10/21/2009 — The Stepped Wedge Design: Outstanding Issues

Tanya Granston

The stepped wedge design of cluster-randomized trials has been growing in popularity and becomes increasingly relevant due to the need to efficiently evaluate the rollout of interventions that are individually efficacious in a community setting.  The design becomes especially practical for HIV/AIDS prevention and intervention trials, as the areas in most need are in resource limited settings where it is very likely not feasible to introduce interventions all at once and where there are substantial clusters/communities/groups to provide answers regarding intervention effects.  However, there remain limitations and outstanding issues to this design, which could potentially limit its usefulness where it’s most needed, and there are interesting extensions to the design that need to be explored.

10/14/2009 — A Quasi-Likelihood Approach to Genome Wide Association Studies

Matthew Bryan

A common issue in Genome Wide Association Studies (GWAS) is relatedness between subjects due to restricting sampling to a confined homogenous population.  As a result, the genetic data can be correlated across subject which may not be accounted for by common genetic testing methods.  This presentation will discuss a method suggested by Timothy Thornton and Mary McPeek that is designed to account for relatedness between subjects for genetic data.  The method uses a quasi-likelihood model approach to generate a quasi-likelihood score statistic.  This statistic can be shown to be locally most powerful for testing genetic associations with a phenotype.  Thus the method is expected to provide higher power for detecting small genetic effects.

10/07/2009 — Glucose, Insulin, and Incident Hypertension in the Multi-Ethnic Study of Atherosclerosis

Greg Levin

Diabetes mellitus and hypertension commonly coexist, but the nature of this link is not well understood. We examined whether diabetes and higher concentrations of fasting serum glucose and insulin were associated with increased risk of developing incident hypertension in the community-based Multi-Ethnic Study of Atherosclerosis. In addition, we explored whether these associations may be mediated by kidney disease and diminished arterial elasticity. Of 3,513 participants free of hypertension at baseline, 965 (27%) developed incident hypertension over 4.7 years median follow-up. Compared to participants with baseline fasting glucose < 100 mg/dL, relative risk of hypertension was 1.16 [95% CI: 0.96, 1.40] among participants with glucose 100-125 mg/dL and 1.41 [1.17, 1.71] among participants with diabetes, adjusting for age, gender, race/ethnicity, education, physical activity, smoking, alcohol use, body mass index, and waist circumference (p=0.0015). Higher concentrations of fasting glucose and insulin within their normal ranges were also associated with increased risk of incident hypertension.  Further adjustment for serum cystatin C, urine albumin-creatinine ratio, and arterial elasticity measured by tonometry attenuated the magnitudes of these associations by approximately 50%. In conclusion, diabetes and higher concentrations of glucose and insulin were associated with increased risks of developing hypertension. The results also suggest that these associations may be mediated by damage to the kidney and arterial wall, but this hypothesis should be studied further.

In the process of carrying out this study, we encountered several challenging statistical and epidemiological questions. These included discussions about the merits of baseline adjustment in analyses of change (in the presence of measurement error), the choice of regression model in the setting of longitudinal measurements and a dichotomous outcome, and the difficulty in attempting to identify mechanisms in a causal pathway (diagnosing mediation).

06/03/2009 — Identification of Ovarian Cancer Symptoms in Health Insurance Claims Data

Sean Devlin

Women with ovarian cancer have reported abdominal/pelvic pain, bloating, difficulty eating or feeling full quickly, and urinary frequency/urgency prior to diagnosis. While case-control studies have been important in identifying the initial symptoms, this seminar examines how health insurance claims data can further elucidate the association between symptom prevalence and ovarian cancer, and potentially address the need for advocacy at the patient- and/or physician-level. We examine methods to discern the relative prevalence of the four symptoms and describe the association with cancer stage at diagnosis. Lastly, we explore the performance of a hypothetical passive screening tool implemented using insurance claims data.

05/27/2009 — Now you see it, now you don’t: Impact of population screening practice on perceived efficacy of cancer therapies

Maggie Au

Many cancer treatment trials are conducted in unscreened populations. As cancer screening becomes more prevalent the questionable validity of these trial results in the presence of screening is increasingly pertinent. We evaluate the impact of screening on treatment efficacy estimates. We develop expressions for cause-specific cumulative incidence in the presence of screening and use these to simulate results under different settings of lead time, overdiagnosis, treatment efficacy, and other-cause survival. Under screening, we see a reduction in the difference in cumulative incidence of prostate cancer death between the two treatment arms, a corresponding reduction in the power of a study to detect a treatment effect, and an increase in the NNT (number needed to treat). These results clearly indicate that it is important to recognize that, under screening, treatments may not attain the benefits and may cost more than would have been expected based on clinical trials of their effect in the absence of screening. Conversely, treatments shown to be ineffective in trials done in a screened population could turn out to be worthwhile in populations where screening is not performed at the population level. Addressing and taking into account these differences that result from screening is an important next step in developing appropriate clinical practice for a population based on trial results.