Most healthcare applications present clinical data by date or encounter. But, studies have shown a real advantage to a problem-centered workflow for healthcare data retrieval.
No matter what you call it: problem-centric view, problem-oriented view, problem concept maps, relevant data display, it doesn’t matter. Healthcare data displays are generally date-centric, or encounter-centric, and not problem-centric.
What do I mean by that? Quite simply it means that it doesn’t matter what application you are using, whether it’s an electronic health record (EHR), an analytics or care management platform, or a multitude of other healthcare applications, the data per patient is generally presented by; date or encounter. Each entry is likely identified as a particular type of report (i.e. lab, radiology, pathology, etc.) and if you are lucky, the data is tagged by specialty. But if not, the clinical notes will usually identify the specialty of the provider that created them.
The Journal of the American Medical Informatics Association (JAMIA) recently published an article on the advantages of a problem-centered workflow for data retrieval using problem concept maps created by the University of Washington. In this article, the value of aggregating data by condition for review by a clinician for medical decision making is clearly articulated. This article also points out that the need for a problem-oriented medical record has been discussed starting in 1968! While we have advanced our collection, retrieval, and use of data in the healthcare world today, a large gap still remains.
Why take a problem-centric approach to healthcare data?
Increasingly, payers, especially those involved in value-based care arrangements are asking for the ability to aggregate data in a problem-centric way. We often hear from our payer clients that they want to be able to identify all their members that are taking a medication related to a certain condition (e.g. bronchodilators for the treatment of COPD). Or, in the reverse, they want to be able to identify all of the medications commonly used to treat a condition (e.g. Diabetes Mellitus Type II with Peripheral Neuropathy). Similarly, many of our clients want to be able to quickly identify lab results, radiology tests, and other procedures that should be performed when diagnosing or treating a specific condition.
The desire to aggregate data in this way is driven by the need to accurately identify risk, highlight potential gaps in care, support quality related activities, and improve health outcomes of the high-risk, costly patient populations by proactively intervening at the right time.
For instance, a member newly diagnosed with COPD should be evaluated for their lung function through standard Pulmonary function tests such as, spirometry, chest x-rays, and 6-minute walk tests. Members with acute exacerbations of their COPD will likely be given a prescription for a Corticosteroid to control the exacerbation, as well as having a spirometry test, a chest x-ray and even possible supplemental oxygen. Traditionally, payers could comb through their claims databases to understand if these tests, procedures, or medication have been billed out in conjunction with a diagnosis of COPD but that method is retrospective in nature. Claims data is used in analytics only after adjudication which can take weeks or even months if appeals and secondary claims are required.
The important role of condition data mapping
Payers are beginning to get more and more clinical data in the form of clinical notes, radiology reports, summary documents (CDA’s) and HL7v2 messages. While that is great, none of that data is easily parsed and extracted into a meaningful longitudinal, problem-centric displays. For example, if a member has a lifetime problem list of 35 conditions, and a lifetime medication list of 175 medications, payers need to know which subset of the medications were likely used to treat which conditions. If the problem list contains “hypertension” and the medication list contains “Lisinopril,” they need a link between the hypertension record and the Lisinopril record. Ideally, this could be done for all possible and logical relationship linkages such as the procedures performed that link to specific conditions, the lab results that were used to diagnose specific conditions, and the lab results that are linked to the therapeutic intent of medications.
Clinical terminology experts are here to help
Sound like a pie in the sky idea? It probably is if you try to do this without the right experts or a dedicated team of clinical terminologists. As terminology experts with deep domain expertise and access to the Wolters Kluwer content libraries and clinical expertise, we are uniquely positioned to help you organize your data into meaningful building blocks for just this sort of problem. We understand all of the clinical components (labs, meds, procedures, symptoms, findings, devices, and more) related to specific conditions and how to identify each clinical entity using standard terminologies to help aggregate your data into problem-centric views.
The process is simple and easy through a services engagement with Health Language. First, you tell us what conditions are of particular interest to you. We will want to understand your use case and data flows to design the best content that will help you achieve your goals. We then get to work building granular value sets that will identify specific labs, medication and other clinical entities of interest related to the conditions that you identified as most important to you. We will work together with you to make sure that you are successful and provide implementation support and fine tuning once the content is delivered to you.
Ready to get started or interested in learning more? Reach out to us today!