WHAT THIS PAPER ADDS Patient selection should be a key component of carotid surgery reports. A calculator was created where the presence and recentness of symptoms, sex, increasing stenosis severity, and complication rates could all be combined into a single figure. The calculator was tested on real life material and some theoretical inclusion scenarios. Single hospital treatment development over time and benchmarking between hospitals can be monitored when registered data includes these parameters, ultimately leading to improved patient selection.Objective: Considering carotid endarterectomy (CEA), reporting treatment delay, symptom status, and surgical complication rates separately gives an incomplete picture of efficacy; therefore, the aim was to combine these factors and develop a reporting standard that better describes the number of potentially prevented strokes. With a real life cohort and theoretical inclusion scenarios, the aim was to explore the stroke prevention potential of different carotid practices. Methods: Landmark studies for symptomatic and asymptomatic patients were revisited. By using published estimates of treatment effect, a simplified calculator was designed to assess the five year stroke prevention rate per 1000 CEAs (stroke prevention potential [SPP], range 0e478), including the presence and recentness of symptoms, sex, increasing stenosis severity, and complication rates. Patients operated on for carotid stenosis at Helsinki University Hospital (HUH) between 2008 and 2016 were collected from a vascular registry (HUSVASC) and categorised according to the model. The local annual complication rate was re-evaluated and added to the model. The HUH patient cohort was incorporated into the SPP model, and changes over time analysed. Finally, theoretical changes in patient selection were compared in order to explore the theoretical impact of patient selection and shortening of the delay. Results: Fifteen hundred and five symptomatic and 356 asymptomatic carotid stenoses were operated on with stroke plus death rates of 3.6% and 0.3%, respectively. The proportion of CEAs performed within two weeks of the index event increased over the follow up period, being 77% in 2016. The SPP increased from 123 in 2008 to 229 in 2016. Theoretically, 350 ischaemic strokes were prevented in the period 2008e16, with 1861 CEAs. Conclusions: National and international comparison of different CEA series is irrelevant if the inclusion criteria are not considered. A calculator that is easy to apply to large scale high quality registered data was developed and tested. SPP was found to increase over time, which is a probable sign of improved patient selection and an increased number of strokes prevented by the CEAs performed.