Robert J. White

To characterize the regional and national variation in prescribing patterns in the Medicare Part D program using dimensional reduction visualization methods. Using publicly available Medicare Part D claims data, we identified and visualized regional and national provider prescribing profile variation with unsupervised clustering and t-distributed stochastic neighbor embedding (t-SNE) dimensional reduction techniques. Additionally, we examined differences between regionally representative prescribing patterns for major metropolitan areas. Distributions of prescribing volume and medication diversity were highly skewed among over 800,000 Medicare Part D providers. Medical specialties had characteristic prescribing patterns. Although the number of Medicare providers in each state was highly correlated with the number of Medicare Part D enrollees, some states were enriched for providers with > 10,000 prescription claims annually. Dimension-reduction, hierarchical clustering and t-SNE visualization of drug- or drug-class prescribing patterns revealed that providers cluster strongly based on specialty and sub-specialty, with large regional variations in prescribing patterns. Major metropolitan areas had distinct prescribing patterns that tended to group by major geographical divisions. This work demonstrates that unsupervised clustering, dimension-reduction and t-SNE visualization can be used to analyze and visualize variation in provider prescribing patterns on a national level across thousands of medications, revealing substantial prescribing variation both between and within specialties, regionally, and between major metropolitan areas. These methods offer an alternative system-wide and pattern-centric view of such data for hypothesis generation, visualization, and pattern identification.

With hospital-onset Clostridium difficile Infection (CDI) still a common occurrence in the U.S., this paper examines the relationship between unit-wide CDI susceptibility and inpatient mobility and creates a predictive measure of CDI called “Contagion Centrality”. A mobility network was constructed using two years of patient electronic health record (EHR) data within a 739-bed hospital (Jan. 2013 - Dec. 2014; n=72,636 admissions). Network centrality measures were calculated for each hospital unit (node) providing clinical context for each in terms of patient transfers between units (edges). Daily unit-wide CDI susceptibility scores were calculated using logistic regression and compared to network centrality measures to determine the relationship between unit CDI susceptibility and patient mobility. Closeness centrality was a statistically significant measure associated with unit susceptibility (p-value < 0.05), highlighting the importance of incoming patient mobility in CDI prevention at the unit-level. Contagion Centrality (CC) was calculated using incoming inpatient transfer rates, unit-wide susceptibility of CDI, and current hospital CDI infections. This measure is statistically significant (p-value <0.05) with our outcome of hospital-onset CDI cases, and captures the additional opportunities for transmission associated with inpatient transfers. We have used this analysis to create an easily interpretable and informative clinical tool showing this relationship and risk of hospital-onset CDI in real-time. Quantifying and visualizing the combination of inpatient transfers, unit-wide risk, and current infections help identify hospital units at risk of developing a CDI outbreak, and thus provide clinicians and infection prevention staff with advanced warning and specific location data to concentrate prevention efforts.

HIV RNA viral load (VL) is an important outcome variable in studies of HIV infected persons. There exists only a handful of methods which classify patients by VL patterns. Most methods place limits on the use of viral load measurements, are often specific to a particular study design, and do not account for complex, temporal variation. To address this issue, we propose a set of four unambiguous computable characteristics (features) of time-varying HIV viral load patterns, along with a novel centroid-based classification algorithm, which we use to classify a population of 1,576 HIV positive clinic patients into one of five different viral load patterns (clusters) often found in the literature: durably suppressed viral load (DSVL), sustained low viral load (SLVL), sustained high viral load (SHVL), high viral load suppression (HVLS), and rebounding viral load (RVL). The centroid algorithm summarizes these clusters in terms of their centroids and radii. We show that this allows new VL patterns to be assigned pattern membership based on the distance from the centroid relative to its radius, which we term radial normalization classification. This method has the benefit of providing an objective and quantitative method to assign VL pattern membership with a concise and interpretable model that aids clinical decision making. This method also facilitates meta-analyses by providing computably distinct HIV categories. Finally we propose that this novel centroid algorithm could also be useful in the areas of cluster comparison for outcomes research and data reduction in machine learning.

Network models of healthcare systems can be used to examine how providers collaborate, communicate, refer patients to each other, and to map how patients traverse the network of providers. Most healthcare service network models have been constructed from patient claims data, using billing claims to link a patient with a specific provider in time. The data sets can be quite large (10⁶–10⁸ individual claims per year), making standard methods for network construction computationally challenging and thus requiring the use of alternate construction algorithms. While these alternate methods have seen increasing use in generating healthcare networks, there is little to no literature comparing the differences in the structural properties of the generated networks, which as we demonstrate, can be dramatically different. To address this issue, we compared the properties of healthcare networks constructed using different algorithms from 2013 Medicare Part B outpatient claims data. Three different algorithms were compared: binning, sliding frame, and trace-route. Unipartite networks linking either providers or healthcare organizations by shared patients were built using each method. We find that each algorithm produced networks with substantially different topological properties, as reflected by numbers of edges, network density, assortativity, clustering coefficients and other structural measures. Provider networks adhered to a power law, while organization networks were best fit by a power law with exponential cutoff. Censoring networks to exclude edges with less than 11 shared patients, a common de-identification practice for healthcare network data, markedly reduced edge numbers and network density, and greatly altered measures of vertex prominence such as the betweenness centrality. Data analysis identified patterns in the distance patients travel between network providers, and a striking set of teaming relationships between providers in the Northeast United States and Florida, likely due to seasonal residence patterns of Medicare beneficiaries. We conclude that the choice of network construction algorithm is critical for healthcare network analysis, and discuss the implications of our findings for selecting the algorithm best suited to the type of analysis to be performed.

RNA polymerase (pol) III contains a dissociable subcomplex that is required for initiation, but not for elongation or termination of transcription. This subcomplex is composed of subunits RPC3, RPC6 and RPC7, and interacts with TFIIIB, a factor that is necessary and sufficient to support accurate pol III transcription in vitro . Direct binding of TFIIIB to RPC6 is believed to recruit pol III to its genetic templates. However, this has never been tested in vivo . Here we combine chromatin immunoprecipitation with RNA interference to demonstrate that the RPC3/6/7 subcomplex is required for pol III recruitment in mammalian cells. Specific knockdown of RPC6 by RNAi results in post-transcriptional depletion of the other components of the subcomplex, RPC3 and RPC7, without destabilizing core pol III subunits or TFIIIB. The resultant core enzyme is defective in associating with TFIIIB and target genes in vivo . Promoter occupancy by pol II is unaffected, despite sharing five subunits with the pol III core. These observations provide evidence for the validity in vivo of the model for pol III recruitment that was built on biochemical data.

Myotonic dystrophy (DM), the most common form of muscular dystrophy in adult humans, results from expansion of a CTG repeat in the 3′ untranslated region of the DMPK gene. The mutant DMPK messenger RNA (mRNA) contains an expanded CUG repeat and is retained in the nucleus. We have expressed an untranslated CUG repeat in an unrelated mRNA in transgenic mice. Mice that expressed expanded CUG repeats developed myotonia and myopathy, whereas mice expressing a nonexpanded repeat did not. Thus, transcripts with expanded CUG repeats are sufficient to generate a DM phenotype. This result supports a role for RNA gain of function in disease pathogenesis.

Myotonic dystrophy is caused by an expanded CTG repeat in the 3′ untranslated region of the DM protein kinase (DMPK) gene. The expanded repeat triggers the nuclear retention of mutant DMPK transcripts, but the resulting underexpression of DMPK probably does not fully account for the severe phenotype. One proposed disease mechanism is that nuclear accumulation of expanded CUG repeats may interfere with nuclear function. Here we show by thermal melting and nuclease digestion studies that CUG repeats form highly stable hairpins. Furthermore, CUG repeats bind to the dsRNA-binding domain of PKR, the dsRNA-activated protein kinase. The threshold for binding to PKR is ∼15 CUG repeats, and the affinity increases with longer repeat lengths. Finally, CUG repeats that are pathologically expanded can activate PKR in vitro. These results raise the possibility that the disease mechanism could be, in part, a gain of function by mutant DMPK transcripts that involves sequestration or activation of dsRNA binding proteins.