Page updated on July 19th, 2016

We identified the most appropriate conditions and selected the most suitable procedures for the cardiovascular pilot project from the inventory and literature research. Several discussions based on preliminary results presented at three meetings of the Scientific Committee of EUPHORIC led to the setup of the following specific aims of the cardiovascular pilot:

  1. To define a simple set of factors that determine quality of health care outcome (in-hospital case-fatality) in patients who received thrombolysis, underwent coronary angiography, or percutaneous interventions or were treated for myocardial infarction or unstable angina. These indicators will be analyzed in the context of characteristics at individual, hospital and country levels.
  2. To develop a set of tools (mathematical functions) to benchmarking European hospitals by their observed indicators (in-hospital case fatality) according to the expected adjusted risk of the outcome, that provides systematic information to end-users (doctors, health staff, health administration, decision makers, policy makers, EU population and public health stakeholders).
  3. To test the functions that estimate the indicators to information obtained routinely for administrative purposes.
  4. To develop and update a systematic review of the literature on the efficacy of GPIIb-IIIa inhibitors in ACS (pdf).

ACS was the easiest to address since admission is always required. There were many ongoing and well completed population-based registers, and in-hospital procedure use and outcome were relatively easy to monitor. Patients admitted for an ACS receive a discharge diagnosis of MI either with or without q-wave in the electrocardiogram, unstable angina (UA), and, to a lesser extent, stable angina or non-coronary ischaemic disease. ACS does not exist as an entry in the international classification of diseases (ICD), which is used in discharge diagnosis in European hospitals. MI and UA have a high social impact. Moreover, they are easy to identify and clearly defined according to the ICD. Their management includes a number of procedures that suit the objectives of the EUPHORIC study. After thorough discussions and considerations based on the analysis of European register data, we decided to analyze the outcome (in-hospital mortality) for the following condition management and related procedures:

  1. Condition management
    1. General MI management
    2. Unstable angina general management.
  2. Coronary heart disease procedures
    1. Coronary angiography
    2. Thrombolysis
    3. Percutaneous intervention (angioplasty with or without stenting).

Hospital, country, and individual characteristics are probably at the origin of varying outcomes of procedures used in ACS. Therefore, a number of such characteristics were taken into account in the analyses.

Particularly, we chose:

  1. At individual level
    1. Sex
    2. Age
    3. Comorbidity (diabetes, hypertension, and previous history of cardiovascular disease).
  2. At country level
    1. Yearly per capita Gross National Income
    2. Age standardized CHD mortality rate
    3. Life expectancy at birth.
  3. At hospital level
    1. Onsite catheter laboratory
    2. Onsite cardiac surgery
    3. University hospital.

We reached some important conclusions supported by our preliminary and advanced analyses:

  1. 30-day and 6-month events and mortality rates are not suitable for analysis with this database given the high number of missing values, and the low probability that hospitals have such information available. Consequently, in-hospital mortality seems to be the most robust endpoint.
  2. AMI and unstable angina management, as well as coronary angiography, thrombolysis and PCI use are fine procedures to assess using benchmarking functions.
  3. There are no important differences in any model in variable coefficient estimates (as seen in the OR and their 95% confidence intervals) by the entire population as compared with those obtained when hospitals with less than 50 patients were excluded.
  4. There are no important differences in any model in variable coefficient estimates (as seen in the OR and their 95% confidence intervals) by the entire population as compared with those obtained when countries with less than 100 patients were excluded.
  5. Coefficients (ln(ORs)) are suitable for our purpose of attempting to benchmark European hospitals in this pilot phase.
  6. Prospective data will be needed to validate this initial proposal.

Based on the results of the statistical analysis, the IMIM working group, in cooperation with CASPUR and ISS, has been implementing a "Coronary Heart Disease Outcome Benchmarking Function" for European hospitals in which coronary heart disease patients are treated. The function will be presented and discussed within the EUPHORIC project working group and made publicly available after its approval.

Technical reports and Deliverables