Study population included
The IIS system collected 337,311 immunization records for Changzhou, Jiangsu's 2016–2021 birth cohort aged 1–6 years by December 31, 2022. 268,691 (79.66%) met the Changzhou "1–4" Vaccination Program.
Between 2017 and 2022, varicella incidences for children aged 1–6 varied from 179.05 (in 2022) to 1,951.19 (in 2019) per 100,000. In this study, the NNIDSS reported 2,512 cases from the birth cohort of laboratory- or clinical-confirmed varicella. 2,299 (91.52%) met the study's case-inclusion criteria by screening the IIS records.
2,873 individuals were selected from the cohort using the subcohort sampling method, with a sampling ratio of 1.12%. The sample size of the subcohort, which included each case without any overlap, was 5,172. Figure 1 illustrates a clear pattern for subcohort selection.
Characteristics of the case, cohort, and subcohort populations
The control subjects of the subcohort (Group 2) did not have statistically significant differences in the distribution of variables of demographic characteristics from the cases (Group 3), which differed from the characteristics displayed by controls in the total cohort (Group 1). Cumulative VarV coverages in Group 2 were almost 74.45% for the first dose, 55.45% for the second, and 25.55% without VarV. Quite differently, 14.62%, 2.22%, and 85.38% of Group 3 received 1, 2, and 0 doses of the VarV vaccination, respectively. In the subgroups by the cost type, the top three types for Group 2 were: "O-F", "N-N", and "O-N". Group 3 consisted of 85.38% "N-N" type, with the remaining percentages distributed as follows: "O-N", "F-N", and "O-F". Group 3 had a median observation time of 824 days (IQR: 438–1,011 days) for cases, while Group 2 had 1,701 days (IQR:1,242–1,825 days). The relevant information is shown in Table 1.
Survival curves and median observation times
The 0-dose subgroup had a median observation time of 906 days, and the 1-dose subgroup was 1,066 days. VarV improved median observation time by 160 days after 1 dose. After two doses, almost 75% of people were healthy at the observation cut-off (December 31, 2022). The relevant information is shown in Fig. 2A. Median observation times were 906 days for the "N-N" subgroup, 804 days for the "O-N" subgroup, and 1,090 days for the "F-N" subgroup. The remaining subgroups' median observation times were not achieved for the reasons above. The relevant information is shown in Fig. 2B.
In the "N-N" subgroup, "O-N" subgroup, and "F-N" subgroup shown in Fig. 3A (born before 2019), the median observation time for children over four years of age was 906 days, 639 days, and 729 days, respectively. The median observation times for children under four years of age in the "N-N" subgroup, the "O-N" subgroup, and the "F-N" subgroup shown in Fig. 3B (born in and after 2019) increased to 952, 1073, and > 1,088 days, respectively.
Figure 2B indicates that the "O-N" curve is usually below the "N-N" curve and crosses over after 1000 days. The median " O-N " observation time was also 100 days less than the "N-N" subgroup. When we analyzed the subgroups using the age of four as the cut-off among children born before 2019 (Fig. 3A), the "N-N" curve was above the "O-N" and the "F-N" curves, even if the confidence interval for the latter was too big. For children born in and after 2019 (Fig. 3B), these three curves correspond to the orders of VEs calculated on the Cox proportional hazards models below.
Comparison differences in VEs between subgroups
In Model 1, VEs (95% CI) against varicella increased with each dose, reaching 82.54% (78.88%, 85.57%) and 97.91% (96.91%, 98.59%). Those who received two VarV doses had higher VEs against varicella in the subgroups "O-F", "F-F", "O-O", and "F-F" (more than 95%). In subgroups "O-N" and "F-N" who received only one VarV, VEs were weaker—noted that subgroup "O-N" had 18% lower VE than "F-N" (Model 2). Except in the case where both doses were self-paying ("O-O"), children who received the first dose at their own expense ("O-N", "O-F") always had lower VEs than children who received it free of charge ("F-N", "F-O", "F-F"), whether in the 1- or 2-dose subgroups. Details of VEs (95% CI) are shown in Table 2.
Figure 4 shows VEs as Nomograms. Age, doses, and cost types showed single-factor disease risks, while total points showed aggregate hazards, with larger points increasing risk. The Nomograms predicted disease-free survival after 1, 2, and 5 years based on total points. The 1-dose subgroup scored 72 points, while the 2-dose subgroup received 0. The Nomogram (Fig. 4A) quantified the 28% increase in protected points with 2 doses compared to 1 dose. The subgroups "F-O", "O-O", "O-F", "F-N", "O-N", and "N-N" had points of 3, 5, 36, 78, 80, and 100, while "F-F" had 0. Personal age also affected total points. A 2-year-old who received a single dose of a self-paying vaccine between 2018 and 2022 had a 90% chance of remaining varicella-free after 1 year, 75% after 2 years, and 40% after 3 years. The relevant information is in Fig. 4.
Sensitivity analysis for the excluded dataset
We excluded 68,620 individuals from the original data. The characteristic variable distribution in this excluded subset is shown in Appendix 1. The results of Models 3 and 4 were consistent with the main study in Table 2. The significant results of Model 5 showed an intercept of -10.42, indicating that individuals had a minimal probability of developing varicella even with a single dose of vaccine and a delayed immunization. The estimated value of -2.30 for level "L-T" indicated that it was more effective than level "L-N," which had only one delayed immunization dose (Table 2).