Statewide immunization databases in California are the California Immunization Registry, Regional Immunization Data Exchange, and San Diego Immunization Registry; the laboratory system is the California COVID Reporting System (CCRS). In New York, immunization information systems include Citywide Immunization Registry and the New York State Immunization Information System; the laboratory system is the Electronic Clinical Laboratory Reporting System (ECLRS). California data were matched between the immunization and case registries using a probabilistic algorithm with exact match for zip code and date of birth and fuzzy match on first name and last name. New York data were matched to the ECLRS with the use of a deterministic algorithm based on first name, last name, and date of birth. In California, person-level hospitalization data from CCRS and supplementary hospitalization reports were used to identify COVID-19–associated hospitalizations.
Specifically acknowledged are:
Dana Jaffe, California Department of Public Health; Rebecca Hoen, Meng Wu, New York State Department of Health; Citywide Immunization Registry Program, New York City Department of Health and Mental Hygiene.
From whom those data were likely in substantial part obtained.
The CDC also has covidnet, which doesn’t have the data heterogeneity problem so much, but I’m not sure it has the vaccination status breakdown needed for this study. CDC has famously not done a good job breaking that down… https://www.cdc.gov/coronavirus/2019-ncov/covid-data/covid-n...
Anyway, none of this is to say that they’re definitely not doing a cherry picking exercise, but at least one should read the paper and decide if the choices they make about data sources and analysis are reasonable. Not super productive to just immediately assume bad faith, IMO. Especially since the conclusions are consistent with research from other institutions/countries