Maternal Vaccination Perceptions, Attitudes and Practices Among Urban Women Receiving Antenatal Care in Panama City: A Cross-Sectional Study

1. Introduction

Vaccination during pregnancy protects two individuals at the same time. It protects the pregnant woman from infections to which she is particularly vulnerable because of the physiological and immunological changes of pregnancy, and, through transplacental transfer of IgG and immunoglobulin A in breast milk, it gives the infant passive protection during the first months of life, when direct immunization is not yet possible and susceptibility to severe disease is highest [1,2].

The evidence accumulated to date broadly supports these strategies. Influenza vaccination during pregnancy reduces the risk of influenza-associated maternal hospitalization by approximately 40% [3] and protects the infant up to six months of age; pertussis vaccination, given as part of a combined preparation, prevents a potentially fatal disease in the newborn; and maternal vaccination against SARS-CoV-2 reduces severe disease and adverse obstetric events [1]. More recently, the prefusion F protein vaccine against respiratory syncytial virus (RSV), recommended between weeks 32 and 36 of gestation, has shown 81.8% efficacy against medically attended severe RSV disease during the infant's first 90 days of life [4,5].

In Panama, maternal immunization forms part of the public health policy of the EPI and covers influenza, Tdap, COVID-19 and, more recently, RSV [5]. Despite free access and both national and international recommendations, coverage remains suboptimal [6]. Self-reported data from Panama, Peru and Colombia point to a concerning gap: 73% to 80% of women reported having received the influenza vaccine, but only 30% to 39% had received Tdap, well below official targets [7]. Compared with other populations, the decision to be vaccinated during pregnancy appears to be strongly mediated by emotional, family and cultural factors, and by trust in the health system [6]. The few qualitative studies available in urban Panamanian women describe a generally favorable attitude alongside persistent concerns about adverse effects, difficulties in access, and a lack of clear and timely information [8].

Against this background, local quantitative evidence can describe vaccine-related perceptions among women attending care while clarifying how sample composition may shape estimates of acceptance. The contrast between suboptimal national coverage and high acceptance in highly connected urban women raises an important question: whether acceptance estimates represent broader population confidence or unequal access to care and information. Such evidence may inform equitable access to counseling, recommendations and vaccination services.

1.1. Objectives

The general objective was to describe perceptions, attitudes and practices related to vaccination during pregnancy among pregnant women aged 18 years or older accessing prenatal care at participating public and private health facilities and vaccination centers in Panama City, and to explore factors associated with vaccine acceptance in this selected sample. The specific objectives were to quantify acceptance of the influenza, Tdap, COVID-19 and RSV vaccines; to identify sociodemographic, clinical and psychosocial factors associated with acceptance; to describe the main sources of information and their influence on the decision; to explore willingness to receive several vaccines at a single visit; and to identify perceived barriers to maternal vaccination.

2. Materials and Methods

2.1. Study Design and Population

We conducted an observational, descriptive, cross-sectional study. The study population comprised pregnant women aged 18 years or older, in any trimester of gestation, attending public or private health facilities or vaccination centers in Panama City and its metropolitan area. Women who agreed to take part voluntarily, gave verbal informed consent and understood Spanish were included. Women younger than 18 years, those with cognitive or language barriers that prevented completion of the survey, and those with severe complications that precluded participation were excluded.

Sampling was by convenience from January to July 2026 and sought diversity across participating public and private facilities and vaccination centers. Recruitment source, number approached and refusals were not recorded; a site-specific recruitment flow could therefore not be reconstructed. The final analytic sample comprised 204 participants.

2.2. Instrument and Variables

A structured, expert-validated, self-administered and anonymous survey was administered. It comprised closed Likert-type, multiple-choice and dichotomous questions and took approximately 10 minutes to complete. The primary outcome was vaccine acceptance, operationally defined as having received at least one vaccine during the current pregnancy, having received all of them, or planning to receive at least one. Perceptions and knowledge about vaccination were measured with Likert scales from 1 to 10, averaging three items in each case to obtain a summary score. We also recorded sociodemographic variables (age, educational attainment, marital status), clinical and obstetric variables (comorbidities, parity, gestational age, antenatal care), the vaccination recommendation and its source, willingness to receive several vaccines at one visit, vaccines received and planned by type, refusal of specific vaccines and the reasons given, and sources of information and trust.

The instrument asked separately whether the participant had received a recommendation to be vaccinated and who had issued it. During data cleaning we identified internal inconsistencies between the two questions: six participants who reported not having received a recommendation went on to name their obstetrician/gynecologist or another healthcare worker, and six who did report receiving one did not identify a healthcare professional as its source. The variable analyzed was therefore having received a recommendation to be vaccinated, irrespective of who issued it.

2.3. Statistical Analysis

Qualitative variables were summarized as absolute and relative frequencies, and quantitative variables as measures of central tendency and dispersion. For descriptive purposes, perception and knowledge scores are presented as mean ± standard deviation (SD) to aid interpretation. Because both variables showed a marked ceiling effect, however, they were treated as non-normally distributed in the inferential analyses and compared using non-parametric methods.

Associations with vaccine acceptance were first explored in bivariate analyses. Categorical variables were compared using Fisher's exact test. Age was the only continuous variable that approximated a normal distribution and was therefore compared using Welch's t-test; all other continuous variables, including the perception and knowledge scores, were compared using the Mann–Whitney U test.

Variables associated with vaccine acceptance were then evaluated in Poisson regression models with a log link and robust variance, which estimate crude and adjusted prevalence ratios (PRs) with 95% confidence intervals (CIs) and are appropriate when the outcome is common, as vaccine acceptance was in this sample [9]. The multivariable model was adjusted for age. Additional covariates were selected from variables with p < 0.20 in the bivariate analysis, taking their epidemiological relevance into account. Collinearity among candidate covariates was assessed using the variance inflation factor (VIF). The variables “recommendation to receive maternal vaccination” and “source of the recommendation” showed evidence of collinearity; only the former was therefore retained, as it was considered the more representative. Given the small number of participants who did not accept vaccination, the final model was kept parsimonious.

Statistical significance was defined as a two-sided p value < 0.05. All analyses were performed in Python 3 (pandas, SciPy and statsmodels), with coding assistance from a generative artificial intelligence tool used under author supervision and verification (see Section 2.5).

2.4. Ethical Considerations

The protocol was reviewed and approved by the Research Bioethics Committee of The Panama Clinic (code EC-CBITPC-237; reference RESEGIS #4975). The committee approved the use of verbal, unsigned informed consent given the anonymous, minimal-risk nature of the study: before the survey was administered, each participant was read the information on objectives, voluntary participation, confidentiality and the right to withdraw, and her verbal agreement was recorded. Participation was voluntary, anonymous and confidential; no identifiable data were collected, no signature that could link a participant to her responses was requested, and no financial incentive was offered. When risk conditions were identified during the survey, the participant was advised to seek medical care, without this affecting her participation.

2.5. Use of Generative Artificial Intelligence

Generative artificial intelligence tools based on large language models (Claude, Anthropic, San Francisco, CA, USA) were used at two stages of this work. First, in data processing and statistical analysis: the AI assistant supported the cleaning of the survey database and the writing and execution of the Python 3 code (pandas, SciPy, statsmodels) used to produce the descriptive statistics, the bivariate comparisons and the Poisson regression models reported here. Second, in manuscript preparation: drafting, wording, structuring and language editing, including translation of the original Spanish draft into English.

These tools did not design the study and did not collect, generate, fabricate or alter any primary data; the only dataset analyzed is the one obtained from the survey administered by the authors. All AI-assisted output was verified by the authors under a division of responsibilities: J.R.S.G. verified the database and the data-cleaning process, checking the derived variables and the coding decisions against the original survey responses, whereas A.E.D.Á.M. verified the statistical component, reviewing the analysis code, the choice and assumptions of the tests and regression models, and the estimates reported in the text and tables. Every AI-assisted passage of text, as well as all tables and references, was likewise reviewed and corrected by the authors. The authors take full responsibility for the integrity, accuracy and interpretation of the content of this article.

3. Results

3.1. Participant Characteristics

A total of 204 pregnant women were analyzed. Mean age was 29.9 years (SD 5.0; range 18–43), and the 25–34 year age group predominated (71.6%). Educational attainment was high: 88.2% had higher education (73.0% university and 15.2% a master's or doctoral degree). Most had a stable partner (91.2%) and were in their first pregnancy (76.0%). Gestational age was left-skewed and is summarized as median (IQR): 28 weeks (IQR 21.8–33.0); most women were in the third (54.4%) or second trimester (41.2%). All participants reported antenatal care, mainly in the private sector (77.9%). These features describe a selected urban group, not pregnant women in general. One in five reported a comorbidity (21.1%), most frequently overweight (6.9%) and hypertension (6.4%) (Table 1).

3.2. Vaccine Acceptance, Perceptions and Knowledge

Overall vaccine acceptance was 93.1% (190/204). This high estimate reflects self-reported receipt or intention within a selected sample fully engaged in antenatal care and must not be interpreted as population coverage. Perceptions and knowledge were favorable: the mean perception score was 8.75/10 (SD 1.41) and the mean knowledge score 8.61/10 (SD 1.59), with particularly high values for the belief that vaccines protect the mother (9.13/10) and the baby at birth (9.02/10). Both scores showed a clear ceiling effect, with medians of 9.3 and 9.0 out of 10, respectively (Table 2).

A recommendation to be vaccinated was reported by 84.3% of participants (172/204). In 166 of those 172 cases (96.5%) the recommendation came from a healthcare professional, most often the obstetrician/gynecologist (144/204; 70.6% of the sample and 82.0% of those who received any recommendation). Six participants who reported having received a recommendation did not identify a healthcare professional as its source and, conversely, six who reported not having received one did name a professional; because of this inconsistency, the recommendation was analyzed irrespective of its source. The vaccines most frequently recommended were influenza (74.0%) and Tdap (68.6%), followed by RSV (50.0%). As for willingness to receive several vaccines at the same visit, the median number of vaccines accepted was 2 (IQR 2–3); 48.0% would accept two and 27.5% three or four simultaneously.

3.3. Vaccines Received, Planned and Refused

Influenza was the vaccine most frequently received (56.9%), followed by Tdap (40.7%) and RSV (24.5%); the COVID-19 vaccine was the least frequently administered (5.4%). Among the vaccines that participants planned to receive, RSV led the intention (51.5%), followed by Tdap (44.6%) and influenza (38.7%). Most expected to be vaccinated in the public system (79.4%). Almost half of the sample (49.5%) reported a vaccine they did not want to receive, and in virtually all of these cases it was the COVID-19 vaccine (47.1% of the total), far above refusal of influenza (5.4%), Tdap (3.9%) or RSV (3.4%) (Table 3).

Among participants who refused a specific vaccine, the most frequent reason was the perception that vaccines are not safe (22.5% of all participants), followed by not knowing that they should be vaccinated (12.7%), the belief that vaccines do not protect (4.9%), and access barriers such as attending a facility and finding no vaccine available or no one to provide care (4.4%).

3.4. Sources of Information and Trust

Healthcare professionals were by far the main source of information when participants had questions (92.6%) and the most trusted source when deciding about vaccines in pregnancy (98.0%). Artificial intelligence emerged as a relevant source, consulted by almost one third of participants (31.4%), although few identified it as their most trusted source (5.9%). Social media and the circle of family and friends played a marginal role, both as a source of information and in terms of trust (Table 4).

3.5. Factors Associated with Vaccine Acceptance

In the bivariate analysis, none of the maternal sociodemographic or obstetric characteristics (age, educational attainment, marital status, parity, gestational age, private antenatal care setting or comorbidities) was significantly associated with vaccine acceptance; restricted variability in this selected sample may have contributed. Perception and knowledge scores did not reach significance, although favorable trends were observed for perception (p = 0.112) and willingness to receive several vaccines (p = 0.074). The unadjusted factors associated with acceptance were having received a recommendation to be vaccinated (86.8% vs. 50.0%; p = 0.002), having received it from the obstetrician/gynecologist (72.6% vs. 42.9%; p = 0.030), and trusting healthcare professionals (98.9% vs. 85.7%; p = 0.024) (Table 5).

After adjustment for age, having received a recommendation to be vaccinated remained associated with a 22% higher prevalence of acceptance than not having received one (adjusted PR 1.22; 95% CI 1.02–1.45; p = 0.029). Given the cross-sectional design and selection into specialized care, this association cannot identify an isolated causal effect and may mark continuity and quality of care, specialist access and health literacy (Table 6).

4. Discussion

This study describes perceptions, attitudes and vaccination practices in a convenience sample of 204 urban pregnant women actively engaged in prenatal care. It found high reported acceptance (93.1%), favorable perceptions and knowledge, and near-universal trust in healthcare professionals. However, the high education (88.2%), private-care predominance (77.9%) and universal antenatal care make this sample systematically different from women facing gaps in access. Accordingly, its principal contribution is not evidence that maternal vaccine acceptance is high in Panama, but the possibility that favorable acceptance is concentrated among women with better access to quality prenatal care, information and counseling.

Healthcare-provider recommendation is associated with maternal vaccination [2]. In Panama, McDermid et al. reported 78.2% received a healthcare-professional recommendation for pertussis vaccination, yet only 30.5% received Tdap [7]. This discordance suggests that recommendation is a marker of integration into a wider care pathway—specialist access, care continuity and quality, health literacy, vaccination access and reminders—rather than an isolated causal determinant. Our result supports repeated counseling but cannot isolate causality. The obstetrician/gynecologist issued roughly eight in ten recommendations, identifying a counseling actor, not an independent predictor.

The near-universal trust placed in healthcare professionals (98.0%) contrasts with the marginal role of social media and the immediate social circle. This is relevant in light of previous qualitative studies of urban Panamanian women, in which safety concerns, difficulties in access and a need for clearer, timely information persisted despite a broadly favorable attitude [8]. In our sample, concerns converged on the COVID-19 vaccine, which was the least frequently received (5.4%) and by far the most frequently refused (47.1%). The main reason for refusal was perceived lack of safety, suggesting focused, vaccine-specific hesitancy rather than generalized rejection of immunization.

Patterns varied across vaccines, cautioning against equating stated acceptance with coverage [10]. Influenza was most frequently received, whereas Tdap uptake (40.7%) and RSV intention (51.5%) were at different stages of adoption. A 2024 PAHO behavioral and social drivers study of 452 pregnant women in Argentina likewise reported high influenza intention (77.2%) but limited current COVID-19 (6.6%) and RSV (5.5%) coverage, with provider recommendation and practical access barriers differing by vaccine [11]. Although these regional data cannot be generalized to Panama, they illustrate that intent and coverage can diverge and that vaccine-specific perceptions, safety concerns, public discourse and implementation maturity may shape behavior beyond recommendation alone.

From a health-systems perspective, national maternal vaccination gaps may reflect not only vaccine hesitancy, but also differences in access to antenatal care, vaccination, reliable information and reminder systems, factors identified in Panama by McDermid et al. [7]. Equitable counseling and vaccine delivery are especially needed in public-sector and underserved settings. Standardized counseling across care settings may extend the benefits observed here. RSV introduction should include implementation and access strategies, because stated intention is not actual coverage.

4.1. Limitations

These results require cautious interpretation. Convenience recruitment through health-related settings, universal antenatal care, high education and predominantly private care create substantial selection bias; these features probably influenced both acceptance estimates and the associations with recommendation and trust in health professionals. The sample cannot represent national or general urban obstetric populations, and recruitment sources, numbers approached and refusals were not recorded, preventing formal assessment of participation. Acceptance may be lower and its correlates different in public-sector, rural, less educated or underserved groups. Because only 14 participants did not accept vaccination, estimates were imprecise and the multivariable model had to remain parsimonious; non-significant associations indicate absence of evidence, not evidence of absence. Ceiling effects reduced the discriminative capacity of perception and knowledge scores. Inconsistencies between recommendation reports and source identification affected twelve participants and limited interpretation by professional type. Finally, the cross-sectional design and self-reported data preclude causal inference and may introduce recall and social desirability bias.

5. Conclusions

In this selected urban sample, pregnant women engaged in antenatal care showed high reported vaccine acceptance. Recommendation was associated with acceptance, but may represent broader access to continuous, high-quality care and counseling; this cross-sectional study cannot determine an isolated causal effect. Concerns were concentrated on the COVID-19 vaccine.

Maternal vaccine policies should combine active, systematic and clear recommendations with equitable access to quality antenatal care, consistent evidence-based counseling and vaccination opportunities across public, private and underserved settings. RSV introduction should be evaluated against actual coverage rather than stated intention. Probabilistic population-based studies with adequate representation of public-sector and non-urban populations are needed to confirm these patterns and quantify coverage gaps.