One way to tease out the effects of daylight saving time is to look at
the health records of people near a time zone boundary, as you get a
one-hour civil-time difference for free.
This could, of course, be confounded somewhat in boundary cases where there are mixed economic ties. A friend who grew up near the Dearborn/Franklin County line in southeastern Indiana (prior to statewide observance of DST starting in 2006) spoke of his home operating on a combination of "School Time" (EST) and "Fast Time" (EDT) whenever clocks differed. It seems likely that folks faced with such a situation would pick one system as their primary, while simply contending with the apparent seasonal changes in the other. I'd think that type of disruption could be just as taxing.
As one data point, the Eastern/Central boundary in the US is effectively invisible when it comes to prime-time television broadcasts. Does the "late local news" report coming on an hour earlier affect bed times, and correspondingly when public affairs and activities are conducted in the evenings? If strong ties to the east coast in certain areas were to affect work and school times, too, it would nullify the "social time difference", but the health effects described could actually be exacerbated in the western extremities of a two-hour-wide "cultural time zone".