Optimal strategy to identify incidence of diagnostic of diabetes using administrative data

Table 1 New cases of diabetes and kappa agreement by clearance period and exclusion criteria

	Exclusion criteria for prevalent case
	One hit method			NDSS* method
Clearance period (years)	Number of new cases	Incidence proportion**	Kappa agreement***	Number of new cases	Incidence proportion**	Kappa agreement***
1	91872	2.9%	0.427 (0.420-0.432)^†,§	91872	2.9%	0.482 (0.479-0.485)^†,§
2	52508	1.7%	0.739 (0.731-0.742)^†,§	52508	1.7%	0.810 (0.801-0.822)^†,§
3	43473	1.4%	0.849 (0.845-0.851)^†,§	45525	1.5%	0.898 (0.896-0.900)^†,§
4	39595	1.33%	0.900 (0.900-0905)^†,§	42791	1.4%	0.936 (0.933-0.937)^†,§
5	37473	1.23%	0.935 (0.933-0.937)^†,§	41261	1.3%	0.957 (0.956-0.958)^†,§
6	36111	1.16%	0.957 (0.955-0.958)^†,§	40218	1.29%	0.971 (0.970-0.973)^†,§
7	35094	1.14%	0.973 (0.972-0.974)^†,§	39475	1.28%	0.982 (0.981-0.983)^†,§
8	34400	1.11%	0.985 (0.983-0.986)^†,§	38975	1.26%	0.990 (0.989-0.991)^†,§
9	33932	1.09%	0.993 (0.992-0.994)^†,§	38627	1.25%	0.996 (0.995-0.996)^†,§
10	33498	1.08%	100	38342	1.24%	100

*: National Diabetes Surveillance System case definition in Canada.
**: Incidence proportion = (Number of incident cases of diabetes for the year 2002/population at risk of diabetes for the year 2002)*100; number of diabetes cases for the year 2002 using NDSS case definition criteria is 263,213; population of 20 years old or older for the year 2002 is 5,686,255;
***: Agreement between 10 years (expected) and observed clearance period according to the clearance period and exclusion criteria for prevalent case;
^†: 95%Confidence interval;
^§: P < 0.0001, Chi-square test statistic for homogeneity.

ISSN: 1471-2288