*4.2. Limitations*

Our study had some limitations related to the design and implementation. The intentional selection of study units with high proportions of zero-dose children means that our findings are not necessarily representative of the countries, or even states/provinces within the countries, and these findings should not be interpreted to represent the most common or typical barriers to vaccination, but rather the barriers faced by those most excluded from access to quality immunization services. Because we sought to tailor the study design for the context of each country, comparing or synthesizing the findings across countries should be treated with caution. We did not originally design the study with the intersectionality lens in mind, and as such, we missed the opportunity to explore specific intersections and interactions during data collection. We were able to identify and interview caregivers of zero-dose children in most study settings, but not the Edo province in Nigeria. Similarly, the study was not implemented in regions with refugees, displacement, or conflict-affected populations, which we know face many barriers to vaccination. We designed the study to be implemented rapidly to inform timely program design, but the short duration of the data collection period limited the number of respondents interviewed in each community, which may mean the findings are biased. In Nigeria, we used FGDs instead of in-depth interviews with caregivers and community members to optimize the limited time available, but FGDs may have consequences on the type of information shared, particularly for sensitive information. Similarly, because we prioritized the ability to validate most findings in co-creation workshops immediately following data collection, we did not have time to analyze quantitative data during the rapid study period.
