[Relationship between hospital volumes and health outcomes: epidemiological evidence supporting the revision process of the Ministry of Health Decree no. 70 of 2 April 2015.]

Abstract

In this perspective, this work intends to establish whether statistical correlations exist between volumes of activity and outcomes of hospital care. The scope is limited to several clinical conditions and therapeutic procedures for which specific indicators are provided by the National Healthcare Outcomes Programme (Programma Nazionale Esiti - PNE). For each condition or procedure, the analysis shows the volume-outcome relationship by hospital centre by means of the Levenberg-Marquardt algorithm (software: XLSTAT). The existence of breakpoints is assessed through the use of segmented models (software: "segmented" R-Package). The results show a statistical correlation for the following: acute myocardial infarction (breakpoint: 91 hospitalizations per year; 95% CI: 81-101; p<0.0001); repair of an unruptured abdominal aortic aneurysm (breakpoint: 69 procedures per year; 95% CI: 52-86; p=0.146); lung cancer (breakpoint: 96 procedures per year; 95% CI: 60-132; p<0.01); knee arthroplasty (breakpoint: 91 procedures per year; 95% CI: 51-131; p=0.484). Conversely, the statistical analysis did not allow to accurately highlight a breakpoint for the isolated aorto-coronary bypass, percutaneous transluminal coronary angioplasty and hip arthroplasty. These results represent a useful knowledge contribution to support the revision process of the above-mentioned Decree. As regards the procedures that may not be currently assessed through this statistical analysis method, literature data is referred to that confirm that the current regulatory thresholds are in the safe range.