Turning Medical Records and Claims Data into Better Asthma Care: A Rutgers Data Analytics Partnership.

About the Client

RITMS serves as Rutgers’ academic home for clinical and translational research, providing infrastructure and training across Rutgers Biomedical and Health Sciences and its partner institutions statewide, and supporting hundreds of clinical trials at any given time. This engagement supported one of RITMS’s population health research initiatives, focused on closing the gap between what severe asthma science had made possible and what was actually reaching patients in everyday care.

Industry Context

The treatment landscape for severe asthma had transformed substantially in the years leading up to this study. Omalizumab, a monoclonal antibody targeting immunoglobulin E, received FDA approval in 2004, followed by mepolizumab in 2015 and reslizumab in 2016, both targeting interleukin-5 for late-onset eosinophilic asthma. Together these biomarker-guided therapies made roughly half of all severe, uncontrolled asthma patients eligible for a monoclonal antibody treatment they may never have been offered — underscoring why translating this evidence into frontline primary care, rather than leaving it siloed with specialists, represented such a high-value target for RITMS’s study.

Current Situation

RITMS designed a study to assess the costs and benefits of providing high-quality training to physicians treating SUA patients and supplying novel biologic medicines to patients meeting specific eligibility criteria, securing grant funding from Horizon Blue Cross Blue Shield of New Jersey, Novartis, and Rutgers University. Four organizations collaborated on the work: RITMS, Horizon, RWJBarnabas Health, and Iknow, with Rutgers and Horizon jointly defining the SUA patient criteria, Horizon identifying the qualifying population and extracting claims data, RWJBarnabas Health providing medical records, and Iknow asked to provide the project’s data analytics capabilities.

• Effective therapies going underused. Biomarker-guided biologic treatments remained underutilized by providers and poorly understood by patients, keeping ED visits and hospitalization rates static since the 1990s.
• High, avoidable cost burden. Patients with severe or uncontrolled asthma accrued direct medical costs five times higher than those with mild or well-controlled asthma — a quantifiable opportunity cost tied directly to underuse of available tools.
• Fragmented, multi-source data. The evidence needed to evaluate the intervention lived in two entirely separate systems — Horizon’s medical claims data and RWJBarnabas Health’s electronic medical records — in different formats, with no existing structure to combine them.
• Strict privacy and security constraints. Any technical environment handling this data needed to satisfy HIPAA privacy requirements and RITMS’s institutional security standards while still supporting exploratory, iterative analysis.
• A heterogeneous patient population. Severe or uncontrolled asthma patients manifest dramatic heterogeneity in etiology, pathogenesis, and treatment response, meaning the data model had to support meaningful subgroup characterization rather than a single aggregate view.

• Provide the data analytics capabilities for RITMS’s demonstration study.
• Design and implement a technical environment meeting Rutgers’ security and HIPAA privacy requirements.
• Integrate Horizon’s claims data and RWJBarnabas Health’s electronic medical records into a unified, standards-based data model.
• Translate RITMS’s working hypotheses into defined, executable analyses, including data elements, population sets, time periods, and output formats.
• Produce comprehensive analyses characterizing the patient and physician cohort and evaluating the effects of provider education and biomarker-guided biologic prescribing.

How Iknow Structured the Work

The project operated as a four-organization collaboration, with each partner contributing a distinct capability: Rutgers and Horizon defined the SUA patient criteria, Horizon extracted claims data, RWJBarnabas Health provided medical records, Rutgers delivered provider education, and Iknow handled data integration, analytics, and visualization end to end.

Key Activities & Decisions

• Technical Environment Design. Iknow developed the recommended hardware and software environment, implemented within Rutgers’ Office of Advanced Research Computing to satisfy RITMS’s security and HIPAA privacy requirements. The environment comprised Tableau for analysis and visualization, Talend for data integration, and a healthcare Common Data Model implemented in Microsoft SQL Server. Iknow selected the Observational Medical Outcomes Partnership (OMOP) CDM as the project’s database schema — an open standard purpose-built for combining claims and EMR data from different source systems into one analyzable format.
• Hypothesis-to-Analysis Translation. Iknow worked directly with the head of RITMS to translate working hypotheses into defined analyses, identifying data elements, population sets and subsets, time periods, and other constraints, and agreeing on analysis types and output formats before any data work began.
• Data Integration and Preparation. Iknow identified the data fields necessary for the planned analyses, then performed data transformation and cleansing for integration into the CDM. Data arrived from Horizon and RWJBarnabas Health in multiple file formats, spanning demographics, hospitalization records, lab test orders, prescriptions, and medical procedures. Iknow developed and executed the integration scripts, completed data validation testing, and integrated the CDM directly with Tableau.
• Analysis and Visualization. Iknow analyzed the data to understand the effects of focused provider education and of prescribing state-of-the-art, biomarker-guided therapies, producing more than 40 analyses capturing baseline characteristics of the patient and physician cohort — demographics, provider specialty and practice distribution, geographic distribution, and cost patterns by age — alongside the study’s outcomes.

Stakeholders & Collaboration

Iknow worked directly with RITMS leadership throughout the engagement and coordinated closely with Horizon Blue Cross Blue Shield of New Jersey and RWJBarnabas Health as the project’s two data-source partners, within a study funded by Horizon, Novartis, and Rutgers University.

Integrating Two Structurally Different Data Sources Without Compromising Privacy or Fidelity

Claims data and electronic medical records come from fundamentally different systems, formats, and organizations, creating real risk of an inconsistent combined dataset if integration were handled loosely. Iknow addressed this by standardizing everything into the OMOP Common Data Model — an open, healthcare-research-standard schema — implemented within a HIPAA-compliant environment inside Rutgers’ own research computing infrastructure, and validated every conversion through dedicated data validation testing before any analysis began.

Making a Large, Unfamiliar Dataset Usable for Iterative Clinical Research

A large, newly integrated clinical dataset risked becoming a one-time analytics deliverable rather than a resource RITMS’s own researchers could explore over the course of a multi-year study. Iknow addressed this by translating open-ended clinical hypotheses into concretely defined, executable analyses from the outset, then integrating the CDM directly with Tableau so the research team could explore the growing library of analyses interactively rather than relying on static reports.

Operational Outcomes

• Produced more than 40 distinct analyses characterizing the SUA patient cohort and their treating physicians, including demographics, geographic distribution, provider specialty mix, and cost patterns by age.
• Integrated multi-format data from two independent organizations — spanning demographics, hospitalization records, lab orders, prescriptions, and procedures — into a single, standards-based clinical data model.
• Sustained a HIPAA-compliant analytics environment within Rutgers’ Office of Advanced Research Computing across a multi-year demonstration study.
• Multiple articles published in academic journals.

Strategic and Organizational Outcomes

The project demonstrated that targeted education for primary care physicians, combined with biomarker-guided use of state-of-the-art biologic therapies, could improve quality of care and reduce total cost of care — measured as a decrease in ED and hospital utilization rates in claims data. While the study focused on Horizon subscribers with severe persistent or uncontrolled asthma, the underlying education program and data analytics approach were explicitly designed to generalize across health systems and managed care programs, and to support similar evidence-generation efforts in other disease categories.

Timeline to Impact

Iknow’s engagement spanned more than five years, reflecting the longitudinal nature of the underlying demonstration study rather than a single-point deliverable, with findings materializing progressively as the analytics environment matured and cohort data accumulated across the study period.

Core Skills

• Healthcare Data Integration & Interoperability
• Clinical and Claims Data Analytics & Visualization
• Research Computing Environment Design
• Multi-Stakeholder Research Program Support

Methods & Frameworks

• OMOP Common Data Model implementation
• Hypothesis-to-analysis translation methodology
• Data validation and quality assurance testing
• HIPAA-compliant technical environment design

Technologies & Tools

• Tableau (data analysis and visualization)
• Talend (data integration)
• Microsoft SQL Server and the OMOP Common Data Model
• Rutgers Office of Advanced Research Computing (OARC) infrastructure