Most of Gregg Dinse’s research focuses on developing improved statistical methods for analyzing data from animal carcinogenicity studies. Several of his recent projects include the following:
Distinguishing treatment effects on the number of induced tumors from treatment effects on the tumor detection times in cancer chemoprevention experiments.
Incorporating explanatory variables, historical data and expert judgements when analyzing tumor incidence in survival/sacrifice studies.
Developing simple yet flexible tumor incidence estimators for studies with limited sacrifice data and no information on cause of death or tumor lethality.
Making simultaneous inferences about tumors at multiple sites, while accounting for informative censoring and within-animal correlations among the tumor onset times.
Adapting order-restricted inference techniques to develop an improved alternative to a conventional survival-adjusted quantal response test.
Extending this new survival-adjusted quantal response test so that it incorporates data from historical control animals.
Using kernel smoothing techniques to estimate the hazard function when some cause-of-death indicators are missing at random.
Adjusting for covariates when some censoring indicators are missing at random.
Jointly modeling animal growth and tumor onset to separate the direct effect of treatment on tumor incidence from its indirect effect via changes in body weight.
Using Bayesian methods to improve the efficiency of inferences about age-specific hazard rates subject to shape constraints.
Selected Publications
Dunson, DB and Dinse, GE: Distinguishing effects on tumor multiplicity and growth rate in chemoprevention experiments. Biometrics 56: 1068-1075, 2000. [Abstract]http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=11129462
Dunson, DB and Dinse, GE: Bayesian incidence analysis of animal tumorigenicity data. Journal of the Royal Statistical Society, Series C 50: 125-141, 2001.
Parise, H, Dinse, GE, and Ryan, LM: Flexible estimates of tumor incidence for intermediately lethal tumors in a typical long-term animal bioassay. Journal of the Royal Statistical Society, Series C 50: 171-185, 2001.
Dunson, DB and Dinse, GE: Bayesian models for multivariate current status data with informative censoring. Biometrics 58: 79-88, 2002. [Abstract]http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=11890330
Peddada, SD, Dinse, GE, and Haseman, JK: A survival-adjusted quantal response test for comparing tumor incidence rates. Journal of the Royal Statistical Society, Series C 54: 51-61, 2005.
Peddada, SD, Dinse, GE, and Kissling, GE: Incorporating historical control data when comparing tumor incidence rates. Journal of the American Statistical Association (in press).
Wang, QH, Dinse, GE, and Liu, C: Hazard function estimation with cause-of-death data missing at random (submitted to the Journal of the American Statistical Association).
Wang, QH and Dinse, GE: Regression analysis with censoring indicators missing at random.
Dinse, GE and Dunson, DB: Causal inferences in carcinogenicity studies.
Dunson, DB and Dinse, GE: Bayesian analysis of constrained hazard functions.