An algorithm to identify patients with treated type 2 diabetes using medico-administrative data

Table 2 Sensitivity (%) and Specificity (%) of the DIABECOLUX algorithm for various possible proportions (T2P) of type 2 diabetes in the diabetic population (%).

T2P^†	TP^†	TN^†	T2P.POPtot	(1-T2P).POPtot	SE = TP/(T2P.POPtot)	SPE = TN/((1-T2P).POPtot)	PPV = TP/Pos	NPV = TN/Neg
Min: 92.8%	20,581	1,110	20,581	1,597	100	69.5	97.7	100
93%	[20,581-20,626]	[1,067-1,110]	20,626	1,552	[99.8-100]	[68.7-71.5]	[97.7-97.9]	[96.1-100]
94%	[20,581-20,847]	[845-1,110]	20,847	1,331	[98.7-100]	[63.5-83.4]	[97.7-99.0]	[76.1-100]
95%	[21,036-21,068]	[731-1,109]	21,069	1,109	[99.8-100]	[65.9-100]	[99.8-100]	[65.9-99.9]
96%	[21,063-21,068]	[731-887]	21,291	887	[98.9-99.0]	[82.4-100]	100	[65.9-79.9]
Max: 96.7%	21,068	[731-732]	21,446	732	98.2	[99.9-100]	100	65.9

T2P: Proportion of type 2 diabetes in the treated diabetic population; TP: True Positive; TN: True Negative; POPtot: Total diabetic population (POPtot = 22,178); SE: Sensitivity; SPE: Specificity; PPV: Positive Predictive Value; Pos: Population included (Pos = 21,068); NPV: Negative Predictive Value; Neg: Population excluded (Neg = 1,110).
^†: pooled from age classes.

ISSN: 1472-6947