Visualizing nationwide variation in medicare Part D prescribing patterns.

BACKGROUND: To characterize the regional and national variation in prescribing patterns in the Medicare Part D program using dimensional reduction visualization methods. 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. RESULTS: 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. CONCLUSIONS: 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.

Predicting hospital-onset Clostridium difficile using patient mobility data: A network approach.

OBJECTIVE: To examine the relationship between unit-wide Clostridium difficile infection (CDI) susceptibility and inpatient mobility and to create contagion centrality as a new predictive measure of CDI. DESIGN: Retrospective cohort study. METHODS: A mobility network was constructed using 2 years of patient electronic health record data for a 739-bed hospital (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 (ie, edges). Daily unit-wide CDI susceptibility scores were calculated using logistic regression and were compared to network centrality measures to determine the relationship between unit CDI susceptibility and patient mobility. RESULTS: Closeness centrality was a statistically significant measure associated with unit susceptibility (P < .05), highlighting the importance of incoming patient mobility in CDI prevention at the unit level. Contagion centrality (CC) was calculated using inpatient transfer rates, unit-wide susceptibility of CDI, and current hospital CDI infections. The contagion centrality measure was statistically significant (P < .05) with our outcome of hospital-onset CDI cases, and it captured the additional opportunities for transmission associated with inpatient transfers. We have used this analysis to create easily interpretable clinical tools showing this relationship as well as the risk of hospital-onset CDI in real time, and these tools can be implemented in hospital EHR systems. CONCLUSIONS: 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 inform prevention efforts.

Predicting acute kidney injury at hospital re-entry using high-dimensional electronic health record data.

Acute Kidney Injury (AKI), a sudden decline in kidney function, is associated with increased mortality, morbidity, length of stay, and hospital cost. Since AKI is sometimes preventable, there is great interest in prediction. Most existing studies consider all patients and therefore restrict to features available in the first hours of hospitalization. Here, the focus is instead on rehospitalized patients, a cohort in which rich longitudinal features from prior hospitalizations can be analyzed. Our objective is to provide a risk score directly at hospital re-entry. Gradient boosting, penalized logistic regression (with and without stability selection), and a recurrent neural network are trained on two years of adult inpatient EHR data (3,387 attributes for 34,505 patients who generated 90,013 training samples with 5,618 cases and 84,395 controls). Predictions are internally evaluated with 50 iterations of 5-fold grouped cross-validation with special emphasis on calibration, an analysis of which is performed at the patient as well as hospitalization level. Error is assessed with respect to diagnosis, race, age, gender, AKI identification method, and hospital utilization. In an additional experiment, the regularization penalty is severely increased to induce parsimony and interpretability. Predictors identified for rehospitalized patients are also reported with a special analysis of medications that might be modifiable risk factors. Insights from this study might be used to construct a predictive tool for AKI in rehospitalized patients. An accurate estimate of AKI risk at hospital entry might serve as a prior for an admitting provider or another predictive algorithm.

Properties of healthcare teaming networks as a function of network construction algorithms.

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.