Validation in Alberta of an administrative data algorithm to identify cancer recurrence

Area of Science:

• Oncology
• Epidemiology
• Health Informatics

Background:

• Improving cancer recurrence surveillance is crucial for women of reproductive age.
• Population-based data can enhance cancer research and patient monitoring.

Purpose of the Study:

• To evaluate the validity of administrative data for identifying cancer recurrence in women of reproductive age.
• To establish a reliable method for cancer recurrence detection in this demographic.

Main Methods:

• A cohort of 200 women diagnosed with cancer and with a history of pregnancy was randomly selected from the Alberta Cancer Registry (2003-2012).
• Administrative data were linked and analyzed using various definitions of recurrence based on diagnosis and treatment dates.
• Chart review served as the gold standard for confirming recurrence.

Main Results:

• Chart review confirmed recurrence in 26 women.
• The optimal definition for recurrence involved "2 or more cancer diagnosis codes 10 or more months from the diagnosis date," achieving 80.8% sensitivity and 81.0% specificity.
• This definition yielded a positive predictive value of 38.9% and a negative predictive value of 96.6%.

Conclusions:

• Administrative data can capture cancer recurrence in reproductive-aged women with moderate validity.
• Findings suggest administrative data is a useful tool but requires cautious interpretation due to potential limitations.