An unintended pregnancy has been defined as the kind of pregnancy that is reported to be either unwanted or mistimed [
In Ghana, it has been identified that about 37 percent of all pregnancies are unintended, comprising 23 percent mistimed and 14 percent unwanted pregnancies [
Unintended pregnancy has been found to have severe implications for both the child and the mother. For instance, unintended pregnancy is found to have negative effects on prenatal visits and care, physical health status, labour experience, pain during labour, and psychological status in the early postpartum period [
Even though myriads of extant studies have examined the levels and determinants of unintended pregnancy globally in the past few decades, information on this phenomenon is quite limited in Ghana. To the best of my knowledge, the few extant studies are clinic-based and focused on only the rural population [
This is a cross-sectional study that used data from the 2014 Ghana Demographic and Health Survey (GDHS). The 2014 GDHS is the sixth round of the global DHS program series, which was conducted by the Ghana Statistical Service (GSS) and the Ghana Health Service (GHS) and was financially supported by the United States Agency for International Development (USAID) among other organisations. It is a nationally representative survey and it comprised 9,396 women aged 15-49 who had been selected from 11,835 households using a two-stage sampling design [
The outcome variable for this study is named “current pregnancy wanted” in the dataset. In this regard, women who participated in the survey were asked whether they were currently pregnant and those who responded in the affirmative were further asked whether they wanted this pregnancy. This generated three responses such as “wanted then,” “wanted later,” and “not at all.” To generate the outcome variable for unintended pregnancy, a dichotomous outcome was created by recoding the original variable “wanted then” as “intended(0)” and “wanted later” and “not at all” as “unintended(1)”. Twelve sociodemographic predictor variables were used in the analysis, which comprised the age of respondent, education level, religion, ethnicity, wealth status, marital status, parity, work status, and knowledge of ovulatory cycle, including control variables such as the type and region of residence and unmet need for contraception. These predictor variables have been included in this study primarily based on their effect in the extant literature concerning the subject.
The R programming language (version 3.4.3) was used to process the data. The data were imported from the GDHS Stata file into R. Descriptive and bivariate analyses were performed to ascertain the level of unintended pregnancy prevalence among the predictor variables. Furthermore, logit regression models were fitted to predict the determinants of unintended pregnancy among Ghanaian women. A logit regression model was chosen because it is suitable for binary outcomes and provides the opportunity to estimate the risk factors of a phenomenon while also providing the opportunity to generate odds ratios for them. Three logit models were fitted where model 1 examines the sociodemographic predictors of unintended pregnancy whereas model 2 examines whether the sociodemographic predictors may complement spatial factors such as type and region of residence to predict unintended pregnancy by nesting model 1 within model 2. Model 3 was used to control for exposure to the need for contraception. A complex survey design was applied to the results to produce nationally representative findings. In doing this, a survey design was generated where the primary sampling units, the strata, and the weight were nested in the final dataset used for the analysis.
This study comprised pregnant women at the time of the survey and a summary of their background characteristics is presented in Table
Sociodemographic characteristics and prevalence of unintended pregnancy.
Variables | Number of women | Percent | Unintended pregnancy |
---|---|---|---|
Age | (%) | ||
15-19 | 64 | 9.4 | 71.7 |
20-29 | 301 | 44.3 | 36.9 |
30-39 | 278 | 40.9 | 34.7 |
40-49 | 36 | 5.4 | 50.2 |
Education level | |||
No education | 161 | 23.8 | 26.1 |
Primary education | 113 | 16.6 | 53.9 |
Secondary/higher | 405 | 59.6 | 41.6 |
Religious affiliation | |||
Christian | 542 | 79.8 | 43.3 |
Muslim | 102 | 15.1 | 21.3 |
Traditionalist | 19 | 2.8 | 44.1 |
Others | 16 | 2.3 | 42.7 |
Ethnicity | |||
Akan | 346 | 51.0 | 43.1 |
Ga-Dangme | 30 | 4.4 | 35.7 |
Ewe | 109 | 16.1 | 60.0 |
Mole-Dagbani | 99 | 14.6 | 24.0 |
Others | 95 | 13.9 | 23.4 |
Wealth status | |||
Poor | 248 | 36.6 | 44.0 |
Middle | 140 | 20.6 | 48.5 |
Rich | 291 | 42.8 | 32.4 |
Marital status | |||
Never married | 104 | 15.3 | 68.9 |
Married/living together | 563 | 82.9 | 34.4 |
Widowed/divorced/separated | 12 | 1.8 | 48.2 |
Parity | |||
0 | 144 | 21.2 | 46.8 |
1-3 | 380 | 55.9 | 31.9 |
4+ | 155 | 22.9 | 53.3 |
Work status | |||
Working | 507 | 74.6 | 37.4 |
Not working | 172 | 25.4 | 47.5 |
Ovulatory cycle knowledge | |||
Yes | 287 | 42.2 | 39.8 |
No | 392 | 57.8 | 40.0 |
Unmet need for contraception | |||
No unmet need | 408 | 60.0 | 000.0 |
Unmet need | 271 | 40.0 | 100.0 |
Type of residence | |||
Urban | 337 | 49.7 | 33.1 |
Rural | 342 | 50.3 | 46.7 |
Region | |||
Western | 73 | 10.7 | 27.6 |
Central | 75 | 11.1 | 51.5 |
Greater Accra | 134 | 19.7 | 36.9 |
Volta | 45 | 6.6 | 76.2 |
Eastern | 71 | 10.4 | 56.6 |
Ashanti | 106 | 15.7 | 46.6 |
Brong Ahafo | 60 | 8.8 | 36.9 |
Northern | 71 | 10.5 | 12.3 |
Upper East | 29 | 4.3 | 16.9 |
Upper West | 15 | 2.2 | 23.6 |
Total prevalence | 40.0 |
The majority of the respondents were married or living together (82.9%) whereas the least were widowed, separated, or divorced (1.8%). More than half had 1-3 children (55.9%) while more than one-fifth had no children (21.2%). The majority were working (74.6%) while the minority were not working (25.4%). Also, more than half did not know their ovulatory cycle (57.8%) while 42.2% knew their ovulatory cycle. Furthermore, 60.0% of the respondents had no unmet need for contraception while 40.0% had unmet need for contraception. A little more than half of the respondents were from rural settings (50.3%) while the remaining women were from urban settings (49.7%). Lastly, about one-fifth of the respondents were from the Greater Accra (19.7%), 15.7% were from the Ashanti region, and 11.1% were from the Central region while only 2.2% were from the Upper East region.
Table
Women who were never married (68.9%) also had the highest prevalence of unintended pregnancy while close to half of widowed, separated or divorced women (48.2%) had an unintended pregnancy, with married or women in union (34.4%) having the lowest. Additionally, women who had 4 or more (53.3%) children had the highest prevalence with more than half of them having unintended pregnancy while women having 1-3 children (31.9%) had the lowest. Nonworking women (47.7%) also had a higher prevalence of unintended pregnancy than working women (37.4%) whereas women who knew their ovulatory cycle (39.2%) and women who did not know (40.0%) almost had the same prevalence of unintended pregnancy. Quite expectedly, all the women who had unmet need for contraception (100.0%) had unintended pregnancy whereas none of the women who had no unmet need for contraception had any unintended pregnancy.
In addition, unintended pregnancies were more prevalent among rural residents (46.7%) than urban residents (33.1%). With regard to variations in regional prevalence, unintended pregnancies were more prevalent in the Volta region (76.2%), Eastern region (56.6%), and the Central region (51.5%) compared to the other regions in the country, but lowest in the Northern region (12.3%).
In this section, I assess the predictors of unintended pregnancy among women using three nested logit models, and the results are presented in Table
Logit regression analysis of unintended pregnancy among pregnant women.
| | | |
---|---|---|---|
Age of woman | | | |
15-19 (Ref) | 1 | ||
20-29 | 3.42[1.27, 9.22] | 3.62[1.56, 8.34] | 1.02[0.68, 1.51 ] |
30-39 | 5.81[2.04,16.55] | 6.59 [2.59, 16.73] | 1.05[0.67, 1.63] |
40-49 | 4.58[1.21, 15.96] | 4.85[1.48, 15.84] | 1.04[0.59, 1.81] |
Level of education | |||
No education (Ref) | 1 | ||
Primary | 0.44[0.20, 0.94] | 0.51[0.25, 0.98] | 0.95[0.73, 1.35] |
Secondary/higher | 0.50[0.25, 0.99] | 0.50[0.26, 1.01] | 0.76[0.73, 1.29] |
Religious affiliation | |||
Christianity (Ref) | 1 | ||
Islam | 1.47[0.72, 2.97] | 1.46[0.74, 2.89] | 0.96[0.69, 1.35] |
Traditional/ | 0.52[0.14, 1.91] | 0.33[0.06, 1.67] | 1.06[0.61, 1.82] |
Spiritual | |||
Other | 1.16[0.28, 4.77] | 1.14[0.30, 4.31] | 0.97[0.52, 1.78] |
Ethnicity | |||
Akan (Ref) | 1 | ||
Ga/Dangme | 1.07[0.41, 2.77] | 1.44[0.49, 4.18] | 1.04[0.60, 1.79] |
Ewe | 0.42[0.25, 0.73] | 0.68[0.32, 1.47] | 0.94[0.64, 1.36] |
Mole-Dagbani | 1.64[0.80, 3.36] | 0.77[0.33, 1.79] | 0.98[0.67, 1.45] |
Other | 2.29[1.50, 7.24] | 1.11[0.53, 2.31] | 0.98[0.68, 1.40] |
Wealth status | |||
Poor (Ref) | 1 | 1 | |
Middle | 0.88[0.52, 1.51] | 0.86[0.48, 1.55] | 1.04[0.78, 1.32] |
Rich | 1.65[0.95, 2.84] | 1.65[0.81, 3.37] | 0.96[0.70, 1.33 ] |
Marital status | |||
Never married (Ref) | 1 | ||
Married/Cohabiting | 3.30[1.50, 7.24] | 3.83[1.67, 8.75] | 1.01[0.71, 1.43] |
Widowed/divorced/Separated | 2.18[0.50, 9.58] | 1.69[0.33, 8.62] | 0.98[0.46, 2.06] |
Parity | |||
0 (Ref) | 1 | ||
1-3 | 0.64[0.31, 1.32] | 0.56[0.26, 1.18] | 0.98[0.74, 1.30] |
4+ | 0.15[0.06, 0.35] | 0.13[0.05, 0.32] | 0.95[0.64, 1.39] |
Work status | |||
Working (Ref) | 1 | ||
Not working | 0.81[0.49, 1.34] | 0.75[0.44, 1.28] | 1.01[0.81, 1.25] |
Ovulatory cycle knowledge | |||
Yes | 1 | ||
No | 1.03[0.66, 1.59] | 1.02[0.64, 1.61] | 1.01[0.83, 1.22] |
Type of residence | |||
Urban (Ref) | 1 | ||
Rural | 0.63[0.35, 1.12] | 0.97[0.75, 1.24] | |
Region of residence | |||
Western | 1 | ||
Central | 0.30[0.10, 0.90] | 0.95[0.60, 1.50] | |
Greater Accra | 0.22[0.08, 0.61] | 1.09[0.74, 1.61] | |
Volta | 0.11[0.03, 0.31] | 1.03[0.63, 1.69] | |
Eastern | 0.22[0.10, 0.50] | 0.97[0.68, 1.38] | |
Ashanti | 0.21[0.09, 0.52] | 0.99[0.7, 1.37] | |
Brong Ahafo | 0.46[0.17, 1.19] | 0.98[0.68, 1.42] | |
Northern | 2.12[0.75, 5.98] | 0.89[0.58, 1.37] | |
Upper East | 1.63[0.45, 5.79] | 0.84[0.47, 1.49] | |
Upper West | 0.83[0.24, 2.81] | 0.82[0.53, 1.26] | |
Unmet need for contraception | |||
No unmet need (Ref) | 1 | ||
Unmet need | 7.13[1.57, 8.91] | ||
| |||
Log Likelihood | -395.411 | -367.130 | -311.281 |
AIC | 832.822 | 796.259 | 764. 463 |
Note: OR=odds ratios; CI= confidence intervals; Ref= reference category.
Also, respondent’s education level had a significant and negative relationship with unintended pregnancy in models 1 and 2 when spatial factors were controlled, but these disappeared when the unmet need for contraception was controlled. The odds of unintended pregnancy were 49% lower for respondents with primary school education (CI=0.25-0.98) and 50% lower for respondents with secondary school education or higher (CI=0.26-1.01) compared to respondents without any formal education. However, religious affiliation had no significant relationship with unintended pregnancy in all the models, though Muslim women and Others had higher odds of unintended pregnancy compared to the Christians. Similarly, wealth status had no significant effect on unintended pregnancy in all the models. The ethnic background of the respondent was found significant in only model 1, with Ewe respondents having 58% lower odds whereas Other ethnic minorities had 129% higher odds compared to Akan respondents.
Additionally, there was a significant relationship between marital status and unintended pregnancy for the first two models, but the significance disappeared in the third model. The odds of unintended pregnancy were higher for respondents who were married or cohabiting (3.83 times, CI=1.67-8.75) compared to respondents who had never been married or in a union. The parity of the respondents also shows a significant relationship with unintended pregnancy, with women who had 4 or more children significantly having 87% lower odds compared to women who had no children. However, work status, knowledge of ovulatory cycle, and type of residence show no significant effect on the risk of unintended pregnancy across all the models.
Furthermore, the results show that the region of residence had a significant effect on unintended pregnancy in model 2 but not in model 3, after controlling for unmet need for contraception. Respondents from regions such as Central (OR=0.33, CI=0.10-0.90), Greater Accra (OR=0.33, CI=0.08-0.61), Volta (OR=0.11, CI=0.03-0.31), Eastern (OR=0.22, CI=0.10-0.50), and Ashanti region (OR=0.21, CI=0.09-0.52) all had significantly lower odds of having unintended pregnancy compared to their counterparts from the Western region. Ultimately, in model 3, only unmet need for contraception had a significant effect on unintended pregnancy with those having unmet need having 7.13 odds of unintended pregnancy (CI=1.57-8.91) compared to those without an unmet need for contraception.
One of the objectives of this study was to determine the prevalence of unintended pregnancy among women in Ghana. This study also sought to establish the factors that predict unintended pregnancies among women in Ghana. The results show that the prevalence of unintended pregnancy among women in Ghana is excessively high. About 40% of pregnant women reported that they either wanted the pregnancy later or did not want it at all. This is just in the range of the prevalence found by other studies [
In examining the factors that influence the high level of unintended pregnancy among pregnant women in Ghana, the study found some sociodemographic characteristics to be significantly associated with unintended pregnancy, net of unmet need for contraception. The age of the woman is found to have predicted the risk of unintended pregnancy among pregnant women in Ghana, net of unmet need for contraception. The relationship between age and unintended pregnancy is curvilinear where the risk of unintended pregnancy increases steadily with age until it reaches its peak among women aged 30-39, before waning among women in their 40s. Thus, even though the prevalence of unintended pregnancy was highest among women aged 15-19, the risk of unintended pregnancy is actually higher among older women in Ghana. This is consistent with the findings of extant studies which also found age to be a significant determinant of unintended pregnancy [
The results further show that unintended pregnancy is significantly determined by the education level of the Ghanaian woman. The effect of educational attainment on unintended pregnancy is quite evident in the extant literature [
According to the findings of this study, marital status also plays a significant role in the risk of unintended pregnancy among Ghanaian women independent of the unmet need for contraception. It is evident that married Ghanaian or cohabiting women are substantially more likely to be at risk of unintended pregnancy than those who never married and those who are previously married. In this context, the higher risk of unintended pregnancy among Ghanaian married women may be due to nonuse of contraceptives or contraceptive failure as many Ghanaian women may believe that contraceptive use has serious side effects while some may also believe that it is a sin. The significance of the marital status and unintended pregnancy nexus is also established by Ikamari et al. [
Additionally, the risk of unintended pregnancy is found to be predicted by the parity of the Ghanaian woman, net of unmet need for contraception. In this study, higher parity women unexpectedly have a significantly lower risk of unintended pregnancy compared to those with lower parity. In context, this may be a consequence of uptake of family planning programmes among this category of Ghanaian women who might have decided to stop childbirth. On the contrary, even though many studies have documented the effect of parity on the risk of unintended pregnancy [
Moreover, the region of residence is significantly associated with unintended pregnancy when the unmet need for contraception is not controlled. In effect, women in the Central, Greater Accra, Volta, Eastern, and the Ashanti region have a significantly lower risk of unintended pregnancy than their counterparts from the other regions. These are virtually southern sector regions of Ghana where much of the socioeconomic development is found; therefore, women in these regions are most likely exposed to family planning facilities and services compared to their counterparts from the northern sector. Women from the Northern region and the Upper East region are at higher risk of unintended pregnancy in the country. It is noteworthy that the role of the region of residence in the risk of unintended pregnancy is not well-documented in the literature. However, a few studies have observed the significance of region of residence as a determinant of unintended pregnancy among women [
Ultimately, this study shows that net of background characteristics such as age, education, and marital status, among others, unmet need for contraception independently has a significant effect on the risk of unintended pregnancy among Ghanaian women. This is evident in model 3 where the significance of all significant factors disappeared after controlling for unmet need for contraception. The implication is that, irrespective of the background characteristics, unmet need for contraception is a principal predictor of unintended pregnancy among Ghanaian women. In this regard, women who have an unmet need for contraception in Ghana have a significantly higher risk of unintended pregnancy than those without an unmet need for contraception. This significant positive relationship between unmet need for contraception and unintended pregnancy is consistent with extant literature [
This study provides evidence that the prevalence of unintended pregnancy among pregnant women in Ghana is excessively high. This is significantly predicted by a number of background characteristics such as the age of the woman, educational attainment, marital status, parity and region of residence, independent of unmet need for contraception. However, unmet need for contraception shows a significant independent effect on the risk of unintended pregnancy among Ghanaian women, irrespective of their background characteristics. The older, the uneducated, the married or cohabiting, and the zero parity women, as well as women in the Western and some Northern regions, all have a higher risk of unintended pregnancy in the country. Public policy decisions must be focused on substantially reducing unmet need for contraception among Ghanaian women. Using the background risk factors as a basis, unmet need for contraception can be tackled by greatly improving access to all forms of contraception methods and family planning services among both Ghanaian women and men. Also, policy options should be considered to tackle and reduce any attitudinal resistance to effective contraceptive use among Ghanaian women. These would help to improve effective contraceptive use and prevent many unwanted pregnancies among Ghanaian women. Unlike this study, future research can focus on including both individual and aggregate level factors that have an effect on the risk of unintended pregnancy in Ghana including the contextual factors.
The dataset used to support the findings of this study is available at the DHS Program data repository.
Data for this study was obtained from the DHS Program.
The author declares that there are no conflicts of interest regarding the publication of this paper.