Background and Importance Therapeutic angioplasty (TA) allows the dilatation of coronary stenosis. This minimally invasive procedure, often combined with diagnostic angiography, requires various costly medical devices (MD).

Aim and Objectives This study was conducted to highlight if there is a difference in interventional cardiology (IC) practices between physicians and estimate the associated costs.

Material and Methods First, data (characteristics of cardiac procedures (CP), number and type of MD (nMD)) were extracted from the cardiovascular information software (Cardioreport^®) on a period of three months. Then, Excel® was used to calculate the average MD cost (AC) per procedure for each operator. Finally, data were explored in RStudio® using the Multiple-Regression, Clustering with K-means and Ward’s method, in order to classify the similitudes and visualise the differences in practices.

Results Our sample of 74 CP includes 11 TA and 63 combined procedures. These, concerned 63 patients (average age 68 years), 26% of females and 76% of males among whom 10 had at least 2 CP. The AC estimated per procedure is €1125 of which €602 is not covered by additional payments (NCA) while €523 is covered (CA). Five physicians A/B/C/D/E operate in IC with a respective percentage of activity of 5%/9%/12%/22%/51%.

Multiple-Regression shows that cost of CP is explained at 89% by nMD and NCA cost as significant variables (adjusted R²=0.891 with P-value <5% (1.465e-12) so null-hypothesis can be rejected). Clustering and Ward’s dendrogram grouped procedures with common characteristics and showed that there were differences in practices among physicians. After excluding operators, A and E, clustering shows that operator C is singular in his practices (with a higher rate of complex procedures defined as longer than 90 minutes for combined procedures), while B and D have similarities in terms of choice of MD.

Conclusion and Relevance The hospital pharmacist, as MD expert plays a central role in managing consumption analysis. Expanding the sample to confirm the results would be more relevant. Thus, it would be interesting to explore the impact of communicating this work to physicians in order to homogenise their practices.

Conflict of Interest No conflict of interest