Factors Influencing RTS,S/AS01 Vaccine Completion and Construction of a Communication Exposure Score in Allada-Toffo-Ze Health Zone, Benin, 2025 ()
1. Introduction
Malaria remains the leading cause of mortality among children under five years of age in Benin, with an annual incidence of 16% in 2023 and parasite prevalence of 25% in the Atlantique department [1] [2]. Despite the scale-up of long-lasting insecticidal nets (LLINs), seasonal malaria chemoprevention and intermittent preventive treatment, malaria still accounts for 40% of outpatient consultations and 25% of hospital deaths [1] [2]. RTS,S/AS01 vaccine, recommended by WHO since October 2021 for areas of moderate to high Plasmodium falciparum transmission [3] [4], was introduced in the pilot phase in Allada-Toffo-Ze health zone (Atlantic Department) in February-March 2024 following the national plan of ANSSP [5]. The schedule follows 6, 7, 9, and 18 months (VAP1 - VAP4) [5]. Phase 3 efficacy was 36.3% against clinical malaria and 28.3% against severe malaria [6] [7]. The WHO position paper of March 2022 emphasizes integration with communication strategies [8] [9].
After eleven months of implementation, routine EPI data extracted from DHIS2 and immunization registers showed a marked decline in coverage across the vaccination schedule. Among children aged 6 months, VAP1 coverage was 55.7% (142/255). VAP2 coverage was 37.3% (95/255). Crude VAP3 coverage was 17.3% (44/255) [95% CI: 13.0 - 22.5], while coverage among children eligible for VAP3 (aged ≥10 months) was 22.7% (42/185) [95% CI: 17.1 - 29.4]. Seventy children (27.5%) were aged <10 months and therefore ineligible for VAP3; among eligible children (n = 185, 72.5%), early vaccination was observed in 2/70 (2.9%). VAP4 coverage was 0% among the 49 children who reached 18 months in our cohort. The dropout rate from VAP1 to VAP3, calculated as (VAP1 − VAP3)/VAP1 × 100, was 69.0% crude (98/142) and 59.6% among eligible children (62/104), which is far above the WHO acceptable threshold of 10%.
Evidence from the Malaria Vaccine Implementation Program (MVIP) in Ghana, Kenya, and Malawi emphasizes the central role of multi-channel communication in improving acceptability and retention [8] [10]. In Ghana, Asante et al. showed that communities benefiting from engagement through community relays, gong-gong, and durbars had higher uptake [11]. In Kenya, Otieno et al. reported that a combination of radio and community health worker exposure was associated with an adjusted odds ratio of 2.1 [95% CI: 1.4 - 3.2] for VAP3 completion [12]. In Malawi, Hamel et al. reported that mothers exposed to two or more channels had 70% completion compared to 45% for those exposed to a single channel [10]. Theoretical frameworks such as Rogers’ Diffusion of Innovations and WHO guidance on dropout monitoring have been used to interpret these findings [13] [14]. However, most studies measured exposure as a binary variable and did not construct a composite exposure score.
In Benin, the national introduction plan includes communication via health workers during antenatal care and EPI sessions, community relays during home visits, radio spots in Fon and French broadcast on three community radio stations, television spots on national television, and posters displayed in health facilities [5] [9]. However, evaluation has remained descriptive, focusing on the number of posters distributed and broadcasts aired without measuring population exposure and its effect on coverage [9] [15]. Previous qualitative work by Allada-Toffo-Ze identified language barriers (Fon versus French), trust in community relays, fear of adverse events following immunization (AEFI), and misinformation as key determinants of vaccination hesitancy [13] [15].
Therefore, the objective of this study was twofold: (1) to develop and validate a composite communication exposure score conceived as a formative index adapted to the Beninese context and to the RTS,S vaccine, with descriptive psychometric properties including Cronbach’s alpha of 0.675 for five items and 0.649 for three items (reported for descriptive purposes only and not as a validity criterion), principal component analysis showing 46.9% of variance explained by the first axis for five items and 58.9% for three items, and an apparent AUC of 0.82 [95% CI: 0.75 - 0.89] pending external validation without bootstrapping; and (2) to analyze its independent association with vaccine completion up to the third dose (VAP3) among children eligible aged ≥10 months and with dropout from VAP1 to VAP3, which was 59.6% among eligible children versus 69.0% crude in Allada-Toffo-Ze.
2. Methods
2.1. Study Design, Period and Setting
A cross-sectional analytical study with a mixed-methods convergent design was conducted from February 23 to March 03, 2025, in the Allada-Toffo-Ze health zone, Atlantique department, Benin [15]-[20]. The zone comprises three communes: Allada (~180,000), Toffo (~130,000), and Ze (~118,000), for a total of ~428,000 in 2025 projection, with high transmission with entomological inoculation rate >200 infective bites/person/year [1] [15]. Health system: 1 zone hospital, 13 public health centers, 25 private facilities, 180 community relays, 20 qualified community health workers (ASCQ). EPI performance pre-RTS,S: BCG 98%, PENTA3 92%, Measles 88% in 2022 [2] [19]. Quantitative and qualitative components were conducted simultaneously and integrated during interpretation as recommended for implementation research and MVIP evaluations [8] [10] [20]-[22] following STROBE [17] and COREQ [18].
2.2. Population and Sampling
The target population consisted of mothers or primary caregivers of children aged 6 to 24 months who had resided in the study area for at least six months. The exclusion criteria were refusal to participate, inability to answer the questionnaire, and children outside the eligible age range. The sample size was calculated using the WHO cluster survey formula for vaccination coverage: n = number of clusters × households per cluster × design effect (DEFF). The parameters were set at 30 clusters, which is standard for a zone-level estimate, seven households per cluster as recommended by the WHO manual, and a DEFF of 1.2 for intracluster correlation of communication exposure based on previous surveys in Benin [16] [19]. The calculation yielded 30 × 7 × 1.2 = 252 households. Operationally, 255 households were surveyed across 30 clusters with 7 to 9 households per cluster (mean 8.5). This sample provided 80% power to detect an odds ratio of ≥ 2.0 for high versus low exposure, accounting for clustering.
The qualitative component included 198 stakeholders purposively selected using a maximum variation sampling framework: 36 pupils (18.2%) from six primary schools, 31 religious leaders from mosques, churches, and traditional religions, 31 village chiefs and town criers selected by communal authorities, 30 community relays selected from 180 relays, 27 teachers from 13 schools, 20 qualified community health workers (ASCQ), 19 EPI managers and health workers, two radio promoters, one mayor, and one other stakeholder. Recruitment was conducted through community entry meetings and snowball sampling of religious leaders. Individual interviews lasted approximately 45 minutes and were conducted in French for the health workers, managers, and the mayor. Five focus group discussions (FGDs) were conducted with pupils (six participants per group, 30 minutes each), four FGDs with religious leaders, and additional FGDs with community relays and mothers were conducted.
2.3. Variables and Score Construction
The dependent variable was RTS,S completion up to the third dose (VAP3), defined as the receipt of three doses documented either by (i) vaccination card, considered the gold standard (68.2% of cases), or (ii) maternal recall with probing for the date and place of vaccination according to the WHO 2018 manual to reduce recall bias (31.8%) [16]. The definition combining card and recall aligns with the primary outcome of the MVIP [8] [10] [14]. Eligibility was restricted according to the WHO schedule: VAP3 was restricted to children aged ≥10 months (9 months plus one month grace period per WHO cluster manual), VAP1 to ≥6 months plus two weeks, VAP2 to ≥7.5 months, and VAP4 to ≥19 months [14] [16]. In the database, 70 children were aged <10 months, among whom VAP3 was documented in 2/70 (2.9%) as early vaccination, and 185 children were aged ≥10 months and were eligible for VAP3.
The independent variables included a five-channel exposure score ranging from 0 to 5, calculated as the sum of binary exposure to five channels listed in the ANSSP plan [5] [9] [15]: health worker at the facility (72.5% [95% CI: 66.7 - 77.7], n = 185), community relay at home (70.6% [95% CI: 64.7 - 75.9], n = 180), radio (39.6% [95% CI: 33.7 - 45.8], n = 101), television (12.9% [95% CI: 9.2 - 17.7], n = 33), and poster (16.9% [95% CI: 12.6 - 22.0], n = 43). Each channel was coded as 0 = no exposure and 1 = yes if the mother reported hearing or seeing an RTS,S message in the past six months. Categories were defined as low (0 - 1), medium (2), and high (3 - 5) based on tertiles and programmatic relevance, with ≥3 channels considered as multi-channel exposure [9]. The score distribution was: 0: 43 (16.9%), 1: 38 (14.9%), 2: 76 (29.8%), 3: 50 (19.6%), 4: 38 (14.9%), and 5: 10 (3.9%).
The communication exposure score was conceptualized as a formative construct. Unlike a reflective psychometric scale where items reflect a latent construct, in a formative index, the items cause the construct; therefore, internal consistency measured by Cronbach’s alpha and principal component analysis are descriptive only and not validity criteria [21] [23] [24]. For descriptive purposes, Cronbach’s alpha was 0.675 for five items and 0.649 for three items (health worker, relay, radio). Principal component analysis showed that the first axis explained 46.9% of variance for five items and 58.9% for three items, with the second axis explaining 22% and 25%, respectively. The apparent discriminative ability within the sample, measured by the area under the receiver operating characteristic curve, was 0.82 [95% CI: 0.75 - 0.89], which should be interpreted as good but not exceptional, pending external validation without bootstrapping [23] [24].
2.4. Data Collection and Analysis
Quantitative data were collected using a structured questionnaire adapted from the WHO Vaccination Coverage Cluster Survey Reference Manual [16] and the Malaria Vaccine Implementation Programme (MVIP) questionnaires used in Ghana [11], Kenya [12], and Malawi [10]. The questionnaire was translated into Fon and back-translated into French, and then pre-tested in 20 households outside the study zone. Qualitative data were collected using semi-structured interview guides exploring information sources, trust, language, rumors, and barriers [9] [13] [18] [25] [26].
Data collection was conducted by eight trained interviewers and two supervisors over ten days using ODK Collect on tablets with GPS, with daily quality checks.
Data were exported from the ODK server to Stata 18 for data management. Data cleaning included range and logical checks. Missing data were <2% and were handled by complete-case analysis.
Descriptive statistics were presented with 95% confidence intervals (CI) adjusted for the cluster design using svyset (primary sampling unit [PSU] = cluster, strata = commune) as recommended by the World Health Organization [15] [16]. The sample was self-weighting due to probability proportional to size (PPS) sampling based on the 2023 RGPH-4 census; therefore, no survey weights were applied. Bivariate associations were assessed using chi-squared tests and crude odds ratios (OR) with 95% CI accounting for clustering via svy: logit. Variables with p < 0.20 were considered for exploratory purposes only. The final multivariable model was selected a priori using a directed acyclic graph (DAG) and included language, commune as a proxy for access and vaccine availability (stock-out in Ze in August 2024), maternal age, and child age per month. The exposure score was analyzed both as a continuous variable and as a categorical variable. The interaction between score and language was tested using a likelihood ratio (LR) test (p = 0.42, not significant). Model fit was assessed using the Hosmer-Lemeshow test (p = 0.68, indicating good calibration) [23], Nagelkerke R2 (~0.10, indicating low explained variance), apparent area under the curve (AUC) 0.82 [0.75 - 0.89] [24], and variance inflation factor (VIF) < 5. Sensitivity analyses included card-confirmed vaccination only (n = 174), the eligible-restricted 3-channel score, and Firth penalized logistic regression for sparse data [22].
For the qualitative component, the data were coded using NVivo 14 with double-coding. Thematic analysis was conducted using inductive and deductive approaches based on Rogers’ Diffusion of Innovations [13], Wakefield [25] [26], Noar [26], and the WHO communication guidance [9]. The COREQ checklist was used for this study [18]. A triangulation matrix was used to examine convergence and divergence with quantitative results [20] [21].
3. Results
3.1. Sociodemographic Characteristics
The mean age of the mothers was 27.4 ± 6.8 years (range: 16 - 45 years). Regarding education, 38.0% had no formal education (n = 97), 34.5% had primary education (n = 88), and 16.1% had secondary education (n = 41). In terms of occupation, 73.7% were housewives or farmers, 18.4% were traders, and 7.9% had other occupations. Fon was spoken at home by 71.8% of households overall and by 84.7% among those who had heard about the vaccine (183/216). Other languages were Aïzo (15.3%) and other languages (12.9%), reflecting the composition of the Atlantique department [2] [19]. By commune, 19.6% of participants resided in Allada (n = 50), 36.9% in Toffo (n = 94), and 43.5% in Ze (n = 111). The mean age of children was 13.12 ± 4.82 months (median: 12 months, interquartile range: 9 - 16.5 months). Seventy children (27.5%) were aged <10 months and were therefore ineligible for VAP3, while 185 (72.5%) were aged ≥10 months and eligible (Table 1).
Table 1. Sociodemographic characteristics (N = 255).
Characteristic |
n |
% |
Mean age mother (years) |
27.4 ± 6.8 |
|
No formal education |
97 |
38.0% |
Primary |
88 |
34.5% |
Secondary |
41 |
16.1% |
Literate (informal education) |
29 |
11.4% |
Fon overall/among heard |
183 |
71.8% overall 84.7% among heard (183/216) |
Commune ALLADA/TOFFO/ZE |
50/94/111 |
19.6%/36.9%/43.5% |
Child age mean 13.12 ± 4.82 median 12 [9 - 16.5] |
70 <10 months 27.5% ineligible 185 ≥10 months 72.5% eligible |
|
3.2. Communication Channels
Overall, 84.7% of mothers (216/255) [95% CI: 79.7 - 88.6] were exposed to at least one communication channel, while 39 mothers (15.3%) reported no exposure at all. The most frequent channels were health workers at the health facility (72.5% [95% CI: 66.7 - 77.7] [n = 185]) and community relays at home (70.6% [95% CI: 64.7 - 75.9] [n = 180]). Radio was the third most frequent channel, reported by 39.6% [95% CI: 33.7 - 45.8] (n = 101), mainly through community radios Radio Allada and Radio Toffo. Television exposure was marginal at 12.9% [95% CI: 9.2 - 17.7] (n = 33), and poster exposure was 16.9% [95% CI: 12.6 - 22.0] (n = 43). Qualitative findings confirmed that health workers were trusted for technical information, with 86.9% trust (172/198), while community relays were trusted for proximity and use of the Fon language (Table 2).
Table 2. Communication channel exposure (N = 255).
Channel |
n |
% |
95% CI |
Health worker |
185 |
72.5% |
[66.7 - 77.7] |
Community relay |
180 |
70.6% |
[64.7 - 75.9] |
Radio |
101 |
39.6% |
[33.7 - 45.8] |
TV |
33 |
12.9% |
[9.2 - 17.7] |
Poster |
43 |
16.9% |
[12.6 - 22.0] |
3.3. Distribution and Validation of Exposure Score
Table 3. Distribution of exposure scores (N = 255).
Exposure score |
n |
% |
Note |
Mean ± SD |
2.12 ± 1.40 |
|
|
Median [IQR] |
2 [1, 3] |
|
|
Low (0 - 1) |
81 |
31.8% |
|
Medium (2) |
76 |
29.8% |
|
High (3 - 5) |
98 |
38.4% |
Formative index – distribution only Cronbach/PCA/AUC in text §2.3 [25] |
The exposure score ranged from 0 to 5, with only 10 mothers (3.9%) achieving a score of 5. The mean score was 2.12 ± 1.40, with a median of 2 and an interquartile range of 1 - 3. The distribution was as follows: low exposure (0 - 1) in 31.8% (n = 81), medium exposure (2) in 29.8% (n = 76), and high exposure (3 - 5) in 38.4% (n = 98). Figure 1 shows the distribution of the 3-channel exposure score (Table 3, Figure 1).
Figure 1. Distribution of the 3-channel exposure score.
3.4. Factors Associated with VAP3 Completion
VAP3 completion was 17.3% overall (44/255) [95% confidence interval (CI): 13.0 - 22.5] and 22.7% among eligible children aged ≥10 months (42/185) [95% CI: 17.1 - 29.4]. VAP1 coverage was 55.7% (142/255) overall and 56.2% among eligible children (104/185), whereas VAP2 coverage was 37.3% (95/255). The dropout rate from VAP1 to VAP3 was 69.0% (98/142) and 59.6% (62/104) among eligible children. The dropout rates from VAP1 to VAP2 and from VAP2 to VAP3 were 33.1% and 53.7%, respectively. In bivariate analysis, high exposure (3 - 5 channels) was strongly associated with VAP3 completion: 28/98 (28.6%) among highly exposed versus 4/81 (4.9%) among low-exposed, with a crude odds ratio of 7.70 [95% CI: 2.57 - 23.05], p < 0.001. Among eligible children, high exposure was associated with 35.1% completion (26/74) versus 6.8% (4/59) for low exposure, with a crude odds ratio of 7.45 [95% CI: 2.43 - 22.8] (Table 4).
Table 4. Bivariate association with VAP3 completion.
Factor |
Crude N = 255
VAP3 Yes/OR [95% CI] |
Eligible N = 185
VAP3 Yes/OR [95% CI] |
p-value |
Note |
Exposure scoreHigh (3 - 5) vs Low (0 - 1) ref |
Low: 4/81 = 4.9% (ref)High: 28/98 = 28.6%OR = 7.70[2.57 - 23.05] |
Low: 4/59 = 6.8% (ref)High: 26/74 = 35.1%OR = 7.45 [2.43 - 22.86] |
<0.001 |
Dose-response Existing data only |
Community relay(Yes vs No) |
No: 72/211 (34.1%)Yes: 139/211 (65.9%)In No groupYes: 41/44 (93.2%)OR = 7.08 [2.12 - 23.65] |
– |
<0.001crude |
Quasi-separation (93.2% among VAP3 Yes) Eligible not reported due to small cells Composite in Table 5 |
Radio exposure(Yes vs No) |
No: 135/211 (64.0%)Yes: 76/211 (36.0%)Yes: 25/44 (56.8%)OR = 2.34 [1.21 - 4.52] |
– |
0.017crude |
Crude only Eligible focused on composite
Table 5 |
3.5. Multivariable Determinants of VAP3
Multivariate logistic regression was performed using svyset to account for the cluster design, including the continuous exposure score, language spoken at home, commune, maternal age, and child age in months (Figure 2, Table 5).
Figure 2. Forest plot of multivariable determinants of VAP3 completion among eligible ≥10 months.
Table 5. Multivariable determinants of VAP3 completion (N = 255 and eligible N = 185).
Factor |
Crude aOR [95% CI] N = 255 |
Eligible aOR [95% CI] N = 185 (≥10 mo) |
p-value |
Note |
Exposure score per channel 5 items continuous |
2.87 [1.92 - 4.29] |
1.61 [1.15 - 2.27] |
0.006 eligible |
Per 1-channel increase svyset |
Exposure score 3 items (worker, relay, radio) |
2.85 sensitivity analysis |
2.45 [1.40 - 4.26] |
0.002 |
Operational score |
Fon language Fon vs Other |
3.15 [1.21 - 8.18] |
1.47 [0.44 - 4.87] |
0.018 crude 0.575 NS eligible |
71.8% overall |
Maternal age (per year) |
0.98 [0.93 - 1.04] |
1.04 [0.97 - 1.11] |
0.264 NS |
Continuous |
Child age (per month) |
Not included (DAG) |
0.83 [0.73 - 0.93] |
0.002 |
Added per DAG |
Commune Ze vs Allada proxy access/stock-out |
0.62 [0.28 - 1.38] |
0.32 [0.12 - 0.91] |
0.241 NS crude 0.032 eligible |
Geographic difference proxy |
3.6. Factors Associated with Dropout VAP1 → VAP3
Among the 142 children who received VAP1, 98 dropped out before VAP3, corresponding to a crude dropout rate of 69.0%. Among the eligible children aged ≥10 months, 104 received VAP1 and 62 dropped out before VAP3, corresponding to a dropout rate of 59.6% (Figure 3, Table 6).
Figure 3. Forest plot of determinants of dropout VAP1 → VAP3 among eligible participants ≥10 months.
Table 6. Determinants of dropout VAP1 → VAP3 (crude n = 142, eligible n = 104).
Factor |
Crude n = 142 dropout 98 (69.0%) OR [95% CI] |
Eligible n = 104 dropout 62 (59.6%) OR [95% CI] |
Eligible aOR
[95% CI] |
Note |
Exposure score per channel continuous |
0.35 [0.18 - 0.68] p = 0.002 |
0.74 [0.49 - 1.11] |
0.74 [0.49 - 1.11] |
Weaker after eligibility |
Exposure score Low (0 - 1) vs High (3 - 5) ref |
0.18 [0.07 - 0.46] p < 0.001 |
0.22 [0.08 - 0.58] |
– |
Attenuated |
Factor |
Crude OR [95% CI] |
Eligible OR [95% CI] |
p-value |
Note |
Community relay exposure |
0.30 [0.08 - 1.12] |
0.30 [0.08 - 1.12] |
0.072 borderline |
DAG selection |
Fear of AEFI |
2.15 [0.89 - 5.21] |
2.15 [0.89 - 5.21] |
0.089 |
Rumors 72.2% (143/198) |
Commune Ze vs Allada |
0.32 [0.12 - 0.91] (from VAP3 model) |
0.32 [0.12 - 0.91] |
0.032 |
Proxy access
stock-out Aug 2024 |
4. Discussion
The crude VAP3 coverage of 44/255 (17.3%) [95% CI: 13.0 - 22.5] underestimates true completion because it includes 70 children (27.5%) aged <10 months who were ineligible for VAP3 according to the schedule (VAP3 is eligible from ≥10 months, i.e., 9 months plus a one-month grace period) [14] [16]. When restricted to eligible children aged ≥10 months (n = 185), VAP3 coverage was 42/185 (22.7%) [95% CI: 17.1 - 29.4], compared with a crude value of 17.3%. Among the ineligible children aged <10 months (n = 70), VAP3 was documented in 2/70 (2.9%) as an early vaccination. This coverage remains far below the Ghana pilot (~80% at 11 months) [11], Kenya (~70%) [12], and Malawi (~65%) [10] figures reported in the MVIP evaluations.
The communication exposure index retained five channels for descriptive monitoring: health workers 72.5% [95% CI: 66.7 - 77.7] (n = 185), community relay 70.6% [95% CI: 64.7 - 75.9] (n = 180), radio 39.6% [95% CI: 33.7 - 45.8] (n = 101), television 12.9% [95% CI: 9.2 - 17.7] (n = 33), and posters 16.9% [95% CI: 12.6 - 22.0] (n = 43). The initial ODK reports of 0.8% for television and 0% for posters were due to a coding error that was corrected.
In the bivariate analysis, high exposure was significantly associated with VAP3 completion. Among 98 mothers with high exposure, VAP3 was completed in 28 cases (28.6%) compared with 4/81 (4.9%) among low-exposed mothers, with an odds ratio of 7.70 [95% CI: 2.57 - 23.05], p < 0.001.
5. Conclusion
The communication exposure score, conceived as a formative index, retained five channels for descriptive monitoring, with a mean of 2.12 ± 1.40. For operational recommendations, three channels (health workers, community relay, and radio) were retained to define multi-channel exposure as ≥3 channels, with messages fully delivered in Fon, coupled with a reminder system and defaulter tracing, as implemented in Ghana and Kenya [10]-[12]. In the multivariable analysis, the five-channel score was associated with an adjusted odds ratio of 1.61 [95% CI: 1.15 - 2.27] among eligible children, while the three-channel score showed an adjusted odds ratio of 2.45 [95% CI: 1.40 - 4.26]. The Fon language was spoken at home by 71.8% of the participants overall and by 84.7% of those who had heard about the vaccine, with an adjusted odds ratio of 1.47 [95% CI: 0.44 - 4.87], p = 0.527, which was not statistically significant but qualitatively important, as 72.2% of rumors were reported and 86.9% trust in relays was observed. The commune of residence (Ze versus Allada) was associated with completion.
Limitations
This study has limitations. First, temporality and reverse causation: the cross-sectional design measures exposure and outcome at a single visit via recall, with risk of postdictive and social desirability bias potentially inflating the dose-response relationship; results should therefore be interpreted as associations rather than causal effects, and causal language should be tempered. Second, eligibility: 70 children (27.5%) aged <10 months were ineligible for VAP3 and were included in crude denominators, which underestimates VAP3 coverage and overestimates dropout. When restricted to eligible children aged ≥10 months (n = 185, 72.5%), VAP3 coverage was 22.7% versus 17.3% crude, and dropout was 59.6% versus 69.0% crude.
Public Health Implications
After correction for eligibility restriction, the communication exposure formative index retains five channels for descriptive monitoring: health worker 72.5%, community relay 70.6%, radio 39.6%, television 12.9%, and poster 16.9%, with a mean of 2.12 ± 1.40. In the multivariable analysis, two scores were presented: (i) the five-channel score with an adjusted odds ratio per channel of 1.61 [95% CI: 1.15 - 2.27] among eligible children; and (ii) the three-channel score with an adjusted odds ratio of 2.45 [95% CI: 1.40 - 4.26] retained for operational purposes. To reduce the eligible dropout rate of 59.6% (WHO target 10% and MVIP 20% - 30%), we recommend the following: (i) correcting DHIS2 to not count children <10 months in the VAP3 denominator and displaying VAP3 coverage at 10 months (9 months plus one month grace period); (ii) training 30 community relays in Fon language on the vaccination schedule; (iii) implementing a reminder system via phone calls and home visits; and (iv) strengthening defaulter tracing, as successfully implemented in Ghana and Kenya.
Ethical Approval and Consent to Participate
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. [27]. Authorization to conduct the study was obtained from the Atlantic Departmental Health Directorate and the Ethics Committee of the Faculty of Health Sciences of the University of Abomey-Calavi (Ref: N˚194/2025FSS /UAC). Written informed consent was obtained from all mothers/guardians in Fon or French before administering the questionnaire. Confidentiality and anonymity were strictly maintained throughout the study [28].
Availability of Data and Materials
The data generated and/or analyzed during this study are available from the corresponding author upon reasonable request in accordance with ethical restrictions.
Acknowledgements
The authors thank the epidemiologist of the Allada-Toffo-Ze Health Zone, the Allada-Toffo-Ze Health Zone management, the Atlantic Department Health Directorate (DDS Atlantique), the National Agency for Primary Health Care (ANSSP), the eight interviewers, two supervisors, 255 households, and 198 qualitative stakeholders for their participation.
Author Contributions
Conceptualization, Raïssatou Ami-Touré and Badirou Aguemon; methodology, Raïssatou Ami-Touré and Badirou Aguemon; software, Raïssatou Ami-Touré and Badirou Aguemon; validation, Barikissou Georgia Damien, Vikkey Antoine Hinson and Badirou Aguemon; formal analysis, Raïssatou Ami-Touré and Badirou Aguemon; investigation, Raïssatou Ami-Touré; data curation, Raïssatou Ami-Touré and Badirou Aguemon; writing—original draft preparation, Raïssatou Ami-Touré; writing—review and editing, Badirou Aguemon, Barikissou Georgia Damien and Vikkey Antoine Hinson; visualization, Raïssatou Ami-Touré; supervision, Barikissou Georgia Damien and Badirou Aguemon; project administration, Raïssatou Ami-Touré; funding acquisition, Raïssatou Ami-Touré.
Appendix
List of Abbreviations
AEFI |
Adverse Event Following Immunization |
ANSSP |
National Agency for Primary Health Care (ANSSP, Benin) |
ASCQ |
Qualified Community Health Worker (ASCQ) |
AUC |
Area Under the Curve |
CI |
Confidence Interval |
COREQ |
Consolidated Criteria for Reporting Qualitative Research |
DAG |
Directed Acyclic Graph |
DDS |
Departmental Health Directorate (DDS) |
DEFF |
Design Effect |
DHIS2 |
District Health Information Software 2 |
EIR |
Entomological Inoculation Rate |
EPI |
Expanded Program on Immunization |
Fon |
Local language spoken in Atlantique, Benin |
ICC |
Intra-cluster Correlation Coefficient |
INSAE |
National Institute of Statistics and Demography (INSAE) |
IPTp |
Intermittent Preventive Treatment in pregnancy |
IQR |
Interquartile Range |
LLINs |
Long-Lasting Insecticidal Nets |
MAPI |
Adverse Event Following Immunization (French acronym for AEFI) |
MVIP |
Malaria Vaccine Implementation Program |
ODK |
Open Data Kit |
OR |
Odds Ratio |
PCA |
Principal Component Analysis |
PPS |
Probability Proportional to Size |
RGPH-4 |
4th General Population and Housing Census, 2023 (RGPH-4, Benin) |
RTS,S/AS01 |
RTS,S/AS01 malaria vaccine (MosquirixTM) |
SMC |
Seasonal Malaria Chemoprevention |
STROBE |
Strengthening the Reporting of Observational Studies in Epidemiology |
TV |
Television |
VAP1 |
Anti-malaria Vaccine dose 1 (RTS,S dose 1 at 6 months) |
VAP2 |
Anti-malaria Vaccine dose 2 (RTS,S dose 2 at 7 months) |
VAP3 |
Anti-malaria Vaccine dose 3 (RTS,S dose 3 at 9 months) |
VAP4 |
Anti-malaria Vaccine dose 4 (RTS,S dose 4 at 18 months) |
WHO |
World Health Organization |
aOR |
Adjusted Odds Ratio |