Michael J. Daniels

My statistical interests are quite varied. My main research interest so far has been on hierarchical modelling, specifically as it can be used in health services research. Much health services data exist in a hierarchical framework: for example, patients within hospitals, hospitals within states, states within regions. Thus, a natural way to model this type of data would be a hierarchical model on which higher levels are partially determined by the lower levels. The usual approach to fitting these models is Markov Chain Monte Carlo techniques, which can often be quite computer intensive. As a result, I am currently working on approximations and alternative ways to fit these models with other members of the department. In addition, I am exploring various prior distributions for covariance matrices, with goal of reducing the number of parameters in the aforementioned models and obtaining more stable parameter estimates.

I have also engaged in some applications of hierarchical modelling to evaluate potential surrogate markers in clinical trials. Good markers might be used to help determine the value of pursuing therapies further, would allow therapies to be evaluated more rapidly and cost-effectively with subject earlier in the cycle of infection, and might be used by regulatory bodies to allow preliminary approval of a drug based on marker effects. However, most clinical trials do not provide enough information to assess the adequacy of the potential surrogate marker. I am currently exploring ways to draw conclusions about the potential markers using the information from many clinical trials through meta-analytic techniques.

Some Related Publications:

Daniels, M., Devlin, B., Roeder, K. (1997) 'Revisiting The Bell Curve's Nature/Nurture Notions' in Galton Redux: Eugenics, Intelligence, Race, and Society: A Review o f The Bell Curve: Intelligence and Class Structure in American Life (tentative title), editors S. Fienberg, B. Devlin, D. Resnick, and K. Roeder.

Daniels, M.J. and Gatsonis, C.A. (1996) 'Hierarchical Polytomous Regression Models for Health Care Data,' Submitted.

Daniels, M.J. and Hughes, M.D. (1996) 'Meta-analysis for the Evaluation of Potential Surrogate Markers,' Submitted.