Given NDSCs market footprint, market entrants seeking differentiation, link the requirement for indication capture to the design of our qualified Clinical Decision Support Mechanism (qCDSM). Linking concerns related to structured indications/high rates of free text to ‘the NDSC approach.’
Some qCDSM producers have represented the ability to ‘derive’ indications to automate CDSM consultations, eliminating the need to enter indications for advanced imaging exams, and cite an example (discussed below) as to how this might ‘work.’
This approach creates compliance risks and will not create a comprehensive framework for quality improvement.
Can a qCDSM automate indication selection when placing orders for Medicare Outpatient Imaging?
Free text indications are routinely entered for every single outpatient advanced imaging order placed today. A consultation of Appropriate Use Criteria (AUC) replaces this free text with indication selection, as defined by the AUC.
As with any change management, the steps to implement a qCDSM requires change of behavior, namely consultation of the mechanism, which require consultation of AUC. Each criteria for an advanced imaging exam linked to a series of indications, or clinical questions to answer.
The automation of this process has been suggested as an alternative to the process of indication selection (or even free text entry).
Can the indication selection for an outpatient advanced imaging order be automated?
As an example; MRI of the Lumbar Spine is appropriate (sic) if the patient has had a previous encounter for low back pain three months prior, received an order for a NSAID, and was recently referred to physical therapy the exam is appropriate.
Let’s dissect how this would be reflected in an order entry decision support workflow. All such tools rely upon the Clinical Document in order to determine the indication for imaging, such that an AUC can be accessed.
- “Recent referral to physical therapy” may not be a discrete data element, such as a lab or medication order. It may instead be part of a free text note, it might also be simply a phone call. It is (supposed to be) documented in the physician’s note. It could also have been sent to another scheduling application or EMR scheduling module, in which case it may not be part of the CDA.
- Medication orders are simple, although here the medication is often OTC (over the counter); ibuprofen is an OTC NSAID. Thus, the med may not be documented outside of a free text note.
- Encounter for LBP three months ago could be in the problem list, but it may not have been documented.
The indication noted, is one of many possible indications for MRI of the Lumbar Spine, and a subset of the more than 500 indications required for reasonable and complete coverage of the Priority Clinical Areas (PCAs). Only a minute subset of the myriad of reasons why an advanced imaging order might be placed.
So using the above example, its worth highlighting and asking the question, “does the organization routinely and consistently document their encounters in their EMR in such a way that reliable information extraction can occur?”
To understand how advanced imaging is utilized and to leverage the opportunity associated with Imaging Decision Support to answer critical questions about savings and appropriate utilization, accurate encounter data and a specific indication is required.
Indication Selection is not a limitation of the Decision Support Tool.
The short story is that if there is a consistent way a health care organization documents and organizes this data, CareSelect™ EHR data extraction is fully capable to extract, parse and analyze this data; our engine can encode the necessary business and clinical logic, that we could automatically map to a specific indication.
- We routinely extract and parse this data (using AI) as part of our delivery of InterQual AutoReview™, where countless, complex data elements are parsed to automate chart review to determine the appropriate admission level.
- This is exactly the way the CareSelect Lab and Patient Blood Management (PBM) solution work. The majority of our content delivery requires zero user interaction, enabling surveillance coupled with data driven, targeted interventions across the entire service line.
The free text indication, currently in use within the EMR coveys important clinical information, the structured indication entry to access the AUC coveys more to the radiologist. To provide the best possible answer through imaging, a clinical question must be asked. The indication, coupled with the EHR data form this question. Specifically and accurately.
Incorrectly doing an ‘automated consultation of a CDSM’ also has implications. Radiology has less information than before, and the provider might be incorrectly assigned outlier status. What if no indication can be reliably be derived? In these cases mechanism interaction would be required such that AUC consultation can occur. So what are the choices when an indication cannot be derived?
- Capture a single indication.
- Interactively ask clarification questions in the workflow to derive the indication.
- Force re-work downstream on the furnishing service as they will need to clarify the indication to properly furnish the service.
Finally the CareSelect Imaging focused implementation approach limits implementation to cover exams and indications within the Priority Clinical Areas. Requiring some interaction within the scope of the PCA and zero CDSM interaction for other exams and indications. This ensures high quality data regarding imaging utilization; enabling clinical transformation, targeted practice improvement and cost savings, while still maintaining regulatory compliance.
In summary, the CareSelect Platform has more than enough capability to automate the selection. The challenge of framing the clinical question in the healthcare providers mind as they decide to place an order for advanced imaging requires asking the question through an indication selection.
For More Information: Reason-for-Exam Robots: the Rub