Quantifying a rare disease in administrative data: the example of calciphylaxis

Area of Science:

• Nephrology
• Epidemiology
• Medical Informatics

Background:

• Calciphylaxis is a rare, severe condition in chronic dialysis patients with unknown epidemiology.
• Lack of a unique ICD code hinders identification in large databases like the USRDS.
• This study aimed to develop an algorithm to identify calciphylaxis cases and analyze incidence and mortality.

Purpose of the Study:

• To develop and validate an algorithm for accurately identifying calciphylaxis cases.
• To investigate the incidence and mortality rates of calciphylaxis in chronic dialysis patients.
• To provide a model for rare disease research using administrative data.

Main Methods:

• Developed an algorithm combining ICD-9 code 275.49 with skin biopsy procedure codes.
• Validated the algorithm using natural language processing and medical record review in a cohort of 11,451 chronic hemodialysis patients.
• Applied the validated algorithm to the USRDS to determine calciphylaxis incidence and mortality.

Main Results:

• The algorithm demonstrated high accuracy: sensitivity 89.2%, specificity 99.9%, AUC 0.96.
• Identified 649 incident calciphylaxis cases in the USRDS, showing a rising incidence, particularly after 2006-2007.
• Calciphylaxis patients experienced 2.5-3 times higher mortality compared to the general chronic hemodialysis population.

Conclusions:

• A novel algorithm effectively identifies calciphylaxis, revealing a significant increase in its incidence.
• The study highlights the high mortality associated with calciphylaxis, underscoring the need for further research.
• The developed algorithm serves as a valuable tool for calciphylaxis research and a template for studying other rare diseases.