Faculty Bibliography

The most common indication for readmission among Medicare patients is congestive heart failure (CHF). Prior studies underscore the use of residents to bolster hospital-wide programs and reduce CHF readmissions. The authors assessed the effectiveness of a novel online training program designed to improve resident documentation and knowledge related to CHF. The findings suggest that despite a significant increase in knowledge scores following the online educational course, there was only a slight increase in documentation scores. Additional teaching modalities need to be identified to foster resident education and create sustained behavior change.

Many techniques of functional data analysis require choosing a measure of distance between functions, with the most common choice being L2 distance. In this article we show that using a weighted L2 distance, with a judiciously chosen weight function, can improve the performance of various statistical methods for functional data, including k-medoids clustering, nonparametric classification, and permutation testing. Assuming a quadratically penalized (e.g., spline) basis representation for the functional data, we consider three nontrivial weight functions: design density weights, inverse-variance weights, and a new weight function that minimizes the coefficient of variation of the resulting squared distance by means of an efficient iterative procedure. The benefits of weighting, in particular with the proposed weight function, are demonstrated both in simulation studies and in applications to the Berkeley growth data and a functional magnetic resonance imaging data set.

BACKGROUND: There are few data characterizing temporal changes in hospitalization for recurrent acute myocardial infarction (AMI) after AMI. METHODS AND RESULTS: Using a national sample of 2 305 441 Medicare beneficiaries hospitalized for AMI from 1999 to 2010, we evaluated changes in the incidence of 1-year recurrent AMI hospitalization and mortality using Cox proportional hazards models. The observed recurrent AMI hospitalization rate declined from 12.1% (95% CI 11.9 to 12.2) in 1999 to 8.9% (95% CI 8.8 to 9.1) in 2010, a relative decline of 26.4%. The observed recurrent AMI hospitalization rate declined by a relative 27.7% in whites, from 11.9% (95% CI 11.8 to 12.1) to 8.6% (95% CI 8.5 to 8.8) versus a relative decline in blacks of 13.6% from 13.2% (95% CI 12.6 to 13.8) to 11.4% (95% CI 10.9 to 12.0). The risk-adjusted rate of annual decline in recurrent AMI hospitalizations was 4.1% (HR 0.959; 95% CI 0.958 to 0.961), and whites experienced a higher rate of decline (HR 0.957, 95% CI 0.956 to 0.959) than blacks (HR 0.974, 95% CI 0.970 to 0.979).The overall, observed 1-year mortality rate after hospitalization for recurrent AMI declined from 32.4% in 1999 to 29.7% in 2010, a relative decline of 8.3% (P<0.05). In adjusted analyses, 1-year mortality after recurrent AMI hospitalization declined 1.8% per year (HR, 0.982; 95% CI 0.980 to 0.985). CONCLUSIONS: In a national sample of Medicare beneficiaries hospitalized for AMI from 1999 to 2010, hospitalization for recurrent AMI decreased, as did subsequent mortality, albeit to a lesser extent. The risk of recurrent AMI hospitalization declined less in black patients than in whites, increasing observed racial disparities by the end of the study period.

We propose a penalized spline approach to performing large numbers of parallel non-parametric analyses of either of two types: restricted likelihood ratio tests of a parametric regression model versus a general smooth alternative, and nonparametric regression. Compared with naively performing each analysis in turn, our techniques reduce computation time dramatically. Viewing the large collection of scatterplot smooths produced by our methods as functional data, we develop a clustering approach to summarize and visualize these results. Our approach is applicable to ultra-high-dimensional data, particularly data acquired by neuroimaging; we illustrate it with an analysis of developmental trajectories of functional connectivity at each of approximately 70000 brain locations. Supplementary materials, including an appendix and an R package, are available online.