Let's not confuse the meaning of a statistic with a performance measure for the individual team. If an automated clock is used that stops at a time that correlates well (over thousands of cases!) with the patient experience, then that is good enough for the statistic. That's because the purpose of the statistic is to measure and improve the patient experience. If you reduce the mean driving time, then the mean "at patient" time goes down as well. Every statistic can be improved. You say that the "time to scene" has no clinical significance, i.e. little correlation with outcome. That's an excellent argument. Often, it is what we call an extrinsic measurement, that provides a more "real-life" assessment of overall performance. For example, you can look at survival of heart attacks. However, when you do that, there will be a lot of factors that influence the outcome, not just how fast and how well-trained you guys are, but also, I'm guessing, whether there's a cath lab nearby and stuff like that. So, plenty of room for "finger-pointing". Now, as for a measure of an individual EMT's performance, or even the performance of a group of ambulances, such a measure is ill-suited for the reasons you have repeatedly stated in your blog. It is an abuse of statistics to apply the measure designed to contribute to a mean to an individual or a local team, especially if you compare the measure to a grand mean, as is done in your case. (e.g., you might have more public housing and more congested streets). So, while you're very right about criticising unsuitable measures, it's important for one's mental health to not abuse such a measure to assess performance during an individual call.