Chop‐Lump Tests for Vaccine Trials

This article proposes new tests to compare the vaccine and placebo groups in randomized vaccine trials when a small fraction of volunteers become infected. A simple approach that is consistent with the intent-to-treat principle is to assign a score, say W, equal to 0 for the uninfecteds and some postinfection outcome X > 0 for the infecteds. One can then test the equality of this skewed distribution of W between the two groups. This burden of illness (BOI) test was introduced by Chang, Guess, and Heyse (1994, Statistics in Medicine 13, 1807-1814). If infections are rare, the massive number of 0s in each group tends to dilute the vaccine effect and this test can have poor power, particularly if the X's are not close to zero. Comparing X in just the infecteds is no longer a comparison of randomized groups and can produce misleading conclusions. Gilbert, Bosch, and Hudgens (2003, Biometrics 59, 531-541) and Hudgens, Hoering, and Self (2003, Statistics in Medicine 22, 2281-2298) introduced tests of the equality of X in a subgroup-the principal stratum of those "doomed" to be infected under either randomization assignment. This can be more powerful than the BOI approach, but requires unexaminable assumptions. We suggest new "chop-lump" Wilcoxon and t-tests (CLW and CLT) that can be more powerful than the BOI tests in certain situations. When the number of volunteers in each group are equal, the chop-lump tests remove an equal number of zeros from both groups and then perform a test on the remaining W's, which are mostly >0. A permutation approach provides a null distribution. We show that under local alternatives, the CLW test is always more powerful than the usual Wilcoxon test provided the true vaccine and placebo infection rates are the same. We also identify the crucial role of the "gap" between 0 and the X's on power for the t-tests. The chop-lump tests are compared to established tests via simulation for planned HIV and malaria vaccine trials. A reanalysis of the first phase III HIV vaccine trial is used to illustrate the method.

[1]  M E Halloran,et al.  Study designs for evaluating different efficacy and effectiveness aspects of vaccines. , 1997, American journal of epidemiology.

[2]  J. Robins,et al.  Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models , 1999 .

[3]  T. Smith,et al.  Attributable fraction estimates and case definitions for malaria in endemic areas. , 1994, Statistics in medicine.

[4]  R. Aster,et al.  Heparin-induced thrombocytopenia associated with fondaparinux. , 2007, The New England journal of medicine.

[5]  M. Halloran,et al.  Randomization and baseline transmission in vaccine field trials , 2006, Epidemiology and Infection.

[6]  Mitchell J. Mergenthaler.  Nonparametrics: Statistical Methods Based on Ranks , 1979 .

[7]  J. Heyse,et al.  Reduction in burden of illness: a new efficacy measure for prevention trials. , 1994, Statistics in medicine.

[8]  M. Hudgens,et al.  Sensitivity Analysis for the Assessment of Causal Vaccine Effects on Viral Load in HIV Vaccine Trials , 2003, Biometrics.

[9]  D. Ankerst,et al.  Does Finasteride Affect the Severity of Prostate Cancer? A Causal Sensitivity Analysis , 2008, Journal of the American Statistical Association.

[10]  R. Steinbrook.  One step forward, two steps back--will there ever be an AIDS vaccine? , 2007, The New England journal of medicine.

[11]  D. Mehrotra,et al.  A Comparison of Eight Methods for the Dual‐Endpoint Evaluation of Efficacy in a Proof‐of‐Concept HIV Vaccine Trial , 2006, Biometrics.

[12]  Michael G Hudgens,et al.  On the analysis of viral load endpoints in HIV vaccine trials , 2003, Statistics in medicine.

[13]  D. Rubin,et al.  Principal Stratification in Causal Inference , 2002, Biometrics.

[14]  W. Tu,et al.  A Wald test comparing medical costs based on log-normal distributions with zero valued costs. , 1999, Statistics in medicine.

[15]  P. Gilbert,et al.  HIV-1 virologic and immunologic progression and initiation of antiretroviral therapy among HIV-1-infected subjects in a trial of the efficacy of recombinant glycoprotein 120 vaccine. , 2005, The Journal of infectious diseases.

[16]  James M. Robins,et al.  An analytic method for randomized trials with informative censoring: Part II , 1995, Lifetime data analysis.

[17]  M. Pallansch,et al.  Vaccine policy changes and epidemiology of poliomyelitis in the United States. , 2004, JAMA.

[18]  P A Lachenbruch,et al.  Comparisons of two‐part models with competitors , 2001, Statistics in medicine.

[19]  Andrea Rotnitzky,et al.  Sensitivity Analyses Comparing Outcomes Only Existing in a Subset Selected Post‐Randomization, Conditional on Covariates, with Application to HIV Vaccine Trials , 2006, Biometrics.

[20]  Kenneth H Mayer,et al.  Placebo-controlled phase 3 trial of a recombinant glycoprotein 120 vaccine to prevent HIV-1 infection. , 2005, The Journal of infectious diseases.

[21]  R H Lyles,et al.  Natural history of human immunodeficiency virus type 1 viremia after seroconversion and proximal to AIDS in a large cohort of homosexual men. Multicenter AIDS Cohort Study. , 2000, The Journal of infectious diseases.

[22]  Donald S. Burke,et al.  Human HIV Vaccine Trials: Does Antibody-Dependent Enhancement Pose a Genuine Risk? , 2015, Perspectives in biology and medicine.

[23]  M. Kieny,et al.  A review of human vaccine research and development: malaria. , 2007, Vaccine.

[24]  Joshua D. Angrist,et al.  Identification of Causal Effects Using Instrumental Variables , 1993 .