TThe objective of translational medicine is to take discoveries from basic science, clinical research and population studies and apply them in clinic to impact human health. This objective, to some extent, has always been the goal of scientists, researchers and practitioners. Translational medicine is not so much a completely new discipline as it is an effort to take a wholistic view of the complete workflow from biomedical discovery to clinical application and enhance its efficiency and effectiveness.
Enhancing efficiency and effectiveness in the wholistic translational medicine workflow requires maturing from loosely integrated, disparate tools (a different tool for each team member) to one tightly integrated suite of tools that pulls all the team members into a common environment with a seamless workflow. TransMed BIS provides a tool suite that integrates, in an open, collaborative, common environment, the following activities: data aggregation, data access, basic reporting, sophisticated statistical analysis, and translation into clinical application.
Familiar, Self-Service User Experience
Cohort Explorer provides self-service, single point of access to all healthcare data, including: clinical, molecular, cost, survey, etc. Cohort Explorer’s user experience is intuitive because it’s familiar; if you know how to find a file on a computer using Microsoft Windows Explorer™, then you know how to find your healthcare data using Cohort Explorer. This familiar user experience makes Cohort Explorer quick and easy for business users such as: clinicians, investigators, and quality analysts to learn because it requires little training.
Cohort Explorer quickly and easily identifies cohorts of patients with similar bioclinical traits. For example, it is possible to identify the cohort of male smokers diagnosed after 6/15/12 with BIS 02lung cancer who received drug A. Analogous to files in Windows Explorer, the tens of thousands of healthcare data elements are organized into folders that reflect your terminology. There is a powerful search capability that searches folder names, trait names and choice fields to quickly find data elements matching the search criteria. Cohort Explorer provides a capability to ‘Group’ which stratifies and returns summary counts of the current cohort based on the selected data element. Cohort Explorer also has a ‘Filter’ capability which limits the patients in the cohort to just those that match the specified filter criteria. Cohort Explorer supports combinatorial searches that use hot breadcrumbs to track the successive filters applied to the current cohort. The resulting cohort can be saved for future reference and cohort filters can be saved and rerun.
Clinical Pattern Matching
Cohort Explorer supports a temporal or longitudinal pattern filter. This allows a cohort to be identified with patients who have had events occur in a specified sequence. For example, identify all breast cancer patients who received a breast conserving surgery followed within 12 months with a mastectomy. Clinical Pattern Matcher has a drag and drop user experience that allows events to be graphically, ordered on a timeline. The clinical pattern can be saved and used as a cohort filter to stratify patients that do and do not match the clinical pattern criteria.
Clinical Pattern Matching
Cohort Explorer has a summary view that presents filter results as simple patient counts matching the filter. The details view is a tabular representation of each matching patient including columns for all data elements that are part of the cohort filter.
Familiar, Self-Service User Experience
Cohort Reporter follows the same Windows Explorer user experience paradigm as Cohort Explorer. Multiple tools adhering to such a ubiquitous user experience greatly reduces the learning curve and increases the adoption rate and value of TransMed BIS.
Quickly & Easily Fulfill Report Request
Cohort Reporter’s familiar user experience empowers business users such as clinicians and investigators with self-service access to fulfill their own report request thus reducing their dependence on data analysts who, due to workload, are otherwise the critical path in this process. Through Cohort Reporter, the business user identifies the data elements to be used for reporting and generates a dataset of just that data. The resulting dataset is presented in a tabular, Microsoft Excel™-like format which is familiar to the business user thus easy for them to learn and use.
Configurable & Interactive Reports
Cohort Reporter provides an Excel-like pivot table so a user can ‘slice and dice’ the data in multiple different ways to produce basic plots such as line graphs, bar charts, pie charts, etc. Pivot tables can represent a multi-dimensional stratification of data, for example, the business user can create a table stratifying patients by surgery type and then age at surgery within that stratification and then also stratify them by their ethnicity and gender. This stratified table of data can then be used to graphically represent the data in bar charts, pie charts, etc. The pivot table is also interactive allowing the user to drill down into the detailed data of one or more cells of the pivot table.
Cohort Reporter is an open environment intended to integrate with an organization’s existing technology. With appropriate user functional permissions, business users can export a dataset in multiple formats including: Excel, CSV, R, SAS, SPSS, etc. thus allowing data to be utilized in commercial statistical and reporting software.
Familiar, Intuitive User Experience
Cohort Analyzer is a straightforward user experience that enables clinicians and investigators to accomplish many initial statistical tasks with minimal support from statistical experts and no expertise in the programmatic use of most statistical packages.
Integrated with the Validated R Statistics Platform
Cohort Analyzer is tightly integrated with the open source R statistics platform thus providing users a simple user experience to a robust and validated statistical analysis capability suitable for publishing of results. Cohort Analyzer, through its R integration, includes, but is not limited to, standard statistics such as: survival analysis, t test, chi squared, ANOVA, regression, linear and non-linear modelling along with advanced statistics for use in genomic analysis such as: LIMMA and gene set enrichment analysis. R is continuously updated with the newest statistical methods and latest research laboratory technicques. Like the Adaptive HUB, Cohort Analyzer expects and therefore easily adapts to technology advances.
The statistics available to users is fully extensible. Cohort Analyzer provides an easy mechanism to add additional statistical algorithms built using R but made available through Cohort Analyzer’s simple user experience. R tools can be easily built by any statistician familiar with R, an industry standard and skillset readily available in the marketplace. These R tools can be standardized for the organization and easily made available to all business users through the simple Cohort Analyzer user experience. While the business user must understand the right application of statistical methods, this does enable the business user to accomplish most initial statistical analyses without requiring the deep expertise of R programming and with minimal consultation with biostatistical experts.
Export Plots and Data
Cohort Analyzer is an open environment. Data can be easily exported for consumption into any commercial statistical package including SAS, SPSS, and STATA. Plots resulting from analyses can also be exported in multiple different formats including publication quality vector graphics formats provided by R.
Save and Share Work Artifacts
Translational research requires collaboration of a multidisciplinary team involving: biomedical research, clinicians, clinical trials staff, life sciences, computer science, biomedical statistics-informatics, Institutional Review Boards, etc. Project Collaborator provides cross-functional team members a common place to securely share and review work artifacts. All BIS tools utilize Project Collaborator for this purpose: for example, Cohort Explorer saves cohorts, cohort filters, choice lists, clinical patterns, etc.; Cohort Reporter saves reports and dataset exports; Cohort Analyzer saves analyses and exported plots. This eliminates the need to email datasets and analyses or transfer them onto shared drives; instead collaborators can look directly at these work artifacts with confidence they are viewing the latest revision.
Organize Work Artifacts
Project Collaborator, similar to Windows Explorer, allows users to create folders and sub folders in their Project Workspace in order to organize all the work artifacts associated with the project. Project Collaborator also provides a search capability in order to quickly find work artifacts.
Project Collaborator performs Project Workspace level auditing which logs certain activities users perform while they have the Project Workspace open. Events recorded for audit purposes include: accessing a workspace, executing a filter or group by command, saving a cohort, creating or downloading an export, etc.
Access in ID or De-ID Mode
Project Collaborator respects the data and functional entitlements granted to users through Project Administrator. Only users granted permission can access a Project Workspace. Further, with appropriate permissions, the same Project Workspace can be opened in identified mode by one user and opened in de-identified mode by another user. Project Collaborator’s tight integration with Cohort Explorer, Reporter and Analyzer appropriately limits each user’s access to PHI even while both users access the same Project Workspace. This fosters collaboration amongst all the team members to share artifacts while appropriately managing visibility to PHI.