Spousal Violence and Unwanted Fertility in Malawi *

Identifying predictors of high fertility in Malawi has been of great concern among stakeholders. One area that has not received much attention is the relationship between spousal violence and unwanted fertility. Therefore, the objective of this paper is to establish the relationship between spousal violence (aggregate of physical, emotional and sexual) and unwanted fertility among women in Malawi. A subset of 4,137women who were selected for domestic violence module and who gave birth within the last five years were extracted from the 2010 Malawi Demographic and Health Survey. Bivariate and multivariate logistic regression analyses were used to examine the relationship between unwanted fertility and background variables. The overall prevalence rate of spousal violence and unwanted fertility among the study population are 31% and 46% respectively. Spousal violence was found to be a significant predictor of unwanted fertility among women in Malawi. The results of bivariate and multivariate logistic regression models indicate that unwanted fertility is higher among women who experienced spousal violence than among those who did not (OR=0.765). Other factors that were found to be significantly related to unwanted fertility are age at first birth, number of children ever born, spousal agreement on ideal number of children household wealth index and region of residence. The study recommends that gender equality should be promoted and violence against women should be discouraged in order to accelerate fertility transition in Malawi.


Introduction
Gender-based violence (GBV) is defined as violence directed at an individual based on his or her biological sex, gender identity, or perceived adherence to socially defined norms of masculinity and femininity (USAID, 2012).GBV is a common practice that has been with humankind in most societies of the world from time immemorial.It may be psychological or mental which is also variously known as domestic violence (Odimegwu, Okemgbo & Ayila, 2010).
Gender-based violence has been a neglected public health issue until recently that much concern are been given in public health discourse (Kishor & Johnson, 2004).Apart from the economic, social, psychological and human right implications of this type of violence, GBV is associated with adverse reproductive health, general poor health outcome and many other negative consequences on the wellbeing of the victims who are mostly women and girls (Kishor & Johnson, 2004;Bibi, Ashfaq, Shaikh & Qureshi, 2014;Solanke, 2014;Wekwete, Sanhokwe, Murenjekwa, Takavarasha & Madzingira, 2014).
Gender-based violence which has no social or cultural barrier as it cut across all human strata (Heise, Pitanguy and Germain, 1995;WHO, 2005) has been on without much attention, but has now become a global issue (Bamiwuye and Odimegwu, 2014).Though with different estimates and patterns many studies have reported that up to 70% of all women worldwide had experienced one form of violence or the other at one time of their lives (Heise, Ellsberg & Guttemoeller, 1999;Oyediran & Isiugo-Abanihe, 2005; Garcia-Moreno, Jansen, Ellsberg, Heise & Watts, 2006) and appreciable number of this violence are perpetrated by their spouses or intimate partners (Hindin, Kishor & Ansara, 2008;NSO/ICF Macro, 2011).
Severally, within the last two decades that the issue of violence against women came to the international discourse, reports of this type of violence most especially by their intimate partners have been worrisome in all regions of the world from less developing nations to the developed countries.Using ownership of property as a yardstick, Kes (2008) conducted a survey in Kerela, India and reported as high as 49% of women without property ever experienced violence.According to Galindo & Serrano, (1994) (see Agnihotri, Agnihotri, Jeebun & Purwar, 2006) in Spain women were often the victims of violence and of domestic violence where 86 violent deaths and 200,000 cases of abuse by partners were recorded in 1993.Studies have also revealed high rate of domestic violence against women in United Kingdom (McGibbon, Cooper & Kelly, 1989;Stanko, Crisp, Hale & Lucraft, 1997;Mirrlees-Black, 1999).
GBV against women in general and pregnant women in particular is a neglected area of research worldwide, particularly in Malawi (Chasweka, Chimwaza, Maluwa & Odland, 2012) and this is not being fair considering the high prevalence of gender-based domestic violence (GBDV) in the country.It was mentioned in the 2010 Malawi Demographic and Health Survey (MDHS) final report that close to 30% of all women, pregnant women inclusive suffer GBDV.Whenever any study mentions GBDV, the negative reproductive outcome is mentioned, while those many pregnancies and childbirths as a consequence of coercion (physical, sexual and/or emotional violence) from the spouses go unmentioned.Search for significant predictors of high fertility in Malawi has been on for decades without satisfactory, appreciable or appropriate solution in sight, despite the success story of high contraceptive prevalence rate in the country amidst other nations in African continent (Chasweka, Chimwaza, Maluwa & Odland, 2012).

Statement of the problem
Fertility in Malawi remains high although there is some evidence of a decline.
Available statistics based on census data and other surveys indicate that total fertility rate (TFR) in Malawi has declined from 7.6 children per woman in 1977 to 5.6 children per woman in 2010 (Malawi Government, 1984Government, , 1994aGovernment, , 1994bGovernment, , 2002Government, , 2006Government, , 2011)).The factors responsible for this decline are not fully understood, and again there is yet no explanation for the mismatch between the observed fertility levels and Contraceptive Prevalence Rate (CPR) in Malawi .

Objectives
Mainly, the present study examines the relationship between experience of spousal violence and unwanted childbirth in Malawi and specifically examines the socio-economic and demographic variables relating to the incidence of unwanted childbirth in the country.

Methodology Data
The study used secondary data extracted from the most recent and credible 2010 Malawi Demographic Health Survey (MDHS) conducted by the National Statistical Office with financial and technical assistance from ICF Macro (Malawi Government 2011).The survey was the fourth Demographic and Health Survey (DHS) ever conducted in Malawi and involved a multi-stage probability sampling to select 23,020 women of reproductive age.Detailed report of the methodologies involved in data collection can be obtained from the appropriate survey report (Malawi Government, 2011).However, the current study is based on the analysis of a weighted sub-sample.The subsample were selected on two levels approach.1.Those respondents that were interviewed for domestic violence module were first kept and 2. Those respondents who ever had a birth were equally kept for the analysis.The specific constructed weight used for domestic violence module in Measure Demographic Health Survey data was applied and this now gave us a total size of 4,137 (18% of the total respondents) women who gave birth within the last 5 years and those that were included in domestic violence module.

Methods of Analysis
Three approaches were used in the analysis.Firstly, descriptive univariate analyses were performed to inspect the frequency distributions of the variables.
Secondly, bivariate analysis was employed to examine the relationships between the independent variables (listed below) and unwanted fertility (unwanted/mistimed childbirth).Chi-square tests of independence were conducted for categorical variables.Significant differences were determined using chi-square at p<0.05.Lastly, binary logistic regression was used to examine the impact of socio-economic and demographic factors on unwanted fertility in Malawi.The use of the logistic regression is based on the fact the dependent variable was dichotomous.Binary logistic regression was performed to explore the association between unwanted fertility and spousal violence and finally to assess the net effect of independent variables on unwanted fertility.

Dependent/Outcome variable
The dependent/outcome variable for the study considered women who had experienced unwanted fertility (V367, either the respondent reported childbirth as unwanted or mistimed).The dependent variable "unwanted fertility" was conceptualized in the study; for those who had ever given birth within the last five years it was asked if the last birth was wanted, wanted later or wanted no more.If wanted 0 was allocated and 1 if wanted later or wanted no more.The aggregate of the two categories (wanted later and wanted no more) gave us the total number of women who had ever experienced unwanted fertility.

Independent variable
The main independent variable describes women who have experienced spousal violence.Spousal violence was conceptualized in the study in such a way that it captured the three dimensions of violence: physical, sexual and emotional.This variable was constructed by combining a series of questions that were asked to determine whether or not the respondents experienced physical, sexual and or emotional violence by the spouse.The variable "physical violence" was determined from the response to questions asking whether the respondent had experienced at least one of the six subtypes of physical violence: whether or not spouse ever pushed, shook or threw something; spouse ever slapped; spouse ever punched; spouse ever attempted to strangle or burn; spouse ever attacked with knife, gun or other weapon.The emotional violence variable comprised three variables: spouse ever threatened with harm; spouse ever humiliated and spouse ever insulted or made to feel bad.Emotional violence variable was obtained by summing the three items.
The sexual violence comprised of three variables: ever sexual violence, ever physically forced sex and ever forced to perform sexual acts.Each item was given a score of 1 if she experienced abuse and 0 if she did not experience abuse.The sums of all items for each individual were obtained.If the total score was "0" the individual was categorized not to have experienced spousal violence otherwise any score different from "0" meant that the individual had experienced spousal violence.

Control or co-founding variables
Control variables used in this study were age, educational attainment, wealth index, place of residence, region of residence, marital status, religion, family type, age at first birth, partner's educational level, children ever born and agreement in number of children wanted by both spouses were selected as the intervening variables.While many of these variables have been found to have a significant relationship with fertility in a number of studies (Odimegwu et. al., 2015), some were selected due to the discretion of the authors.
All the independent variables were obtained from the various sections on the women questionnaire.The categorizations of the selected variables were recorded as follows Age (15-24, 25-29, 30-34, 35-39, 40-44

Socio-economics and demographic characteristics of respondents
Respondents were relatively young with a mean age of 28 years of which the majority was in the youngest reproductive age group (15-29 years) and lowest frequency in the age group 45-49 years.Educational status of women was limited with no less than 85% having no formal or primary educational level and about 15% having secondary or higher education.Table 1 further indicates that the majority of the three sub-groups of wealth index among the study population (43%) were poor and less than four in every ten respondents were in the rich category of the wealth index.While 85% of the respondents were rural dwellers, slightly less than five in every ten sampled women were from Central and Southern regions of Malawi and the remaining 11% were from the Northern part of the country.In terms of religion, eight in every ten were Christian, 15% and 1% were Muslim and Traditional faith believers respectively.Evidence from the data as presented in Table 1 reveals a high rate of early childbirth in Malawi with (71% having their first child birth before 20 years of age).Partner's educational attainment showed that seventy-one per cent of the respondents' partners had less than secondary education while just two per cent had higher education.Majority (70%) of the sampled respondents reported having 4 or less children as at the time of the survey though most of them (respondents) were still in their prime years of reproduction.One-wife and one-husband (monogamous) type of family reigns than multiple wives to a single man system in Malawi.Spousal agreement on number of wanted children was another variable examined in the study and as presented in Table 1 it was shown that while fifty-six per cent of the respondent wanted the same number of children as did their partners, twenty per cent reported never discussed such theme with partners and fifteen per cent claimed that husband wanted more children than they wanted.Table 1 also presents the prevalence of spousal violence and unwanted fertility as thirty-one per cent and forty-six per cent respectively.

Bivariate analyses (Association between selected socio-economic and demographic background characteristics and unwanted fertility among women who were interviewed for domestic violence in Malawi)
Bivariate analyses were conducted to study the relationship between unwanted fertility and selected socio-economic and demographic background variables of respondents.Table 2 shows the number and percentage distributions of women who had experienced unwanted childbirth/fertility by some selected background characteristics.A significant association exists between a number of independent variables and occurrence of unwanted fertility among which are age, region, number of children ever born, spousal agreement on number of ideal children, marital status, type of family and experience of spousal violence.Older women, who had more than five children, never married or divorced, from polygamous family setting were significantly likely to experience unwanted fertility than their other counterparts.Furthermore, mothers who ever experienced spousal violence and who reside in central or southern regions were also likely to experience unwanted or mistimed fertility more than their other counterparts (p<0.001).

Odds of ever-experience unwanted fertility by spousal violence experience and selected background characteristics of respondents. (Multivariate analysis)
In order to determine the relationship between unwanted fertility and spousal violence a four model binary logistic regression was conducted.The results of this exercise are given in Table 3.In the first model, the unadjusted odds of unwanted fertility was significantly higher among women who had ever experienced spousal violence than those other women who never experienced spousal violence.It was discovered that women who never experienced spousal violence were 23% less likely to experience unwanted fertility than those who ever experienced spousal violence.In model 2, the inclusion of some selected demographic background characteristics of respondents, which serve as the control variables did not have a pronounced impact on the interactions of experience of spousal violence and the likelihood of unwanted fertility among the respondents.The odds ratio of unwanted fertility remained almost the same when no control variable was introduced in model 1.Of all the demographic variables tested in this second model, only region, number of children ever born and spousal agreement on ideal number of children had significant effects on the relationship between spousal violence and the occurrence of unwanted fertility.
The third model shows the interaction between spousal violence and unwanted fertility while controlling for the socio-economic background characteristics.
The odds of unwanted fertility was significantly still higher among women who experienced spousal violence than those women who never experienced spousal violence.It was revealed that women who never experienced spousal violence were 21% less likely to experience unwanted fertility than those who ever experienced spousal violence.Surprisingly, none of the controlling socioeconomic variables had a significant co-founding effect on the relationship.
Model 4 controlled for the net effect of both the demographic and socioeconomic variables all together.While the incidence of unwanted fertility was still significantly higher among women who had experienced spousal violence, only region, age at first birth, number of children ever born, spousal agreement on ideal number of children and household wealth index indicated significant effect in the relationship.

Discussion
Violence in every corner of the family unit and within the four walls of the place of abode is a common, but hidden phenomenon everywhere world over.
While many go unreported, many that were reported went away without appropriate or commensurate punishment for the offenders.Women and girl child, the so called "weaker vessels" were most often than not always at the receiving end of many of the domestic violence.Despite the fact that most communities and societies frown at violence against women, the reality is that violations against women's human rights are often sanctioned under the garb of cultural practices and norms, or through misinterpretation of religious tenets (UNICEF, 2000).Unfortunately, appreciable number of the domestic violence are perpetrated by women's husbands or close partners (spousal violence) (Hindin, Kishor & Ansara, 2008;Malawi Government, 2011;Bibi, Ashfaq, Shaikh & Qureshi, 2014), this makes proper prosecution difficult most times.
Maybe if appropriate punishment has been meted out against the abusers to serve as deterrent to others, the high rate of GBDV could have been reduced.
Considering the socio-economic, demographic, health and psychological negative implications of GBDV (spousal violence) one would know that much is still desired to be done in bringing the menace to its knee.Malawi, despite her success story on the uptake of contraceptive among all other Africa countries has been on the search for the significant predictors of the high fertility rate for the country.One area that has not been explored is the unwanted pregnancies and childbirth that were likely to be as a result of coerciveness (physical and emotional) and forceful sexual act from the spouses.
Using 4,137 sampled women who gave birth within the last five years to the survey and who were selected for domestic violence module in 2010 Malawi Demographic and Health Survey, our study found a prevalence rate of 30.7% ever experienced spousal violence in Malawi and this is in consonant with other nationally reported figure of between 15% and 71% (Garcia-Moreno, Jansen, Ellsberg, Heise & Watts, 2005;Hindin, Kishor & Ansara, 2008;Bibi, Ashfaq, Shaikh & Qureshi, 2014;Bamiwuye & Odimegwu, 2014;Wekwete, Sanhokwe, Murenjekwa, Takavarasha & Madzingira, 2014) in some other countries.On the basis of unwanted fertility, our study found a 46% prevalence rate among these sub-sampled women in Malawi.This is almost at par with the national rate of unwanted pregnancies and mistimed of the last birth and pregnancies put together in the country (Malawi Government, 2011).
Findings from the multivariate analysis across all the four models revealed a significantly higher prevalence of unwanted fertility among women who had experienced spousal violence than those who never experienced spousal violence.Women, who reside in northern region, who had 4 or less children and who had first birth at 20 years or more were significantly less likely to have unwanted fertility than the occurrence among their other counterparts.
Those ones who were poor and in the middle class of wealth index and whose partners want fewer children were significantly more likely to have unwanted fertility than their other counterparts respectively.
Difference in ideal number of children between spouses might be one of the factors responsible for this spousal's misunderstanding and violence.When women or partners considered they have enough or need more children, but either of the spouses could agree with the decision or consideration, then this disagreement could erupt violence.And again, this result might not be so surprising because the multivariate analysis further showed a higher likelihood of spousal violence and occurrence of unwanted childbirths among women who were also of low or no formal educational attainment, women who were from poor or middle class and women from Southern part of Malawi.The chances are high that these set of women were less empowered and have no say either in the reproductive health matters or even in the general decision of the house.The extent of unintended and unwanted fertility due to spousal violence shows the low status of women's reproductive rights, and the low degree of women's autonomy in Malawi.This is in tandem with UNICEF submission (UNICEF, 2000).This finding again could in a way collaborating the final report of 2010 Malawi Demography and Health Survey which stated that women from Southern part of Malawi were found to be of the lowest educational level and have the highest proportion of persons in the lowest wealth quintile for the country.It is paramount to identify these predictors of unwanted fertility, to enable policymakers and program planners channel efforts specifically to the women who are most likely to experience this problem.
In conclusion, unwanted fertility rate among ever married women in Malawi is very high and it was seen that this high rate of unwanted fertility is closely and significantly related to spousal violence in the country.Our study has an important implication for policy, programme and strategy aiming at reducing fertility in this part of the world.Thus, there is need for women empowerment such that violence from their spouses could be reduced among them.
and 45-49); Educational attainment (No education, Primary, Secondary and Higher); Place of residence (Urban and Rural), Region (Northern, Central and Southern), Family type (Monogamy and Polygamy); Wealth (Poor, Middle and Rich), Marital status (Never married, Married, Widowed and Divorced/Not living together), Religion (Christian, Muslim and Traditional/Others), Age at first birth (19 or less and 20 or more), Children ever born (4 or less and 5 or more).Others are Partners' educational attainment (No education, Primary, Secondary and Higher); Spousal agreement on number of children wanted (Both want same, Husband wants more, Husband want fewer and Never discuss it).