Determinants of Fertility Desire among Women Living with HIV in the Childbearing Age Attending Antiretroviral Therapy Clinic at Jimma University Medical Center, Southwest Ethiopia: A Facility-Based Case-Control Study

Background High fertility and aspiration to have more children are a normal phenomenon in many developing countries including Ethiopia. The desire of people living with HIV/AIDS (PLWHA) to have children can have significant public health implications. Despite the growing number of women living with HIV/AIDS, the issues of fertility and childbearing intention have not been widely studied in Ethiopia. Objective To identify determinants of fertility desire among women living with HIV in the childbearing age attending antiretroviral therapy clinic at Jimma University Medical Center, Southwest Ethiopia. Methods A facility-based case-control study was conducted in March 2019. Cases were women living with HIV who had fertility desire, and controls were those who had not. Data was collected using a face-to-face interview using a pretested questionnaire. The data was entered into EpiData 3.1 and exported to SPSS Version 24 for analysis. Bivariate and multivariable logistic regression analyses were used to identify candidate and independent determinants of fertility desire, respectively. Independent determinants associated with fertility desire were assessed using AORs with their corresponding 95% CIs at P value < 0.05 cutoff point. Results. Three hundred forty-four (115 cases and 229 controls) were included into the study with a 100% response rate. Age categories 15-24 (AOR: 4.1; 95% CI: 2.0, 8.4) and 25-34 (AOR: 2.3; 95% CI: 1.3, 4.2) years, not using family planning (AOR: 2.3; 95% CI: 1.4, 4.0), and having a sexual partner (AOR: 1.9; 95% CI: 1.1, 3.2) were independent predictors of fertility desire. Conclusions Age of women, family planning, and sexual partner were found to be the independent predictors of fertility desire among women living with HIV/AIDS. Policymakers and health care providers who are working on an ART clinic should try to consider the effects of these factors for women living with HIV while developing HIV/AIDS interventions and discussing on sexual and reproductive health issues with their clients, respectively.


Introduction
Globally, an estimated 36.7 million people were living with HIV/AIDS; over 25.5 million of them are in Africa where 76% of all HIV-positive women live in sub-Saharan Africa.
AIDS-related illnesses remain the leading cause of death among women of reproductive age (15-49 years). It is the second leading cause of death among young women aged 15-24 years globally, and the first in Africa [1]. In Ethiopia, HIV prevalence among women of the reproductive age (15-49 years) was 0.9% and the most affected group are those who are sexually active and economically productive falling within the 25-49 age group [2].
Women's desires for fertility can put them on danger against the preventive strategies for vertical transmission of HIV/AIDS. Worldwide, over 90% of HIV infections among young children are due to mother-to-child transmission (MTCT). In the absence of any appropriate intervention, transmission of HIV/AIDS ranges from 15% to 45%. This transmission rate can be reduced to under 5% with effective interventions during the times of pregnancy, labor, delivery, and breastfeeding [3].
The UNICEF report estimated that MTCT of HIV was 9% in Latin America and Caribbean, 15% in Eastern Europe and Central Asia, 16% in Middle East and Northern Africa, 18% in West and Central Africa, and 8% in East and Southern Africa [4]. In Ethiopia, a finding from systematic review showed that the pooled prevalence of MTCT of HIV was 9.93%. Furthermore, the study also showed that almost one in every ten HIV-exposed infants became HIV positive [5]. Moreover, Ethiopia was characterized by low PMTCT service coverage. For example, Ethiopian public health institute (EPHI) reports indicated that more than 30,818 (8%) mothers need PMTCT of HIV/AIDS that cannot be addressed in 2018 [6].
Evidence indicates that women living with HIV continue with desire of having more children at varying degrees. For example, two studies conducted in Canada [6] and Malawi [7] showed that the proportions of HIV-positive women who wanted to have children in the future were 69% and 17% [7,8], respectively. In Ethiopia, different studies indicated different levels of fertility desire among HIV-positive women. A study done in Nekemt Town [9], Tigray, [10] and Addis Ababa [11] demonstrated that the proportions of women living with HIV who had fertility desire were 42.1%, 45.5%, and 56.2%, respectively.
In Ethiopia, limited studies have pointed out different factors that determine fertility desire of HIV-positive women [9,10,12,20,22]. However, no study has been conducted before in our study area to identify fertility desire of women living with HIV. Moreover, most of the previous studies does not assess HIV pregnancy-related knowledge and household wealth status, so this study strengthens the previous findings by addressing the above gaps.
The main aim of this study was to identify determinants of fertility desire among women living with HIV in Jimma University Medical Center, Southwest Ethiopia. So, this study had an implication for PMTCT of HIV/AIDS, the need for counseling to facilitate informed decision-making about childbearing, and the future demand for services for children born to infected women by identifying factors that determine fertility desire of women.

Methods
2.1. Study Setting, Design, and Period. An institution-based case-control study was conducted at Jimma University Medical Center (JUMC) in March 2019. Geographically, JUMC is located in Jimma City, which is 352 km to southwest of Addis Ababa. Currently, it is the only teaching and referral hospital in the southwestern part of the country, providing services to approximately 15,000 inpatient, 160,000 outpatient attendants, 11,000 emergency cases, and 4500 deliveries in a year. A total of 3221 HIV-positive individuals were on ART follow-up at JUMC ART clinic, of whom 1956 were females and 1775 were women in the reproductive age. Every stable client attends clinics every month, but clients with severe illness have frequent appointments within a month. The clinic provides counseling, tracing, ART treatment, and nutritional and medical treatment services.
2.2. Eligibility Criteria. All consecutively selected women living with HIV in the reproductive age (15-49) who had a follow-up in the ART clinics of Jimma University Medical Center were included in this study. Women who were pregnant, have severe mental illness, had hearing difficulties, and/or with confirmed infecundity were excluded.

Sample Size Determination and Sampling Technique.
Sample size was estimated using two population proportion formulas for unmatched case-control study using Epi Info version 7 StatCalc function considering assumptions of 95% CI, 80% power, and a case-control ratio of 1 : 2. From a similar study conducted in Addis Ababa, health status of the women living with HIV was taken as the main predictor of outcome fertility desire. P 1 representing the proportion of cases who are exposed to very good current health status was 60%, and P 2 representing the proportion of controls who are exposed to very good current health status was 44.4% [11]. Then, after considering a 10% nonresponse rate, the total sample size was estimated to be 344 (cases = 115, controls = 229). Study participants were grouped as cases if they had fertility desire and controls if they did not have fertility desire among women living with HIV on ART follow-up. The actual participants were selected through a consecutive sampling technique based on the criteria of the case definition.

Measures and Operational Definitions
2.4.1. Fertility Desires. It is a psychological state in which someone has the personal motivation to have a child. Those who have motivation to have more children in the future have fertility desire, and those who did not have motivation to have more children were considered not to have fertility desire during the study period [26]. The first category is TND which means target not detected or the virus is not at a detectable level. The second and third categories are the detectable level which is <20 copies and ≥20 copies HIV RNA copies per milliliter of blood, respectively [30].
2.6. Data Management and Analysis. Data completeness was checked manually, then entered into EpiData 3.1 and exported to SPSS Version 24 for analysis. Pretest of the questionnaire was conducted in Shenen Gibe Hospital, which had similar services to the study hospital. Training was given to the data collectors and supervisors on the objective of the study, maintenance of ethical standards, and methods of data collection; hence, the data collectors were familiarized with data collection tools with respect to the study with practical exercises. Data collectors were blinded about the outcome of the study so that they did not distinguish cases and controls.
The study participants were identified/classified after data collection based on unique identification codes. Data coding and cleaning were performed by cross-checking to printout data for possible errors. Missing values and outliers were checked before analysis by running descriptive analysis. The household wealth index was computed using the PCA method by considering locally available household assets which were dummy coded. Before running the PCA, the requirements of factor analysis were checked.
The descriptive statistics were presented in frequency tables and graphs. Normality assumption was checked for age. Multicollinearity between the independent variables was checked. The association between fertility desire and each covariate was assessed first by bivariate logistic regression to identify candidate variables for the final model. Variables with P value < 0.25 were taken into multiple logistic regression analysis to identify independent predictors of fertility desire.
The backward likelihood ratio with 0.1 probability removal was used to develop the model. Goodness of fit of the final model was checked using the Hosmer-Lemeshow test of goodness of fit considering good fit at P value > 0.05, omnibus likelihood test < 0:05, and model classification of accuracy was checked. Finally, independent predictors of fertility desire were declared at P value < 0.05 cutoff point, and strength of the association was assessed using AORs with their corresponding 95% CIs.
2.7. Ethical Statement. The study was approved by the Research and Ethics Committee, of Jimma University Institute of Health. Then, relevant offices were communicated for their cooperation with a formal letter issued by the Research and Ethics Committee. The purpose of the study was verified briefly to the study participant, and confidentiality was assured. Finally, written consent was obtained from study participants before conducting the interview. For women, who were married and less than 18 years of age, in addition to the assent, permission of their husbands was obtained to include them in the study. But for study participants, who are less than 18 years of age, were not married, and are living independently, as there is no one to sign on behalf of the parent or guardian, only assents from the participant were obtained.

Sociodemographic and Economic Characteristics.
A total of 115 cases and 229 controls were included in the study with a 100% response rate. Three-fourth of the 87 (75.7%) cases and majority 185 (80.8%) of the controls were urban residents. Concerning the age of the study participants, the mean age of cases was 28:5 ± 5:6 SD and the mean age of controls was 32:7 ± 6:5 SD. Closely half 54 (47.0%) of the cases and 101 (44.1%) of the controls were Orthodox religion followers. Concerning marital status of the study participants, 53 (46.1%) of the cases and nearly half 111 (48.5%) of the controls were married.

International Journal of Reproductive Medicine
In bivariate analysis, from all sociodemographic and economic characteristics, age of the respondents and household wealth status were potential candidates for multivariable logistic regression (see Table 1).

Reproductive Characteristics.
Majority of the cases 95 (83.3%) reported to have planned to have one or two children where more than half 67 (58.3%) of the cases wished to have children within a two-year duration. Among controls, the main reason not to have children for about 80 (34.9%) were due to fear of mother-to-child transmission of HIV (see Figure 1).
Almost all 113 (98.2%) of the cases and 227 (99.1%) of the controls reported to have had sexual intercourse in their lifetime. About three-fourth 86 (76.1%) of the cases and closely two-third 139 (61.2%) of the controls were sexually active in the last 12 months preceding the study. More than half 65 (57.5%) of the cases and three-fourth (77.1%) of the controls had children. Majority of 89 (78.8%) cases and 161 (70.9%) controls had no history of abortion. Regarding family planning option, slightly below three-fourth of 82 (72.6%) cases had not used family planning and 126 (55.5%) controls used family planning. About two-third 75 (65.2%) of the cases and slightly above three-fourth 177 (77.3%) of the controls had no any history of sexual violence.
In bivariate analysis, from all reproductive characteristics, sexual activity in the last 12 months, having a sexual partner currently, lack of child, history of abortion, family  International Journal of Reproductive Medicine planning, family support, family pressure to have children, and history of sexual violence were potential candidates for multiple logistic regression (see Table 2).    Table 3). In bivariate analysis from all clinical characteristics, CD4 count, viral load, and knowledge were potential candidates for multiple logistic regression (see Table 4).

Determinants of Fertility
Desire. From the total variable entered in the multivariable regression, three variables were found to be independently associated with fertility desire among women living with HIV (see Table 5). The odds of fertility desire among the age group 15-24 years and 25-34 were four times (AOR: 4.1; 95% CI: 2.0, 8.4) and two times (AOR: 2.3; 95% CI: 1.3, 4.2) higher as compared to age range from 35 to 49 years, respectively.
The odds of fertility desire among women who had not used family planning were two times (AOR: 2.3; 95% CI: 1.4, 4.0) higher as compared to those women living with HIV who use family planning. The odds of fertility desire among women who had sexual partners were about two times (AOR: 1.9; 95% CI: 1.1-3.2) higher when compared to those women living with HIV who had no sexual partners (see Table 5).

Discussion
Understanding the desire for fertility among women living with HIV/AIDS has a significant role to reduce MTCT of HIV with the introduction of ART, which changes the views of childbearing despite having the disease [31]. Age of the women was significantly associated with fertility desire. Women of younger age had higher odds fertility desire as compared to older women. This finding is similar to that of studies from Brazil, Canada, USA, Guinea, Ghana, and Rwanda which indicate that young women living with HIV experience significantly more fertility desire than HIVinfected, older women [7,13,[15][16][17]25]. In Ethiopia, some studies demonstrated the presence of an association between age and fertility desire [10,12,20]. The possible similarity is that it is quite normal to have a fertility desire during this period b/c it is the peak of reproduction. In many societies, women are active in producing their offsprings during this period [32]. On the contrary in another study, being young was reported to be inversely associated with fertility desire [33]. The possible reason for the inconsistency may be sociocultural or methodological differences between study participants.
In the current study, the odds of fertility desire among women, who had not used family planning, were higher when compared with those who used family planning. This finding is supported by a study done in Uganda and Guinea [25,34], which revealed that women who have not used any family planning are more likely to have fertility desire than those who used. A study in Ethiopia has reported the same finding [35]. This might be due to the fact that they have not attained their desired family size and unmet need of family planning service [36]. In contrast, having no family planning is negatively associated with fertility desire [37]. The possible difference might be due to poor economic status having less fertility desire. The health workers providing their chronic AIDS care must be made aware of this, and a range of contraceptive options should be made available, as well as all the associated information about these options.
Women who had sexual partners had higher odds of fertility desire as compared to women who had no sexual partners. This result is also in line with the study done in Tanzania and Guinea [23,25] which stated that odds of fertility desire were high on women who had sexual partners than their counterparts. A study in Ethiopia reported the same finding [20]. Probably, this may be due to the fact that having sexual partners may play a role in preconception planning behaviors. Moreover, having male partners minimizes her economic, social, and psychological burden of HIV/AIDS [12]. Thus, policymakers and health care providers should consider the role of male partners while developing policy and running an HIV/AIDS intervention.

Study Limitation.
Men were not included in the study, in spite of the essential role they can play in deciding about their family size. This study is also limited to women who are receiving chronic, long-term care for their HIV infection, despite fertility desire differing from women who are not Abbreviations AIDS: Acquired immune deficiency syndrome AOR: Adjusted odds ratio ART: Antiretroviral therapy COR: Crude odds ratio EDHS: Ethiopia Demographic and Health Survey HIV: Human immunodeficiency virus JUMC: Jimma University Medical Center MTCT: Mother-to-child transmission PLWHA: People living with HIV/AIDS PMTCT: Preventing mother-to-child transmission TND: Target not detected