Optimal size and throughput of tertiary services: A Case Study in Renal Transplant and Cardiac Surgery in NSW, CHERE Discussion Paper No 5

The research evidence on the relationships between patient throughput and costs and outcomes indicate that, for cardiac and vascular surgery, outcomes are better in units with high rates of patient activity. However, the gains to increasing throughput are most substantial at rates of activity below those currently being achieved in New South Wales (NSW). Research evidence on economies of scale is sparse and unimpressive but suggests that further cost savings are minimal at rates of throughput in excess of 100 cases per annum. The impression from practice in NSW is that larger units have lower average costs. However, this is not necessarily a scale effect and does not constitute an argument for centralising services. Projected increases in the rates of activity will ensure that whatever economies of scale which do exist are realised without the need for reorganisation. In the short term, gains in efficiency may be better realised by management audit within units shown to have higher costs or poorer outcomes. The available research evidence on cardiac surgery suggests that units should be performing at least 300 open heart procedures per annum. It does not support arguments that units should be performing a minimum of 900 procedures per annum unless this includes closed heart procedures in addition to open heart surgery. For renal transplantation, the evidence is less conclusive and clinical opinion more divided. There are gains to be made in centralising surgical procedures but the extent of these gains is small in comparison to the total costs and outcome of managing the patient in end stage renal failure. Patient costs are dominated by the need for ongoing immuno-suppressive therapy and it has been argued that final outcome is influenced more by the medical management of the patient before and after surgery. The effect of centralisation of surgical services on organ supply also needs to be considered. The small benefit of centralised surgery may be more than offset by a reduction in organ donation. On the evidence available there is little to suggest that at current rates of activity, scale effects on costs and outcomes should influence decisions about the size or location of tertiary services. Economies of scale in cardiac surgery will be realised naturally. Variations in cost and outcome do exist in both cardiac surgery and renal transplantation but do not appear to be related to patient throughput in ways which can be exploited through centralisation. Differences in the performance of units appear independent of scale. The reasons for these differences (efficiency or case-mix) warrant closer examination by management or peer-review in-hospital rather than by the macro-approach adopted here. Two extensive sources of data containing information on the outcomes of surgery (the databases held by the National Heart Foundation and ANZDATA) could not be fully exploited for reasons of confidentiality. The feasibility of linking these data sets to estimates of cost and subjecting the information to independent confidential analysis should be explored.