Search

Background A lack of studies with large sample sizes of patients with rotator cuff tears is a barrier to performing clinical and genomic research. Objective To develop and validate an electronic medical record (EMR)-based algorithm to identify individuals with and without rotator cuff tear. Design We used a deidentified version of the EMR of more than 2 million subjects. A screening algorithm was applied to classify subjects into likely rotator cuff tear and likely normal rotator cuff groups. From these subjects, 500 likely rotator cuff tear and 500 likely normal rotator cuff were randomly chosen for algorithm development. Chart review of all 1000 subjects confirmed the true phenotype of rotator cuff tear or normal rotator cuff based on magnetic resonance imaging and operative report. An algorithm was then developed based on logistic regression and validation of the algorithm was performed. Results The variables significantly predicting rotator cuff tear included the number of times a Current Procedural Terminology code related to rotator cuff procedures was used (odds ratio [OR] = 3.3; 95% confidence interval [CI]: 1.6-6.8 for ≥3 vs 0), the number of times a term related to rotator cuff lesions occurred in radiology reports (OR = 2.2; 95% CI: 1.2-4.1 for ≥1 vs 0), and the number of times a term related to rotator cuff lesions occurred in physician notes (OR = 4.5; 95% CI: 2.2-9.1 for 1 or 2 times vs 0). This phenotyping algorithm had a specificity of 0.89 (95% CI: 0.79-0.95) for rotator cuff tear, area under the curve (AUC) of 0.842, and diagnostic likelihood ratios (DLRs), DLR+ and DLR- of 5.94 (95% CI: 3.07-11.48) and 0.363 (95% CI: 0.291-0.453). Conclusion Our informatics algorithm enables identification of cohorts of individuals with and without rotator cuff tear from an EMR-based data set with moderate accuracy.

Background: As targeted cancer therapies and molecular profiling become widespread, the era of “precision oncology” is at hand. However, cancer genomes are complex, making mutation-specific outcomes difficult to track. We created a proof-of-principle, CUSTOM-SEQ: Continuously Updating System for Tracking Outcome by Mutation, to Support Evidence-based Querying, to automatically calculate and display mutation-specific survival statistics from electronic health record data.Methods: Patients with cancer genotyping were included, and clinical data was extracted through a variety of algorithms. Results were refreshed regularly and injected into a standard reporting platform. Significant results were highlighted for visual cueing. A subset was additionally stratified by stage, smoking status, and treatment exposure.Results: By August 2015, 4310 patients with a median follow-up of 17 months had sufficient data for survival calculation. As expected, epidermal growth factor receptor (EGFR) mutations in lung cancer were associated with superior overall survival, hazard ratio (HR) = 0.53 (P Interpretation: CUSTOM-SEQ represents a novel rapid learning system for a precision oncology environment. Retrospective studies are often limited by study of specific time periods and can lead to incomplete conclusions. Because data is continuously updated in CUSTOM-SEQ, the evidence base is constantly growing. Future work will allow users to interactively explore populations by demographics and treatment exposure, in order to further investigate significant mutation-specific signals.

195 Background: Breast cancer patients face difficult decisions about their surgical care without a full understanding of their options. The learning health system goal is to use information from the care of prior patients to inform the care of future patients. We aim to apply this concept to generate data-driven surgical paths, develop interactive path visualizations to inform patients, and evaluate their impact. Methods: We used cancer registry and administrative CPT codes for women diagnosed with stage 0-III breast cancer between FY2010-14 at a comprehensive cancer center. We generated surgical event sequences and visualized them using interactive Sankey diagram path visualizations. We will run a prospective educational intervention this winter to evaluate their impact on the shared decision making process. A web-based application will be available to patients prior to, during, and after their surgical clinic visit; we will survey their reaction pre-visit, post-visit, and post-surgery. Results: 1556 patients had 1951 surgical events in the registry and 48% started their surgical care with a breast conserving surgery while 52% began with a mastectomy. Mastectomy paths are presented in Table 1. We have developed interactive visualizations for patients to view, will be conducting our prospective educational intervention this winter, and will be ready to present preliminary results in February. Conclusions: We have been able to develop interactive, data-driven surgical path visualizations for breast cancer patients from cancer registry and administrative data. We will be conducting a prospective educational intervention to evaluate our implementation of this learning health system concept. [Table: see text]

The use of illicit drugs places a large burden on the American healthcare system due to their detrimental effects on patients and their families. Network pharmacology is a novel approach to investigate drug actions. This study seeks to investigate illicit drugs and their targets in the context of a drug-protein network to elucidate the molecular mechanisms of abused drugs. We extracted 188 illicit substances from the DrugBank database. Among them, 86 drugs had 74 unique human targets, which formed a drug-target network specific for illicit drugs. We then analyzed the network by superimposing enriched pathways of these targets, drug categories, and network clusters. We found that four sub-networks corresponded to three major medication categories: depressants, stimulants, and narcotics & analgesics. The results not only provide novel insights on illicit drug actions but also illustrate that the methods utilized here can identify important drugs and targets. This study represents the first systematic investigation of drug-target interactions in drug addiction.

Objective To try to lower patient re-identification risks for biomedical research databases containing laboratory test results while also minimizing changes in clinical data interpretation. Materials and methods In our threat model, an attacker obtains 5–7 laboratory results from one patient and uses them as a search key to discover the corresponding record in a de-identified biomedical research database. To test our models, the existing Vanderbilt TIME database of 8.5 million Safe Harbor de-identified laboratory results from 61 280 patients was used. The uniqueness of unaltered laboratory results in the dataset was examined, and then two data perturbation models were applied—simple random offsets and an expert-derived clinical meaning-preserving model. A rank-based re-identification algorithm to mimic an attack was used. The re-identification risk and the retention of clinical meaning for each model's perturbed laboratory results were assessed. Results Differences in re-identification rates between the algorithms were small despite substantial divergence in altered clinical meaning. The expert algorithm maintained the clinical meaning of laboratory results better (affecting up to 4% of test results) than simple perturbation (affecting up to 26%). Discussion and conclusion With growing impetus for sharing clinical data for research, and in view of healthcare-related federal privacy regulation, methods to mitigate risks of re-identification are important. A practical, expert-derived perturbation algorithm that demonstrated potential utility was developed. Similar approaches might enable administrators to select data protection scheme parameters that meet their preferences in the trade-off between the protection of privacy and the retention of clinical meaning of shared data.

e17624 Background: Monitoring quality metrics in near real time can help healthcare providers and organizations improve care delivery, engage patients, and comply with reporting requirements of acc...

Humans have an uncanny ability to estimate haptic features of an object such as haptic shape, size, texture and material by visual inspection. A significant computer vision problem is that of estimating haptic features from visual images. While explorations have been made in estimation of visual features such as visual texture, work on estimation of haptic features from video is still in its infancy. We present a methodology to establish ground truth for estimation of haptic features from visual images. We assembled a visio-haptic database of 48 objects ranging from nonsense objects to everyday objects. The variation was controlled in objects by systematically varying haptic features such as shape and texture, and the physical and perceptual ground truth of visual and haptic features was documented. This database provides visio-haptic features of objects and can be used to develop algorithms to estimate haptic features from visual images. Finally, a tactile cueing experiment is presented demonstrating how visio-haptic ground truth can be used to assess the accuracy of a system for visio-haptic conversion of image content.

Background Drug addiction is a complex and chronic mental disease, which places a large burden on the American healthcare system due to its negative effects on patients and their families. Recently, network pharmacology is emerging as a promising approach to drug discovery by integrating network biology and polypharmacology, allowing for a deeper understanding of molecular mechanisms of drug actions at the systems level. This study seeks to apply this approach for investigation of illicit drugs and their targets in order to elucidate their interaction patterns and potential secondary drugs that can aid future research and clinical care. Results In this study, we extracted 188 illicit substances and their related information from the DrugBank database. The data process revealed 86 illicit drugs targeting a total of 73 unique human genes, which forms an illicit drug-target network. Compared to the full drug-target network from DrugBank, illicit drugs and their target genes tend to cluster together and form four subnetworks, corresponding to four major medication categories: depressants, stimulants, analgesics, and steroids. External analysis of Anatomical Therapeutic Chemical (ATC) second sublevel classifications confirmed that the illicit drugs have neurological functions or act via mechanisms of stimulants, opioids, and steroids. To further explore other drugs potentially having associations with illicit drugs, we constructed an illicit-extended drug-target network by adding the drugs that have the same target(s) as illicit drugs to the illicit drug-target network. After analyzing the degree and betweenness of the network, we identified hubs and bridge nodes, which might play important roles in the development and treatment of drug addiction. Among them, 49 non-illicit drugs might have potential to be used to treat addiction or have addictive effects, including some results that are supported by previous studies. Conclusions This study presents the first systematic review of the network characteristics of illicit drugs, their targets, and other drugs that share the targets of these illicit drugs. The results, though preliminary, provide some novel insights into the molecular mechanisms of drug addiction. The observation of illicit-related drugs, with partial verification from previous studies, demonstrated that the network-assisted approach is promising for the identification of drug repositioning.