No abstract

ULTIPLE SCLEROSIS (MS) IS a chronic disease that often affects people in the prime of their lives. A key feature of MS is clinical progression of the disease over time manifested by the accumulation of disability. Interferon beta drugs are the most widely prescribed disease-modifying drugs (DMDs) approved by the US Food and Drug Administration for the treatment of relapsing-onset MS, the most common MS disease course. Although a substantial reduction in brain lesion development, as evidenced by magnetic resonance imaging (MRI), 1 and a one-third relative reduction in relapse frequency were demonstrated in the pivotal clinical trials of interferon beta for relapsing-remitting MS, 2 there is a lack of well-controlled longitudinal studies investigating the effect of in-terferon beta on disability progression. Typically, drug efficacy (as established through randomized clinical trials conducted under optimal conditions) is greater than drug effectiveness (as measured in "real-world" settings). 3 Patients participating in clinical trials tend to be highly selected in For editorial comment see p 290.

No abstract

No abstract

We consider Bayesian sensitivity analysis for unmeasured confounding in observational studies where the association between a binary exposure, binary response, measured confounders and a single binary unmeasured confounder can be formulated using logistic regression models. A model for unmeasured confounding is presented along with a family of prior distributions that model beliefs about a possible unknown unmeasured confounder. Simulation from the posterior distribution is accomplished using Markov chain Monte Carlo. Because the model for unmeasured confounding is not identifiable, standard large-sample theory for Bayesian analysis is not applicable. Consequently, the impact of different choices of prior distributions on the coverage probability of credible intervals is unknown. Using simulations, we investigate the coverage probability when averaged with respect to various distributions over the parameter space. The results indicate that credible intervals will have approximately nominal coverage probability, on average, when the prior distribution used for sensitivity analysis approximates the sampling distribution of model parameters in a hypothetical sequence of observational studies. We motivate the method in a study of the effectiveness of beta blocker therapy for treatment of heart failure.