Dr. Heller's current research interests include the analysis of phase 2 clinical trial data, survival analysis, and the design of animal studies. Phase 2 clinical trials are undertaken to provide initial evidence of treatment efficacy. A test statistic that accounts for the heterogeneity in the patient population is proposed. It is hypothesized that adjusting for individual patient risk will improve the accuracy of the test for treatment efficacy relative to conventional tests. In survival analysis, a measure based on the concordance index is being developed to assess the predictive accuracy of the proportional hazards model. Prediction models have become important tools in the assessment of patient risk and in the subsequent determination of treatment strategy. Currently, a reliable measure of the predictive accuracy for right censored regression models does not exist. In the area of small sample animal designs, power/sample size calculations are typically undertaken using the asymptotic distribution of the test statistic. When the comparative study includes more than two groups, the asymptotic approximation is not accurate. An alternative approach, using a rank based test statistic, is an exact power calculation. When the number of groups is greater than two, the number of calculations required to perform an exact power calculation will often exceed the limit of viability in the current computing environment. An intermediate approach, using Monte Carlo resampling, is under development to approximate the exact power function of a rank test statistic under the family of Lehmann alternative hypotheses.
Selected Bibliography:
Heller G. Incorporating follow-up time in M-estimation for survival data. Lifetime Data Analysis 2004; 10:51-64.
Heller G, Venkatraman ES. A nonparametric test to compare survival distributions with covariate adjustment. Journal of the Royal Statistical Society, Series B 2004; 66: 719-733.
Gonen M, Heller G. Concordance probability and discriminatory power in proportional hazards regression. Biometrika 2005; 92: 965-970.
Heller G. Power calculations for preclinical studies using a K-sample rank test and the Lehmann alternative hypothesis. Statistics in Medicine 2006; 25: 2543-2553.
Heller G. Smoothed Rank Regression with Censored Data. Journal of the American Statistical Association 2006; (in press).
