02/04/2009 — Estimating information growth in longitudinal clinical trials and settings leading to nonmonotonicity

Abby Shoben

In group sequential clinical trials, it is necessary to estimate the amount of information at an interim analysis relative to the amount that would be present at the final analysis. If only one measurement is made per individual, this is often the ratio of sample sizes of the interim and final analyses. However, as discussed by Wu and Lan (1992), when the statistic of interest is a change over time with longitudinal data, such an approach overstates the information. In this talk, we discuss other problems that can result in overestimating the information, such as heteroscedasticity. We further explore situations in which the true information is nonmonotonic. For example, when using an inefficient estimator, imbalance during the conduct of the trial can lead to nonmonotonic information growth. We demonstrate the consequences of these scenarios and provide suggestions for future studies.

01/28/2009 — Survival increases with CPR before defibrillation of out-of-hospital ventricular fibrillation or ventricular tachycardia: Observations from the Resuscitation Outcomes Consortium and difficulties in analysis.

Erin Gabriel

Context: Immediate defibrillation is the traditional approach to resuscitation of cardiac arrest due to ventricular fibrillation or tachycardia (VF/VT). Delaying defibrillation to provide chest compressions may improve survival.

Objective: To estimate the effect of the duration of Emergency Medical Services (EMS) cardiopulmonary resuscitation (CPR) prior to the first defibrillation attempt on survival in patients with out-of-hospital VF/VT.

Design, Setting, and Patients: Prospective multi-center observational study of EMS treated out-of-hospital non-traumatic cardiac arrest with first recorded rhythm VF/VT or “shockable” from December 2005 to May 2007.

Outcome Measure: Survival to hospital discharge as a function of EMS CPR duration prior to first shock.

Results: Of 13,601 EMS-treated cardiac arrests, 3,292 (24%) had VF/VT/shockable rhythm.  Of these, 1,661 (60%) had complete data for analysis.  Excluded were cases witnessed by EMS (n=143) or with first CPR duration > 315 seconds (n=184). Included patients were aged 0-100 years, 79% were men, 31% occurred in public locations, and 49% received bystander CPR. Time to arrival of first EMS unit was median 5:18 ( IQR 4:06, 6:48 ) minutes. Compared to the reference group of first EMS CPR duration < 45 seconds, the odds of survival was greater among patients who received between 46 seconds to 195 seconds of EMS CPR before first shock.  Odds of survival were more than 35% greater than the reference group during much of this EMS CPR interval.  An optimal EMS CPR duration was not identified within this range and no individual EMS CPR duration achieved statistical significance.  The benefit of EMS CPR before defibrillation was reduced when the duration of CPR exceeded 195 seconds.

Conclusion: In this observational analysis of VF/VT arrest, up to 195 seconds of EMS CPR prior to defibrillation was associated with a trend toward improved survival compared to < 45 seconds. Randomized trials are needed to evaluate this association given the potential for residual confounding in our analysis and to assess the impact on all initial rhythms.

01/21/2009 — Estimating the strength of association between a single nucleotide polymorphism in PTPN22 and type 1 diabetes – methods and results of a case-control study

Marlena Maziarz

Type 1 diabetes (T1D) is an autoimmune disease, with 15,000 youth diagnosed annually in the US alone. At diagnosis, an individual’s insulin-producing beta cells of the pancreas have been destroyed by their immune system, committing the patient to a lifetime of insulin injections to regulate their blood sugar level. Elucidating the underlying genetic risk factors of T1D, and the mechanisms associated with disease progression, are instrumental to treatment development aimed at delaying, or even preventing, the onset of T1D.

Our current focus is on a single nucleotide polymorphism in PTPN22, a gene encoding an important negative regulator of T-cell activation. This SNP has been identified in a genome-wide association study as a potential risk factor for several autoimmune diseases, including T1D. In this talk I will describe the methods and results of a logistic and polytomous logistic regression analysis to investigate the association of PTPN22 with T1D in the Swedish population, and examine whether this association is potentially modified by other known genetic risk factors or autoantibody status.

1/14/2009: Multi-state Markov models with applications to dementia: one approach

Amy Laird

Multi-state models are appealing tools for analyzing data about the progression of a disease over time.  In a 2007 Statistics in Medicine paper, Salazar et al consider a multi-state Markov chain with two competing absorbing states: dementia and death, and three transient non-demented states: congnitively normal, amnestic mild cognitive impairment,  and non-amnestic mild cognitive impairment. Using a polytomous logistic regression model with shared random effects, the authors derive the likelihood function and estimates for the effects of the covariates on transitions. The presence of the shared random effect complicates the form and maximization of the likelihood function. Three approaches to likelihood maximization are compared via simulation: one based on Gauss quadrature, one on importance sampling, and one on Taylor expansion. The approach with the best performance is used in an application to a longitudinal study on a cohort of cognitively normal subjects, followed annually for conversion to mild congnitive impairment of dementia, conducted at the Sanders Brown Center on Aging at the University of Kentucky. In this talk, I will review their work and discuss possible future directions.

1/7/2009: Spatial Clustering of MDS in the Seattle-Puget Sound Region

Michelle Ross

The myelodysplastic syndromes (MDS) are a group of clonal proliferative bone marrow disorders that result in dysmyelopoiesis and peripheral blood cytopenias. Despite the serious health outcomes of MDS, little is known about its causes.  The few known risk factors include radiation or chemotherapy treatment for a previous malignancy, and exposure to benzene and possibly other solvents.  It has also been suggested that the initial event in MDS may be infectious. The aims of our study were to investigate geographic clustering of MDS and first primary MDS (i.e., MDS cases with no previous cancer diagnoses) and to identify the location of potential clusters, for cases reported to the SEER program in the Seattle-Puget Sound region of Washington State from 2002-2006. Spatial modeling was performed using Poisson regression with non-spatial and spatial random effects. The Besag and Newell and Kulldorff methods were used for cluster detection. Our analyses did not indicate significant spatial dependence (clustering) either among all MDS cases or among first primary MDS. However, several local clusters were identified in Island and Pierce counties. Further investigation into potential causes of the identified clusters could potentially shed light on environmental risk factors for MDS.