Fertility Intentions among HIV-positive and HIV-negative mothers in Zambia: Analysis of 2013-14 and 2018 DHS data

Background: HIV infection is a strong predictor of fertility as it might inuence one’s fertility desire. With advent treatment, HIV-infected mothers are now living longer and healthier just like the HIV-negative mothers. Zambia is among the developing countries in the region with high desire for more children in spite of government efforts to reduce the fertility rate. However, little is known on the inuence of HIV status on the desire for children among mothers of reproductive ages in Zambia. Methods: The analyses are based on mothers with linkable information on HIV testing and fertility preferences from the 2013-14 and 2018 Zambia and Demographic Health Survey data, with national representative samples of 16,411 and 13,683 women, respectively. HIV data was merged with each woman’s individual le, which also contained household variables to create an analytical le. Of the total sample; 11,683 mothers in 2013-14 and 9,172 mothers in 2018 were eligible for the study. Cross-tabulations with statistical tests were conducted to ascertain the crude relationship and nally multiple logistic regression analysis were employed to identify the major predictors of fertility intentions among HIV positive and HIV negative mothers using Stata software version 14.0. Results: Findings indicate that fertility intentions among women either HIV-positive or HIV-negative reduced as parity increases. Generally, in both data sets, there is a signicant difference between HIV-negative mothers preferring more children compared to HIV-positive mothers. Interestingly, in the 2018 survey, HIV-positive mothers residing in urban areas were 33 percent more likely to prefer more children compared to HIV-negative mothers. Conclusion: According to this study, predictors of fertility intentions among women of reproductive age are different by HIV status. Other signicant variables are age, education level, marital status, and parity, which were found to be the major predictors of fertility intentions among HIV-positive. Among the HIV-negatives, age, education level, parity, marital status, employment status, wealth quintile and region explained their fertility intentions. The fact that many HIV-infected mothers expect to have more children has important implications for the prevention of vertical and horizontal transmission of HIV. There is, therefore, the need for comprehensive and continuous expansion of family planning, voluntary counseling, and prevention of mother-to-child transmission HIV treatment care 1.2 and 1.1 times more likely to prefer more children, respectively. Results indicate that HIV-positive women in 2013-14 drawn from other provinces were more than 10% more likely to want children compare to those from Central province. Conversely, those form 2018 survey with exceptions from Luapula [OR:1.24, C.I: (0.93–1.65)] and Northern [OR: 1.20, C.I: 0.89–1.61)] provinces were less likely to want more children.


Introduction
The human devastation being shaped by the Human Immune Virus and Acquired Immune-De ciency Syndrome [HIV/AIDS] in sub-Saharan Africa (SSA) and other regions, and the ripple effect that this epidemic is having on communities around the world presents one of the greatest challenges to global public health. Nearly 40 years after the rst case, HIV has continued to spread world over and it is now rmly established as an important public health issue and can still be considered a strong predictor of fertility world over. HIV is one of the most signi cant causes of illness and death in human life history [1]. The burden is, however, heaviest in SSA [19]; with approximately 58% of adults living with HIV and 53% of all adult deaths in the region are women [20]. In Zambia, for example, (14.2%) of women 15-49 compared to (7.5%) of men in the same age group are living with HIV (Zambia Statistical Agency [25]. Since a larger proportion of these women are in the reproductive age , they risk infecting their newborn babies and sexual partners and thus face di cult choices about childbearing. Over the past two decades, the HIV epidemic in many parts of the world has drastically reduced, but in Africa the situation is very different. In 2010 UNAIDS reported that SSA still remains the region most heavily affected by HIV, accounting for 67% of the global prevalence of infections with women in the reproductive age being the majority of the population living with HIV in the region (UNAIDS, 2010). In 2015, WHO highlighted that Africa is the leading continent and remains the center of HIV epidemic in the world due to high poverty levels, lack of innovation, lack of capacity, and government commitments to ght against HIV [23]. Since the beginning of the epidemic, it is estimated that more than 70 million people have been infected with HIV in Africa and about 35 million people have died due to HIV infection in this continent. In 2008, UNAIDS estimated that three quarters of the 3.3 million pregnant women infected with HIV who gave birth each year lived in Africa, where most (700,000) of the annual new infections of HIV in children occurred [21]. Another report by the UNAIDS in 2012 showed that about 70% of the 34 million people living with HIV globally resided in SSA, with women comprising 58% of persons HIV infected [18].These disproportionately higher rates of HIV among women of reproductive ages in SSA have implications not only for health but also life course transitions such as childbearing.
Generally, the effects of HIV infection on fertility have been extensively studied in generalized HIV epidemic settings in SSA [4]. Principally, this has been to help forecast the demographic impacts of hyperendemic HIV [8] and because HIV prevalence among pregnant women has been widely used for estimating general population HIV prevalence levels and trends [8]. Existing literature, largely based on cross sectional data, has demonstrated that the relationship between HIV infection and fertility depends strongly on age. For instance, studies conducted by Chen et al., 2010 on the impact of HIV on fertility in Africa using cross sectional data revealed that younger (15-24 years) women had higher antenatal care (ANC) when compared to other women in the reproductive ages. Their ndings further indicated that higher ANC was associated with pregnancy and HIV risks which mostly occurred among the subset of women who are sexually active while among older age groups (25-49 years), the Fertility Rate Ratio (FRR) of the HIV-positive women tended to increase lower relatively to their HIV-negative counterparts. This understanding is critical, because from this pivotal study, it is indicative that in some way HIV have a negative in uence on fertility desires especially to those living with HIV.
Another study by Marston (2016) and colleagues was carried out to determine the effects of HIV on fertility by infection duration in Uganda using evidence from the population cohorts among the HIV-positive women.
Interestingly, their ndings revealed that longer duration of infection was signi cantly associated with greater relatively fertility reduction for HIV-positive women (Marston, 2016). Marston's ndings were in tandem with those by Chen et al., 2003 as they all postulated that HIV infection tend to reduce fertility desires among HIV-positive women which may also be suggesting possible in uence of HIV infection on fecundity. In this regard, the current study considers these study ndings critical in its assessment of fertility intentions among HIV-positive and treats HIV infection a strong in uence on the fertility desires among those living with HIV.
Other studies have also indicated that without intervention, HIV has between a 25% to 50% chance of infecting an HIV-positive mother's baby [13]. These ndings are crucial since they establish important information on the implication of HIV infection on the general population. Further, this information can help family planning programme planners at country level assess the desire for children, the extent of mistimed and unwanted pregnancies, and the demand for contraception to space or limit births. Although several studies have generated important nexus between HIV and fertility desires among those living with HIV, other existing research on fertility intentions of HIV-infected women in SSA shows considerable diversity in perspectives on future fertility intentions due to the advent of Anti-Retroviral Therapy [ART].
In SSA, for instance, some studies conducted have indicated that with the advent of ART, the quality of life for people living with HIV has also improved, potentially impacting fertility intentions among HIV-infected individuals [12,1,10,7]. From these study ndings, it can be deduced that the introduction of free ART means having HIV is no longer a big barrier to having children. Although this may be anticipated, the change seems to have enormous implications for reproductive decisions and behaviors of HIV-infected individuals. Further, issues of fertility and childbearing among HIV-positive women have received relatively little attention in Zambia and information on this topic is still scanty. For example, available evidence from all the six DHSs in Zambia collected information from women age 15-49 about their preferred total number of children in their lifetime. Even though this information is based on a hypothetical situation, it provides two important measures. Firstly, for women who have not yet started a family, the data provides an idea of future fertility. Secondly, for older and high-parity women, the excess of past fertility over ideal family size provides a measure of unwanted fertility. Results in gure 1 indicate that in 1992 DHS the average number of desired children was highest while the average number of desired children remained the same between 1996 and 2018. On the other hand, gure 2 show the trend on the percentage of who want no children from the rst ZDHS to the most recent one. Results show that women age 15-49 who want no more children increased from 24% in 1996 to 38% in 2018 [26].
Though Zambia recorded some improvements in reducing fertility desires among women over the two decades, investigating the in uence of HIV status on the desire for children among women of reproductive ages in Zambia is critical to strengthen family planning. Therefore, this study assesses the fertility intentions among HIV-positive and HIV-negative women using evidence from latest Zambia Demographic Health Surveys, 2013-14 and 2018. Speci cally, the study seeks to identify the socio-economic characteristics of the HIV-positive women and their fertility intentions, and nally, compares the fertility intentions between the infected and non-infected women of the reproductive ages.
Considering that Zambia's long-term vision (known as Vision 2030) is to become a prosperous middle-income nation by 2030, a research of this nature will be of great bene t, especially in the academic spheres and in the facet of reproductive health in the country as it is a crucial part of the general health. Additionally, the results of this study will contribute to update existing knowledge and inform the healthy policy makers and programmes to promote efforts for provision of safer and healthier reproductive options among HIV positive women in Zambia. It is also envisaged that the study would be relevant to HIV/AIDS programmes especially on interventions on preventing mother-to-child transmission (PMTC) of the HIV virus and serve as a guide in designing possible interventions for people living with HIV in Zambia as well as enriching demographic literature on fertility intentions, contraceptive use and HIV/AIDS in Zambia and sub-Saharan Africa.

Data And Methods
The study used secondary data from the 2013-14 and 2018 Zambia Demographic and Health Surveys (DHS) that had linkable information on HIV testing and fertility preferences. The DHS studies used a multistage sampling design that rst selected a random sample of enumeration areas and then selected a random sample of households systematically from a household listing of all households in the enumeration area. All eligible women in the sampled households were approached and asked to participate in the interview. The DHS is nationally representative survey which included a total sample of 16,411 women  in 2013-14 and 13,683 women in 2018. The respondents of this survey were interviewed from households using women's questionnaire [26]. This analysis only included 9,688 mothers of the reproductive age from the 2013-14 and 9,172 from 2018 surveys respectively with linkable information on HIV status either positive or negative and were eligible to answer questions on fertility preferences while women who did not know their HIV status were excluded from analysis.
In terms of fertility preference, only those who reported having a child, wanting to have more children or do not want to more children were included in the study. However, because some of the eligible women did not consent to provide blood sample for HIV testing, the le for reproductive data does not exactly match with HIV les [25].
For this study, the non-matching cases were excluded from the analytical le during data analysis.
The outcome variable for this study is fertility intention which was measured using a question asked in the DHS on whether women wanted to have more children or not. Analysis was done at three levels: descriptive, bivariate, and multivariate regression analysis. Frequency and percent distributions of selected background characteristics were presented by HIV status. At bivariate level, cross-tabulations with Chi-square tests were used to analyze the association between dependent variables and the selected independent variables. Multivariate binary logistic regression analysis was conducted to assess the in uence of HIV status on fertility intention. All the analysis was performed using Stata software version 14.0. Table 1 contains a summary of information on the number of women from the 2013-14 and 2018 Zambia Demographic and Health Survey Datasets. The distribution show that a total of 16,411 women (15-49 years) were captured during the 2013-14 survey while 13,683 in the 2018 survey. Of the total number captured, 71% (11,683) and 67% (9,172) had linkable information on HIV status and fertility preferences, respectively. Further, results indicate that highest (n = 9,847 and n = 7,702) numbers of respondents of respectively were HIV-negative mothers with more than 15% of the HIV-positive mothers from each survey. Regarding contraceptive use, only 38% (n = 711 and n = 550 respectively) of the HIV positive women reported using contraceptives from each survey while the vast majority (85%) of their HIV negative counterparts indicated using contraceptive from both surveys.     Table 3 show results of women's HIV status and the association between selected independent variables and the outcome variable (fertility intention) without controlling for the effects of the respondents' selected characteristics in bivariate analysis. In bivariate analysis, the following factors were signi cantly associated with fertility intentions; age (p < 0.000), education level (p < 0.003), wealth index (p < 0.017), marital status (p < 0.000), parity (p < 0.000) and employment status (p < 0.000). Overall, results show that majority of the HIV-negative women had intentions of having more children with 65% coming from 2013-14 and 60% from the 2018 DHS.

Descriptive Analysis
Similarly, results indicate that over two-fths of HIV-positive mothers from both surveys (50% and 45% respectively) wanted more children. Regarding age of the respondent, results reveal that fertility intentions tended to reduce by age with younger (15-24 years) women preferring to have more children regardless of their HIV status in both surveys. In terms of residence, most (53% and 47%) of HIV-positive women preferring more children resided in urban areas while over half of the HIV-negative women living in rural areas.
Regardless of HIV status and survey, it is shown that women with secondary and tertiary education comprised highest proportions of women wanting more children in future. Results in Table 3 also shows that majority (55% and 71% respectively) of

Logistic Regression Analysis of fertility intentions and HIV status
To identify the major predictors of fertility intentions, use by HIV status, multivariate binary logistic regression was used. Results of the logistic regression presented in Table 5

Logistic Regression Analysis of Contraceptive use and HIV status
The results in Table 6 Table 6 show that the odds of using contraceptives among HIV-positive married women in 2013 DHS were 2.2 times higher compared to the never married. In contrast, the odds of using contraceptives for HIV-positive women from 2018 DHS were 1.4 more likely using contraceptives when compared to the never married women. Equally, HIV-negative married women were more than 2 times likely to using contraceptives when compared to their never married counterparts.
During analysis, women with zero parity were dropped in the study as it only considered women with one or more children. In this regard, results show that HIV-positive women with more than 3 children (4-6

Discussion
This study aimed at assessing fertility intentions of HIV-positive and HIV negative mothers in Zambia. The study also sought to answer research questions on whether the HIV status and contraceptive use among women of reproductive age any in uence on fertility intentions has considering that the growing availability of ART may change the relationship. The predictors of fertility intention are discussed around identi ed major variables in the study. Correspondingly, details of factors associated with contraceptive use among HIV-positive and HIV negative mothers is also explained among the major determinants.
Fertility data in Zambia for the period 1992 to 2018 show that women generally have been having high preference for more children. For instance, the ZDHSs show that that the average number of desired of children ranges from 4 to 5 over from 1996 to 2018. Moreover, women living in rural areas preferred to have more children than those living in urban areas. This shows that the demand for large family size is higher in rural settings compared to urban settings [16]. The ndings of this study reveal that the prevalence of fertility intentions of HIVnegative women in 2018 is higher than that of HIV-positive (61% compared to 45%). These results are in a way consistent with studies from South Africa but higher than reports from Uganda (29% and 24%) respectively [1,7,22]. The ndings of this study show the fertility intention of HIV-positive women in Zambia is higher than those from other countries. However, the observed differences in fertility intention might be related to study sites, time, This secondary analysis of the 2013-14 and 2018 DHS reveals that education and parity are the strongest predictors of fertility intentions among HIV positive and HIV negative women. For instance, the fertility intentions for not wanting more children were increasing as education level increased. This entails that as one gets more educated there is a chance that knowledge of HIV also increases which may result into removal of the fertility desires and intentions for children especially after knowing the HIV status. These ndings are in tandem with a study conducted by Myer in 2017 and friends in South Africa and Ethiopia [15]. They also conform to Harriers' et al (2007) ndings in Kenya who used DHS data from several countries in SSA to assess fertility desires and intentions among HIV-infected individuals. The results also show that women with higher parity (above 3 children) are more likely to stop childbearing after learning their HIV status.
The study also examined the relationship between contraceptive use and fertility preference among HIV positive and HIV negative women. The prevalence of contraceptive use among HIV infected women was 40% in 2013-14 and 38% in 2018. This nding is lower than the reported overall prevalence of contraceptive use (49%) in 2013-14 and (50%) in 2018 among women of reproductive age in Zambia. It is also lower than that reported in the studies conducted in South Africa (53%) and Addis Ababa, Ethiopia (78%) [15]. The lower prevalence of contraceptive among HIV positive women in Zambia might be explained by the high levels of intention for more children (53%) among the married women. The other possible reason for the low prevalence of contraceptive use among HIVpositive women could be due to unavailability or lack of the preferred types of contraceptive methods at clinics offering ART.
The nding that the majority of HIV-positive women not using contraceptives in Zambia translates into high levels of HIV infected women. The situation is worrying considering that a number of family planning programmes in Zambia have been implemented not only to reduce fertility rate but also to help slow down the transmission of HIV. Moreover, strengthened family planning program efforts are cost effective points for HIV prevention (Feldman & Maposhere, 2003). However, the results may be re ecting the fact that HIV-positive women are expecting more children in future. Further, this evidence in Zambia is different from what Dube (2012) and friends found in Malawi [5]. In multivariate analysis, education level, marital status, parity were found to have more in uence on contraceptive use. HIV-positive women with primary, secondary, and higher education were more than two (2) times likely to use contraceptives compared to those with no education. These results may in a way suggesting the in uence of education on contraception and level of knowledge one would have acquired. Further, they are in line with the reported fertility intentions of HIV-positive women in this study which indicates that as education levels increase, the fertility intention reduces.
The study has revealed that even though HIV-negative women have higher fertility preference compared to HIVpositive women, the proportion of HIV positive women who prefer more children is unexpectedly high in Zambia.
In terms of policy implication for PMTCT service provisions, the ndings may be pointing out the need for strengthening the support for HIV prevention services especially for this group of women considering their high demand for children.

Conclusion
This study has examined fertility intentions of mothers in Zambia in the context of HIV pandemic. Although most HIV prevention programmes seem to focus on prevention of pregnancy among HIV-positive women, this study has demonstrated that the intention to have additional children among this group of women cannot be disregarded and ignored. Based on the ndings of this study, the major predictors of fertility intentions among HIV-positive women include; age, education level, marital status, parity, and type of contraceptive method in Zambia. The study has found out to some extent that availability of ART and other prevention services can still raise hopes for women living with HIV and they can believe to have a normal birth. The study concludes that use of contraceptive among these women of reproductive group is driven by marital status, parity, ethnicity, age and education. The study may have con rmed to some extent that availability of ART and other prevention services can still raise hopes for individuals living with HIV and they can believe to have a normal birth. It has been noted that the prevalence of contraceptive use among HIV-positive women is still low. It has been noted that the demand for additional children was lower among HIV-positive mothers compared HIV-negative mothers. This is an indication that maternal and child health care programmes focusing prevention of mother to child transmission are effectively working in disseminating health education information to women of reproductive age.

Declarations Competing Interests
The Authors declare that they have no competing interests. Figure 1