Health insurers’ data is inadequate for healthcare research

Wolfgang Greiner

Wolfgang Greiner

The professor concluded that this data source should only be used very cautiously and with reasonable diligence.

‘Health services research is in vogue in Germany,’ he said, when presenting the study at this year’s German Capital Congress. However, this research area is still in the early stages in this country. The first annual Congress of Health Services Research was held in Cologne in 2002 The German network of health services research was founded in 2006.

Unlike clinical randomised controlled trials, health services research focuses on scientific analyses of the daily routine of treatment and medication, using a trial design specially suited to the everyday work of physicians and all their patients, without knockout criteria, as in conventional clinical trials, it also factors in a cost-benefit calculation.

That’s why one problem for this country’s health services research is encountered in the data acquisition. The data has to figure out effectively followed medical care in everyday life, not in a special clinical setting, and the group to be evaluated statistically is a good deal larger than in clinical trials. Consequently the costs for these specially designed health services studies, including extra data acquisition, are very high (primary studies). Also, there is still a lack of interest in that kind of research here and a reserved attitude due to the high costs, Prof. Greiner pointed out.

Therefore the routine data of health insurance funds can be regarded as a valued asset for health services research. Improved technical requirements and a change in the mindset within the management of the health insurance funds now allow the use the funds’ data for health services research (secondary studies). The patients can be made anonymous and assigned to treatments and medications over years. ‘As a by-product of administrative billing and reimbursement, the health insurance funds data reflect the direct treatment and prescription process and therefore they can be seen as most comprehensive data base in general healthcare,’ Prof. Greiner said.

In 2009, Prof. Greiner closed the study* that analysed whether health insurance funds routine data are suitable for research. In cooperation with the German health insurance fund Techniker Krankenkasse and supported by the pharmaceutical company Bayer-Schering, he and his team examined the follow-up therapy of 14,001 patients after their first implantation of hip or knee joint endoprosthesis. The retrospective study surveyed the follow-up therapy for three months after surgery, using data from 2006 and 2007. The result of the study is astonishing: Two in three patients did not receive a common thrombosis prophylaxis. ‘This result does not reflect reality or, in other words, it is very unrealistic. But it points out the constraints of the data,’ said Prof. Greiner.

The research team revealed that many patients received thrombosis prophylaxis in the post-operative phase of their hospital stay and it is continued within the often following rehabilitation in a rehabilitation centre.

Indeed, the German health insurance funds only collect data from out-patient care (ambulatory treatment). Prof. Greiner concluded: ‘The analysis of routine data of health insurers requires high caution and diligence. We need an increased set up of theory and methodology to balance the lack of a scientific purpose of data and to offset their limitations,’

In July, Prof. Greiner was appointed to the German Advisory Council on the Assessment of Developments in the Healthcare System. In this position, he will possibly provide political decision-makers with more understanding of the situation and what is needed for health services research in Germany.

* Study to be published in Gesundheitsökonomie + Qualitätsmanagement (Thieme-Verlag).
 

07.09.2010

• health insurance (42)
• Health Insurers (35)
• research (3026)