Tornadoes are very "streaky." That is, any given area can go long periods with few or no tornadoes, then a favorable weather system can produce a whole lot in a short time. A single outbreak can produce an average month's or even year's worth of tornadoes for an area in a matter of days.
Both large-scale factors (e.g., where the jet stream is) and small-scale factors (e.g., boundaries from previous storms) have a significant impact on whether and where tornadoes are produced. And even those large-scale factors can shift around a lot.
Because of this, exceptions are the norm when it comes to tornadoes, and it's difficult to identify patterns in the data. That's not to say it can't be done - it just would take a lot of careful work. (Including that the researcher would need to account for how tornado reporting has changed over time, as mentioned elsewhere). And therefore, one month of unusual tornado activity in the Midwest doesn't tell us much one way or the other, because "unusual tornado activity" is kind of always happening.
That said, there are long-term factors that could affect where tornadoes are most likely. For example, Boris made a very pertinent suggestion earlier that a shift in temperatures has the possibility of moving where tornadoes are more likely. And indeed, global temperatures can be changed by both natural causes (e.g., the Milankovitch cycles that affect how much energy the Earth receives from the sun) and man-made causes (e.g., the increase in greenhouse gases in our atmosphere from burning fossil fuels, which reduces heat radiated to space). In the geological record, we can see that Earth's temperature has varied a lot over its history: there are times with glacial sediments at the equator, and times with palm trees at the poles.
So is it possible that tornado probabilities are shifting location? Absolutely. But are we actually seeing that today? Significant work would need to be done to determine whether that's the case (and it's even possible that we don't have good enough data to really determine that).
"Streaky" is an excellent word here, and you see it from the synoptic to mesoscale level.
And sometimes it is non-linear. I always say in wx, "when it is good, it can be *really* good!" Meaning it can goes off the scale as to high-end and relentless activity (not just w/ tornadoes, any type of event).
Case in point, April 2011. Sure, the Superoutbreak at the end of the month everyone recalls, but outside of the event, there were over 500 other tornadoes that month, giving by *far* the most tornadoes in any month (#2 is 216 less than April 2011 759 tornadoes). No April up to this time had had 300 tornadoes, let alone over 700! And we did not have another April w/ at least 300 tornadoes until 2024.
Or look at that wild 9-day period May 3-11, 2003, in the Plains/Midwest -- 363 tornadoes w/ 62 F2+.
Even within an record event, you can get other not as apparent records that can go over the top. For instance, between midnight and noon April 27, 2011, 76 tornadoes (5 EF3s) occurred in the South w/ a single QLCS. This is the biggest morning tornado outbreak on record by a sizable margin, and then you had the main show later that day. That shows truly how ideal and epic the set up was that day.
Or those days where it did not look synoptically evident until the last minute. May 3, 1999 in OK is an excellent example, and being a more local "superoutbreak" that far exceeded anyone's expectations.
Given the virtually endless comnbinations of the global circulation, internal and external factors, and all other vagaries large and small, it is not too surprising we see what we see at times. Within normal climatic variation, the law of large numbers and averages dictate that such things are going to happen, and not necessarily w/ a regular frequecy. And looking at the entire wx range picture for a location on the globe like the CONUS, we get a remarkable range of all wx types, owing to our geography and latittude, and other less apparent factors come into play. For instance, the Atlantic is the most variable tropical basin in the world when it comes to TC activity over the long term.
Some of the best chase days have been slight risk days w/ no apparent focus or synoptic forcing. Perhaps the best example of this was the Chapman KS EF4 5/25/2016. Slight risk day, and for many who caught this tornado, it was there #1 chase all-time!
