Kim Jinnett writes,

Chris Paulse suggested that I contact you about the use of mixed models (or lack thereof) in health economics research.

About 10 years ago Nigel Rice co-wrote a paper on the underuse of multilevel models in health economics [“Multilevel models and health economics”, with AM Jones, Health Economics, 1997; 6: 561-75.]. I’m finding that there still is not as much uptake of multilevel techniques as I would expect in the health economics field. There continues to be use of Huber correction and fixed effect modeling, but I don’t see much incorporation of mixed models that include both fixed and random effects.

Other than the multilevel listserve, are their groups/listserves specific to economists that use multilevel methods? I’m interested for a number of reasons, 1) because we’ll be issuing a research fellowship in the Fall and would like to direct it toward economists who use multilevel methods and 2) because I continue to seek out sources for explaining MLM to economists and for deepening my own understanding of the similarities and differences between traditional econometric modeling and MLM.

Any advice on newer papers that specifically address this topic (similarities and differences) and groups that might be inclined to have economists using MLM are very much appreciated.

My reply:

Multilevel models have been successfully in heath services research, I believe. I’m most familar with the work of Alan Zaslavsky, Sharon-Lise Normand, and others at the Dept of Health Care Policy at Harvard Medical School, but I’m sure there’s lots of other stuff too.

For teaching and learning multilevel models, I recommend our recent book. We don’t have health economics examples, but we have lots of other examples and also have much discussion of connections between multilevel and classical methods.

Finally, I don’t know of any listservs of multilevel modeling; perhaps someone else can help here.

P.S. Kim writes,

I am aware of a multilevel listserve that is very active, it doesn’t seem to have a lot of health economists on it however. Multilevel modelling discussion list [[email protected]]. It is a great list with terrific discussion around model application and statistical advances, though.