Detecting Healthcare’s Data Dilemma
By David Lareau
David Lareau is CEO of Medicomp Systems of Chantilly, VA.
“It is a capital mistake to theorize before one has data.” – Sherlock Holmes, “A Study in Scarlett” (Arthur Conan Doyle).
The great detective Sherlock Holmes understood the important role that data plays in decision-making. Whether you’re sleuthing or delivering patient care, you need data in order to make sense of things.
Not long ago, before EHRs were pervasive across health systems, providers struggled to obtain the data they needed for good clinical decision-making. Today healthcare has an abundance of clinical data, along with a new data dilemma: finding the right data at the right time.
In a recent webinar, our team asked 76 healthcare IT professionals and physicians about their biggest data-related challenges. According to 43 percent of the respondents, the top struggle was not a lack of data, but finding the right data at the right time. An additional 25 percent claimed they did not have access to the data they needed; 9 percent said they did not have enough data; and 6 percent complained of having too much data.
In other words, providers are challenged by the inability to access the data they need when and where they need it.
Consider what happens when a physician sees a patient and lacks ready access to their medical history, problem lists, medications, and test results. If the physician does not have access to the results of a critical test, the provider may re-order the identical test, possibly wasting healthcare resources and creating confusion about the accuracy of the patient’s records.
Interoperability advances are making it easier to share data, but true interoperability continues to be a struggle, thanks to a lack of standards, inconsistent data, and inadequate monetary incentives. We asked our webinar participants about their biggest barriers to achieving interoperability and 42 percent pointed to the challenge of data exchange. An additional 35 percent expressed difficulty applying data and making it actionable, while 20 percent reported difficulties organizing the data they received.
New exchange standards, such as FHIR, are making it easier to send data, but typically the incoming data is highly disorganized and not stored in an easily searchable format that adds value for clinical decision-making. In fact, the flood of incoming, unorganized data is creating new concerns about potential medical liability risks for providers. For example, a physician that inadvertently overlooks a critical abnormal finding that’s hidden within an incoming record could be held accountable for any ensuing patient complications.
Despite new standards and APIs that facilitate the exchange of data, much of the data exists in unstructured formats that are difficult to organize, include too many duplicates, and are not easy to search. In fact, an estimated 80 percent of all health data is stored in unstructured formats, such as free-text or scans.
Many in healthcare are optimistic that new technologies such as natural language processing (NLP) and artificial intelligence (AI) can be leveraged to convert dictated chart notes to free-text, and free-text to data that is in a format that is actionable for clinicians. The reality is that these solutions are still not sufficiently mature for most healthcare applications: the error rates for converting speech to text to data are, at best, between 8 and 10 percent, which is not reliable enough to support clinical decision- making.
Not all data is created equal – and not all data is equally usable for advanced analytics or for clinician use at the point of care. In order to be actionable for providers and usable for AI and analytics applications, data must be structured and organized in a way that facilitates viewing across clinical domains. One way to do this is to leverage technology that intelligently identifies, interprets, and links medical concepts and maps them to standard nomenclature, such as ICD-10, SNOMED, RxNorm, and LOINC.
Once disorganized data is converted to structured, actionable formats, it becomes more accessible to clinicians, allowing them to easily find the information they need during patient encounters and within their normal workflows. The structured data is also properly formatted for input into AI systems that use advanced algorithms to deliver clinical insights.
To ensure the ongoing creation of high-quality, structured data, we need to give clinicians the ability to capture coded clinical data as a byproduct of the documentation process and within their normal workflows. With more usable data, physicians can more readily access actionable information at the point of care. Organizations can more easily exchange quality information, and not just chunks of data that must be manually interpreted and organized. And, health systems are better equipped to harness the power of AI and the advanced analytics that enhance the delivery of patient care.
Detective Holmes understood that he could not optimally perform his job without data. To optimally deliver healthcare, providers need more than just data, which is why the industry must embrace technologies that make it easy to access the right data at the right time.