Validation of Diabetes Classification Algorithms from Cases Identified Through the National Diabetes Surveillance System in the Under 20 Population

Description

The National Diabetes Surveillance System (NDSS) uses a case definition that identifies Canadians with diabetes mellitus based on administrative coding data. While this definition has shown a high sensitivity in several validation studies, it does not distinguish between diabetes type. In adults, type 2 diabetes mellitus (T2DM) accounts for the large majority of cases. In children and adolescents however, type 1 diabetes mellitus (T1DM) is most often observed. Even so, rates of T2DM in those under the age of 20 are reportedly on the rise. These are separate diseases, and our approaches to their prevention and management should reflect their distinct characteristics. For children and adolescents, strategies using linkable data sources should be explored to distinguish NDSS case type.

Diabetes cases among children and adolescents under the age of 20 in British Columbia will be identified based on NDSS definitions. Drug usage for a subset of these cases will be derived, and provisional algorithms reflecting these patterns will be developed. Validation of these algorithms will then be performed against clinical datasets, using linkable prescription population-based drug databases available in British Columbia. Concurrently, a provisional algorithm relying on hospital discharge diagnostic coding will be explored.

The proposed algorithms will assist us to update our prevalence estimates for T1DM or T2DM in the under 20 population. In addition, accurate linkable prescription drug data and hospital discharge coding classification strategies would expand the data sources available to the NDSS and enrich its definition.


Research Team

Jeffrey Johnson, Saskia Vanderloo, Sumit Majumdar, Hude Quan, Clark Mundt

Research Funding

New funds for this project have been obtained from a CIHR Operating Grant "Using prescription drug information in NDSS: Validation of a classification algorithm for type 1 and type 2 diabetes"

Funding period December 2006 to March 2010.

 

Website Content Copyright © Alliance for Canadian Health Outcomes Research in Diabetes | Web Design By: Backstreet Communications