Determining the feasibility of linked claims and vaccination data for a Covid-vaccine pharmaco-epidemiological study in Germany – RiCO feasibility study protocol

Abstract

AbstractIn Germany, there has been no population-level pharmaco-epidemiological study on the safety of the Covid-19 vaccines. One factor preventing such a study so far relates to challenges combining the different relevant data bodies on vaccination with suitable outcome data, specifically statutory health insurance claims data. Individual identifiers used across these data bodies are of unknown quality and reliability for data linkage. As part of a larger pharmaco-vigilance study on the COVID-19 vaccines, called RiCO (German "Risikoevaluation der COVID-19-Impfstoffe”, Englisch "Risk assessment of COVID-19 vaccines”), a feasibility study is being conducted to determine the overall confidence level with which existing data can be analysed in relation to the safety of the COVID-19 vaccine. This RiCO feasibility study will establish a dataflow combining claims data and vaccination data for a sub-sample of the total German population, describe data quality for each data set from the various sources, estimate the proportion of the different linkage errors and will develop various approaches for linking the data in addition to the simple form of linkage using a common identifier in order to reduce possible linkage errors. These last three points are the core objective of the feasibility study. A secondary objective is to test the viability of the required dataflow involving multiple stakeholders from different parts of the healthcare system. Results will be published and used to plan the actual pharmaco-vigilance study on the COVID-19 vaccines for Germany, as well as future research on the role of COVID vaccines as risk or protective factors for long-term COVID-19 effects.Strength and limitationsPotential for a population-level pharmaco-epidemiological study on the safety of the Covid-19 vaccines for Germany, based on vaccination data combined with statutory health insurance claims data.Introduction and estimation of quality metrics pertaining to the linkability of the various data bodies existing in Germany.Direct measurement of linkage error based on the available identifiers is not possible, proxy metrics and descriptive analytics need to be used.An attempt at linkage with the vaccination data could only be made using data from one smaller statutory health insurance, which may limit the extent to which the data can be analysed.