Claims Data Analysis

What can we learn from them?

Tue, 2020 / 12 / 01
Nowadays claims data reach increasing importance, as they represent a comprehensive source of data which is often complementary to conventional methods.

A study of German claims data on the economic burden of ANCA-associated vasculitides (AAV) (click here to view the publication), a group of potentially life-threatening rare autoimmune diseases, was recently published under co-authorship of SKC. But why are analyses of claims data so important? What can we learn from them?

Claims data are the data on billing and reimbursement services provided by the 105 statutory health insurances (SHI) currently found in Germany. They are administered in pseudonymised databases of varying sizes. The identification of patients with a specific disease is based on their diagnosis according to ICD-10-GM coding. This coding has been evaluated as reliable in independent studies. In addition to demographic data (age, gender, region of residence), information on the prescription of drugs, outpatient and inpatient treatment, morbidity and mortality is also available. Health care data represent so-called "real-world data".

In comparison to clinical studies or register-based data collection, claims databases offer a significantly larger cohort of usually several million insured persons. This allows for a detailed and reliable retrospective data collection, especially in the case of orphan diseases. However, direct access to the databases is not permitted due to data protection regulations. Usually, after a detailed examination of the respective research project, a permission is granted, and the associated analyses of the data are carried out. Finally, the applicant only receives the completely anonymised results of these analyses for further use.

The above-mentioned study is based on analyses of the InGef claims database, which contains a total of around 6.3 million insured persons from approximately 60 regionally and nationally operating SHI. In addition to investigations on the hospitalisation of AAV patients, the study quantifies the costs of this disease for the German health system in terms of outpatient and inpatient expenses, medication, and sickness benefit payment. Based on these data, health care costs for an AAV patient in the first year after diagnosis are many times as high as for a correspondingly age-adjusted person in the German general population. This understanding of the economic burden enables a significant improvement in the transparency and efficiency of health care provision.

Due to the large number of characteristics recorded in the claims data, a plenty of possible analytical questions arises. These go far beyond the economic aspects of a disease, as extensive studies on the morbidity of patients can also be carried out. However, the design of the question is limited by the lack of detailed clinical data, such as on quality of life or laboratory findings. The InGef database has been comprehensively validated. In contrast, particularly small databases or those based on only one SHI just operating in a certain region are not representative of the German general population. In this context, the organisation and public availability of a single, Germany-wide claims database has often been requested. Further, the use of claims data is subject to a latency of about 2 years until they are made available.

Analyses of German claims data are valuable and reliable, and should be considered for a variety of questions. They are also useful as a source of information in research and to support political and economic decision-making processes concerning the German health care system, although the underlying database should be chosen with care.

Sources:
  • Hellmich et al. 2020 "Die wirtschaftliche Belastung durch ANCA-assoziierte Vaskulitiden in Deutschland – eine Versorgungsdatenstudie", DOI: 10.1055/a-1275-1636
  • Neubauer et al. 2016 "Access, use, and challenges of claims data analyses in Germany", DOI: 10.1007/s10198-016-0849-3
  • Kreis et al. 2016 "Status and perspectives of claims data analyses in Germany – A systematic review", DOI: 10.1016/j.healthpol.2016.01.007
  • Andersohn & Walker 2016 "Characteristics and external validity of the German Health Risk Institute (HRI) Database", DOI: 10.1002/pds.3895
  • Carnarius et al. 2018 "Diagnosenkodierung in deutschen Arztpraxen aus klassifikatorischer Sicht: Eine retrospektive Studie mit Routinedaten", DOI: 10.1055/s-0043-125069
to the top