Jean-Pierre Molès, Institut National pour la Sante et la Recherche Médicale, France
Robyn Eakle, President's Emergency Plan for AIDS Relief, United States
ORCID Wilbroad Mutale
This is an open-access article distributed under the terms of the
Pregnant and breastfeeding women at substantial risk for HIV infection in sub-Saharan Africa can benefit from biomedical interventions such as pre-exposure prophylaxis (PrEP). We estimated the benefit that pregnant and breastfeeding women may derive from PrEP service delivery in order to guide PrEP roll-out in the target population in Zambia.
Between September and December 2021, we conducted a discrete choice experiment (DCE) among a convenient sample of 389 pregnant and breastfeeding women not living with HIV in Lusaka, Zambia. Women aged 18 years or older, with a documented negative HIV result in their antenatal card responded to a structured questionnaire containing 12 choice sets on service delivery attributes of PrEP: waiting time at the facility, travel time to the facility dispensing PrEP, location for PrEP pick-up, health care provider attitude and PrEP supply at each refill. Mixed logit regression analysis was used to determine the participant's willingness to trade off one attribute of PrEP for the other at a 5% significance level. Willingness to wait (WTW) was used to determine the relative utility derived from each attribute against waiting time.
Waiting time at the facility, travel time to the facility, health care provider attitude and amount of PrEP supply at each refill were important attributes of PrEP service delivery (all
Patient-centered approaches can help to inform the design and implementation of PrEP services among pregnant and breastfeeding women. In this study, we found that a reduction in clinic visits—including through multi-month dispensing of PrEP—could improve uptake of services in antenatal and postnatal settings.
Pregnant and breastfeeding women are disproportionately affected by HIV in sub-Saharan Africa (
PrEP was first introduced in Zambia in 2016 as a component of HIV prevention services in line with the World Health Organization guidelines at the time. National campaigns have been implemented in the years that followed—including Zambia Ending AIDS (
Discrete choice experiments (DCEs) elicit preferences for health care service delivery and can be used to optimize health services. It is a quantitative technique that is grounded in the economic theory of utility maximization and can be useful in eliciting key modifiable attributes among potential users (
We conducted a DCE study to quantify service delivery preferences for PrEP among pregnant and breastfeeding women not living with HIV in Lusaka, Zambia. In order to elicit an initial list of attributes and attribute levels to include in the DCE, we conducted in-depth interviews using a semi-structured interview guide as part of our formative work. These interviews were conducted with 24 women who were either pregnant or breastfeeding. We also reviewed existing literature on PrEP service delivery attributes. Based on this literature review—as well as findings from the formative study, and nominal group technique (
Attributes and levels of PrEP service delivery.
| Attributes | Levels |
|---|---|
| Location | Health facility |
| Home | |
| Travel time | Less than 1 hour |
| 1 hour or more | |
| Waiting time | 1 hour |
| 3 hours | |
| 5 hours | |
| Attitude of health care provider | Negative |
| Positive | |
| Supply at each refill | 1 month |
| 2 months | |
| 3 months |
We conducted this study at Chipata Level 1 hospital, a government facility that is located in peri-urban Lusaka. This is a high volume facility that has a catchment population of over 100,000 and an antenatal clinic that attends to about 400–450 new ANC attendees each month. Given this information, we were confident that we would be able to recruit the required sample size for the study from this single site. The HIV prevalence among pregnant women attended to at this health facility is approximately 16%, a rate that is similar to the national prevalence for women (
To calculate the study's sample size, we used the formula:
where
Study personnel described the study and obtained written informed consent from eligible candidates who were willing to participate. We then collected information about sociodemographic attributes and PrEP knowledge questions using a structured questionnaire developed by the study team. This was followed by questions on preferences for PrEP service delivery. To ensure that participants understood the choice sets, a study team member described the attributes and attribute levels that participants were going to encounter in the DCE. This was done using a sample DCE question which had one choice set. Participants were later presented with 12 choice sets and were asked to choose between the different scenarios which contained the following service delivery attributes of PrEP: (1) place where PrEP is dispensed; (2) waiting time at the facility; (3) travel time to the facility dispensing PrEP; (4) attitude of health care provider and (5) supply at each refill. An example of the choice sets presented to women in our study is illustrated. In all the 12 tasks, the attributes remained constant while the attribute levels varied. The instrument was pre-tested to evaluate feasibility and clarity of questions. The final version was translated into Nyanja and Bemba (commonly spoken local languages) and independently back-translated to English for verification. Trained study personnel administered the questionnaire using face-to-face interviews via paper forms.
We summarized the background characteristics of participants using frequency (percentage) for categorical variables and median (interquartile range) for continuous variables. The primary outcome or utility was women's preferences for PrEP service delivery. We used McFadden's choice model (
We determined the relative importance of the attributes using the significance of the coefficients
The study received approval from the University of Zambia Biomedical Research Ethics Committee in Lusaka, Zambia (934–2020) and the Human Research Ethics Committee at the University of the Witwatersrand in Johannesburg, South Africa (M200564 MED20-02-145). Additional approvals were obtained from the Zambia National Health Research Authority and the Lusaka District Medical Office prior to study activation. Members of the study team went through the information sheet together with study participants in the participants’ preferred language (English, Nyanja or Bemba). After going through the information sheet, participants were given an opportunity to ask questions and seek clarity on any aspects of the study. Participants who agreed to participate in the study were asked to sign consent forms prior to their participation in the study. Those who were unable to provide written consent substituted their signature with their thumb-print. All interviews were conducted in a private place, and participants were assured that their information would be kept confidential. Participants were given a snack in form of a drink and biscuit and transport reimbursement at the end of the interview.
We conducted face-to-face interviews with 389 pregnant and breastfeeding women receiving care at the study site's Maternal and Child Health Clinic between September and December 2021.
Sociodemographic characteristics of pregnant and breastfeeding women (
| Characteristic | |
|---|---|
| Age in years, median (IQR) | 26 (22–30) |
| Marital status | |
| Single | 44 (11.3) |
| Married (cohabiting with partner) | 327 (84.1) |
| Married (not cohabiting with partner) | 18 (4.6) |
| Education | |
| No formal education | 22 (5.7) |
| Primary | 130 (33.4) |
| Secondary | 226 (58.1) |
| Tertiary | 11 (2.8) |
| Employment status | |
| Not working | 269 (69.2) |
| Employed for wages | 27 (6.9) |
| Self employed | 93 (23.9) |
| Maternal status | |
| Pregnant | 172 (44.2) |
| Breastfeeding | 217 (55.8) |
| Gestational age (pregnant women only, |
|
| First trimester | 10 (5.8) |
| Second trimester | 70 (40.7) |
| Third trimester | 92 (53.5) |
| Condom use in past 30 days | |
| Never | 338 (86.9) |
| Sometimes | 36 (9.2) |
| Always | 15 (3.9) |
| Lifetime sexual partners, median (range) | 2 (1–25) |
| Partner HIV status | |
| Known | 318 (81.8) |
| Unknown | 71 (18.2) |
| HIV risk perception | |
| No risk | 158 (40.6) |
| Low risk | 102 (26.2) |
| Moderate risk | 80 (20.6) |
| High risk | 49 (12.6) |
| PrEP protects against HIV ( |
|
| Yes | 98 (69.5) |
| No | 43 (30.5) |
Among women who reported being knowledgeable about PrEP prior to the interview.
Each patient in the study responded to 12 choice sets; cumulatively, 4,668 choice sets were presented to study participants. Results from our mixed logit regression analysis (
Attribute's weight in decision making about PrEP uptake during pregnancy and breastfeeding.
| Attribute | Coefficient ( |
Robust std. error | Odds ratio | Willingness to wait (h) |
|
|---|---|---|---|---|---|
| Location | |||||
| Facility | 0.00 | – | – | 1.00 | 0.00 |
| Community | 0.00 | 0.04 | 0.955 | 1.00 | 0.13 |
| Travel time | |||||
| ≥1 hour | 0.00 | 1.00 | 0.00 | ||
| <1 hour | 0.27 | 0.07 | <0.001 | 1.30 | 0.60 |
| Time (h) |
−0.57 | 2.00 | <0.001 | 0.56 | |
| Attitude | |||||
| Negative | 0.00 | 1.00 | 0.00 | ||
| Positive | 0.97 | 0.19 | <0.001 | 2.64 | 1.96 |
| Supply at refill | |||||
| 1 month | 0.00 | 1.00 | 0.00 | ||
| 2 months | 0.87 | 0.16 | <0.001 | 2.38 | 1.77 |
| 3 months | 1.68 | 0.54 | <0.001 | 5.38 | 3.33 |
Estimated by the ratio of each coefficient to time coefficient.
Measured on a continuous scale.
Participants had 30% increase in the odds of preferring travel time that is less than 1 h compared to travel time of 1 h or more (OR = 1.30,
We simulated alternative strategies of delivering PrEP in order to determine which model would offer the greatest utility to women who choose to initiate PrEP during pregnancy and breastfeeding.
Estimated preference utilities and additional time willingness to wait for change in PrEP delivery.
| Waiting time | Location | Travel time | Attitude | Supply | Benefit score | Additional time willingness to wait |
|---|---|---|---|---|---|---|
| Base case | ||||||
| 3 h | Health facility | 1 h or more | Negative | 1 month | – | – |
| Changing only one attribute of PrEP service delivery | ||||||
| 1 h | – | – | – | – | 0.54 | 2.00 |
| 5 h | – | – | – | – | −0.54 | −2.00 |
| – | – | Less than 1 h | – | – | 0.27 | 0.93 |
| – | – | – | Positive | – | 0.97 | 3.40 |
| – | – | – | – | 2 months | 0.87 | 3.04 |
| – | – | – | – | 3 months | 3.38 | 11.78 |
| Changing two attributes of PrEP service delivery | ||||||
| 1 h | – | Less than 1 h | – | – | 0.81 | 2.93 |
| 1 h | – | – | Positive | – | 1.51 | 5.4 |
| 1 h | – | – | – | 2 months | 1.41 | 5.04 |
| 1 h | – | – | – | 3 months | 3.92 | 13.78 |
| – | – | Less than 1 h | Positive | – | 1.24 | 4.33 |
| – | – | Less than 1 h | – | 2 months | 1.14 | 3.97 |
| – | – | Less than 1 h | – | 3 months | 3.65 | 12.71 |
| – | – | – | Positive | 2 months | 1.84 | 6.44 |
| – | – | – | Positive | 3 months | 4.35 | 15.18 |
| 5 h | – | Less than 1 h | – | – | −0.27 | −1.07 |
| 5 h | – | – | Positive | – | 0.43 | 1.40 |
| 5 h | – | – | – | 2 months | 0.33 | 1.04 |
| 5 h | – | – | – | 3 months | 2.84 | 9.78 |
| Changing three attributes of PrEP service delivery | ||||||
| 1 h | – | Less than 1 h | Positive | – | 1.78 | 6.33 |
| 1 h | – | Less than 1 h | – | 2 months | 1.68 | 5.97 |
| 1 h | – | Less than 1 h | – | 3 months | 4.19 | 14.71 |
| 1 h | – | – | Positive | 2 months | 2.38 | 8.44 |
| 1 h | – | – | Positive | 3 months | 4.89 | 17.18 |
| – | – | Less than 1 h | Positive | 2 months | 2.11 | 7.37 |
| – | – | Less than 1 h | Positive | 3 months | 4.62 | 16.11 |
| 5 h | – | Less than 1 h | Positive | – | 0.70 | 2.33 |
| 5 h | – | Less than 1 h | – | 2 months | 0.60 | 1.97 |
| 5 h | – | Less than 1 h | – | 3 months | 3.11 | 10.71 |
| 5 h | – | – | Positive | 2 months | 1.30 | 4.44 |
| 5 h | – | – | Positive | 3 months | 3.81 | 13.18 |
| Changing four attributes of PrEP service delivery | ||||||
| 1 h | – | Less than 1 h | Positive | 2 months | 2.65 | 9.37 |
| 1 h | – | Less than 1 h | Positive | 3 months | 5.16 | 18.11 |
| 5 h | – | Less than 1 h | Positive | 2 months | 1.57 | 5.37 |
| 5 h | – | Less than 1 h | Positive | 3 months | 4.08 | 14.11 |
Alternative ways of providing PrEP delivery to pregnant and breastfeeding women yielding the highest benefit score.
We compared alternative scenarios that changed two attributes. Based on our responses, women would derive more value from using PrEP if they received 3-months’ supply of PrEP at each refill and encountered a health care provider with a positive attitude (
Our study revealed that waiting time at the facility, travel time to the facility dispensing PrEP, health care provider attitude and amount of PrEP supply at each refill were important attributes of PrEP service delivery among pregnant and breastfeeding women. These findings on preferred attributes of PrEP service delivery are consistent with WHO guidance (
In this study, participants demonstrated a relatively strong preference for a 3-month supply of PrEP. In fact, this was the most preferred single attribute likely to influence PrEP uptake among pregnant and breastfeeding women in our study. There are a number of possible explanations. For example, it is possible that women preferred the 3-month supply (as opposed to shorter refill periods) to avoid costs associated with frequent clinic visits for PrEP refills. Multi-month refills can address other key barriers to PrEP uptake such as transport expenses and lost wages arising from long waiting hours at the facility. However, there may be disadvantages to such an approach. Caution should be taken to ensure that multi-month scripting of PrEP is not offered to women who intend to use PrEP for a short duration, who are uncertain of their period of risk, and/or have medical conditions that require more intensive monitoring (
Health care provider attitude was the second most preferred single attribute of PrEP service delivery among study participants, a finding similar to other studies in the region. In Kenya and Uganda, for instance, women highlighted the importance of health care providers with respect to PrEP messaging and adherence (
Although circumstances surrounding uptake of PrEP may differ from those around HIV treatment, both PrEP and ART are biomedical interventions and lessons learnt from ART studies could inform PrEP delivery. In a DCE that was conducted in Zambia to quantify facility-based preferences for ART services, patients were willing to wait 19 h more to see nice rather than rude providers. Patients were willing to accept a facility located 10 km from home (as opposed to 5) that required 5 h of waiting per visit (as opposed to 1 h) and that dispensed 3 months of medications (instead of 5) in order to access nice (as opposed to rude) providers (
Waiting time and travel time to the facility dispensing PrEP were equally important attributes that influenced the likelihood of PrEP uptake among women interviewed in our study. Participants preferred less waiting time and shorter travel time to the facility. These results are supported by several other studies that reported long waiting hours and long travel time to health facilities as health systems barriers to uptake of services. In one study, living far from the clinic created difficulties to visit the clinic while lengthy clinic visits were viewed as extremely disruptive and demotivating, lack of consideration for participants’ time (
Our findings have important public health implications. Although waiting time, travel time to the facility, amount of PrEP supply and health care provider attitude all influenced the likelihood of PrEP uptake in our study population, addressing all barriers of service uptake amidst resource constraints may require adopting an incremental approach. This approach would require an understanding of the value derived from each of the preferred attributes of PrEP service delivery. Our study quantifies the trade-offs that women are willing to make in order to derive value from using PrEP during pregnancy and breastfeeding. In view of our findings, prioritizing pregnant and breastfeeding women for multi-month dispensing of PrEP as well as improving client-health care provider relationships, through improved health care provider attitude may have a positive impact on PrEP uptake in antenatal and postnatal settings. Increased uptake of PrEP in the target population could significantly reduce the risk of maternal HIV infections and thereby avert the negative impact of HIV infections on women, including increased risk of maternal morbidity and mortality, reduced fecundity, miscarriages and still births (
Effective delivery of HIV prevention to pregnant and breastfeeding women is essential. Women who newly acquire HIV face significant lifetime risk for morbidity and mortality; when this occurs during the antenatal and postnatal periods, there is further concern for onward vertical HIV transmission. PrEP is an important component of comprehensive HIV prevention, but its delivery within maternal-child health platforms can present challenges. Patient-centric approaches—such as differentiated service delivery to PrEP implementation—may hold promise. These strategies respond to client needs and preferences, and can overcome known individual and health system barriers to PrEP uptake and continuation (
Despite the strengths of our work, we note several limitations. First, theoretical uptake of PrEP used in our study limits our understanding of preferences for PrEP delivery in real-life settings. At the time of study implementation, PrEP was not yet integrated into maternal and child health platforms, and so study findings were meant to inform programs that would do so. Second, we limited the study to daily oral PrEP, since that was the only formulation available at the time that the study was conducted. Findings on preferences could be different with long-acting formulations such as injectable cabotegravir or the dapivirine vaginal ring; however, at the time of data collection, neither was approved for HIV prevention in Zambia. Third, we limited PrEP service delivery attributes to only five and therefore were unable to measure other attributes that could potentially influence women's decision making process, such as location for pick-up of PrEP within the health facility (e.g., ART department, out-patient department, maternal and child health department). However, presenting more attributes and attribute levels is associated with greater cognitive difficulty of completing a DCE for participants (
Our study shows that waiting time at the facility, travel time to the facility dispensing PrEP, health care provider attitude and supply at each refill are key attributes of service delivery likely to influence uptake of PrEP among eligible pregnant and breastfeeding women. As national programs seek to implement differentiated service delivery for PrEP—especially initiatives for pregnant and breastfeeding women—such efforts to prioritize individual needs and preferences are urgently needed.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The studies involving humans were approved by University of Zambia Biomedical Research Ethics Committee, Wits Human Research Ethics Committee. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
TEH: Conceptualization, Formal Analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. WM: Conceptualization, Methodology, Supervision, Writing – review & editing. JI: Supervision, Writing – review & editing, Conceptualization, Methodology. SB: Formal Analysis, Writing – review & editing, Supervision. OA: Writing – review & editing, Formal Analysis. BHC: Conceptualization, Funding acquisition, Methodology, Supervision, Writing – review & editing.
The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This work was funded with support from the Fogarty International Center [grant number: D43 TW010558] and the National Institute of Allergy and Infectious Diseases [grant numbers: K24 AI120796, R01 AI131060].
We extend our sincere gratitude to the UNC-UNZA-Wits Partnership for HIV and Women's Reproductive Health program for supporting this work. Further gratitude goes to the Zambian Ministry of Health and also to our study participants for their time and insights.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.