Hongwei Liu, Qian Zhang, Mingshuang Li, Lance Rodewald, Dan Wu, Tingting Yan, Zhijie An, Zundong Yin, Hui Zheng
City-specific implementation of PCR-enabled confirmatory reporting was associated with a substantial reporting increment concentrated in laboratory-confirmed reports. The predicted counterfactual remained above historical reference levels, indicating that the broader 2024 increase in reported incidence was not explained by differential local implementation alone.
BACKGROUND: In December 2023, China revised its pertussis case definition to allow laboratory confirmation based on polymerase chain reaction (PCR) positivity alone.
OBJECTIVE: To quantify the reporting increment associated with city-specific implementation of PCR-enabled confirmatory reporting after the 2023 protocol revision.
METHODS: We analyzed monthly surveillance data from 58 prefecture-level cities in four Chinese provinces during 2018-2024; 48 implemented PCR-enabled confirmatory reporting and 10 remained never-treated. We used the Callaway-Sant'Anna staggered difference-in-differences (CS-DID) analysis to estimate implementation-associated changes in total reported, laboratory-confirmed and clinically diagnosed pertussis incidence among children aged ≤5 years.
RESULTS: The joint pre-treatment test for the prespecified -6 to -1 month window did not reject the null, whereas the -18 to -1 and -24 to -1 month windows showed evidence of pre-treatment differences. During the January 2023-December 2024 primary window, the observed treated-post mean was 33.32 per 100,000, and the aggregate CS-DID-predicted counterfactual mean was 14.85 per 100,000. The average treatment effect on the treated (ATT) was 18.47 per 100,000 (95% CI 10.31-26.63; P<0.001) for total reported incidence. The corresponding estimates were 17.98 per 100,000 (95% CI 10.15-25.81; P<0.001) for laboratory-confirmed incidence and 0.49 per 100,000 (95% CI -0.80 to 1.77; P=0.459) for clinically diagnosed incidence.
CONCLUSIONS: City-specific implementation of PCR-enabled confirmatory reporting was associated with a substantial reporting increment concentrated in laboratory-confirmed reports. The predicted counterfactual remained above historical reference levels, indicating that the broader 2024 increase in reported incidence was not explained by differential local implementation alone.