How accurately do nephrologists predict the need for dialysis within one year?

Area of Science:

• Nephrology
• Clinical Prediction
• Chronic Kidney Disease Management

Background:

• Accurate prediction of dialysis initiation is crucial for patient counseling and timely vascular access placement in chronic kidney disease (CKD).
• Current methods for predicting dialysis need rely heavily on clinician judgment.

Purpose of the Study:

• To assess the accuracy of clinician judgment in predicting the need for dialysis within a 12-month timeframe for CKD patients.
• To compare the accuracy of predictions made by different levels of clinical expertise.

Main Methods:

• Nephrologists in a pre-dialysis clinic predicted the time to dialysis initiation for 184 CKD patients.
• Predicted versus actual time to dialysis was compared, analyzing prediction accuracy across clinician grades.
• Multivariate logistic regression identified clinical parameters correlating with predicted and actual dialysis timelines.

Main Results:

• Clinician judgment demonstrated high sensitivity (95%) but moderate specificity (62%) for predicting dialysis within 12 months.
• Consultants achieved 71% accuracy, while trainees achieved 68% accuracy.
• Estimated glomerular filtration rate (eGFR) was the sole independent correlate for both predicted and actual dialysis timelines, with age and pruritus also associated with actual dialysis need.

Conclusions:

• Clinical judgment is a sensitive but not fully specific tool for predicting dialysis initiation in CKD patients.
• Enhancing clinician education could potentially improve the specificity of these predictions.
• Improved accuracy in prognostic information delivery will benefit patient counseling and preparation for dialysis.