Data value beyond the traditional
Given the need for quick adaptations and the increased pace of recent circumstances, the world of healthcare data is advancing at warp speed.
Data alone has no value if it’s not informing decision-making or improving an action. Yet, many still encounter limitations when it comes to turning data into insight. Everyone wants to empower physicians and patients with information to improve outcomes. But, insight requires someone to put it into a relevant context.
That’s where Guardian Research Network excels.
Healthcare Data Questions
Whether within a health system, life sciences group or other third party, leaders want to understand and quantify the potential for healthcare data, electronic health or medical records (EMR/EHR), and/or publicly sourced data. Often because of limited experience with specific data formats, disjointed analytic tools or competing investment priorities, they lack the mission-critical skills, tools and abilities to harmonize and integrate data. Without the right data and data science acumen, insights can be difficult to uncover and may curtail speed to market, margin growth and improved patient outcomes.
GRN’s records are sourced from our Partner health systems and other third-party referral groups. We capture timely, complex information—from rigorous EMR/EHR documentation to genomic profiling. That means GRN’s data goes deeper than traditional sources, capturing the clinical information most valuable and up-to-date for diagnostics and therapeutics. Often internal systems or vendor systems have diverse interoperability, creating data points that do not connect well, are limited by missing elements, are stale and not current, or include data not easily integrated into a useful view of the patient.
Relative to others, GRN has the capability to extract substantially deeper insights from timely EMR data (including unstructured data and genomic data). Though originally cancer-focused, GRN is expanding into other therapeutic areas not limited to oncology disorders or de-identified patient data. While the majority of our data comes from health system Partners, our healthcare data knowledge and expertise produces a rich, in-depth dataset for a fuller, more clearly defined picture of a patient’s or a population’s experience and opportunities for outcome improvement.
The amount of healthcare data accessible is ever increasing, as are the different types of data available. GRN collects clinically-relevant data on 21 cancer sub-types and other life-threatening conditions. Generally our data includes information on more than 2 million patients that goes well beyond traditional EMR/EHRs to include the following and more.
- Tumor morphology and histology
- Labs, specialized tests and imaging
- Advanced & molecular biomarkers
- Genetic profiles & genomic sequencing
- First, second, third line, etc., course of treatment
- Patient characteristics and status
- Tele-health, pharmacy, insurance and hospice details
Since patient identities are securely retained according to HIPAA standards, GRN has a far greater ability to provide actionable clinical insight than other clinical data providers.
EMR or EHR data comes in a wide variety of formats, from strictly formed relational databases to your last scan or x-ray image. In all its different formats, data is divided into two main categories, structured and unstructured. Structured data can be stored in data warehouses, and displayed in a consistent, organized, searchable manner. Typical examples are names, addresses, dates, diagnosis codes, lab results, drug Rx, and so on.
Unstructured data, on the other hand, is collected in its native format and lacks the inherent organization and quantitative nature of structured data. Often, unstructured data is stored in a data lake and is tedious to process, analyze and understand as there is no given data model to use. Typical examples are physicians’ written notes, histology status, line of therapy, genetics, clinician judgments, images, scans, hospice information and more.
There are other options—Natural Language Processing (NLP), Machine Learning (ML) and Artificial Intelligence (AI)—that can turn unstructured data into useful structured formats. GRN deals with unstructured data regularly and is very familiar with the information gleaned from hand-written physician notes.
While our data acumen and advanced technology enable much of what we do, GRN does include a human touch. Our clinically verified patient review allows for nuances in unstructured data to be reviewed and captured by clinical professionals. This enables judgment, timing and other qualitative considerations to be factored into a decision. Over the years, this “white glove” approach has made a significant difference in the number of eligible patients we have on a “watch list” for trial inclusion, enabling our no-patient-left-behind approach. This also provides a richer picture of a patient’s journey for RWE use.
To get to valuable data, GRN removes all Protected Health Information (PHI) from structured and unstructured data. Structured data is then aggregated and included as rounded numbers or categories (e.g. age at diagnosis 40-49 years). Unstructured data, which makes up 80% or more of all data, holds a growing amount of valuable intelligence and requires a different approach.
GRN’s approach includes a multifunctional, HIPAA-compliant secure software platform that enables healthcare providers to more efficiently manage their internal operation activities. Our tools exploit proprietary NLP and AI heuristics to perform clinically meaningful searches of raw EMR records, Limited Data Sets (LDS) and fully de-identified datasets to make preliminary clinical trial matches for Network Partner patients.
Generally, biopharma, CROs, diagnostic companies and others have used GRN to support all phases of clinical trials, product development studies, commercial and clinical assessments, and supplemental external comparators. With most Life Sciences projects, our goal is to support the end game of FDA approval whether in drug, device or diagnostic development, however, we are happy to discuss other opportunities for transforming data into cures.
GRN uses anonymized, de-identified data to minimize the risk of a breach of privacy and necessary standards are always observed. Before de-identified data sets are released to third parties, the data is anonymized, deidentified, and reviewed by a series of evaluators to assure adequate de-identification (e.g., the GRN Compliance Department). Only after review and confirmation of sufficient de-identification occur will it be electronically transmitted to a third party.
EMR patient data flows from Network Partners to GRN via secure, HIPAA-compliant electronic transmission systems. Once received our specially trained and certified Honest Brokers de-identify patient data according to a HIPAA-compliant Business Associate Agreement (BAA) and/or a Data Use Agreement (DUA). Once a clinical trial match has been identified, only GRN Honest Brokers have the ability to re-identify the patient and securely notify Partner clinical trial personnel.
Specifically, patient reidentification occurs only if GRN identifies a trial match and only to inform the Network Partner. Our Honest Brokers will never share fully identifiable PHI with any other GRN employee, researcher or third party unless expressly permitted by the Network Partner. GRN Honest Brokers similarly ensure patient privacy for the use of pseudonymized biological samples (e.g., blood and/or tumor biospecimens) when “genetically matching” patients to clinical trials. Finally, GRN Honest Brokers are never directly involved in the research or final reporting.
No, only fully de-identified/pseudonymized data may be shared with a drug trial sponsor prior to patient enrollment in a drug trial or use of biospecimens in a device trial. According to the terms of a BAA or a DUA, GRN may grant third parties a license to de-identified or pseudonymized data sets (falling outside the scope of HIPAA) or disclose LDS data to facilitate healthcare operation activities on behalf of the Network Partner and/or for the proper management and administration of GRN or to carry out its legal responsibilities. As required by HIPAA, any such third party shall be bound by restrictions, terms and conditions that are no less restrictive than those existing between GRN and the Network Partner.
GRN is subject to the 2018 Revised Common Rule that regulates human subject research conducted or supported by one of the federal departments or agencies (approx. 20 agencies including the FDA and HHS) that have codified the Common Rule in regulation or agreed to comply with all parts of the Rule. Under the Common Rule, any non-exempt human subject research project conducted or supported by these departments and agencies must obtain informed consent (or a waiver of consent) from all research subjects, receive approval from an Institutional Review Board (IRB), and certify compliance with the Common Rule before beginning research.
GRN can help you translate genomic tissue and longitudinal molecular data into value by providing tissue matches with biomarker and/or whole-exome sequencing with real world data. Because specimens are prospectively consented and collected under a separate and distinct IRB protocol, we protect PHI and matching longitudinal datasets. DNA, RNA and whole-exome sequencing from paraffin-embedded, surgical tumor specimens are linked with clinical data on mutation status, side effects, treatment response and more. You benefit from our reputation as a pioneering leader in DNA and whole-exome sequencing and active participation in clinical research to bring more patients new treatment options. Depending on tissue handling requirements for block-cutting and slide prep, you may benefit from limited internal coordination including no patient contact and very limited coordinator requirements for Phase I, II or III studies. This biobank tissue offers more detail than previously available from old biobanks which only provided tissue blocks by cancer type with minimal age and anatomic details. With GRN, you get the same level of detail by using molecular profiles to determine eligibility, i.e. histology, genomics, cancer stage, metastases, labs, radiology reports, comorbidities, treatment response and more