The 2011 Deming Student Scholar award recipients are:

Dungang Liu, Rutgers University,

Matthew White, University of Pennsylvania, School of medicine, and
 Kai Zhang, University of Pennsylvania, Wharton.
They will be giving a brief presentation on their work on Monday, December 5, 2011 at 4 pm.
Abstracts:
Exact meta-analysis approach for the common odds ratio of 2 by 2 tables with rare events
Dungang Liu, Regina Liu and Minge Xie
Meta-analysis is considered by some as the only approach for drawing reliable inference in the analysis of similar 2X2 tables with rare events. An often occurring situation in rare event studies is that a non-negligible portion of the studies may observe zero events in both arms, generally referred to as zero total studies. In this situation, conventional meta-analysis approaches either exclude zero total studies from the analysis or add some arbitrary correction factor, such as 0.5, to the zero events. But both practices are known to have undesirable consequences in inference. In this paper, we propose an exact meta-analysis approach for the common odds ratio, which can incorporate all 2X2 tables in the analysis without using any artificial corrections for zero events. This approach combines significance functions obtained from an exact test associated with each of the 2X2 tables. This is different from conventional approaches of combining point estimates. For zero total studies, the proposed approach can reflect the appreciable difference between the impacts of large trials and small trials. For example, the impact of a zero total study with 1000 cases and 1000 controls is clearly different from that with 10 cases and 10 controls. We also provide an explicit formula for computing the actual test size, which enables us to estimate the actual test size from the data. Moreover, we show that the approach is efficient in large sample settings. Numerical studies using both simulated and real data show that the proposed approach is superior to the Mantel-Haenszel method and Peto method in the presence of rare events.
Adjustment for Measurement Error in Evaluating Diagnostic Markers by Using an Internal
Reliability Sample
Matthew T. White and Sharon X. Xie
Repeated measurements of a biomarker within subjects are obtained with measurement error due to imperfect lab conditions or temporal variability in subjects. Using an internal reliability sample of the biomarker, we propose a bias-correction approach for estimating a variety of diagnostic performance measures including sensitivity, specificity, the receiver operating characteristic curve, positive and negative predictive values, and positive and negative diagnostic likelihood ratios when the biomarker is subject to measurement error. We derive the asymptotic properties of the proposed estimators and show that they are consistent and asymptotically normally distributed.
We demonstrate through extensive simulations that the proposed approach removes the bias due to measurement error and performs much better than the naïve approach (which ignores the

measurement error) in estimating these diagnostic measures. We also derive the theoretical bias of naïve estimates. The proposed method has broad biomedical applications, such as the evaluation of candidate biomarkers in the drug discovery process and in clinical trials. Using data from a biomarker study in Alzheimer's disease, we show that naïve estimates of the diagnostic measures are biased towards estimates produced when the biomarker is ineffective (i.e., when sensitivity equals 1 - specificity). Thus, in the presence of measurement error, naïve estimates could give the erroneous impression that the biomarker is ineffective. We recommend collecting an internal reliability sample during the biomarker discovery phase in order to adequately evaluate the performance of biomarkers with careful adjustment for measurement error.
Using Split Samples and Evidence Factors in an Observational Study of Neonatal Outcomes
Kai Zhang, Dylan S. Small, Scott Lorch, Sindhu Srinivas, Paul R. Rosenbaum
During a few years around the turn of the millennium, a series of local hospitals in Philadelphia closed their obstetrics units, with the consequence that many mothers-to-be arrived unexpectedly at the city’s large, regional teaching hospitals whose obstetrics units remained open. Nothing comparable happened in other US cities, where there were only sporadic changes in the availability of obstetrics units. What effect did these closures have on mothers and their newborns? We study this question by comparing Philadelphia before and after the closures to a control Philadelphia constructed from elsewhere in Pennsylvania, California and Missouri, matching mothers for 59 observed covariates including year of birth. The analysis focuses on the period 1995-1996, when there were no closures, and the period 1997-1999 when five hospitals abruptly closed their obstetrics units. Using a new sensitivity analysis for difference-in-differences with binary outcomes, we examine the possibility that Philadelphia mothers differed from control mothers in terms of some covariate not measured, and perhaps the distribution of that unobserved covariate changed in a different way in Philadelphia and control-Philadelphia in the years before and after the closures. We illustrate two recently proposed techniques for the design and analysis of observational studies, namely split samples and evidence factors. To boost insensitivity to unmeasured bias, we drew a small random planning sample of about 26,000 mothers in 13,000 pairs and used them to frame hypotheses that promised to be less sensitive to bias; then these hypotheses were tested on the large, independent complementary analysis sample of nearly
240,000 mothers in 120,000 pairs. The splitting was successful twice over: (i) it successfully identified an interesting and moderately insensitive conclusion, (ii) by comparison of the planning and analysis samples, it is clearly seen to have avoided a exaggerated claim of insensitivity to unmeasured bias that might have occurred by focusing on the least sensitive of many findings. Also, we identified two approximate evidence factors and one test for unmeasured bias: (i) factor 1 compared Philadelphia to control before and after the closures, (ii) factor 2 focused on the years
1997-1999 of abrupt closures and compared zip codes with closures to zip codes without closures,
(iii) and the test for bias focused on the years 1995-1996 prior to closures and compared zip codes which would have closures in 1997-1999 to zip codes without closures in 1997-1999 --- any

ostensible effect found in that last comparison is surely bias from the characteristics of Philadelphia zip codes in which closures took place. Approximate evidence factors provide nearly independent tests of a null hypothesis such that the evidence in each factor would be unaffected by certain biases that would invalidate the other factor.