Comparing outcomes in renal replacement therapy: how should we correct for case mix?

Area of Science:

• Nephrology
• Public Health
• Biostatistics

Background:

• Evaluating clinical practice effectiveness is vital for justifying healthcare costs.
• Risk stratification aids in identifying patients most likely to benefit from renal replacement therapy (RRT).
• Existing complex risk stratification methods create small risk groups, hindering center comparisons.

Purpose of the Study:

• To compare the effectiveness of three simple risk stratification methods for patients undergoing renal replacement therapy.
• To identify the optimal method for predicting mortality and discriminating between risk groups.

Main Methods:

• A cohort of 1,407 patients commencing RRT across five European countries over 7 years was analyzed.
• Three methods were compared: Method 1 (age >55 & diabetes), Method 2 (age >70 & comorbidities), and Method 3 (number of comorbidities).
• Kaplan-Meier survival curves and Cox's regression models were used to assess survival differences and mortality relationships.

Main Results:

• All three methods showed significant differences in patient survival between risk groups.
• Cox's regression analysis indicated Method 2 offered the greatest discrimination between risk groups.
• Method 2 demonstrated superior sensitivity (84%) and specificity (80%) in predicting mortality compared to Methods 1 and 3.

Conclusions:

• Method 2, incorporating age and comorbidities, is the most effective simple risk stratification method for renal replacement therapy patients.
• This method shows promise for comparing survival data and presenting RRT outcomes.
• Further validation in diverse populations via prospective studies is recommended before widespread clinical adoption.