The impending adoption of ICD-10 – now set for Oct. 1, 2014, according to the final rule just issued by the U.S. Department of Health and Human Services (HHS) – has stimulated an increased interest by the American healthcare industry in computer-assisted coding (CAC) systems. Everyone is keenly aware of the staggering increase in code complexity and specificity demanded by ICD-10, and the anticipated impact on coder productivity. The benefits of concurrent CAC have not been demonstrated sufficiently yet.
CAC (even concurrent models), however, has certain recognized limitations. Coding, computer-assisted or otherwise, is obviously dependent on complete, precise and accurate clinical documentation. Those in charge of clinical documentation improvement (CDI) programs long have appreciated this and have served as important promoters of HIM coding integrity. CDI initiatives progressively have moved toward concurrent and comprehensive processes as the need for presence on admission and continuous confirmation requirements have been mandated by payer review entities.
With the implications of ICD-10, CACs and the recovery audit contractors (RAC) in mind, this convergence of forces now has made the consideration of computer-assisted CDI a pressing reality. While the projected impact of ICD-10 on coding performance is undisputed, a corresponding burden to meet the enhanced requirements for specificity unquestionably will face CDI specialists as well. But CDI needs not only are shared with those of coders; they extend to significant clinical prerequisites that mandate credible supporting documentation, beyond that which defines an ICD-10 code itself. Thus the contribution that CAC provides is only one final benefit in the coding process. CDI additionally requires a pre-CAC, computer-assisted analytic system to identify clinically implied but omitted documentation.
CAC systems all are based on various implementations of natural language programming (NLP). This technology has a long history in the healthcare industry. It has been employed most commonly in the automated coding of clinical records, and it is virtually synonymous with CAC. But it has much broader utility than simply identifying terminology that can be coded in the clinical record. NLP can form the basis of a CDI expert system.
Expert systems typically are constructed on rules-based protocols that attempt to replicate human thought processes in arriving at appropriate diagnostic deductions and therapeutic interventions. In the case of clinical documentation improvement, a CDI expert system also can generate appropriate clinician-to-clinician clarifications with a high level of clinical confidence.
But the expectations of such advanced NLP-driven systems only can be met by integrating existing terminology with existing clinical data elements. Such integration should produce the same logical inferences of clinically plausible diagnoses and procedures that a CDI specialist would generate. The salient advantages of making such a process data-driven and automated are speed and consistency.
In the inpatient setting, CAC can be a helpful tool within the context of a larger coding and documentation process and workflow, but only when CAC isn’t your first course of action.
While CAC historically has been viewed as a coder productivity enhancement tool, the benefits of CAC can go well beyond this narrow focus, helping improve data integrity, consistency and payment accuracy – but only when CAC is implemented in tandem with a health organization’s clinical documentation improvement program.
In the inpatient setting, CAC alone won’t make your source data any better, and productivity gains from CAC are not as impressive as seen in the outpatient arena.
No matter how sophisticated the CAC tool, the unavoidable, fundamental fact is that a CAC program only assigns codes based on documentation in the record. Clinical documentation needs to be accurate in order to get the most out of CAC tools.
The role of the documentation specialist in the setting of concurrent CAC still requires validation and investigation of CAC-suggested diagnoses. The CDS still needs to determine if the diagnosis is clinically supported, consistent and determined in the absence of conflicting documentation. The CDS must determine if the documentation will hold up under repayment scrutiny, demands from the audit contracting industry and compliance requirements.
Computer-assisted coding can provide an important support tool for all clinical documentation improvement (CDI) programs. The prerequisites for success include creating an established, successful CDI program and defining the collaborative, tandem workflow processes involving CAC, concurrent documentation review and final coding.
About the Author
Melinda Tully MSN, CCDS, CDIP is the senior vice president at J. A. Thomas and Associates. Mel has extensive experience as a provider in multiple healthcare arenas, a clinical manager in a large academic facility, and as an expert in clinical documentation improvement (CDI). She has played an important role in the development and expansion of advanced CDI for the past 13 years. She is recognized for her expertise, vision, and promotion of CDI. Mel is a recognized national speaker for compliance, clinical documentation, and the Hospital Quality Initiative.
Contact the Author
To comment on this article please go to firstname.lastname@example.org