Smoking and SIDS

A grad student in Biology at McMaster University believes he and his team have discovered why smoking is related to SIDS. It sounds like it may also help explain why campaigns like Back to Sleep were so successful, since as the researcher says, "When a baby is lying face down in bed, for example, it should sense a reduction in oxygen and move its head. But this arousal mechanism doesn't work as it should in babies exposed to nicotine during pregnancy."

Again, I go to the lowest hanging fruit research opportunities when I read something like this, but I wonder if anyone thought to look at the effect of smoking on SIDS using the Big Tobacco legislation in the 1990s that resulted in significant price increases on cigarette packs as instruments for smoking. I did a quick google search, but I can't find anything, which means nothing since I tried two different searches. I could sick my grad students on this, but something tells me I'll lose interested in the project as soon as I hit publish.

To do that kind of study, what would you need? Preferably, monthly data on cigarette prices, monthly data on child mortality (probably the NBER Natality Data Files is what you'd want), and use the months after the legislation ended as the treatment variables, which I'm suspecting caused increases in prices of packs due to the huge settlement by Tobacco corporations. Also, the data would need to be geographically disaggregate, preferably by county. You would estimate a two-stage least squares model with price, which is endogenous, instrumented using the aforementioned treatment variables, which are then included as predicted values in the second stage of the regression in a child mortality regression. You would be looking for negative correlations between tobacco prices (per pack) and child mortality. You would to net out the Back to Sleep campaign, which was national, using year dummies.