Quantifying the Impact of Switching From Rotarix to Rotavac Rotavirus Vaccine in Ghana

Abstract

Background

Ghana introduced the Rotarix vaccine in April 2012 and switched to the Rotavac vaccine in January 2020. Evaluating the health impact of this change is essential for national policy and can inform decisions in other countries.

Methods

We applied a previously validated mathematical model of rotavirus transmission in Navrongo, Ghana, to quantify the health impact of switching from Rotarix to Rotavac among children under 2. We first simulated the model from January 2007 to December 2019 using previously estimated model parameters from pre- and post-Rotarix surveillance data. We then refitted the model to data from January 2020 to March 2023 to estimate Rotavac vaccine effectiveness parameters. To assess overall vaccine effectiveness (OVE), we compared model simulations under the 2 vaccination scenarios (Rotarix and Rotavac) with a counterfactual scenario assuming no vaccination.

Results

The model accurately captured both the trend and seasonality of postvaccination rotavirus cases, with a Spearman correlation of 0.70. Compared to the no-vaccination scenario, the estimated OVE against moderate-to-severe rotavirus cases was 21.7% (95% prediction interval [PI]: 17.4%–26.1%) in children under 2 years in Navrongo. Annual OVE ranged from 8.4% (PI: 1.0%–21.0%) to 63.5% (PI: 39.0%–87.0%). Following the switch from Rotarix to Rotavac (2020–2023), we estimated a mean relative OVE of 37.1% (PI: 27.5%–46.5%).

Conclusions

Our model offers a robust tool to support policymakers in Ghana and similar settings by evaluating current rotavirus vaccine effectiveness, guiding vaccine switching decisions and informing introduction strategies in countries yet to adopt immunization.

With several rotavirus vaccines available, countries must make informed decisions on whether to introduce a vaccine or switch products. Our model offers a robust framework to evaluate vaccine impact and support decision-making related to vaccine introduction or switching between products.

Rotavirus vaccination protects against severe morbidity, hospitalization, and mortality caused by rotavirus-associated gastroenteritis (RVGE) among infants <5 years of age. Consequently, the World Health Organization (WHO) has recommended the inclusion of rotavirus vaccines in routine childhood immunization programs since 2009 [1]. Currently, 4 WHO-prequalified oral live-attenuated rotavirus vaccines (Rotarix, RotaTeq, Rotavac, and Rotasiil) are available globally [2, 3]. Following widespread routine introduction, there has been a consistent and marked decrease in global RVGE–related hospitalizations and mortality [4–9]. Given substantial evidence of rotavirus vaccine effectiveness, it is imperative that countries, especially those with a significant rotavirus burden, consider introducing the vaccine if they have not yet implemented it.

The availability of multiple rotavirus vaccines with varying prices and cold chain requirements presents a range of options for decision-makers. Countries yet to introduce the vaccine must decide which product to adopt, while those that have already introduced one may consider switching to a more cost-effective profile. Vaccine supply disruptions can also necessitate a switch. For instance, the 2019 withdrawal of RotaTeq from the Gavi market forced several African countries to transition to alternative vaccines [10–12]. Despite these switches, only 1 study to date has evaluated the impact of changing rotavirus vaccines [13]. Continued assessment of the postintroduction impact of available rotavirus vaccines is therefore essential to inform and support national immunization programs.

Ghana introduced the Rotarix vaccine into its routine immunization program in April 2012, with a 2-dose schedule administered at 6 and 10 weeks of age [14]. Previous studies have reported moderate and heterogeneous Rotarix vaccine impact in Ghana [15–19]. Since 2013 (1 year after introduction of the vaccine), Ghana has maintained a high vaccination coverage rate of over 86% [15, 20]. In 2020, Ghana switched from the Rotarix vaccine to the 3-dose Rotavac vaccine administered at 6/10/14 weeks of age [21]. The rationale for the switch was due to lower vaccine costs and smaller cold chain space. A prior study demonstrated that switching from Rotarix to Rotavac vaccine would be cost-effective [21]. It is worth noting that the study assumed both vaccines had comparable efficacy and coverage, and therefore, the findings were significantly influenced by differences in vaccine price. Thus, there is a need to evaluate the potential impact of this vaccine switch on the rotavirus burden in Ghana.

Here, we aimed to evaluate the impact of switching from the Rotarix vaccine to the Rotavac vaccine on the rotavirus burden using surveillance data from War Memorial Hospital in Navrongo, Ghana. We utilized a previously developed and validated mathematical model of rotavirus epidemiology for the rollout of the Rotarix vaccine in Ghana [18, 19]. We predicted rotavirus cases in the absence of the switch from Rotarix to Rotavac vaccine from January 2020 to March 2023. We then refitted the model for the Rotavac rollout period (from January 2020 to March 2023) to estimate model parameters that describe vaccine effectiveness. We compared the 2 models to evaluate whether there has been a significant change in the rotavirus burden following the switch from Rotarix to Rotavac in Ghana.

METHODS

Rotavirus Surveillance Data

We utilized data on monthly confirmed RVGE cases from surveillance studies at War Memorial Hospital in Navrongo, Ghana (Supplementary Figure 1), between July 2007 and March 2023. Stool samples were collected from infants <2 years presenting with acute gastroenteritis (defined as 3 or more watery stools per day lasting <7 days) and tested using standard enzyme-linked immunosorbent assay (ELISA) methods [22, 23]. Following Rotarix vaccine introduction in April 2012, the immunization status of all tested cases was determined from vaccination cards. The dataset was divided into 3 parts: prevaccine (July 2007–May 2011), the Rotarix era (January 2013–December 2019), and the Rotavac era (January 2020–March 2023). Ghana began transitioning from Rotarix to Rotavac in January 2020, completing the switch by the end of that year [24]. Data from June 2011 to December 2012 were missing. Vaccination coverage data were extracted from the WHO and UNICEF estimates of national immunization coverage (WUENIC) database [20]. We assumed equal coverage for both rotavirus doses based on similar Rotarix dose 1 and 2 uptake and minimal differences between DTP1 and DTP3 coverage in Ghana [20 , 25 ].

Transmission Model Description

We used a previously developed age-structured compartmental model of rotavirus transmission for Navrongo, Ghana, incorporating local demographic data [18, 19] (Supplementary Figure 2). The model was previously fitted to RVGE surveillance data from July 2007 to September 2016 and validated with data from October 2016 to December 2019 [18, 19].

We stratified the modeled population into 42 age groups (monthly intervals for infants <2 years, yearly intervals for ages 2–4, 5-year intervals from 5 to 70 years, and those over 70). Homogeneous mixing was assumed across age groups. This model has been extensively used to identify key drivers of rotavirus epidemiology and quantify vaccine impact across different settings [26–29]. A detailed description of the model and parameter values is provided in the Supplementary Data. Briefly, individuals are born with maternally acquired immunity, which wanes over time, making them susceptible to primary infection. Susceptible individuals are infected at a rate defined by the force of infection, with a significant proportion of first infections leading to moderate-to-severe RVGE. After recovery, individuals gain temporary immunity. Upon waning, reinfection can occur at a reduced rate, making them less likely to develop moderate-to-severe RVGE and recover at a faster rate. The model allows for multiple infections, assuming that infections beyond the second are typically mild or asymptomatic.

We incorporated rotavirus vaccination into the model by adding compartments for vaccinated individuals. The rate of transition from unvaccinated to vaccinated compartments is determined by first-dose coverage. A single dose provides partial protection for those who seroconvert, with vaccine-induced immunity waning over time. After waning, individuals become susceptible to primary infections after the first dose, secondary infections after the second dose, and tertiary infections after the third dose. Transmission rates and infectious periods postwaning are assumed to be the same as those of unvaccinated individuals in the corresponding unvaccinated compartments (Supplementary Figure 2). As the model uses monthly age groups for children <2 years, we approximated age of vaccination to the nearest month, consistent with data on the timeliness of vaccination [30].

Model Fitting Procedure

We used the same model, assumptions, and fitting procedures as described in [18, 19]. We simulated the model using previously estimated mean parameters (see Table 1 [18, 19] and Supplementary Table 1) for the pre- and post-Rotarix vaccine era (up to December 2019). We then refitted the model to data from the Rotavac era (from January 2020 to March 2023), estimating the vaccine response rate for each dose (assuming a heterogeneous vaccine response) and the duration of vaccine-induced immunity. We assumed infants who do not respond to the first dose are less likely to respond to subsequent doses. Vaccine protection was estimated using a maximum likelihood framework, assuming the observed number of RVGE cases is Poisson distributed with a mean equal to the model-predicted monthly number of RVGE cases multiplied by the time-varying reporting fraction. To account for temporal variation in care-seeking behavior (due to changes in healthcare utilization and other potential drivers of diarrheal illness), we multiplied the mean estimated reporting fraction by the ratio of the 24-month moving average of confirmed test-negative cases to the overall average number of test-negative cases during the study period (see Figure 1A for the time series and relative reporting rate). We estimated 95% CIs using the profile likelihood method. Model performance was assessed using Spearman correlation, root mean square error, and the percentage of months in which observed RVGE cases fell within the model's 95% prediction intervals (PI).

Table 1.

Estimated Model Parameters Obtained After the Model Was Fitted Separately to the Monthly Rotavirus Surveillance Data From Navrongo, Ghana, for the Prevaccination, Post-Rotarix, and Post-Rotavac Periods

Parameter	Definition	Values	Source
R0	Basic reproductive number	31.529 (30.36–32.78)	[18]
1/ωm	Average duration of maternal immunity (months)	1.890 (0.39–3.18)	[18]
b1	Amplitude of annual seasonal forcing	0.999	[18]
ϕ1	Annual seasonal offset (months)	1.05	[18]
b2	Amplitude of biannual seasonal forcing	6.63E-08	[18]
ϕ2	Biannual seasonal offset (months)	5.125	[18]
h	Proportion of moderate-to-severe diarrhea cases reported	0.012	[18]
d3	Proportion of subsequent infections that are severe	2.54E-05	[18]
Parameters describing Rotarix vaccine protection
SC1RV1	Proportion who responded to the first dose	0.612 (0.36–0.87)	[19]
SC2RV1	Proportion who responded to the second dose	0.19 (0.11–0.48)	[19]
1/ωvRV1	Duration of vaccine-induced immunity (months)	5.465 (3.16–8.33)	[19]
Parameters describing Rotavac vaccine protection
SC1	Proportion who responded to the first dose	0.811 (0.69–0.85)	Estimated
SC2	Proportion who responded to the second dose	0.250 (0.19–0.31)	Estimated
SC3	Proportion who responded to the third dose	0.154 (0.06–0.30)	Estimated
1/ωv	Duration of vaccine-induced immunity (months)	20.247 (15.63–28.25)	Estimated

Values in parentheses indicate the 95% CIs. The references to previously estimated model parameters (both prevaccination and post-Rotarix) are provided.

Figure 1.

Confirmed rotavirus cases, vaccination coverage, and comparison of model-simulated and observed rotavirus cases in Navrongo. A, Monthly time series of laboratory-confirmed rotavirus-positive (black) and rotavirus-negative (red) cases for Navrongo, with annual rotavirus vaccination coverage indicated by blue boxes. The dark red line indicates the relative reporting fraction, used to account for potential changes in care-seeking behavior and other drivers of diarrheal illness. B, The observed number of rotavirus-associated gastroenteritis cases per month is plotted in black, while the model predictions for the scenarios with no vaccination, Rotarix vaccination, and Rotavac vaccination are plotted in green, red, and blue, respectively. The green, red, and blue shaded regions correspond to the fitted model projections accounting for both parameter uncertainty and the observation process. The vertical lines indicate the 4 time periods: prevaccine fitted era (July 2007–May 2011), post-Rotarix fitted era (April 2012–September 2017), post-Rotarix prediction era (October 2017–December 2019), and post-Rotavac fitted era (January 2020–March 2023).

Finally, we quantified the impact of the vaccination program by estimating overall vaccine effectiveness (OVE) using model simulations. We simulated 2 counterfactual scenarios to estimate the RVGE cases that would have occurred: (1) in the absence of vaccination and (2) with continued use of Rotarix (ie, no switch to Rotavac). The first scenario was used to estimate annual OVE, while the second was used to estimate the relative effectiveness of Rotavac compared to Rotarix. The OVE is defined as the reduction in RVGE cases in the vaccinated population compared to the unvaccinated population [31] and was estimated as follows:

$$OV{E}_y=1-\frac{{\mathrm{\Sigma }}_{mϵy}{Y}_{vacc,m}}{{\mathrm{\Sigma }}_{mϵy}{H}_{novac,m}}$$

where Y_vacc,m represents either the observed or model-simulated rotavirus cases for each month (m) in year y, and H_novac,m denotes the model-simulated rotavirus cases in the absence of vaccination. The relative vaccine effectiveness of Rotavac compared to Rotarix (baseline) was estimated using the continued use of Rotarix as the denominator. To quantify uncertainty in OVE and the counterfactual scenarios, we sampled 1000 parameter sets from the estimated 95% CIs using bootstrapping with replacement. For each parameter set and time point, “observed” cases were simulated from a Poisson distribution with a mean equal to the model prediction. The 2.5th and 97.5th percentiles of these simulations were used to generate 95% PIs for the counterfactual scenarios and corresponding 95% CIs for the OVE.

Sensitivity Analyses

To assess the potential impact of coronavirus disease 2019 (COVID-19) restrictions on rotavirus transmission, we refitted the model to data from January 2020 to March 2023, estimating only R₀ while using the estimated parameters for Rotarix vaccine (ie, assuming the same effectiveness for Rotarix and Rotavac). We also assessed model sensitivity to the identifiability of the probability of responding to each vaccine dose by simulating 2 vaccine response scenarios. For Scenario 1, we assume S_C1 = 50% and S_C2 = 20%, whereas for Scenario 2, we assume S_C1 = 30% and S_C2 = 40%. Finally, we fitted a full-dataset model that estimated separate vaccine-impact parameters for the Rotarix and Rotavac periods and a reduced model for post-Rotavac in which the duration of Rotavac-induced immunity was assumed to be equal to that estimated for Rotarix (Supplementary Data).

RESULTS

Rotavirus transmission in Navrongo, Ghana, is highly seasonal, with nearly all cases occurring between October and March (Figure 1). This strong seasonal pattern has persisted despite over a decade of vaccination. Among infants <2 years, the peak monthly number of confirmed RVGE cases ranged from 10 to 49 in the prevaccine era, 3 to 56 during the Rotarix era, and 4 to 19 during the Rotavac era (Figure 1). Trends in the confirmed rotavirus-negative cases remained consistent before, during, and after the COVID-19 pandemic (Figure 1A). Vaccination coverage has remained high, increasing from 42% in 2012 to 95% in 2023, with no substantial reduction during the COVID-19 pandemic (Figure 1A).

The fitted model reproduced confirmed RVGE cases during the Rotavac period (January 2020–March 2023) (Figure 1B). The Spearman correlation between observed and model-predicted cases was 0.70, with a root mean square error of 5.9 cases, indicating a good agreement. About 93% of observed monthly cases fell within the model's 95% PIs. In contrast, predictions using Rotarix-era vaccine parameters consistently overestimated RVGE cases for the same period (Figure 1B). This discrepancy is attributed to the slightly higher vaccine response rate and longer duration of immunity estimated for Rotavac compared to Rotarix (Table 1). The third dose of the Rotavac vaccine provides an additional opportunity for infants to receive another dose, potentially leading to an improved immune response compared to the 2-dose Rotarix vaccine. Both vaccines show a similar trend in differential response rates between the first and subsequent doses (Table 1).

The reduction in RVGE cases after Rotavac introduction cannot be explained solely by reduced transmission during the COVID-19 pandemic. The estimated R₀ for 2020–2023 was 22.2 (95% CI: 19.8–27.0), lower than the prevaccination estimate (Table 1). Despite this reduced R₀, the refitted model overestimated RVGE cases compared to the model assuming different parameters for Rotavac (Supplementary Figure 3). Additionally, the model showed limited sensitivity to the identifiability of the vaccine response parameters (Supplementary Figure 4). Despite differences in estimated vaccine effectiveness parameters, the full-dataset model yielded results comparable to the main model (Supplementary Figure 5 and Tables 2 and 3). Likewise, although the reduced model produced higher response rate estimates for the second and third dose of Rotavac, its Akaike information criterion was similar to that of the main model (75 vs 73).

Using the no-vaccination scenario as a counterfactual, the average annual OVE from 2012 to 2023 among children <2 was 21.7% (95% PI: 17.4%–26.1%) (Figure 2A). The annual OVE ranged from a minimum of 8.4% (95% PI: 1.0%–21.0%) to a maximum of 63.5% (95% PI: 39.0%–87.0%). Effectiveness was higher during the Rotavac period (2020–2023) compared to the Rotarix period (2012–2019). During the Rotarix era, OVE ranged from 8.4% (95% PI: 1.0%–21.0%) to 16.8% (95% PI: 3.0%–43.6%), while during the Rotavac era, it ranged from 26.5% (95% PI: 10.4%–44.5%) to 63.5% (95% PI: 36.6%–87.5%).

Figure 2.

Annual time series of OVE against moderate-to-severe rotavirus-associated gastroenteritis cases among children <2 y of age. A, The OVE was estimated using model simulations, with the no-vaccination scenario serving as the baseline. The vertical dashed line indicates the year of the vaccine switch (2020). B, The relative OVE was estimated by comparing model-simulated Rotavac outcomes to Rotarix as the baseline, for the post-Rotavac introduction period (2020–2023).

Using the Rotarix model projections as the baseline, we estimated a substantial relative OVE following the switch to Rotavac during 2020–2023 (Figure 2B). The average annual relative OVE was 37.1% (95% CI: 27.8%–46.3%), ranging from 14.3% (95% CI: 7.0%–22.9%) to 58.4% (95% CI: 31.5%–83.0%).

DISCUSSION

With multiple rotavirus vaccines available, countries face different policy decisions depending on whether they have already introduced rotavirus vaccination. Mathematical models capable of evaluating the impact of various vaccination strategies are essential for guiding rotavirus vaccination policy. Our study demonstrates the health benefits of rotavirus vaccination and highlights the importance of maintaining routine immunization in Ghana. More importantly, we estimated a substantial increase in overall vaccine effectiveness following the switch to the 3-dose Rotavac vaccine in Navrongo, Ghana. We have shown that our model can assess and quantify the impact of rotavirus vaccine switching and offers a flexible platform for countries to quantify the health impact and optimize the implementation of different rotavirus vaccination strategies.

The switch from Rotarix to Rotavac was associated with a 37% relative increase in OVE between January 2020 and March 2023. Our result contrasts with the only other published study, which found comparable effectiveness between the 2 vaccines following the switch from Rotarix to Rotavac in Palestine [13]. Despite these differences, switching from Rotarix to Rotavac has been found to be cost-effective in both Ghana and Palestine [21, 32], largely due to lower vaccine prices, as previous analyses assumed equal effectiveness. Our findings of higher relative effectiveness in Navrongo suggest that the switch may be even more cost-effective than previously estimated in this setting. Moreover, our model offers a robust tool for generating counterfactual scenarios to inform future cost-effectiveness analyses.

Although this is not a head-to-head vaccine effectiveness analysis, the difference in dosing schedules (3 doses for Rotavac vs 2 for Rotarix) may help explain the estimated variation in relative vaccine effectiveness. The improved impact of the 3-dose schedule could be due to (1) individuals who failed to seroconvert after 1 or both Rotarix doses potentially achieving seroconversion with a third dose; (2) infants who missed a dose in the 2-dose schedule having the opportunity to receive an additional dose in the 3-dose schedule; and (3) a higher response rate to the first dose of Rotavac compared to Rotarix, potentially due to reduced interference from maternal antibodies against the Rotavac strain—a human-bovine reassortant initially isolated from asymptomatic neonates. Clinical trials in Navrongo and other low- and middle-income countries (LMICs) have shown higher response rates with a 3-dose Rotarix schedule compared to 2 doses [14, 33, 34]. Our previous modeling work also suggests that adding a third Rotarix dose at 14 weeks could yield a greater impact than the then-implemented 6-/10-week schedule [19]. However, further studies are needed to quantify the benefit of an additional Rotarix dose.

It is also important to note that the impact of Rotarix vaccine in Navrongo was modest compared to 2 other sites in Ghana [18, 19]. For instance, using model simulations with no vaccination as the baseline, we estimated an average annual OVE of 44.4% (95% PI: 35.0%–53.9%) among children <5 years in Accra between 2012 and 2015 (Supplementary Figure 6). One possible explanation for the estimated increase in OVE for Navrongo from 2018 onward is the progressively higher proportion of vaccinated children <5 years of age. At the time of Rotarix introduction, transmission was primarily driven by unvaccinated children outside the recommended vaccination age. Over the years, as coverage increased, nearly the entire under-5 population became vaccinated, likely enhancing indirect protection through herd immunity and contributing to the higher estimated OVE. Given the substantial within-country heterogeneity in rotavirus vaccine effectiveness, it is important for countries to establish additional surveillance sites to better quantify vaccine impact. Further research is also needed to identify the factors underlying the modest effectiveness of rotavirus vaccines in LMICs.

Over a decade after rotavirus vaccine introduction in Ghana, strong seasonal transmission patterns persist in Navrongo, which differ from the seasonal shifts observed in high-income countries postvaccine introduction [35–38]. These persistent strong seasonal patterns could be explained by high birth rates and the suboptimal performance of the rotavirus vaccine, which allows for the continued presence of susceptible individuals during the peak season (November–March). Despite evidence of strongly seasonal patterns across the Sudan–Sahel region of West Africa [39–42], its influence on vaccine impact remains poorly understood. More data are needed to explore infection rates among infants vaccinated before, during, and after the rotavirus season to better evaluate potential seasonal variation in vaccine effectiveness.

A key strength of this study is the use of a previously calibrated and validated model based on local surveillance data, which has been shown to capture realistic rotavirus transmission patterns [18, 19]. It is worth noting that since our model accounts for changes in healthcare-seeking efforts, we can capture potential disruptions in care-seeking trends over time. For example, although the Rotavac rollout occurred during the COVID-19 pandemic, care seeking for acute gastroenteritis remained stable, as indicated by trends in test-negative cases (Figure 1A). The strong seasonal epidemics observed in Navrongo are consistent with patterns across the Sudan–Sahel region of West Africa [40, 42, 43], where test-positive cases are rare outside the rotavirus season, but test-negative cases remain stable year-round. Additionally, this model successfully predicted the impact of Rotarix vaccination in Malawi before, during, and after the COVID-19 pandemic [29].

While COVID-19 restrictions have the potential to reduce overall rotavirus transmission, several factors likely limited their impact on rotavirus dynamics in Navrongo. First, rotavirus transmission is highly seasonal in Navrongo, peaking between January and March, while most COVID-19 measures in Ghana (eg, lockdowns and mobility restrictions) were implemented outside this period in 2020 (March–September) [44]. The only major intervention overlapping with the rotavirus season was school closures, but schools reopened in January 2021, coinciding with the seasonal peak, likely minimizing their effect. Second, rotavirus infections in Navrongo predominantly occur early in life and often result in severe dehydration requiring intravenous treatment [18]. It is therefore unlikely that care-seeking behavior for such severe cases significantly declined during the COVID-19 period. Third, results from our sensitivity analysis suggest that a reduction in R₀ (used as a proxy for COVID-19–related restrictions) alone does not fully explain the observed decline in rotavirus cases. Finally, in Ghana, evidence suggests that COVID-19 restrictions had a limited impact on childhood immunization, including rotavirus vaccination, with an estimated 6% reduction in coverage during the brief restriction period between March and June 2020, followed by a return to prepandemic levels by 2021 [45]. Together, these findings suggest that the improved vaccine impact observed after 2020 is more attributable to the switch from Rotarix to Rotavac than to pandemic-related restrictions.

This study has several important limitations. First, limited post-Rotavac data constrained our ability to fully validate the model. Second, the relative vaccine effectiveness comparison between Rotavac and Rotarix is based on data from a single site, limiting generalizability. Third, post-Rotavac data were not age stratified, so the model was fitted only to overall time series data, potentially reducing parameter identifiability compared to age-stratified fits in the prevaccine and post-Rotarix periods (Supplementary Figure 7). Additionally, the small number of observed cases per month may introduce statistical noise, which could affect the robustness of model estimates. Finally, our method of controlling for changes in rotavirus epidemiology over time using trends in test-negative cases may not fully capture key shifts in transmission drivers, including those resulting from COVID-19 restrictions. Further work and data from additional sites in Ghana are needed to quantify the impact of switching from Rotarix to Rotavac and to obtain more robust estimates of vaccine effectiveness.

In summary, our modeling study demonstrates that, despite suboptimal rotavirus vaccine performance, there has been a considerable reduction in moderate-to-severe RVGE cases, particularly following the switch from Rotarix to Rotavac vaccine in Ghana. Although the vaccine impact improved after the switch, it is imperative to sustain rotavirus surveillance efforts to monitor the post-Rotavac vaccine impact in Ghana. While vaccine choice is sometimes constrained by supply limitations, this study is critically important for countries outside of, or transitioning from, Gavi support, as they must independently assess the health impact of available rotavirus vaccines and determine which to adopt. Our model can provide a robust framework for quantitative evaluation of the effects of rotavirus vaccine switching to better inform policy decisions in countries considering a transition to an alternative rotavirus vaccine.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions . Ernest O. Asare (Conceptualization, Investigation, Methodology, Visualization, Writing–original draft, Writing–review & editing), Virginia E. Pitzer (Conceptualization, Investigation, Methodology, Supervision, Writing–review & editing), Belinda L. Lartey (Methodology, Writing–review & editing), and George E. Armah (Conceptualization, Methodology, Supervision, Writing–review & editing)

Financial support. This work was supported by the National Institutes of Health/National Institute of Allergy and Infectious Diseases grant R01AI112970 (V. E. P.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Data availability . All data and code needed to evaluate the conclusions in the paper are available in the paper and/or the Supplementary Data and are available at Dryad (https://datadryad.org/stash/dataset/doi:10.5061/dryad.4mw6m90kv).