Eventually, algorithms will be broadly used to assess the appropriateness of diagnostic imaging orders by clinicians, but not by January 2020. That's the date CMS has mandated that such algorithms will be deployed as part of PAMA (Protecting Access to Medicare Act). This topic was discussed in a recent article (see: Docs Brace for Medicare 'Appropriate' Imaging Rule). It's long so link to it if you are vitally interested. Below is a short excerpt from it:
As the medical community braces for implementation of the Protecting Access to Medicare Act (PAMA) by the Jan. 1, 2020 deadline, some wonder if it's even feasible or if another program delay is on the horizon. The policy, aimed at reducing unnecessary testing, mandates that all advanced diagnostic imaging orders go through an algorithm that provides key confirmation codes required when Medicare is billed later on for the service. Dubbed a "clinical decision support mechanism" (CDSM), this software processes each CT, MRI, nuclear medicine, and PET order before spitting out its verdict to the ordering professional: "appropriate," "maybe appropriate," or "rarely appropriate".... Eventually, Medicare will not pay claims or advanced imaging without these codes. The program has been delayed several times, and the 2020 deadline equates to a "soft launch," as improper imaging claims won't be rejected until the following year, with no firm deadline for additional penalties beyond that. Still, many physician organizations have raised concerns about the policy...."The way PAMA requires AUC [Appropriate Use Criteria] to be implemented is very prescriptive, very limiting to physicians, with the potential for unintended consequences" [according to a physicians' spokesperson].[Physician groups are] asking that the Centers for Medicare and Medicaid Services (CMS) postpone PAMA implementation "until we come up with an alternative way to apply AUC and practice value-based imaging."
The article correctly identifies an algorithm to identify a priori inappropriate imaging orders as a type of clinical decision support mechanism (CDSM). The mandatory use of the algorithm should reduce healthcare costs for CMS. I have recently blogged about clinical decision support software (see, for example: PCPs Responsible for Allocating About 90% of Total Hospital Costs). This proposed algorithmic software could also be classified as an example of diagnostic/predictive analytics (see: Diagnostic and Predictive Analytics and Their Possible Link to the Future of the LIS). The broad use of predictive and diagnostic algorithms is surely on the horizon and will be the basis for algorithmic medicine (see: What’s In The Box? Issues In Algorithmic Medical Care). However, the chance that CMS could internally develop and release such an algorithm in the proposed time frame is near zero. The chance that it could successfully outsource the development of such a tool to a medical software company is also close to zero.
So what does the future hold for this plan? Possibility #1 is that the release of the software will be postponed but the discussion of the possibility of the future release of the software may have the desired effect of proactively causing radiologists and clinicians to be more cautious about ordering advanced imaging studies in questionable cases. Possibility #2 is that CMS will release, perhaps on a delayed basis, such an algorithm that functions as blunt tool to achieve the desired goal, denying confirmation based on obvious criteria. However, it could exclude the most obvious cases of inappropriate imaging orders. The experience with such software could also form the basis for the release of a more sophisticated tool in the upcoming years based on the real world experience that would be gained.
Comments