The most common data flying around about ASDs is that if currently is estimated to occur in 1 in 150 births each year.  It seldom results directly in death, but presents usually as a serious handicap to social integration for those who "contract" it.  These difficulties can remain pronounced for a lifetime although inroad are being made towards recovery (although percentages of effective recovery stated vary considerably).

As a basis for comparison, I've been trying to get some pre-vaccine statistics on the various childhood diseases and am finding this surprisingly difficult (at least just from home on the web).  The raw data I can find comes from Canada (The Public Health Agency of Canada Website).

Here is what I've found so far:

Measles:  "Before measles vaccine became available, virtually all children contracted measles; as estimate 135 million cases with about 7-8 million deaths globally each year... Currently, more than 30 million people are affected each year by measles. Globally, in 2003, it was estimated that there were 530 000 measles deaths, the majority of them children." (source url:  http://www.phac-aspc.gc.ca/tmp-pmv/info/measles_e.html#profile)  This calculates to a global pre-vaccine rate of death per case of about = 1 in 19 and a current rate of death per case of about = 1 in 56.  However, statistics regarding deaths from complications of the vaccine were not mentioned.

Mumps:  "Since the approval of vaccine in 1969, the number of reported mumps cases has decreased by greater than 99% from an average of 34,000 cases reported per year in the early 1950s to under 400 cases per year in the early 1990s."  (source url: http://www.phac-aspc.gc.ca/im/vpd-mev/mumps-eng.php).  According to Statistics Canada, the estimated population of Canada in 1969 was about 21 million people (making the rate of incidence about 1 in 617 overall); however, a breakdown regarding infant risk was not available.  The PHAC website also indicates that "About 40% of those infected develop acute parotitis [i.e. glandular swelling], which is unilateral in about 25% of cases." and goes on using similar "big" medical words which should all be looked up in a dictionary before deciding - it seems to me that some of the words being used sound more threatening that the actual condition being described.

Rubella: "An MMR immunization program for all infants was introduced in Canada in April 1983.  The average number of rubella cases reported decreased from approximately 5,300 (1971-1982) to fewer than 30 cases per year (1998-2004).  The average annual incidence decreased from 0.08 per 100,000 in 1998 to 0.03 per 100,000 in 2004 (range 0.02-0.09 per 100,000 per year)." (source URL: http://www.phac-aspc.gc.ca/im/vpd-mev/rubella-eng.php).  No, these are not a typo.  The population of Canada in 1977 (the last year on the StatsCan Table I found) was about 23 million.  Either something is wrong with the data or the incidence of rubella in Canada was extremely low.

Conclusion - I think that better raw data should be available to the public in an easy to understand format so that the comparative risks of vaccinating vs. not vaccinating can be readily determined.  At present, it seems to me that it is extremely difficult for any member of the public to make a truly informed decision on this matter.