TY - GEN
T1 - The gap between the planned and real-life implementation of the vaccination programme: Timely and effective BCG vaccination in Guinea-Bissau
AU - Vedel, Julie Odgaard
PY - 2025/5/12
Y1 - 2025/5/12
N2 - IntroductionThe World Health Organization recommends Bacillus Calmette-Guérin (BCG)vaccination at birth to protect against tuberculosis in low-income countries and countries with high tuberculosis burden, including Guinea-Bissau. However, in Guinea-Bissau, BCG vaccination is often delayed. One of the reasons for the delays is the restrictive vial-opening policy (RVOP): BCG is supplied in 20-dosevials and any unused doses must be discarded 6 hours after opening and reconstitution. This has led to practices of not opening a BCG vial unless a fixed number of children are present for vaccination. Field observations indicate that togather enough children to open the multi-dose vials, BCG vaccinations are only made available at predefined days. Even on these days, the vaccines are only opened if the predefined number of children are present. Similar practices are in place for the measles vaccine and yellow fever vaccine, both delivered in 10-dosevials. Beyond tuberculosis protection, the timing of BCG vaccination is relevant as BCG has beneficial non-specific effects. Randomised trials from Guinea-Bissau and Uganda have found that early BCG vaccination lowers all-cause mortality and non-tuberculosis morbidity. Beyond timeliness, vaccination technique is important: BCG is administered intradermally, and a correctly administered vaccine usually causes a scar at the injection site. Yet, scarring rates vary widely between study cohorts. Among BCG-vaccinated children, having a BCG scar is associated with lower mortality. Nevertheless, currently neither timing nor quality of BCG vaccinations are reflected in the widely used performance indicator, the 12-month coverage estimates. The aim of this thesis was to assess the real-life implementation of the childhood vaccination programme in rural Guinea-Bissau, focusing on BCG and the implications of the RVOP, and identify better BCG-vaccination performance indicators. MethodsThe aim was explored through five studies, building on two different data collection platforms.The Bandim Health Project (BHP) runs a nationally representative rural health and demographic surveillance system (HDSS), monitoring pregnancies, birth outcomes, and childhood vaccinations through repeated household visits in an open cohort of more than 25,000 women and 23,000 children below 5 years of age.The BHP HDSS was the foundation for three observational studies.One of the studies based in the HDSS and two other studies were conducted within the context of the cluster-randomised CS-BCG trial. The aim of CS-BCG is testing whether increasing the availability of BCG at rural health facilities (HFs)can lower early infant mortality. In three rural regions, half of the HFs (n=25) were supplied with extra BCG vials to allow opening a BCG vial every time just one eligible child was present. The other half were randomised to the control group,continuing the usual practice of RVOP. The randomisation of the HFs was crossed-over after one year, ensuring that all HFs served as both intervention and control HFs. Individual-level information on pregnancies, birth outcomes, and childhood vaccinations are collected through repeated household visit in a trial specific platform covering all villages in the trial area. The three studies based on the BHP HDSS aimed to 1) assess BCG coverage and missed BCG-vaccination opportunities among children born in 2013-2019, 2)assess the impact of the COVID-19 pandemic on vaccination coverage and timeliness for all vaccines in the vaccination schedule by comparing children born in2017 and 2020, and 5) assess BCG-scarring rates among children born beforeand after the implementation of refresher training in BCG-vaccination technique for health care workers (HCWs) as part of the CS-BCG trial. Within the CS-BCG trial, we conducted a qualitative study, 3), exploring the effects of the RVOP for HCWs involved in vaccinating children in both RVOP and no-RVOP areas through semi-structured interviews and for caregivers of vaccine eligible children in RVOP areas through focus group discussions. In 4), a mixed methods study, we combined insights from a HF staff survey implemented before the initiation of CS-BCG, the CS-BCG data platform, and semi-structured interviews with vaccinators to assess how the implementation of the vaccination programme and BCG vaccination uptake is affected by the RVOP, and whether vaccinators deem it feasible to abandon the RVOP. Quantitative data were analysed using descriptive statistics, logistic regressionmodels with cluster robust standard errors, and logistic multi-level models; qualitative data using thematic network analysis and a phenomenological approach. ResultsIn study 1), we found that only 40% of at-birth BCG-vaccination opportunities were utilised. If all opportunities were utilised, BCG coverage at birth could havebeen 45% instead of the actual 19%. Study 2) showed that vaccination delays were common but only vaccines affected by stock-outs had markedly lower 12-month coverages in the 2020 cohort than in the 2017 cohort. When excluding vaccines affected by stock-outs, there was no difference in proportions of fully vaccinated children between the two cohorts, which remained at on third. In study3), both caregivers and HCWs described several barriers to timely vaccinations, including the RVOPs. These barriers are interconnected and aggravate each other. Most HCWs perceived the RVOP as a logical result of limited vaccine availability but balancing vaccine uptake and dose wastage negatively affected their sense of purpose and job satisfaction. To optimize the vaccination process, many participants recommended abandoning the RVOP and allocating more resources to the HFs. Study 4) demonstrated that all HFs had RVOPs for BCG, measles vaccine, and yellow fever vaccine. When the HFs were supplied with additional BCG vials to abandon the RVOP for BCG, vaccination coverage and timeliness improved, the median vaccination session sizes decreased, and fewer children had missed BCG-vaccination opportunities. The majority of HCWs described feeling or previously having felt conscious about the increased wastage during no-RVOP periods. All deemed it feasible to continue without the RVOP,but many mentioned the need for addition funding for the extra vaccines. In study5), the analyses show that it seems possible to increase BCG-scar prevalence with a combination of refresher training in BCG-vaccination technique and subsequent regular supervision of vaccination routines at the HFs. ConclusionOur findings highlight the gap between how the childhood vaccination programme is designed and how it is implemented. To identify how the programme can be improved, continued research into real-life vaccination programme implementation in needed. However, the gap is not easily identified through the currently used performance indicators. If the vaccination programme was routinely monitored on indicators associated with improved survival, such as BCG coverage atone week and/or 30 days as well as BCG scar prevalence 4-6 months after BCG vaccination, it could help ensure that the maximum number of children in lowincome countries survive early childhood and get a chance at life.
AB - IntroductionThe World Health Organization recommends Bacillus Calmette-Guérin (BCG)vaccination at birth to protect against tuberculosis in low-income countries and countries with high tuberculosis burden, including Guinea-Bissau. However, in Guinea-Bissau, BCG vaccination is often delayed. One of the reasons for the delays is the restrictive vial-opening policy (RVOP): BCG is supplied in 20-dosevials and any unused doses must be discarded 6 hours after opening and reconstitution. This has led to practices of not opening a BCG vial unless a fixed number of children are present for vaccination. Field observations indicate that togather enough children to open the multi-dose vials, BCG vaccinations are only made available at predefined days. Even on these days, the vaccines are only opened if the predefined number of children are present. Similar practices are in place for the measles vaccine and yellow fever vaccine, both delivered in 10-dosevials. Beyond tuberculosis protection, the timing of BCG vaccination is relevant as BCG has beneficial non-specific effects. Randomised trials from Guinea-Bissau and Uganda have found that early BCG vaccination lowers all-cause mortality and non-tuberculosis morbidity. Beyond timeliness, vaccination technique is important: BCG is administered intradermally, and a correctly administered vaccine usually causes a scar at the injection site. Yet, scarring rates vary widely between study cohorts. Among BCG-vaccinated children, having a BCG scar is associated with lower mortality. Nevertheless, currently neither timing nor quality of BCG vaccinations are reflected in the widely used performance indicator, the 12-month coverage estimates. The aim of this thesis was to assess the real-life implementation of the childhood vaccination programme in rural Guinea-Bissau, focusing on BCG and the implications of the RVOP, and identify better BCG-vaccination performance indicators. MethodsThe aim was explored through five studies, building on two different data collection platforms.The Bandim Health Project (BHP) runs a nationally representative rural health and demographic surveillance system (HDSS), monitoring pregnancies, birth outcomes, and childhood vaccinations through repeated household visits in an open cohort of more than 25,000 women and 23,000 children below 5 years of age.The BHP HDSS was the foundation for three observational studies.One of the studies based in the HDSS and two other studies were conducted within the context of the cluster-randomised CS-BCG trial. The aim of CS-BCG is testing whether increasing the availability of BCG at rural health facilities (HFs)can lower early infant mortality. In three rural regions, half of the HFs (n=25) were supplied with extra BCG vials to allow opening a BCG vial every time just one eligible child was present. The other half were randomised to the control group,continuing the usual practice of RVOP. The randomisation of the HFs was crossed-over after one year, ensuring that all HFs served as both intervention and control HFs. Individual-level information on pregnancies, birth outcomes, and childhood vaccinations are collected through repeated household visit in a trial specific platform covering all villages in the trial area. The three studies based on the BHP HDSS aimed to 1) assess BCG coverage and missed BCG-vaccination opportunities among children born in 2013-2019, 2)assess the impact of the COVID-19 pandemic on vaccination coverage and timeliness for all vaccines in the vaccination schedule by comparing children born in2017 and 2020, and 5) assess BCG-scarring rates among children born beforeand after the implementation of refresher training in BCG-vaccination technique for health care workers (HCWs) as part of the CS-BCG trial. Within the CS-BCG trial, we conducted a qualitative study, 3), exploring the effects of the RVOP for HCWs involved in vaccinating children in both RVOP and no-RVOP areas through semi-structured interviews and for caregivers of vaccine eligible children in RVOP areas through focus group discussions. In 4), a mixed methods study, we combined insights from a HF staff survey implemented before the initiation of CS-BCG, the CS-BCG data platform, and semi-structured interviews with vaccinators to assess how the implementation of the vaccination programme and BCG vaccination uptake is affected by the RVOP, and whether vaccinators deem it feasible to abandon the RVOP. Quantitative data were analysed using descriptive statistics, logistic regressionmodels with cluster robust standard errors, and logistic multi-level models; qualitative data using thematic network analysis and a phenomenological approach. ResultsIn study 1), we found that only 40% of at-birth BCG-vaccination opportunities were utilised. If all opportunities were utilised, BCG coverage at birth could havebeen 45% instead of the actual 19%. Study 2) showed that vaccination delays were common but only vaccines affected by stock-outs had markedly lower 12-month coverages in the 2020 cohort than in the 2017 cohort. When excluding vaccines affected by stock-outs, there was no difference in proportions of fully vaccinated children between the two cohorts, which remained at on third. In study3), both caregivers and HCWs described several barriers to timely vaccinations, including the RVOPs. These barriers are interconnected and aggravate each other. Most HCWs perceived the RVOP as a logical result of limited vaccine availability but balancing vaccine uptake and dose wastage negatively affected their sense of purpose and job satisfaction. To optimize the vaccination process, many participants recommended abandoning the RVOP and allocating more resources to the HFs. Study 4) demonstrated that all HFs had RVOPs for BCG, measles vaccine, and yellow fever vaccine. When the HFs were supplied with additional BCG vials to abandon the RVOP for BCG, vaccination coverage and timeliness improved, the median vaccination session sizes decreased, and fewer children had missed BCG-vaccination opportunities. The majority of HCWs described feeling or previously having felt conscious about the increased wastage during no-RVOP periods. All deemed it feasible to continue without the RVOP,but many mentioned the need for addition funding for the extra vaccines. In study5), the analyses show that it seems possible to increase BCG-scar prevalence with a combination of refresher training in BCG-vaccination technique and subsequent regular supervision of vaccination routines at the HFs. ConclusionOur findings highlight the gap between how the childhood vaccination programme is designed and how it is implemented. To identify how the programme can be improved, continued research into real-life vaccination programme implementation in needed. However, the gap is not easily identified through the currently used performance indicators. If the vaccination programme was routinely monitored on indicators associated with improved survival, such as BCG coverage atone week and/or 30 days as well as BCG scar prevalence 4-6 months after BCG vaccination, it could help ensure that the maximum number of children in lowincome countries survive early childhood and get a chance at life.
U2 - 10.21996/55069759-bc8a-45fc-a448-773dd0f275d4
DO - 10.21996/55069759-bc8a-45fc-a448-773dd0f275d4
M3 - Ph.D. thesis
PB - Syddansk Universitet. Det Sundhedsvidenskabelige Fakultet
ER -