Derivation of a Clinical Model to Predict Unchanged Inpatient Echocardiograms

Area of Science:

• Cardiology
• Diagnostic Imaging
• Clinical Prediction Models

Background:

• Transthoracic echocardiography (TTE) is a frequent healthcare test.
• Repeat TTEs constitute a significant portion (24-42%) of all TTE studies.
• Identifying unnecessary repeat TTEs is crucial for healthcare efficiency.

Purpose of the Study:

• To develop a clinical prediction model for unchanged repeat TTE.
• To identify a subset of repeat TTE studies that may be unnecessary.
• To improve the utilization of cardiac diagnostic imaging.

Main Methods:

• Retrospective cohort study of hospitalized patients undergoing repeat TTE.
• Analysis of 601 repeat TTEs between October 2013 and September 2014.
• Development of the CAVES score based on independent predictors of major TTE changes.

Main Results:

• 37% of repeat TTEs showed major changes.
• Key predictors for major changes included: chronic kidney disease, acute myocardial infarction, valvular disease, ECG changes, and cardiac surgery (CAVES score).
• The CAVES score demonstrated good discrimination (C-statistic 0.78), with low probability of change in low-score groups (5.6% for -1, 17.7% for 0).

Conclusions:

• The CAVES score effectively predicts the likelihood of major changes in repeat TTEs.
• The model shows good discrimination and calibration.
• Further validation could establish the CAVES score for identifying potentially unnecessary repeat TTEs.