Data Analytics: The Cornerstone of Meaningful Use
By Kim Garriott
The execution of a successful enterprise imaging strategy provides a platform to acquire, retain, and make meaningful use of a wealth of clinical information that has previously been unmanaged. More specifically, it affords a unique opportunity to build an archive of clean and uniform imaging data being produced by as many as 40 different service lines within a typical health care organization.
Without a mandate for required data elements and uniform data sets, a vendor neutral archive (VNA) established solely for the purpose of storing data becomes a very expensive missed opportunity. Without these principles, images may be retained, but meaningful use of the images becomes nearly impossible. Careful consideration must therefore be given to data governance principles in the early stages of planning an enterprise imaging strategy with particular attention paid to naming conventions and data requirements.
In order to retrieve precise information from patient records, HIT organizations first must learn to speak the right language, and naming conventions are that language. When planning naming conventions, consider what data components will comprise imaging study descriptions; the department where the images were acquired, the modality used to acquire the image, and an anatomical reference are common choices.
With regard to anatomical references, decisions must also be made as to the granularity of anatomical descriptions that will be used. A dermatologist, for example, might need a greater level of granularity than the majority of other caregivers searching for images within the same patient's EHR.
Another important area of consideration is presentation, or how the images will be displayed to the viewing provider. If the image descriptions are displayed in a worklist format, for instance, the first characters within the naming description should be modality type or department—something relevant to the end user.
Until recently, a standard vernacular of imaging names did not exist. Therefore, each organization developed its own internal vocabulary, including names for imaging and procedural studies. Over the past few years, however, the Radiological Society of North America developed a comprehensive lexicon, RadLex, for the purpose of standardizing the indexing and retrieval of radiology information resources.
RadLex unifies and supplements other lexicons and standards, including the SNOMED-CT, LOINC, and Digital Imaging and Communications in Medicine, or DICOM. Although RadLex provided a turning point in the unification of naming conventions, it is focused strictly on radiology and does not reflect digital photography and other point-of-care imaging studies; thus, additional data are required.
Determining the data that need to be associated with each set of images committed to a VNA must be accomplished early in the planning of an enterprise imaging implementation, well before the first image is presented for retention. Attention should be given to how each set of images, or each study, will be uniquely identified within the system. In radiology, for example, each diagnostic imaging test is assigned a unique identifier called an accession number. This may be unique to radiology, but the same concept applies when thinking about imaging at the enterprise level.
Clearly, careful deliberation is needed when developing naming conventions. A standardized imaging study name should, for example, be a required data element during the acquisition process. Doing so at the point of acquisition ensures that imaging studies are consistently named and that caregivers can easily identify the images they are looking for from a worklist. Without proper naming, searching for images will become a point of frustration for the entire care team. Proper and uniform identifiers will also be commonly used in analytics queries.
If the imaging studies will need to be logically segregated in the VNA, for example, separating research imaging studies from imaging studies that will present within the patient's EHR, specific and uniform data elements will need to be assigned to indicate to which logical partition the study should be committed.
By establishing uniform naming conventions and data sets at the beginning of the enterprise imaging process, health care organizations save themselves tremendous work on the back end. The ultimate goal is not simply organization; it's the analysis of collected data to present meaningful, actionable information to caregivers in the right format at the perfect time.
With such a well-planned, well-executed approach to acquiring, storing, and retrieving medical images, providers will be able to use predictive analytics to better assist physicians. Based on the patient's medical history and the physician's specialty and past clinical behavior, data analytics tools will soon be able to retrieve and compile the imaging studies and reports a caregiver will want to review during a patient's visit, presenting those images directly to the physician at the point of care—the very essence of what meaningful use is all about.
• Naming conventions matter: Establish data governance principles during the earliest stages of an enterprise imaging strategy with particular care taken to uniformly identify naming conventions and data requirements that will work well in a data analytics environment.
• Imagine the outcome: Details are the cornerstone of a well-executed enterprise imaging strategy, and it is those details, coupled with the ability to intelligently retrieve and analyze what has been collected, that will deliver actionable results to improve patient care, clinician workflow, and satisfaction in the end.
— Kim Garriott, principal consultant of health care strategies for Logicalis Healthcare Solutions, helps health care clients develop thoughtful enterprise imaging strategies that maximize the value of their HIT projects, ensuring the expected business outcome is achieved.