goals of women differ. Similarly, continuation rates of contraception uptake after abortion may vary for younger and older women. Young women (including adolescents) may be more willing to accept family planning to prevent future unplanned pregnancies because they may have educational aspirations and other life achievement goals to pursue. It is also likely that young women are inexperienced with modern contraception and perhaps more susceptible to ‘subtle directives’ from health providers to accept post-abortion family planning counselling services compared to older women.

The findings also show that, compared to women who were not in a union, women in a union were more likely to use contraception at any time following termination of pregnancy. This finding is somewhat unexpected although there is evidence to demonstrate that married women initiate post-abortion contraception with their partners’ approval [39]. It is plausible that married women have reached their fertility goals and aspirations, therefore, they prefer using contraception to regulate or stop childbearing [40]. On the other hand, unmarried women may have fears or anxieties about adopting contraception if they believe that using contraceptives will make them infertile before they get married. Support and involvement of male partners at the time of the abortion could also influence the use of contraception after abortion [37].

From the results of the study, place of residence is an important factor in women’s use of modern contraception immediately and subsequently following induced abortion. We found that compared to women residing in urban areas, women living in rural settings were more likely to use modern family planning methods post-abortion. Similar findings have been reported in other contexts among women in their reproductive ages [41,42]. In Ghana for instance, the 2007 GMHS evidences slight differences in modern family planning use by women living in urban and rural areas. Aviisah and colleagues [43] used the 2003, 2008 and 2014 demographic and health surveys to examine patterns in the use of modern contraceptive methods among married women. They found an increase in modern methods of contraceptive use among married women in rural areas compared to urban dwellers. Women living in rural settings in Ghana have greater avenues and opportunities to access, use, and choose from a variety of modern family planning methods due to the implementation of the Community-based Health Planning and Services (CHPS) concept at the district levels which expands and integrates the provision of child and maternal care with reproductive health services [44].

Prior contraceptive use before pregnancy termination is necessary to sustaining efforts aimed at avoiding future unplanned pregnancies. We found that women were more likely to use contraception not immediately but, subsequent after an induced abortion, if they had used contraceptives prior to becoming pregnant. Findings echo results from other settings indicating that contraceptive history predicts post abortion contraceptive use [26,45,46]. These findings can be explained considering women’s desire and intention to reduce their fertility or stop having children as well as discontinuing future unplanned pregnancies. By implication, the previous contraceptive practices of women present an opportunity for health providers to counsel women on the efficacy and benefits of effective long-acting contraceptive methods at the time of abortion.

Interpersonal factors associated with post-abortion contraceptive use

According to the socio-ecological model, an individual’s decision is shaped not only by their own characteristics, but by others including significant others. Kayi [37] and Rominski et al [12] demonstrated that attributes of male partners shaped women’s reproductive decisions. Our results, however, show that partner-related characteristics such as male partner attitude, financial support of male partner and partner-related reasons for the abortion were not associated with immediate and subsequent post-abortion contraception use. These findings provide insights for discussions on autonomy in contraceptive decision-making among post-abortion women who have varying degrees of vulnerability [20]. Though evidence exists to demonstrate male dominance in women’s contraceptive decisions, male power appears to be overshadowed during abortion care.

Structural/Institutional level- health system factors

The findings show that structural level factors specifically, the provision of family planning counselling by health providers before or after abortion is strongly associated with immediate and future use of post-abortion contraception. Our study found that, women were less likely to use contraception following abortion if family planning counselling was not provided by a health provider. This finding suggests that, abortion method, type of facility, and location (or access) are not as important as the role played by health professionals in the delivery of abortion care and family planning services. Health care providers form an integral part of the healthcare system particularly where maternal and reproductive health services are concerned. To a large extent, confidence in the medical expertise of health professionals coupled with a change in provider attitudes on abortion, training and provider skills [47] may have served as strong motivating factors to encourage women’s adoption and uptake of a contraceptive method post-abortion. The results are consistent with existing studies [3,13,32,33,48].