John Sides has a smart piece up looking at claims that enthusiasm, manifested in different turnout rates, was responsible for the Democrats' debacle in 1994. Turns out that turnout, contra James Morone, wasn't the culprit.
Of course, that still doesn't answer the question of whether Democrats should be worried about midterm elections, and what they should do if they are worried. I'm not sure that looking at 1994 tells us very much, however.
What were the ingredients of the GOP sweep in 1994? Gary Jacobson tells us that the Republicans put up good candidates and took advantage of nationalized issues, among them the 1993 budget bill, NAFTA, and a crime bill (which included a gun control measure), but more basically the intense dislike many voters, especially in the South, held for Clinton by 1994. That last bit also hints at one of the key things about the 1994 sweep -- there were a lot of districts that had voted Republican for president for a generation but still elected Democrats to Congress.
Health care in particular seems to play a relatively small role in any of that. On the one hand, Jacobson only refers to it in passing in his article on the 1994 elections. On the other hand, the demise of Clinton's health care bill in September 1994 is associated with the second big dip in Clinton's approval ratings (his approval rating went down about 10-15 points from May to September). But then again, what probably mattered a lot more was the first big dip. Clinton started with historically low approval ratings, and fell to 37% approval in June 1993, thanks to a number of missteps and general poor White House management. It was that dip that led to a series of strategic retirements and candidate recruitment decisions, which meant that Republicans had a lot of good candidate match-ups in 1994. So the 1994 landslide was the result of a good initial playing field for the Republicans; expectations in 1993 that the cycle would be a very good one for Republicans; and then a variety of issues and circumstances, of which health care failure was only one.
What lessons can Democrats learn from all of that? Well, the most important one has already been implemented; Obama had a very good start, compared to Clinton's disastrous transition and first few months. The second one is that it does behoove Democrats to try to keep Obama popular, which would most likely mean helping health care pass. Unfortunately for Obama, there's also evidence that voting with Clinton in high-profile votes was dangerous, and so if Obama is unpopular, marginal Democrats looking to 1994 should vote against it.
In other words, Democrats in marginal districts interested only in their reelection should probably hope that health care passes...but without their vote. Easier said than done, right? Or, as I've said all along, they could try to find a way to have health care pass, but without it becoming a controversial vote (by either getting GOP support or, barring that, making it clear that they were at least trying to find common ground).
Since the playing field is far better for Democrats than it was in 1994, and since so far Obama has avoided the problems that beset Clinton in his first year, I think overall Democrats are unlikely to get clobbered in the 2010 elections. But the point here is that the behavior of marginal Democrats is actually quite logical. They're behaving exactly how we should expect professional politicians to behave.