The title of the editorial article “Health Care Research versus RCTs” in Monitor Healthcare Research 1/2009, pp. 16 – 17, not only compares apples with oranges, but is also misleading in terms of content. It should have been correctly called: RCTs for health services research! There is no meaningful question here: “Which evidence class is the right one and when?”, but one should have emphasized: the highest possible evidence class is the best! – of course also for health services research.

And that is precisely why Memorandum III of the German Network for Health Care Research (DNVF) on the methodological discussion in health care research is of utmost importance. After the first two memoranda had dealt with the description of the initial situation (Badura et al. 2003) and with the “conceptual, methodological and structural requirements of health services research” (Schrappe et al. 2005), statements on methodology could no longer be left behind. The danger was too great that health services research could be misused as a bypass to avoid evidence-based medicine and the EBM-supported description of the benefits of treatment methods: marketing and financing of self-help groups instead of randomized studies.
This is being demonstrated to us especially these days by a sharp debate in the USA; The Infectious Disease Society of America (IDSA) is forced by a court order to withdraw a scientific guideline on Lyme disease that is backed by high-level evidence because patients and other participants feel their rights have been violated (Kraemer et al. 2009). So we have to learn to describe values and preferences scientifically as important implementation factors.
This is an impressive task. While evidence-based medicine could still rely on the tools of clinical epidemiology, the multidisciplinary construct of health services research needs a whole shelf of toolboxes. And – to stay with the image – the shelf is open from every side, each methodology has different access points. Scientific thinking requires accepting and using this.

Problem with “effectiveness”
Health services research has also established itself as a relevant research area in Germany and is becoming increasingly important. However, it must be clear to everyone involved what the task is, what the message is: “Outcomes research – the study of the end results of health services” (Clancy and Eisenberg 1998). The “last mile”, this effectiveness gap between absolute (efficacy) and relative effectiveness (effectiveness), has been the subject of scientific discussion for many years (Pfaff 2003, Bero et al. 1998) and was also the main topic of the 7th German Health Services Research Congress last year in Cologne.
The questions behind this are urgent: why do EBM-backed indicators only explain so little or not at all the variance in results in everyday clinical practice (Werner and Bradlow 2006)? Last but not least, the Advisory Council for the Assessment of Developments in the Health Care System has repeatedly referred to the “effectiveness” problem in its most recent reports.
In the 2007 report he went one step further, namely he made a suggestion for the substrate of health services research: the appropriateness of treatment procedures (Brook et al. 1986). This term was derived from the economic discussion about the terms benefit and objective need and, as a general term, describes those benefit components that fall under the relative effectiveness (Wille et al. 2007, No. 579). The SVR's plea was to scientifically support these benefit aspects to such an extent that they can be incorporated into the allocation process as valid arguments in addition to efficacy.
From a higher-level perspective, it is of utmost importance for the innovative power of biomedical research to know under which conditions the innovation arrives in everyday care and how the innovation changes along the way. In order to come to valid statements here, however, one has to borrow methodologically from economics, psychology, organizational and social sciences, law and philosophy and ethics. This broad interdisciplinary collaboration is actually what health care research is all about and what makes it so fascinating.

No hesitation with Memorandum III
But the methodological standards must be high, so no hesitation with Memorandum III, you could say. All interested parties are invited, all DNVF working groups are generally open, dates can be found upon request at the office.
The question of which study designs are adequate to obtain reproducible results from health services research is another matter. On the one hand, the study designs used depend on the disciplines that work with them: the epidemiologist will use different methods to research than the organizational researcher.
But in any case, the study design must be clearly presented (obviously, actually), the choice of study design must be justified (this is not always the case), and the design must be used in such a way that internal validity is as high as possible. The (double-blind) randomized trial rightly has the pole position, although there are possible questions for which it is still not the first choice, but other methods are better able to produce valid evidence.

EBM and/not versus health services research
In order to prevent possible abuse here, too, the SVR has made the proposal to make the test of absolute effectiveness a necessary condition and to only allow the test of adequacy in procedures for the allocation decision if the absolute effectiveness (in comparison to the standard) is given can be viewed. This should prevent us from confusing objective need with demand.
Finally: unfortunately not only in the last issue of the MVF, but in many places it is rumored or assumed that EBM and health services research are opposites. They are not, they are neighbors and they need each other: “EBM and health services research”. There will be further systematic questions to be discussed; this is an interesting field and there is enough space to approach each other. <<

by: Prof. Dr. M. Schrappe, Prof. Dr. H. Abholz, Prof. Dr. S. Bartholomeyczik, Prof. Dr. G. Glaeske, Prof. Dr. M. Gottwik, Prof. Dr. M. Klinkhammer-Schalke, Prof. Dr. E. Neugebauer, Prof. Dr. C. Ohmann, Prof. Dr. H. Pfaff