Table 1.
Levels and changes in levels in birth related health factors associated with China’s universal two child policy (announced in October 2015), based on data from individual level delivery information records (IDIR)
| Proportion (%) of subgroups | Monthly percentage (%; mean (standard deviation)) | Level change (95% CI; P value) | |||
|---|---|---|---|---|---|
| Baseline level* | Plateau level† | Difference-of-proportions estimate‡ | Segmented regression estimate§ | ||
| Deliveries to multiparous mothers | 46.4 (1.5) | 58.5 (1.0) | 12.2 (11.1 to 13.3; <0.001) | 9.1 (6.4 to 11.7; <0.001) | |
| Deliveries to mothers aged ≥35 | 8.5 (0.3) | 13.5 (0.3) | 5.0 (4.8 to 5.2; <0.001) | 5.8 (5.2 to 6.4; <0.001) | |
| Deliveries to mothers aged ≥35 who were nulliparous mothers | 3.2 (0.2) | 4.0 (0.2) | 0.8 (0.6 to 1.0; <0.001) | 1.4 (1.1 to 1.7; <0.001) | |
| Deliveries to mothers aged ≥35 who were multiparous mothers | 14.7 (0.5) | 20.3 (0.4) | 5.6 (5.2 to 6.0; <0.001) | 7.2 (6.4 to 8.1; <0.001) | |
| Deliveries by caesarean | 39.6 (0.3) | 39.1 (0.4) | −0.5 (−0.9 to −0.1; 0.01) | — | |
| Deliveries to nulliparous mothers that were by caesarean | 39.6 (0.4) | 36.6 (0.5) | −3.0 (−3.5 to −2.5; <0.001) | — | |
| Deliveries to multiparous mothers that were by caesarean | 39.7 (0.2) | 40.9 (0.5) | 1.2 (0.8 to 1.6; <0.001) | — | |
| Deliveries to mothers aged <35 years that were by caesarean | 38.4 (0.3) | 36.9 (0.3) | −1.5 (−1.8 to −1.2; <0.001) | — | |
| Deliveries to mothers aged ≥35 that were by caesarean | 52.5 (0.5) | 53.1 (1.0) | 0.6 (−0.4 to 1.6; 0.20) | — | |
Baseline level of birth related health factors. The baseline period varies across factors: for parity and maternal age, it starts from January 2015 (start of the IDIR data series) to June 2016; for caesarean delivery, it starts from January 2015 to June 2015 (just before the beginning of the trough shown in fig 4).
Plateau level shows the mean monthly percentage during the period February 2017 to December 2017.
Difference-of-proportions estimates of level change show the result of a simple two tailed test of the difference-of-proportions between the baseline level and plateau level columns.
Segmented regression estimates of level change show the estimated level change from segmented regressions (described in the methods and supplement). Segmented regressions were not estimated for caesarean delivery because visual inspection suggested that the data were inconsistent with model assumptions (seen in the temporary trough shown in fig 4).