Here are the two I found in really brief search.
Source: Ho J, Lindquist M: Time saved with the use of emergency warning lights and siren while responding to requests for emergency medical aid in a rural environment.
Prehosp Emerg Care. 2001 Apr-Jun;5(2):159-62.
ESULTS: The 32 responses with L&S averaged 105.8 seconds (1 minute, 46 seconds) faster than those without (95% confidence interval: 60.2 to 151.5 seconds, p = 0.0001). The time difference ranged from 425 seconds (7 minutes, 5 seconds) faster with L&S to 210 seconds (3 minutes, 30 seconds) slower with L&S. CONCLUSION: In this urban EMS system, L&S reduce ambulance response times by an average of 1 minute, 46 seconds. Although statistically significant, this time saving is likely to be clinically relevant in only a very few cases. A large-scale multicenter L&S trial may help address this issue on a national level.
To be fair, I also found this article that I am not sure of the validity of since it had such an awfully small sample size to make such a broad statement and also I also question the standard they are using for "rural" among other things (average posted speed limits on the roads, the terrain and road contour, etc) that are not addressed in the abstract that could cause a significant difference in response times where you are driving strictly adherent to the traffic laws and then when you are violating them when running lights and sirens. Also the fact that they spread 67 studied calls over almost two years seems kind of odd....not sure why they would do that. I'm still a little too tipsy to pull it and look at it in depth, but here you go.....:
Brown LH, Whitney CL, Hunt RC, Addario M, Hogue T: Do warning lights and sirens reduce ambulance response times?
Prehosp Emerg Care. 2000 Jan-Mar;4(1):70-4.
RESULTS: Sixty-seven runs were timed during a 21-month period. The average code 3 response interval was 8.51 minutes. The average code 2 response interval was 12.14 minutes. The 3.63 minutes saved on average represents significant time savings of 30.9% (p < 0.01). Shorter runs had higher time savings per mile than the longer runs. Run distance was the only variable that was statistically significant in affecting time saved during a code 3 response. CONCLUSION: Code 3 operation by EMS personnel in a rural EMS setting saved significant time over code 2 operation when traveling to a call.
The main thing that is lacking here is evidence that the response time is going to be
clinically relevant, as the first abstract mentioned. Also, since we're discussing the benefits of transport to trauma centers, etc by helicopter versus ground, I am going to try to find an article I read several years back that looked at transport time (as opposed to response time) for ground ambulances with and without lights and sirens since that is what is most pertinent here and not looking at a separate situation and trying to extrapolate to another. Sorry if this is a little disjointed....