The story of the fallacy of coverage is made up of five theme areas of evidence that reveals the discrepancies in immunization coverage rates and the reasons behind these discrepancies. It also includes evidence for turning around this fallacy of coverage. The first theme is the challenge of timeliness or age-appropriate immunization. This is evident in the studies presented in this supplement from India , Pakistan [4, 8], and Burkina Faso [9, 10]. By assessing coverage through the analytical lens of age-appropriateness of coverage, these studies bring into question the progress of immunization coverage. In fact, as seen by the evidence, while overall coverage can be deceptively good, the story changes when one takes into account whether or not vaccines are administered on time. In India for example, while immunization coverage has overall increased, the work by Corsi et al.  shows that, nationally, complete age-appropriate coverage is still under 50%. Not only do age-appropriate immunization rates provide a truer picture of actual coverage, but such an approach is also useful for health workers and service providers - as noted in the work by Dugas et al.  and Bicaba et al.  - as it allows them to improve and tailor their immunization strategies to increase coverage. The implications for not providing immunization on time are reduced benefit of the vaccine and increased mortality.
The critical importance of the timely immunization theme was recently detailed in a review of data from 45 low-income and middle-income countries published in the Lancet.
The next key theme area of evidence to better understand the fallacy of coverage is the issue of social and gender equity. Work by Corsi et al.  demonstrates that the progress of immunization coverage in India is hindered by the persistence of gender inequities across all socio-economic levels, resulting in girls having significantly lower coverage rates for bacille Calmette-Guérin (BCG), oral polio vaccine (OPV), DPT and measles vaccination. Interestingly, gender inequities affect who gets immunized or not, but does not affect the decision to immunize on time. Gender also needs to be considered in targeting interventions, as demonstrated through an in-depth ethnographic study by Dugas et al.  in Burkina Faso. The researchers found that in some communities despite the father's decision to vaccinate the children, mothers do not always bring them to be immunized. This gap between decision-making and actual vaccination practice requires interventions that target both parents.
The linkage between poor coverage and increasing inequities is demonstrated in the studies in Pakistan and Burkina Faso. Mitchell et al.  provide evidence on how poor access and mother's education (in urban areas only) were the key equity factors obstructing measles vaccination uptake in the Lasbela district of Pakistan. In the Nouna district of Burkina Faso, Sanou et al.  provide evidence for how the education of both parents along with the economic conditions of households affected immunization coverage. The authors did note, however, that the influence of economic conditions is complex as immunization services are free of charge, thus pointing to the importance of communication about the free services to avoid potential abuses by providers.
In addition to inappropriate coverage and increasing inequities linked to coverage, the Pakistan study uncovered the next key theme area, which is that of vaccine efficacy. Through the development of a communication tool (a "balance sheet" summarizing published evidence on benefits and possible adverse effects of vaccination) to enable communities to balance costs and benefits of measles vaccination, Ledogar et al  uncovered a much lower vaccine efficacy rate in Lasbela, Pakistan, than expected. As such, the rate they found was 41.5% compared to the rate generally found in developed countries of 95% (range of 90-98%) . Reasons for this discrepancy are discussed in the paper. While this "balance sheet" was not used in the randomised controlled trial of a community intervention in Lasbela, Pakistan, such a tool could serve as a web-based reference for project managers and health officials, helping them identify areas of improvement in immunization services.
Many of the articles in this supplement offer a strong demand side perspective, i.e. from existing and potential beneficiaries. This demand side evidence is the next key theme area which looks behind the discrepancies in vaccination rates to help tailor potential solutions. In the Pakistan series of articles, Shea et al.  conducted a systematic review of the literature on the impact of demand side interventions, demonstrating existing gaps and highlighting the need for such operational research. By focusing on the demand side, Dugas et al.  provide an increased understanding of why there is a lower than expected coverage rate in a health district in Burkina Faso. Their study points to the need to tailor interventions such that they take into account parents' perception of childhood illness and to the need to examine local vaccination procedures or requirements. In this particular case, their research found that vaccination procedures served to deter rather than ensure access. In practice, immunization access was conditional on women going for antenatal care and acquiring and preserving a vaccination booklet for their child. In the same vein, Bicaba et al.  argue for the importance of understanding the reasons why some children are still not completely vaccinated.
As part of a demand side analysis, understanding the local context is one of the keys to unravel the fallacy of coverage. Fourn et al.  in Benin ask the question: what are the factors that lead to reticence to vaccination among religious populations? Through the use of qualitative methods, their results suggest that interpretation of religious principles by church-going populations is primary in explaining reticence and that the solutions lie in creating an open dialogue among all actors: reticent parents, their religious leaders and health authorities. Such an intervention is also alluded to by the work in Burkina Faso by Sanou et al.  and national EPI teams have used the results of the study.
The local level analyses also provide valuable information on existing immunization programs/campaigns. For example, Mitchell et al.  note that while other areas in Pakistan have demonstrated the positive impact of mobile vaccination teams, this was not the case in Lasbela district. Similarly, as seen in Burkina Faso, Haddad et al note that Immunization Days did not have any impact on the performance of routine vaccination services. While both authors discuss this finding in their papers, it is worthy to note that these results from South Asia to West Africa further underscore that blueprint national programs/campaigns often do not resonate with communities as their assumptions are disconnected from local realities. Local data can be used to tailor such programs/campaigns to increase their effectiveness.
The studies in this supplement provided evidence from a local/district level and also compared their results to national data sets, thus presenting the last key theme area with regards to the fallacy of coverage. In the article by Cockroft et al.  the authors unravel a complex context where actual coverage rates are masked by national rates, and where there is heterogeneity in vaccination coverage between and within districts, and particularly between urban and rural areas in Pakistan. Despite this heterogeneity, as Cockroft et al.  point out, there is some commonality in the variables associated with vaccination uptake such as the quality of services, mother's education and knowledge of benefits. Local evidence on these issues can then be used to turn around the fallacy of coverage, thereby increasing immunization uptake.