Salient beliefs and intention to use pre-exposure prophylaxis among pregnant and breastfeeding women in Zambia: Application of the Theory of Planned Behaviour

ABSTRACT Primary HIV prevention is a priority for pregnant and breastfeeding women in sub-Saharan Africa; however, such services should be designed to optimise uptake and continuation. Between September and December 2021, we enrolled 389 women who were not living with HIV into a cross-sectional study from antenatal/postnatal settings at Chipata Level 1 Hospital. We used the Theory of Planned Behaviour to study the relationship between salient beliefs and intention to use pre-exposure prophylaxis (PrEP) among eligible pregnant and breastfeeding women. On a seven-point scale, participants had positive attitudes towards PrEP (mean = 6.65, SD = 0.71), anticipated approval of PrEP use from significant others (mean = 6.09, SD = 1.51), felt confident that they could take PrEP if they desired (mean = 6.52, SD = 1.09) and had favourable intentions to use PrEP (mean = 6.01, SD = 1.36). Attitude, subjective norms, and perceived behavioural control significantly predicted intention to use PrEP respectively (β = 0.24; β = 0.55; β = 0.22, all p < 0.01). Social cognitive interventions are needed to promote social norms supportive of PrEP use during pregnancy and breastfeeding.


Introduction
Pregnant and breastfeeding women in sub-Saharan Africa remain at high risk for HIV infection (Drake et al., 2014;Graybill et al., 2020).In 2018 alone, an estimated 140,000 women in sub-Saharan Africa acquired HIV infection while pregnant or breastfeeding (UNAIDS, 2019).Biological, physiological and immunologic alterations that pregnant women experience further contribute to the elevated HIV risk observed during this period (J.Kinuthia et al., 2015;MacIntyre et al., 2015;Masson et al., 2015;Morrison et al., 2014;Sappenfield et al., 2013).This population is also of particular interest because incident maternal HIV infection during pregnancy poses an increased risk for mother-to-child transmission (MTCT) (De Schacht et al., 2014;Drake et al., 2014).Mothers who are acutely infected with HIV are twice as likely to transmit HIV to their children compared to their counterparts with chronic infection (Dunn et al., 1992).In Zambia, maternal seroconversion during late pregnancy and breastfeeding contributes significantly to the paediatric HIV burden (Johnson et al., 2012), accounting for about 13% of child infections (UNAIDS, 2021).Unfortunately, for most pregnant and breastfeeding women, HIV prevention has emphasised condom use, counselling, and education (Homsy et al., 2019;Villar-Loubet et al., 2013).Until recently, access to biomedical interventions has been limited.
PrEP is effective in preventing HIV infection to pregnant and breastfeeding women and could thus contribute towards the prevention of mother-to-child HIV transmission (PMTCT).In accordance with the World Health Organization guidelines recommending offering PrEP to at-risk populations (WHO, 2016), the 2020 Zambia HIV prevention guidelines extended the provision of PrEP to pregnant and breastfeeding women at risk for HIV infection (MOH, 2020).Despite these recommendations and several health benefits of PrEP at the individual and population level, both knowledge and uptake of PrEP among pregnant and breastfeeding women still remain low (Zimba et al., 2019).Unfortunately, there is paucity of knowledge about factors likely to influence uptake in the target population.
Qualitative studies that have explored reasons for low uptake of PrEP among young African women indicated the need for social support from their partners and significant others, and reported fear of their partners reactions about product use (van der Straten, Stadler, Montgomery, et al., 2014).Women also reported concealing use of PrEP and even hiding it for fear of being mistaken to be HIV positive (van der Straten, Stadler, Luecke, et al., 2014).PrEP studies conducted among pregnant and breastfeeding women in sub-Saharan Africa have mainly been integrated within demonstration projects and have been qualitative in nature, with recent studies focusing on how best to integrate PrEP in antenatal settings Most were conducted in Kenya and South Africa (Joseph Davey et al., 2021;Joseph Davey et al., 2022;John Kinuthia et al., 2020), and none of them have determined salient beliefs and their relative importance in influencing pregnant and breastfeeding women's intention to use PrEP.This research gap is even more evident in Zambia where there is limited literature on facilitators and barriers to PrEP uptake in antenatal settings.Determining modifiable antecedents of PrEP uptake, such as beliefs about PrEP and how these influence pregnant and breastfeeding women's intentions to use it may provide useful information that could inform interventions aimed at improving demand for PrEP in the target population.
Overcoming the social and structural barriers to universal access to combination HIV prevention services such as PrEP, requires an understanding of modifiable social cognitive antecedents of the target health behaviour (in this case, PrEP uptake).The Theory of Planned Behaviour (TPB) is a useful framework in understanding such relationships.According to this theory, behaviour is seen as a product of positive, pro-behaviour intentions.These intentions are in turn determined by attitudes toward using PrEP, subjective norms surrounding PrEP use, and perceived behavioural control (the perceived ease or difficulty of taking PrEP) (Ajzen, 1991).Intention can be used as a proximal determinant of behaviour (Francis et al., 2004).In the context of low PrEP uptake, understanding beliefs influencing intention could provide useful information on target behaviours that could influence actual PrEP uptake.The TPB has been used in several health related studies to predict both intention and actual behaviour such as cervical cancer screening (Abamecha et al., 2019), HIV testing (Mirkuzie et al., 2011), human papillomavirus vaccine uptake (Gerend & Shepherd, 2012), condom use (Asare, 2015;Teye-Kwadjo et al., 2017), and PMTCT (Igumbor et al., 2006).We applied the TPB to determine salient beliefs (attitude, subjective norm, perceived behavioural control) and their association with intention to use PrEP among pregnant and breastfeeding women not living with HIV in Zambia.

Study design and setting
We conducted a cross-sectional quantitative study to determine women's attitudes, subjective norms, and perceived behavioural control, and how these theorised constructs influence intention to use PrEP during pregnancy and breastfeeding.This study was conducted at the maternal and child health clinic in Chipata Level 1 Hospital (Lusaka, Zambia).The site was purposively selected based on the volume of pregnant women seen at the facility.Each month, it averages 400-450 new antenatal patients and 900-1000 return antenatal visits.

Study population and procedures
We recruited study participants using convenience sampling.We recruited pregnant and breastfeeding women aged 18 years or older with a documented HIV-negative result in their antenatal record.This study was part of a larger discrete choice experiment (DCE) whose aim was to determine preferences for PrEP delivery among pregnant and breastfeeding women not living with HIV and its sample size (n = 392) was based on those analyses.In this study, we did not offer PrEP to participants as part of the study procedures.However, it was available at the health facility where the study was conducted.Women who reported that they would like to use PrEP for HIV prevention were referred to the maternal and child health clinic within the hospital to discuss their interest to use PrEP with the health care providers.
Study personnel described the study and obtained written informed consent from eligible participants.The questionnaire was translated to Nyanja and Bemba (local languages commonly spoken in the study area) and then back translated to English for validation.It was further pre-tested to evaluate feasibility and clarity of questions.The questionnaire was administered to the women using face-to-face interviews in a private place within the hospital premises.

Measurement
The first part of the questionnaire included questions on the women's sociodemographic and obstetric characteristics, PrEP-related knowledge, engagement in risky sexual behaviour as well as HIV risk perception.The second part of the questionnaire collected information pertaining to the women's beliefs about PrEP use during pregnancy and breastfeeding.To do this, we selected an existing 37-item survey based on the TPB (Francis et al., 2004) and used it to guide the phrasing of questions for our questionnaire.Since the TPB questionnaire has not been validated in the Zambian context, we conducted a separate elicitation study (n = 24) in order to inform the development of PrEP-specific questions that would comprehensively capture beliefs likely to influence pregnant and breastfeeding women's intention to use PrEP, in Zambia.In this formative work, we asked women questions about the most frequently perceived advantages and disadvantages of taking PrEP; the most important people or groups of people who would approve or disapprove of using PrEP; and the perceived barriers or facilitating factors which could make it easier or more difficult to use PrEP during pregnancy and breastfeeding.This information was used to develop items that we used in the current study to measure attitudes towards PrEP, subjective norms about PrEP and perceived behavioural control of PrEP, respectively.This process was informed by the TPB framework.We used the Chronbach's alpha as a measure of internal consistency reliability to select the final items that were used to measure each of the TPB constructs.
The dependent variable, intention to use PrEP, was measured using a 3-item scale which was: 'I intend to take PrEP during pregnancy and breastfeeding'; 'I expect to take PrEP during pregnancy and breastfeeding'; 'I want to take PrEP during pregnancy and breastfeeding' (α = 0.80).Participants responded to each item on a 7-point Likert scale with responses ranging from strongly disagree (1) to strongly agree (7).A composite score for intention was obtained by getting the mean of the three items.The independent variables were attitudes towards PrEP, subjective norms about PrEP and perceived behavioural control of PrEP use.The scale measuring attitudes towards PrEP included four items: 'Overall, I think that taking daily oral PrEP would be: harmful (1) to beneficial (7)'; 'Overall, I think that taking daily oral PrEP would be not helpful (1) to helpful (7)'; 'Overall, I think that taking daily oral PrEP would be bad (1) to good (7)'; 'Overall, I think that taking daily oral PrEP would be wrong (1) to right ( 7)' (α = 0.78).A composite score for attitude towards PrEP was obtained by calculating the mean of the four items.
The subjective norms were measured with the following item: 'It would be expected of me to take daily oral PrEP during pregnancy and breastfeeding'.The perceived behavioural control scale was measured with the following single item: 'I am confident that I could take daily oral PrEP during pregnancy/breastfeeding if I wanted to'.Responses for both subjective norms and perceived behavioural control ranged from strongly disagree (1) to strongly agree (7).For all of the measures (attitudes, subjective norms, perceived behavioural control and intention), higher scores indicate more favourable attitudinal beliefs, greater perceived normative influence, greater perceived efficacy/controllability and more favourable intentions towards PrEP use.

Data analysis
We used descriptive statistics (frequencies and proportions) to describe background characteristics (socio-demographics, sexual behaviour and knowledge about PrEP) of study participants.Means and standard deviations were used to summarise the TPB constructs.Pearson Product Moment Correlation (r) was used to determine the correlation between intention to use PrEP and the three TPB predictor variables.We used simple linear regression to obtain unadjusted estimates for the association between intention to use PrEP and each of the TPB predictor variables, stratified by maternal status.Next, we used multivariable linear regression analysis using robust standard errors to obtain adjusted estimates of the association between intention to use PrEP and the predictor variables, including maternal status (pregnant or breastfeeding), while adjusting for other significant background characteristics (age, education, marital status and employment status).Associations were considered statistically significant at p < 0.05 at the 95% level of significance.We used the r-squared values from the stata output to determine the variance in intention to use PrEP explained by the models.We also assessed multi-collinearity in the final model using the variance inflation factor (VIF).VIFs greater than 10 were suggestive of presence of multi-collinearity.All analyses were conducted using Stata v.16 (StataCorp LLC, Lakeway Drive College Station, Texas 77845, USA).

Ethics approval
The study received approval from the University of Zambia Biomedical Research Ethics Committee (Lusaka, Zambia) and the Human Research Ethics Committee at the University of the Witwatersrand (Johannesburg, South Africa).Additional approvals were obtained from the Zambia National Health Research Authority and the Lusaka District Medical Office prior to study activation.All participants provided written informed consent prior to initiating study activities.

Results
Between September and December 2021, we administered a structured questionnaire to 389 pregnant and breastfeeding women not living with HIV who were receiving care at the study facility.Table 1 shows that participants had a median age of 26 years (IQR: 22-30) and the majority were married and cohabiting with a partner (84.1%).More than half of the participants (58.1%) had acquired secondary education and about one-third (33.4%) had acquired primary education.Very few women had acquired tertiary education (2.8%) and the majority were unemployed (69.2%).More than three quarters (86.9%) of the women never used condoms with their regular sexual partner in the 30 days preceding the survey, and a cumulative one-third of women perceived themselves to be at either moderate risk (20.6%) or high risk (12.6%) of HIV infection.The majority (63.7%) were not aware of PrEP prior to the survey.Among those who knew that PrEP protects against HIV infection, the majority (65.3%) cited health facility staff as their source of information.
Results from Table 2 show that participants in our study had favourable intentions towards PrEP use (mean = 6.01, std.dev.= 1.36) and positive overall evaluation of taking PrEP (mean = 6.65, std.dev.= 0.71).Participants also responded to items on social norms for PrEP and perceived behavioural control about PrEP use with high levels of agreement (mean = 6.09, std.dev.= 1.51 and mean = 6.52, std.dev.= 1.09), respectively.Results in Table 2 also show that all the three predictor variables were positively correlated with intention to use PrEP: attitude (r = 0.37, p < 0.001); subjective norms (r = 0.74, p < 0.001); perceived behavioural control (r = 0.48, p < 0.001).
In simple linear regression, stratified by maternal status, intention to use PrEP was statistically significantly associated with all the three constructs of the TPB both among pregnant and breastfeeding women, separately (all p < 0.01).However, coefficients for the association between PrEP intention and attitude (β = 0.83, p < 0.01) as well as subjective norms (β = 0.72, p < 0.01) were higher among breastfeeding compared to pregnant women (β = 0.55, p < 0.01 and β = 0.61, p < 0.01).
(Table 3).Among both pregnant and breastfeeding women combined (Table 4), intention to use PrEP was statistically significantly associated with: attitude (β = 0.70, p < 0.01); subjective norms (β = 0.67, p < 0.01); perceived behavioural control (β = 0.60, p < 0.01); and maternal status (β = 0.52, p < 0.01).Women who had more positive attitudes, those who held more positive subjective norms and those who felt more confident in their ability to take PrEP during pregnancy and breastfeeding also reported higher levels of PrEP intention.Breastfeeding women reported higher levels of PrEP intention compared to their pregnant counterparts.Subjective norms explained 55% of the variance in intention to use PrEP while perceived behavioural control explained 23% of the variance in intention to use PrEP.Attitude explained 13% while maternal status explained 4% of the variance in intention to use PrEP during pregnancy and breastfeeding.
In the second model (Table 4), we adjusted for all the theoretical constructs of the TPB.In this model, all the TPB predictor variables were positively associated with intention to use PrEP: attitude (β = 0.24, p = 0.03); subjective norms (β = 0.57, p < 0.001); perceived behavioural control (β = 0.21, p = 0.01).This model explained 60% of the variance in intention to use PrEP (p < 0.01).In the third and final model, we adjusted for all the TPB constructs, maternal status and sociodemographic characteristics of participants.All the predictor variables were positively associated with intention to use PrEP: attitude (β = 0.24, p = 0.03); subjective norms (β = 0.55, p < 0.01); perceived behavioural control (β = 0.22, p < 0.01); breastfeeding (β = 0.24, p = 0.01).Participants who scored higher on the attitude, subjective norms and perceived behavioural control scales, as well as breastfeeding women, reported greater intention to use PrEP both in the unadjusted and adjusted models.Subjective norms had the strongest association with intention to use PrEP.The final model explained 61% of the variance in intention to use PrEP among pregnant and breastfeeding women.There was no evidence of multi-collinearity in the final model as defined a priori (i.e.variance inflation factors less than 10 for all the variables included).

Discussion
In this study, we showed that PrEP attitudes, subjective norms, and perceived behavioural control were associated with PrEP use intentions among PrEP-naïve pregnant and breastfeeding women.Further, we demonstrated that pregnant and breastfeeding women not living with HIV have positive attitudes towards PrEP.Women in our study believed that people who are important to them would support their decision to use PrEP during pregnancy and breastfeeding.They also believed that they could easily take PrEP if they desired.Our study demonstrates that women who held favourable beliefs about the benefits of PrEP (attitude) had higher intention to use it during pregnancy and breastfeeding.A similar positive association on intention was observed among women who anticipated PrEP approval from others (subjective norms) and those who were confident that they could take PrEP during pregnancy and breastfeeding if they desired (perceived behavioural control).
One-third of women in our study perceived themselves to be at moderate to high risk of HIV infection, but the majority never used a condom with their regular sexual partners during pregnancy and breastfeeding.This is despite the generally high HIV prevalenceacross multiple populationsreported in the catchment area of the study site.In a population characterised by low education and employment, women may face challenges in negotiating safer sex practices with their male partners.This is exacerbated by the gender and power imbalances within relationships already inherent in such societies that tend to operate in a synergistic manner to further elevate these women's vulnerability to HIV infection (Wingood & DiClemente, 2000).Designing interventions aimed at sensitising women on risk factors for HIV infection could empower them with knowledge to help them better assess their own HIV risk and hopefully enable them to appreciate the importance of taking PrEP in order to protect themselves and their infants against HIV infection.
The majority of women in our study who knew about PrEP cited health facility staff as their source of information.Women and the broader community members who rarely seek health care from health facilities may never get to know about PrEP.This could be even worse for men, who tend to avoid health facilities because health services are considered to be for the sick, weak and generally female and children spaces (Mweemba et al., 2022).Therefore, concentrating PrEP education at health care facilities, especially Maternal and Child Health clinics, may further lower the prospects of men ever getting to know about the benefits of PrEP to the family and the community at large.Without this knowledge, men would be less likely to support their female partners who choose to initiate PrEP during pregnancy and breastfeeding.The more community members know about PrEP, the more they are likely to have positive attitudes towards PrEPand the more they are likely to support women who choose to use PrEP during pregnancy and breastfeeding.
Compared to pregnant women in our study, those who were breastfeeding showed more positive attitudes, subjective norms and control beliefs about PrEP use.They also reported higher PrEP intention.In qualitative studies, pregnant women generally raised persistent concerns about infant safety with antiretroviral exposure (Beima-Sofie et al., 2019;Pintye et al., 2017).Furthermore, pregnant women could have been more sceptical about product use compared to their breastfeeding counterparts due to the value placed in a woman's ability to bear children, especially in our context, where most often than not, failure to have a child is viewed as a threat to the stability of the relationship (Pintye et al., 2017).PrEP messaging should therefore emphasise the importance of initiating PrEP early on in order to stay HIV-negative throughout the maternal period when the risk for seroconversion and onward MTCT is highest.
Our finding of a positive association between women's positive attitude towards PrEP and the intention to use PrEP corroborates results of previous studies.For example, African American women who had positive perceptions and attitudes towards PrEP showed willingness to use PrEP (Ayangeakaa et al., 2022;Hill et al., 2021).Similar observations on the association between attitudes and intention were observed among women who inject drugs (Tran et al., 2021).In the United States, women's perceptions of the beneficial outcomes of PrEP were associated with intention to use PrEP in the next three months (Teitelman et al., 2020).It is imperative that PrEP messaging focuses on factors that may influence PrEP attitudes such as increasing PrEP knowledge and awareness on the benefits of PrEP in HIV prevention, as well as accurate HIV risk assessment in the target population (Aidoo-Frimpong et al., 2020).Behavioural interventions that target these factors could lead to positive attitudes which could translate to improved demand and actual uptake of PrEP by pregnant and breastfeeding women (Ajzen, 1991).
Women's broader social contexts may influence their decision to take PrEP and adhere to it.Young and McDaid argue that these social and cultural factors significantly influence women's perception and decisions regarding their use of HIV prevention methods including PrEP (Young & McDaid, 2014).In our study, we found that women who anticipated positive peer norms or less disapproval from people who are important to them were more likely to intend to use PrEP during pregnancy and breastfeeding.Our findings corroborate those from similar studies conducted among women, that found an association between receiving encouragement and support from loved ones and other women on PrEP and intention to use PrEP (Ayangeakaa et al., 2022;Calabrese et al., 2018;Goparaju et al., 2017;Pasipanodya et al., 2021;Scott et al., 2022;Teitelman et al., 2020).Together, these findings reiterate the need for interventions that advocate for the adoption of social norms that promote PrEP uptake by emphasising PrEP benefits for pregnant and breastfeeding women.
Since male partners are key decision makers in health-related decisions during pregnancy and breastfeeding (Montgomery et al., 2015;Mweemba et al., 2022;Roberts et al., 2020), it is important that male partners are engaged in female-controlled HIV prevention initiatives such as PrEP.This is even more critical for populations such as our study population where the majority of women are unemployed and may be dependent financially on their male partners.A lack of this anticipated approval from significant others may negatively influence pregnant and breastfeeding women's intention to use PrEP and ultimately uptake (Ayangeakaa et al., 2022).PrEP education messages that target broader communities, including couple-based approaches could improve community members' understanding about PrEP.This could address potential barriers to PrEP uptake, such as PrEP stigma, posed by negative social norms about PrEP (Goparaju et al., 2017).
Women in our study felt confident that they could take PrEP if they desired.This is contrary to the socio-culturally acceptable norm where women have to consult and seek approval from their partners on decisions that relate to their health, which is commonplace in Zambia (Mweemba et al., 2022).This also contrasts the common belief that women have limited control over their sexual and reproductive health choices and practices (Blanc, 2001;Darteh et al., 2019).It is possible that women in our study felt confident about their ability to take PrEP because they believed that people who are important to them would approve of their decision to use PrEP during pregnancy and breastfeeding.This confidence could also be based on the premise that PrEP provides women an increased sense of empowerment in directing their own sexual health (Pasipanodya et al., 2021), including control when they cannot control their partner's behaviour or where barrier protection fails (Flash et al., 2014).
Our findings further show a positive association between perceived control beliefs and intention to use PrEP.Women who felt confident about their ability to take PrEP had high intention to use it.These results corroborate findings from similar studies conducted among women (Pasipanodya et al., 2021;Tran et al., 2021).In light of our findings, the importance of empowering women so that they can have more control over their own health cannot be over-emphasised.Women in power-imbalanced relationships tend to depend on their male partners because men usually tend to bring more financial assets (i.e.money, status) to the relationship.As the economic inequity between men and women increases, and favours men, women will be more likely to experience adverse health outcomes (Wingood & DiClemente, 2000).Hence, empowering women economically may give them some level of control over PrEP-related decisions including its uptake.
Our study makes a significant contribution to the body of knowledge on salient beliefs that may influence intention to use PrEP in a population that is disproportionately affected by HIV.To the best of our knowledge, this is the first study to investigate psycho-social correlates of PrEP and how these correlates influence intention to use PrEP among pregnant and breastfeeding women not living with HIV in Zambia.However, we do note some limitations.First, because of limited PrEP availability in antenatal settings, we were only able to determine beliefs associated with intention (as opposed to actual PrEP uptake).According to the belief disparity hypothesis, people hold more positive beliefs about hypothetical behaviours compared to when the behaviour is enacted in real life (Ajzen & Driver, 1992).Therefore, it is possible that participants' perceptions of PrEP could change when given an opportunity to take it.Second, very few women knew about PrEP prior to their participation in this study.Their responses were mainly dependent on the description of PrEP provided by interviewers and could also have been influenced by social desirability.Third, we did not identify factors associated with the observed positive attitude towards PrEP use during pregnancy and breastfeeding.However, findings from a related qualitative study suggested that pregnant and breastfeeding women's attitudes toward PrEP were influenced by the belief that it has the potential to protect them and their infants from contracting HIV.PrEP was viewed as a way of protecting themselves in the event that their partner was cheating or secretly on ART (Hamoonga et al., 2022).Fourth, the cross-sectional nature of the data limits our ability to make causal conclusions on the association between the theorised TPB beliefs and intention to use PrEP since both exposures and outcome were measured at the same time.Despite these limitations, intention has been proven in several studies as a reliable antecedent to actual behaviour.Therefore, our findings are still valid and informative regarding beliefs that may influence uptake of PrEP in antenatal settings.

Conclusion
Consistent with the Theory of Planned Behaviour, our findings show that women's attitudes, subjective norms and perceived behavioural control are important correlates of intention to use PrEP during pregnancy and breastfeeding.Our findings suggest that social cognitive interventions aimed at generating demand for PrEP among pregnant and breastfeeding women should promote positive attitudes, subjective norms and control beliefs about PrEP use.Future studies should investigate how attitudes, subjective norms, perceived behavioural control, and intentions influence actual PrEP uptake in antenatal and postnatal settings.

Table 2 .
Self-reported intention, attitude, subjective norms and perceived behavioural control regarding PrEP use and Pearson correlations among the Theory of Planned Behaviour constructs.
b Mean scores and standard deviation among pregnant women.c Mean scores and standard deviation among breastfeeding women.

Table 3 .
Unadjusted estimates for intention to use PrEP stratified by maternal status.

Table 4 .
Unadjusted and adjusted estimates for intention to use PrEP among HIV-negative pregnant and breastfeeding women.
a Model 2 adjusted for the Theory of Planned Behaviour constructs only (attitude, subjective norms and perceived behavioural control).b Model 3 adjusted for the Theory of Planned Behaviour constructs and maternal status, alongside key characteristics of study participants (i.e.age, marital status, education, employment status, HIV risk perception).