Due to its inherent inconsistency, nearly half of all clinical data isn’t usable for driving business decisions.
How do we know? Diameter Health has spent the last decade focused on how to make clinical data actionable. As a result, we’ve identified significant challenges. Among them:
- Non-standard, erroneous, or missing codes are pervasive in data sourced from EHRs, labs, and HIEs, limiting its use in analytics
- Information is frequently documented in incorrect places in the medical record, meaning it can be lost to downstream applications
- Duplicate entries of the same prescription or diagnosis are common, obscuring medical facts under mountains of information
- Multisource and multiformat patient information across sites of care, laboratories, and data aggregators must be synthesized to understand the patient journey
The solution? Organizations that invest in acquiring clinical data can address its usability in-house, but that route can be costly, demanding, and time consuming. For instance, our customers and partners have learned that terminology mapping and other approaches yield a modest return in terms of data usability and force data scientists to focus on data cleansing, working well below the top of license. Further, the work is never done, as each new data source requires additional work.
Another approach is to simply start Upcycling DataTM, which is what Diameter Health’s scalable, automated Fusion solution is all about.
What is Upcycling Data and why is it needed? Simply put, upcycling is the process of aggregating clinical data from multiple sources and transforming it into a data asset that can be readily used for multiple purposes.
Upcycling starts with making data more useful by normalizing codes to interoperable national standards. It extends to filling in data gaps caused by poor documentation and enriching the data so that it can be analyzed, searched, and shared. Upcycling also includes creating a longitudinal record for a single patient from disparate sources, in which duplicate entries have been removed to make the information less “noisy.” The best part? It’s completely automated. Moreover, upcycling considers the need for integrating clinical data into your infrastructure and exchanging it using standards like FHIR.
Dr. Paulo Pinho, Vice President and Medical Director of Innovation at Diameter Health, offers an example that illustrates the value of Upcycling Data. Consider a prescription for Vicodin, which is documented in the medical record as Vicodin 5/300 with an NDC code of 00074-3041-13. As a first step, Fusion will normalize the code to a national standard, in this case RxNorm code 856987. That matters, because although each brand name (Vicodin, Norco) has a different NDC code, the ingredient combination can be normalized to one single RxNorm code. Normalizing unambiguously to national standards makes it easier to apply the information for clinical decisions. It also readily supports analytics and exchange of data. Consider that now the information can be used to:
- Query across an individual’s record or within a population to identify patterns of usage
- Query the data by base ingredient in the case of an allergy for an individual, or for an adverse event identification across a population
- Establish a crosswalk to a medication that is on a particular health plan’s formulary, reducing costs
- Adhere to interoperability requirements (for example, the FHIR standard requires use of RxNorm in the Medication and Allergy Intolerance resources)
But this clinically precise semantic normalization is only the first step. Other parts of the upcycling process are also critical to delivering value. In our example, Vicodin is decomposed to its component ingredients of acetaminophen and hydrocodone. Using this enriched data, a clinician could avoid prescribing or dispensing Vicodin to someone with an acetaminophen allergy and simply look to prescribe hydrocodone.
The drug is also classified by pharmacologic class as an opioid agonist, and a Schedule II, controlled substance. For a public health initiative, the data can be queried to identify a population on a particular controlled substance without needing to use all brand names, and analytics could reveal patterns indicating over-prescribing or drug diversion behavior in a population.
This post is part of a three-part series. Miss the first blog? Ashley Basile, Chief Product Officer wrote “Clinical Data is a Resource Waiting to be Tapped.” Then sign up below to receive the next in the series discussing the value of UpcycledTM clinical data to your organization.