PCI: No clear correlation between volume of operations and quality of outcomes

On the basis of the studies available, there is no recognisable correlation between the volume of percutaneous coronary interventions (PCI) and the quality of outcomes. For patients with acute myocardial infarction, there are some indications that mortality in hospital decreases with increasing numbers of interventions. However, this only applies to the volume per hospital, not to the volume per surgeon. This is the conclusion reached in the final report by the Institute for Quality and Efficiency in Health Care (IQWiG), which the Institute published on 2 August 2006.

In general, it is still unclear whether there is a change in quality in other outcomes, such as subsequent myocardial infarction or the necessity of bypass operations. As the Cologne scientists report, studies currently available have only investigated the possible correlation between volume of procedures and quality of outcomes for very short periods of time, so that no long-term conclusions are possible.

On 16 November 2004, the Federal Joint Committee ( Gemeinsamer Bundesausschuss, G-BA) commissioned IQWiG to analyse the literature on minimum volumes in percutaneous transluminal coronary angioplasty (PTCA) with and without implantation of a vascular prosthesis (stent), and to present the results in the form of an evidence report. As a stent is now predominantly used in PTCA and the procedure is known internationally under the superordinate term PCI, IQWiG researched and evaluated studies on PCI.

Studies give contradictory results

The studies evaluated by IQWiG examine both indications for PCI - the planned (elective) operation for coronary heart disease (CHD) and the intervention after acute myocardial infarction - but come to contradictory results. This is also the case when the outcomes are differentiated. No consistent tendency was evident for mortality (the most frequently studied outcome) - neither with regard to the volume of procedures per hospital nor per surgeon.

For studies that only investigated patients with PCI after a myocardial infarction, there were several indications that mortality decreases with an increasing number of operations per hospital. However, the IQWiG scientists could not find any reliable correlations for the volume per surgeon.

Mortality after the operation was investigated in all studies. However, only death during the period in hospital was usually recorded. The evidential value of this parameter is limited, as patients treated electively usually spend only a short period of time in hospital, and their mortality risk within this time is relatively low. An analysis of 30-day mortality would, for example, be of greater evidential value. However, only two studies actually recorded mortality data after discharge from hospital.

Other important events may occur after the PCI - for example, patients may suffer myocardial infarction, need a bypass operation, or the PCI may have to be repeated. This was only rarely analysed in the available studies. Therefore no reliable conclusion can be reached on these points. The few studies available give inconsistent or even contradictory results.

On the basis of the available studies, it is not possible to conclude that there is a specific threshold for minimum volumes which, if introduced, would improve the quality of care.

Data from intervention studies are lacking

The studies evaluated in the IQWiG literature analysis were register studies. Some of these analysed data from clinical registers, but most used administrative sources, with data originally collected for other purposes, such as billing. The IQWiG review only included studies that fulfilled specific minimum requirements for consideration of the distribution of risk factors (risk adjustment). The studies currently available with data from clinical registers and some studies with data from administrative data sources not only fulfilled these minimum requirements, but also considered the patient's prior clinical history and adjusted for this. However, apart from this risk adjustment, the quality of the statistical analysis and its documentation is deficient in most cases. Moreover, each study showed at least one deficit in the quality of the report or procedure applied. Data from intervention studies are needed for the evaluation of whether, and how, patient-relevant factors are influenced by specifying minimum volumes.

As most of the studies evaluated patient data from abroad, it is unclear whether the results can be transferred to the German health care system. The only study based on German data at least provided indications that, in this country, the quality of outcomes may be correlated with the volume of procedures. The relevance for Germany could be reliably established if, for example, the data for external inpatient quality assurance at the Federal Office for Quality Assurance (Bundesgeschäftstelle für Qualitätssicherung, BQS) or the available quality registers from professional societies could be analysed for this specific issue.

Contact: info@iqwig.de