Why Clinical Data is Hard to Use
5.11.2023 By Lucy Parente, Director of Strategy at Availity
Clinical and claims data can be powerful, strategic assets that help provider and payer organization to improve patient outcomes and satisfaction, comply with government regulations, control costs, and mitigate onerous administrative burdens. Unfortunately, making complex and diverse clinical data elements actionable and fit-for-purpose is expensive and unwieldy. Electronic health records (EHRs), data warehouses, and other information systems and repositories have enabled healthcare organizations to acquire—and even aggregate, to a degree—massive volumes of clinical and claims information. While these data acquisition and storage systems often represent the most significant technology investments made by payers and providers, their ability to consistently and accurately transforms data into information fit-for-purpose often require costly add-ons in the form of point solutions and technology overlays.
The difficulties of clinical data
Healthcare organizations that want to innovate with clinical data face a number of daunting obstacles.
Data is everywhere and everything. In some ways, healthcare is a victim of its own data acquisition success. In today’s electronic ecosystem, data can be found in multiple systems and departments. Data also appears in multiple formats, including text, codes, paper, digital, and multimedia. (Think of the data associated with a simple leg fracture—from doctor’s notes and x-rays to ICD-10 codes and pain medication prescriptions). Data challenges in this area will only grow more complex as more sources of data, such as patient-generated information from wearables, need to be collected.
Structured and unstructured data. EMRs can standardize data capture but lack of detail/options/nuance can crowd providers into a one-size-fits-all approach to documentation. Moreover, unstructured data, such as free-text clinical notes are typically written in natural language, making them difficult to analyze using traditional data analysis methods. Different healthcare providers may use different language, abbreviations, and terminologies, making it challenging to compare and combine data from different sources.
Inconsistencies and variability. Some data may lack consensus among clinicians regarding definitions and meaning. Additionally, new knowledge regarding medical conditions, such as cancer treatments, can add complexity to aggregation and analysis. Variability of data and its sources also makes standardization a worthy, but incomplete without data management. For example, coding variation within and across sources such as EHRs, labs, and HIEs creates tremendous complexity when it comes to integrating clinical data into an enterprise architecture. Other challenges contributing to the significant variation include information is frequently documented in incorrect places in the medical record. Duplicate entries of the same prescription or diagnosis are also typical, obscuring medical facts under mountains of information.
Completeness. Claims data is standardized, but incomplete without clinical data. To put it into perspective, understanding a patient’s health status and healthcare activities using health-related data is similar to assembling a puzzle. Each piece of claims and clinical data are puzzle pieces that provide information about a specific aspect of the patient’s health. Claims data and multi-source clinical data each have their limitations and cannot provide a comprehensive, accurate, and timely view on their own. To obtain a more complete understanding, it is necessary to integrate and combine both data sources and utilize their respective strengths. The future of health data analysis will depend on effectively assembling these puzzle pieces to create a more comprehensive view of patient’s health status.
Regulations. Evolving and emerging regulations make reporting burdens for providers more difficult and cumbersome. To make large sets of multi-format clinical data usable, healthcare organizations may attempt to leverage in-house resources and teams of data analysts. However, these approaches are costly, not scalable, and yield a modest return in terms of data usability. Further, the work is never done, as each new data source requires mapping, and data volumes continue to explode over time.
How to make data actionable
Upcycling Data™ is our process of integrating clinical data from multiple sources into a data asset that can be readily used for numerous purposes. It starts with normalizing codes to national standards and making sure data is semantically normalized to convey the clinical intent. Upcycling Data also includes adding rich metadata such as drug classes to make it easier to query and analyze. Upcycling Data extends to generating a longitudinal record from multiple patient encounters in which duplicate entries have been removed to streamline the data. All this is entirely automated, in near real-time.
Finally, our upcycling technology, known as Availity Fusion, features APIs and specialized data extracts to flexibly integrate clinical data into your infrastructure and data processing pipelines, support for data interchange standards like Fast Healthcare Interoperability Resources (FHIR®), and generation of data fit for specific business purposes to drive immediate value and ROI. Download “Why Upcycling Data Matters” to learn how to make your investment in clinical data count.