Implementation of a Real-Time Documentation Assistance Tool: Automated Diagnosis (AutoDx).

Area of Science:

• Health Informatics
• Clinical Documentation Improvement
• Healthcare Management

Background:

• Clinical documentation improvement (CDI) programs are standard in healthcare to enhance provider documentation.
• Current methods, like coding queries, are time-consuming and disrupt provider workflows.
• Alternative strategies are needed to improve documentation without adding burden.

Purpose of the Study:

• To evaluate the impact of a real-time documentation assistance tool, AutoDx, on coding query volume.
• To assess AutoDx's effect on encounter-level outcome metrics, specifically case-mix index (CMI).

Main Methods:

• AutoDx was developed using existing electronic health record tools, generating in-note messages based on clinical conditions.
• The tool initially focused on diagnoses like electrolyte deficiencies, obesity, and malnutrition.
• Implementation occurred in Hospital Medicine and the Neuro Intensive Care Unit (NICU).

Main Results:

• Coding queries for targeted diagnoses decreased by 57% (p < 0.0001) in Hospital Medicine.
• In the NICU, the proportion of AutoDx queries decreased from 54.9% to 37.1% (p < 0.0001).
• A significant increase in CMI was observed post-implementation (4.00 to 4.55, p = 0.02).

Conclusions:

• The AutoDx tool effectively reduced coding queries for specific diagnoses in diverse clinical settings.
• The implementation of AutoDx was associated with a significant improvement in CMI.