Sorry for the delay in answering, but needed to give this some

  Thank you for your letter and copy of the MOMA letter on blind
accessibility at two-lane roundabouts.  Accessibility policy at intersections
for the blind is extremely important.  However, traffic signals with or
without a pedestrian phase are not a guarantee of safety.  Your suggestions
are certainly sensible, but raise the question of providing corridors and
networks of accessibility not just a particular intersection.  A roundabout
proposal must not be forced to meet a higher standard of safety than other
types of intersections, although a roundabout proposal may well be the basis
for bringing into question a network or corridor approach.    Ideally, there
would be grade separation of motorized and nonmotorized transportation.

        In terms of roundabouts, there is a study of pedestrians and
accidents before and after at about 300 Australian two-lane roundabouts in
busy commercial areas.   Pedestrian injuries and severity decreased.   That
study would be instructive (I do not have a citation for the study-- Michael
Wallwork would be of help in this case).  Clearly the ideal would be to have
statistically significant data on the blind and their pedestrian accident
experience at traffic signals/stop control/roundabout conditions--probably
unlikely to be developed because of the lack of data.   However, a collection
of data from real-world experience of the blind would be wortwhile at the
national level.

        I am not sure this is the answer you are looking for, but the issue
is one much larger than roundabouts and must be addressed in a far more
general way, not just when an individual roundabout replaces a traffic

        Tony Redington