Nine case definitions will be discussed.  Please refer to the MCHP report for the full list of case definitions investigated.

Note: H = hospital; P = physician, Rx = prescription drug; One-year case definitions are for fiscal year 2002/03, two-year case definitions are for fiscal years 2001/02 – 2002/03, and three-year case definitions are for fiscal years 2000/01 – 2002/03

Crude prevalence for these case definitions ranged from 4.4%-8.2%.  For a given set of case definitions, increasing the number of years of data used resulted in an increase in prevalence estimates. For example, estimated crude prevalence using case definition a (1+H or 1+P, 1 year of data) was 5.8%; this increased to 7.1% when 2 years of data were used (case definition d).

Fewer cases are being captured when a more stringent case definition is used.  For example, changing the case definition so that the number of physician visits resulting in a diabetes diagnosis increases from one to two results in a decrease in prevalence estimates.  This is illustrated through the comparison of case definitions a and b; here prevalence decreases from 5.8% to 4.4%.

Alternatively, expanding the case definition to include another source of data (here prescription usage) increases the estimated prevalence.  This is illustrated, for example, by comparing case definitions a and b to case definition c.  Including the prescription usage data increases estimated prevalence to 6.5%.