*2.6. Outcome*

The primary outcome was the male-to-female ratios (M:F) at enrollment and by antigens among children from the 2019, 2020, 2021, and 2022 birth cohorts enrolled in the SEIR. The M:F ratio was the number of vaccinated males relative to females. Enrollment was defined as the first encounter of the child with the SEIR. We calculated the M:F ratios at the district and UC levels. We adjusted the M:F ratios using the sex ratios at birth (1.055) in Pakistan [21]. We computed the M:F ratios for the up-to-date vaccination coverage for Pentavalent-1, Pentavalent-3, and Measles-1 at 6, 9, 12, 18, and 24 months. Up-to-date coverage was defined as the proportion of 0–24 months children who received vaccinations by the specified months of age. In order to examine timely coverages, we also calculated the up-to-date coverage of Penta-1 at 10 weeks, Penta-3 at 18 weeks, and Measles-1 at 10 months to account for the timeliness criteria used by EPI-Sindh (an additional 4 weeks' time duration beyond the age at which each vaccine is due, as per the WHO-specified EPI schedule). Furthermore, we compared the M:F ratios by maternal literacy level, geographic residential location of the child (urban vs. rural, rural vs. remote-rural and slums vs. non-slums), modality of vaccination (fixed center, outreach, and enhanced outreach), and the sex ratio of vaccinators in the province. As a secondary outcome, we also calculated the Gender Inequality Ratio (GIR) for all the above analyses, where the gender inequality ratio was defined as the proportion of vaccinated males among those who were due for vaccination, relative to the proportion of vaccinated females who were due for vaccination.

#### *2.7. Statistical Analysis*

We reported the median and interquartile range (IQR) of the UCs for the M:F ratios, along with the ranges at the UC level. UCs with no children vaccinated for any particular vaccine were excluded from the analysis for that particular vaccine only. A male-to-female ratio of 0.00 indicated that there were no males or females vaccinated in the particular UC. This was due to the reduced population sizes when we examined our indicators across the sub-categories (maternal literacy and geographic location of vaccination) within a UC.

For our secondary outcome, we computed the GIR by dividing the proportion of males who were due and received vaccinations by the proportion of females who were due and received vaccinations. A GIR of 1.00 implied no differential in coverage rates between females and males, whereas a GIR of above 1.00 indicated inequalities (with higher coverage rates for males relative to females). We performed statistical analyses with Stata, release 17 (StataCorp, College Station, TX, USA). We used digital maps to review the immunization coverage by the district and UC using QGIS (3.16.7-Hannover).
