Minimum volumes for knee replacement (knee TEP) cannot be derived from the routine data available

German health care data can be used to show a statistical correlation between the volume of operations and the quality of the outcomes for total knee endoprostheses (knee TEP). However, generally applicable threshold values cannot be derived from this. Firstly, the evidential value of the available data is too low. Secondly, the data are contradictory. The relationship between case numbers and outcomes is quite different for the different quality indicators. This was the conclusion reached in a report by the Institute for Quality and Efficiency in Health Care (IQWiG), which the Cologne scientists presented on 6 February 2006.

The Federal Joint Committee ( Gemeinsamer Bundesausschuss, G-BA) commissioned the Institute to develop a statistical model that could be used to determine threshold values for knee TEP. Staff members of the IQWiG Department of Medical Biometry evaluated health care data from the Federal Office for Quality Assurance (Bundesgeschäftstelle für Qualitätssicherung, BQS). They used data from 110 349 patients with a first knee TEP in 2004. This was compared with data from 2003.

U-shaped curve for the risk of immobility

Experts in knee TEP were nominated by the Federal Joint Committee and defined the quality indicators "mobility of the knee joint" and "frequency of wound infections after operations". They also defined possible risk factors that might have influenced the outcomes. For example, it was tested whether the age or state of health of patients before the operation distorted the outcomes. The Cologne scientists established that the correlation between the volume of procedures and the risk of immobility was non-linear. Up to a volume of about 400 knee TEP operations per hospital per year, the risk of having only restricted mobility in the knee was reduced. However, from about 500, it increased sharply again. If plotted in a graph, this gives a U-shaped curve. The results with the quality indicator "infection" were quite different. The rate of wound infection dropped evenly with increasing numbers of operations, although the absolute risk reduction is very low. The information from the IQWiG model calculation is thus contradictory. On the basis of the BQS data, higher minimum volumes may be suited for the indication "infection", but should not be recommended for the indication "mobility".

Evidential value of the data is very restricted

There is an additional reason why the IQWiG scientists consider that their results do not provide a suitable basis for drawing conclusions on health care. According to the IQWiG project manager Dr. Ralf Bender PD, "The quality of the currently available data is very unclear, so that the reliability of the conclusions is very limited." The data available to the scientists had been routinely collected and reported by the hospitals themselves to the BQS. For example, the data provided on the quality indicator "immobility" was implausible in many cases, and could not be evaluated. Moreover, it was unclear at which time point mobility was measured. According to Bender, "Even if our calculations had permitted unambiguous statements, these could not have been used to deduce specific minimum volumes." The Cologne scientists are of the opinion that statistical correlations with routinely collected data cannot be used to provide evidence-based proof for threshold values. This is only possible with controlled intervention studies. The final report is about 60 pages long and also contains the statements submitted on the preliminary report. It can be accessed as a PDF file at the IQWiG website under .

Background

Minimum volumes were first introduced into Germany with the Health Modernisation Act (Gesundheitsmodernisierungsgesetz [§137 Social Code Book V]). Doctors and hospitals may only perform specific operations if they do so with a specified frequency. In 2003, the self-administration partners agreed to define minimum volumes for five indications from 2004 onwards (liver, kidney, and stem cell transplantation, as well as complex operations on the oesophagus and pancreas).

In September 2004, the Federal Joint Committee also included coronary operations and total knee endoprostheses (knee TEP) into the so-called catalogue of plannable services, but did not specify numbers. The Federal Joint Committee commissioned IQWiG on 22 December 2004 to determine minimum volumes for knee TEP; this commission was reformulated on 9 June 2005. On 16 August 2005, the members of the Federal Joint Committee responsible for hospital treatment specified a binding minimum volume of 50 knee TEP operations per hospital per year. This regulation came into force on 1 January 2006. The commission to IQWiG was unaffected by this decision.

At the start of December 2005, the Federal Joint Committee commissioned a two-year research project, funded with € 180 000, to investigate the effects of the specified minimum volumes on the quality of treatment and the structure of hospital care in Germany.

Contact: Tel. 0221/35685-0