[Determination of vaccination coverage and disease incidence using statutory health insurance data]

The KV-Sentinel, established in 2004, is a joint project of the Robert Koch Institute and the 17 associations of statutory health insurance physicians (ASHIPs) in Germany. The ASHIPs provide anonymous physicians billing data to the Robert Koch Institute. The aim of this article is to describe methodological approaches for processing these routine data to determine vaccination coverage and incidence of vaccine preventable diseases. Furthermore, we discuss limitations in interpreting these data.

The ASHIPs perform a data query of all vaccinations and of ICD-10 codes for pertussis, measles, mumps, varicella and herpes zoster and send anonymous data to the Robert Koch Institute. We perform routine tests to ensure data quality. Study population is the statutory health insured population (85.5% of the German population). Vaccination coverage is determined by the number of vaccinated persons and the number of statutory health insured persons. Incidence is calculated by the number of diseased persons per 100 000 statutory health insured persons.

All 17 ASHIPs participate in the project. In total, 95 905 605 data records for vaccinations and 4 570 919 data records for pertussis, measles, mumps, varicella and herpes zoster were provided from 2004 to 2007. After performing routine tests with regard to structure and content of data, more than 99% of the data records can be analysed. In 2007, the majority of given vaccinations were monovalent vaccinations against influenza (39%) and tick-borne encephalitis (17%). In 2006 and 2007, 1 893 790 data records for diagnoses were provided. Of these, 75% were acute diagnoses and of these 70% were confirmed diagnoses. Most often, ICD-10 codes for herpes zoster (57%) and varicella (35%) were reported.

Nationwide vaccination coverage of statutory health insured persons by age group can be determined by using billing data. It is possible to validate billing data of vaccinations with available data from other studies. Interpretation of billing data of acute vaccine preventable diseases remains challenging because it is difficult to assess potential under- or overestimation without the possibility of external validation. Therefore, further research is needed.