How AI is Instilling Innovation into CDI Programs

One such example is computer-assisted physician documentation (CAPD).

Artificial intelligence (AI), natural language processing (NLP), and “the engine,” machine learning, began creeping into the clinical documentation integrity (CDI) profession 10 years ago, whereas it previously was a field that depended solely on clinical documentation specialists’ (CDS) and coding professionals’ knowledge and relationship with providers. AI computer assistance accelerated the review of a plethora of notes and mountains of data.

Not unexpectedly, apprehension in the CDI profession arose regarding the security and longevity of CDI professionals; how would the CDS role change? There were also concerns for the accuracy and compliance of AI, and provider acceptance of this new technology. Was this to be an Alice in Wonderland experience? It was suspected that AI could add value to highly specialized CDS reviewers, and do it in a split second – and compliantly.

A decade later, the CDI industry has experienced phenomenal growth, enhancing patient care through complete and accurate clinical documentation. Fiscal outcomes are now congruent with services provided. Artificial intelligence has now successfully assisted CDI practitioners and more, possibly even alleviating some components of physician burnout by supplementing CDI processes with faster, more efficient presentation of data, clinical information, and possible diagnoses, to a high degree of precision.

Learning About CAPD

Computer-assisted physician documentation (CAPD)is a tool used to scan the chart for clinical indicators, then pose a request for clarification to a healthcare provider, in much the same way that a CDS or coder creates a query. The implications of this automated workflow are far-reaching for hospital revenue cycle professionals. With AI as part of a holistic CDI program, the most common CDI questions and concerns can be posed to a physician and resolved, at the time of documentation, thus freeing the CDSs and coders to focus on advanced practice strategies and more challenging case reviews. The analytics derived from AI in revenue cycle can yield targeted improvement plans to make human resources more agile, accurate, and efficient.

From the physician perspective, any additional work, like CDI queries and coding queries, can be time-consuming, representing yet another task to clear from a bloated inbox of messages and chores. Presenting AI-generated suggestions based on evidence within the chart, while in the moment of patient care, with the information being quickly consumable with minimal cognitive burden, is far better than retrospective work – often coming days later, and detached from the inline workflow and knowledge of the current patient condition. In addition, the ability to easily review the data points highlighted by NLP and record diagnoses agreed upon by clinical judgment in the notes – and even the problem list via a single-click interface – is a win-win for providers, the CDI team, and the parent organization. How? Faster and more accurate documentation of patients’ conditions by providers, alleviation of a significant proportion of the massive amounts of charts to be reviewed by CDI, and improved reimbursement and quality metrics for the healthcare system.

About Compliance

CAPD and CDS physician queries are designed to be compliant with applicable law and coding guidelines framed by the American Health Information Management Association (AHIMA). CAPD solutions are informed by a comprehensive understanding of the applicable healthcare program laws and coding guidelines from the U.S. Department of Health and Human Services (HHS), including the following: Medicare regulations on DRG classification for hospital payment; the Uniform Hospital Discharge Data Set (UHDDS) definitions for hospitals to report inpatient data elements in a standardized manner; the uniform Code Set Standards and official coding guidelines adopted by HHS under HIPAA; the American Hospital Association (AHA) Coding Clinic for ICD-10-CM and ICD-10 PCS for official coding advice; and the HHS Centers for Medicare & Medicaid Services (CMS) informal guidance on physician queries.

CAPD compliance best practices include the following oversight and monitoring activities and safeguards for the query process:

• Schedule testing and quality review of CAPD clarifications for clinical evidence and validity for each setting;
• Provide complete transparency and audit reports for physician response and agreement rates for CAPD clarifications. These reports should be readily available to the CDI staff, coding professionals, and quality reviewers. The reports should include sources of evidence and physician response;
• Train physician leadership, CDI staff, and coding staff on AI/CAPD design controls and clinical evidence-based documentation strategies;
• Establish a reconciliation process for CAPD and final coding when discrepancies on DRG assignment occur;
• Establish a multidisciplinary steering group for oversight and compliance monitoring of the CAPD implementation, and follow up on any deviations from that standard with corrective action consistent with the hospital’s internal compliance plan and reporting systems;
• Engage with the hospital chief compliance officer as an active voice and participant in CAPD compliance oversight; and
• Align the CDI program with the hospital’s compliance program for auditing, monitoring, and reporting activities, and provide compliance education to the CDI team.