Phil Plait recently mentioned that a supernova needs to be less than 100 lightyears away to be harmful to our planet/civilization.
This was in context of a discussion about supernovae in the Trumpler 15 cluster, a "collection of thousands of stars packed into a volume of space only a few light years across." It seems to me that such clusters would be hostile to civilizations like ours developing.
If a large proportion of systems which otherwise could give birth to a civilization are in clusters with a high rate of supernovae, we should lower the one of the "fraction of life/civilization" coefficients in the Drake equation.
Additionally, if the worlds with many nearby interstellar neighbors are the same worlds most likely to be affected by supernovae, then the fraction of civilizations that become star faring should be reduced as the civilizations that are safest from supernovae have the largest distances to overcome to become starting.