Drivers and barriers of vaccine acceptance among pregnant women in Kenya

ABSTRACT Maternal vaccination coverage remains suboptimal globally and is lowest in low- and middle-income countries. Attitudes toward maternal vaccines have been characterized in middle-high income settings, however data from African countries are limited. We assessed drivers and barriers of vaccine acceptance among pregnant women in Kenya. We conducted a cross-sectional survey among pregnant women aged 15–49 y. We enrolled a convenience sample of women presenting for antenatal care at seven health-care facilities in four diverse counties (Nairobi, Mombasa, Marsabit, Siaya) of Kenya and from the community in two counties (Nairobi, Siaya). We described frequencies of socio-demographic characteristics of participants and their knowledge, attitudes, and beliefs regarding maternal vaccination. We enrolled 604 pregnant women with a median age of 26.5 y, of whom 48.2% had primary education or less. More than 95% agreed that maternal vaccines are “important for my health” and that getting vaccinated is “a good way to protect myself from disease”. The most commonly cited reason in favor of maternal vaccination was disease prevention (53.2%). Fear of side effects to mother/baby (15.1%) was the most frequently reported potential barrier. Influenza vaccine is not in routine use in Kenya; however, 77.8% reported willingness to accept influenza vaccination during pregnancy. Maternal vaccination is well accepted among Kenyan pregnant women. We identified the provision of adequate vaccine information and addressing safety concerns as opportunities to improve maternal vaccine uptake. The expressed willingness to receive a vaccine not currently in routine use bodes well for implementation of new maternal vaccines in Kenya.


Introduction
Maternal vaccination is an important strategy to prevent maternal, neonatal, and infant disease. [1][2][3] Despite evidence on the safety and effectiveness of maternal vaccines, challenges to achieving high vaccination coverage during pregnancy still exist globally. 4 The World Organization (WHO) recommends tetanus toxoid (TT) vaccination for all pregnant women until they receive required doses to achieve full protection; however, globally, coverage for two or more TTcontaining vaccines among pregnant women is 73%, 5 while the universal target requires ≥80% coverage that is needed to achieve and maintain maternal and neonatal tetanus elimination. 6 The proportion of pregnancies adequately protected against tetanus is lowest in Africa (60%). 7,8 In Kenya, TT is the only maternal vaccine included as part of the Kenya Expanded Programme on Immunization, yet the uptake of two or more TT doses among pregnant women in the last Kenya Demographic and Health Survey (KDHS, 2014) was 51.1%, 9 which is far below the universal target for TT coverage. Although Kenya achieved the elimination of maternal and neonatal tetanus in 2019, 10 this low coverage of maternal TT vaccination presents a substantial risk to the health and well-being of pregnant women and young infants.
Acceptance of vaccines and vaccine hesitancy is influenced by a wide range of factors that may vary by time, place, and type of vaccine. 11,12 Acceptance of vaccines among pregnant women presents additional and more complex issues since both mother and baby are affected. 13 Lower levels of education, lack of knowledge on vaccination during pregnancy and other socioeconomic factors like high parity and low income have been identified as barriers to maternal vaccination in low-and middle-income countries (LMICs); however data on maternal vaccine acceptance from African settings are very limited. 4,[14][15][16] TT is the most widely implemented vaccine among pregnant women in LMIC. 14 In 2012, however, the WHO recommended that countries considering initiation or expansion of seasonal influenza vaccination give the highest priority to pregnant women, 17 and in 2015 noted that maternal vaccination against pertussis is likely to be the most costeffective way to prevent the disease in infants too young for vaccination. 18 Moreover, new vaccines for use in pregnant women, for example, against Group B Streptococcus and respiratory syncytial virus, are under development and have great potential to reduce disease burden in resource-poor settings. [19][20][21][22][23] Thus, a better understanding of factors that may affect the acceptability of currently available and newly introduced maternal vaccines in LMIC, and particularly in Africa, is needed. The purpose of this study was to examine knowledge, drivers, and barriers of maternal vaccine acceptance in Kenya.

Materials and methods
This analysis was part of a larger study examining factors that shape the acceptance of maternal vaccines in Kenya. We conducted a cross-sectional survey of knowledge, attitudes, and beliefs regarding maternal vaccination among pregnant women. The survey was implemented in four counties in Kenya: Marsabit, Nairobi, Siaya, and Mombasa. Map showing the position of the counties of study implementation is shown in Figure 1.
Study site choice was guided by the geographic spread, cultural and religious diversity, and aiming for a mix of urban, sub/peri-urban, and rural settlements in Kenya. The study settings have relatively poor vaccination coverage and high burden of maternal and infant mortality. 24 A summary of key characteristics of the study sites is shown in Table 1.
Women were eligible for study participation if they were pregnant, aged 15-49 y old, were residents of the study counties, and able to provide informed consent. Potential participants were recruited from antenatal clinics at health facilities in all selected counties, and from the community in two counties (Nairobi and Siaya). We aimed to enroll 600 pregnant women across all sites, including 500 from health facilities and 100 from the community. The sample size was computed under the assumption of identifying the most conservative response proportion, 50%, for a given survey variable, using PASS v11 (NCSS, LLC; Kaysville UT), using the confidence interval for one proportion module. To estimate a response proportion of 50% with a 95% confidence interval of ± 5% (i.e., 95% confidence interval 45% to 55%) with 80% power, we would need a total of 402 completed surveys; we increased the target enrollment to 600 in order to conduct certain stratified analyses (for example, among women with a prior pregnancy) with the desired precision.
For facility-based recruitment, study staff enrolled a convenience sample of pregnant women presenting for antenatal care at the main public referral hospitals in each of the counties, as well as two private, nonprofit facilities: Tabitha Clinic (located in Kibera, an informal urban settlement in Nairobi) and St. Elizabeth Lwak Mission Hospital (located in rural Siaya County). For recruitment from the community in Nairobi and Siaya, study staff identified pregnant women registered for antenatal care at a participating facility who did not present for care during the recruitment period; these women were contacted by phone and invited to participate in the study. In addition, participants were encouraged to refer other pregnant women from their communities; interested pregnant women who contacted study staff and met enrollment criteria were invited to participate. Women recruited from the community were given the option to be interviewed at home or at a nearby health facility.
We pilot tested the study questionnaire at two facilities in Nairobi county prior to study implementation. Adjustments were made on the questionnaire based on feedback from the pilot test to ensure reliability. After obtaining written informed consent, the standardized questionnaire was administered. The questionnaire covered participant demographics and obstetric history, prior experience with vaccination (for themselves and their children), sources of vaccine information and recommendations during pregnancy, beliefs about vaccine protection and perceived benefits during pregnancy, and reasons for and against vaccination during pregnancy. Although Kenya does not currently have a maternal influenza vaccination program, the questionnaire also included questions about awareness of influenza disease (including perceived risk), attitudes/beliefs on influenza vaccination and willingness to receive influenza vaccine. The questionnaire was constructed primarily through measurement on threeand five-point Likert scales, responses ranged from agree to disagree and not relevant at all to very relevant or strongly disagree to strongly agree, respectively. The survey instrument was based upon previously used questionnaires shown to have high validity [25][26][27][28] and the compendium of survey questions developed by the WHO Strategic Advisory Group of Experts working group on vaccine hesitancy. 29 The survey questionnaire was translated into local languages (Swahili, Luo, Gikuyu, and Borana) and backtranslated into English to ensure accuracy of translations before administration. Interviews were conducted in private study offices or in households of study participants enrolled in the community. Data were collected electronically using tablets and transmitted real time onto KEMRI servers for management and storage.
Analysis was done using STATA version 13.0 (Stata Corp., College Station, TX). We described the frequencies of sociodemographic characteristics and responses to survey questions.
Ethical clearance for the study was obtained from KEMRI (SSC. 3292) and Emory University (IRB00089673) institutional review boards (IRBs), with CDC reliance on non-CDC IRB (CDC Protocol #6974.0). Informed consent was obtained from all participants before enrollment. Skilled provider includes doctor, nurse, or midwife. b Regional rates used, county rates not available c Includes mothers with two injections during the pregnancy of her last birth, or two or more injections (the last within 3 y of the last live birth), or three or more injections (the last within 5 y of the last birth), or four or more injections (the last within 10 y of the last live birth), or five or more injections at any time prior to the last birth.
Among all participants, 361 (59.8%) had received a recommendation for vaccination during the current pregnancy, and health-care providers (n = 253, 70.1%) were the most frequent source of vaccine recommendation (Table 3). Other common sources of recommendations for vaccination during pregnancy included relatives (n = 114, 31.6%), friends/ neighbors (n = 109, 30.2%), husbands (n = 102, 28.3%), and community health workers (n = 82, 22.7%). The vaccine most commonly recommended was TT (n = 277, 76.7%). The rest  Women could reply to more than one option for the type of complication, so total >82 e Denominator excludes mothers in their first pregnancy (n = 163) and those who were not sure of receiving vaccine in the past pregnancy (n = 8).
of the women did not know (n = 47, 13.0%), could not remember (n = 34, 9.4%) the specific vaccine recommended, or mentioned other products (n = 3, 0.8%) such as iron boosters that were not TT. Overall, 429 (71.0%) reported having received a vaccine in the current pregnancy, and among 433 women with one or more prior pregnancies, 401 (92.6%) reported having previously received a vaccine while pregnant. TT was the most common vaccine received during the current (n = 415, 96.7%) and prior (n = 347, 86.5%) pregnancies. Among those who received TT during the current pregnancy, 82 (19.8%) reported some type of complication; pain (n = 39, 47.6%) was most common. When asked about the number of vaccines they would be willing to receive during pregnancy, 101 (16.7%) participants replied no more than one, 71 (11.9%) no more than two, and 216 (35.8%) reported that they would have no limit and 140 (23.2%) were unsure. When asked about the top three reasons why a woman would receive a vaccine during pregnancy, disease prevention was the most frequent response (first priority for 53.2%, second priority for 26.5%, and third priority for 9.3%), followed by the benefit to the baby in the womb (first priority for 18.9%, second priority for 33.9%, and third priority for 25.8%) and to follow health-care provider recommendations (first priority for 17.2%, second priority for 15.4%, and third priority for 20.5%) ( Table 4). When asked about the top three reasons why a woman would refuse a vaccine during pregnancy, about half answered that they would never refuse vaccination against tetanus during pregnancy (first priority for 51.0%, and second priority for 0.5%). The most frequent reason for refusal when given was if the vaccine caused side effects to a mother or baby (first priority for 15.1%, second priority for 6.1%, and third priority for 1.5%). When asked about whose benefit should be prioritized in deciding whether a pregnant woman should be vaccinated, the baby in the womb was the most common first priority (first priority for 58.8%, second priority for 33.6%, and third priority for 7.6%), followed by the mother as the most common second priority (first priority for 34.1%, second priority for 41.9%, and third priority for 24.0%).
Nearly all the expectant mothers agreed with the statement that vaccines given in pregnancy are "important for my health" (n = 599, 99.2%) and that getting vaccinated is "a good way to protect myself from disease" (n = 596, 98.7%) (Figure 2). More than 90% of the mothers believed that maternal vaccines are effective and all vaccines offered by the government are beneficial. Overall, 36.8% expressed concern about serious adverse effects of vaccines and 27.0% agreed with the statement that new vaccines carry more risks than older ones.
Slightly more than a half (53.5%) of participants agreed that a pregnant woman should be vaccinated if only the baby (not the mother) is protected and 51.3% agreed that a pregnant woman should be vaccinated if it primarily protects the community; 38.1% agreed that an experimental vaccine that could protect the mother or baby should be used on pregnant mothers (Figure 3).
A small proportion of participants (n = 26, 4.3%) reported ever having refused a vaccine for either herself or a child in the past ( Table 5). The most common reasons cited for refusal were not enough information (n = 11, 42.3%) and safety/side effects concerns (n = 9, 34.6%). Rumors about vaccine safety were mentioned by 3 (11.5%) mothers. A total of 26 (4.3%) mothers reported having a negative experience with a vaccine in the past, most of these (61.5%) were swelling at the injection site. Four (15.4%) women who reported a negative experience also reported ever refusing a vaccine in the past. When asked about influenza vaccine (not routinely available in Kenya), 402 (66.6%) agreed that a pregnant woman should be vaccinated, and 470 (77.8%) reported that if given the option, they would accept influenza vaccination (Table 6).

Discussion
We found a high level of acceptance for maternal vaccines among pregnant women in Kenya. Most participants reported having received a vaccine during the current and/or prior pregnancy, and nearly all reported favorable views on the importance and effectiveness of vaccination during pregnancy. The main driver for vaccine acceptance in pregnancy was disease prevention, with protection of the baby in the womb as the highest priority. Few participants (<5%) reported having ever refused a vaccine for either themselves or a child, and the most commonly reported potential barrier to maternal vaccine acceptance was concern about side effects. Although influenza vaccine is not routinely administered to pregnant women in Kenya, nearly 80% of participants reported that they would accept it if offered. The findings demonstrate an important willingness on the part of pregnant women in Kenya to accept both currently available and potential future maternal vaccines. 26    The pregnant women enrolled in this study demonstrated an understanding of the role of vaccine in disease prevention. Most participants agreed that vaccines given in pregnancy are important for their health and that getting vaccines was a good way to protect themselves from disease. Knowledge about vaccines forms a critical component of vaccine literacy, which is the degree to which a person has the capacity to obtain, process, and understand basic vaccine information and services to help them make appropriate health decisions. 30 For maternal vaccines, knowledge of vaccines among pregnant women has been shown to be an important driver of uptake in various settings. 16 In this study, reported refusal of vaccines was very uncommon; however, among the small number (n = 26) that had refused a vaccine, lack of sufficient information was the most common reason for refusal provided. Thus, even in a setting where pregnant women are relatively knowledgeable about vaccines, it is important to ensure that adequate and appropriate information is shared about maternal vaccines. This might be particularly important for the introduction of maternal vaccines.
The primary factors found to facilitate vaccination during pregnancy were disease prevention, vaccine benefit to the baby in the womb, and recommendation by a health-care provider. An understanding of the role of vaccines in disease prevention is critical for vaccine acceptance. Studies in Nigeria 31 and Ethiopia 32 reported that women who had received TT during pregnancy most commonly cited protection against tetanus as the reason for vaccination. Recommendation by a health-care provider has very consistently been shown to impact pregnant women's decision to be vaccinated 4,33-37 although data from other African settings are limited. Nearly 60% of women in this study had received a recommendation for vaccination in their current pregnancy, and health-care providers were the most frequent source of vaccine recommendation. Of note, relatively few women reported having received a recommendation for vaccination from a community health worker, despite efforts by the Kenyan government to have community health workers provide outreach and communication on maternal and neonatal health services, including vaccination. 38,39 Community health workers represent an underutilized resource for promoting maternal vaccination, particularly in areas with limited numbers of health-care providers.
The main potential barrier to maternal vaccination that emerged was concerns about vaccine safety. More than onethird of participants were concerned about serious adverse effects of vaccines. Side effects were also cited as the leading reason why a pregnant woman might refuse vaccination. Concerns about the safety of vaccines in pregnancy is a well-recognized barrier to the implementation of maternal vaccines globally. 13,40 The safety of TT in Kenya was aggressively questioned in 2014 by the Kenya Conference of Catholic Bishops, who alleged that it was being used to sterilize women. 41 Nonetheless, a high proportion of women in this study reported having received TT during a prior and/or current pregnancy. Thus, our findings suggest that the allegations did not substantially affect the acceptance of TT among pregnant women in Kenya. Nonetheless, in designing and implementing maternal vaccination programs, it is important to recognize that vaccine adverse effects are a primary concern among pregnant women.
New maternal vaccines are under development (e.g. against Group B Streptococcus and respiratory syncytial virus) [19][20][21][22][23] and efforts are underway to increase the use of currently available but underutilized vaccines for pregnant women in LMIC (e.g. influenza and pertussis vaccines 17,18 ). Most pregnant women in this study reported a willingness to receive influenza vaccine, which is not currently in routine use in Kenya, but was recommended by the Kenya National Immunization Technical Advisory Group in 2016, pending availability of local data on burden of influenza in pregnancy. 42 Contrarily, 27.0% felt that new vaccines are riskier than old vaccines, and 50.8% reported that "experimental" vaccines should not be used on pregnant women. Also, 16.7% of participants reported they would only accept one vaccine during pregnancy, and 11.9% would accept up to two vaccines. Planning Women could reply to more than one option so total >26. for introduction of new maternal vaccines should take into account perceptions of risk; combined antigen vaccines (i.e. fewer number of doses administer) may facilitate greater acceptance.
The study had several limitations. First, we used convenience sampling technique to reach study participants and had 16% of the women recruited through referrals which presents potential bias for similar knowledge, attitudes, and beliefs. Also, we recruited participants in only four counties and 48.8% of women enrolled were of Luo ethnicity. Nationally, the Kenyan population is made up of 42 tribes, and Luos represents 11% of the population. 43 Although we aimed to enroll 100 women per site from health facilities, low numbers of patients in clinics at some sites (Mombasa and Marsabit), resulted in variability in numbers enrolled per site and per facility. Thus, our findings may not be generalizable to all of Kenya. Some questions involved health practices in past pregnancies and recall may have been inaccurate. Furthermore, participants may have given socially desirable responses, especially for questions regarding TT vaccination, which is recommended in Kenya. The data we captured on receipt of TT are not directly comparable to government statistics on the coverage of two or more TT doses among pregnant women in Kenya, and this study was not intended to assess vaccine coverage. However, it is notable that we observed high levels of vaccine acceptance when compared with the low TT coverage reported among pregnant women in Kenya; 9 this contrast may be due to the limitations in the generalizability of our findings or the potential for socially desirable responses. Because the only maternal vaccine accessible to all pregnant women was TT, we could not reliably assess attitudes and behaviors regarding other maternal vaccines.

Conclusion
This study adds to the body of data on factors which may facilitate or impede vaccine uptake among pregnant women in Africa and other LMICs, where such data remain sparse. Nationally, Kenya has made tremendous effort toward achieving the target coverage of 80% for TT vaccination in pregnancy. Opportunities for further improving uptake of maternal vaccines lie in providing adequate information about vaccines for pregnant women and mitigating safety concerns. Health-care providers are the main source of vaccine recommendation for pregnant women and must be encouraged to consistently advise antenatal patients about vaccination. Community health workers should be better leveraged to reach pregnant women with accurate messages about vaccination. Our findings suggest that new maternal vaccines introduced in Kenya, such as the influenza vaccine, will likely be well accepted by pregnant women, as long as information on benefits and risk is communicated widely and comprehensively.