Skip to main content

View full-text article in PMC

. 2022 Nov 1;10(11):e36711. doi: 10.2196/36711

Table 6.

Performances of the Damerau-Levenshtein distance (DLD)–based matching algorithm per birth country and per university hospital.

University hospital	Sample size	Se^a,b BIF^c, % (95% CI)	Se^b BOOF^d, % (95% CI)	P value Fisher exact test	Sp^eb BIF, % (95% CI)	Sp^b BOOF, % (95% CI)	P value Fisher exact test
Nantes	Se, BIF: 2000 Se, BOOF: 1660 Sp, BIF: 2000 Sp, BOOF: 2000	85.8 (84.3-87.3)	74.9 (72.8-77)	<.001	99.3 (98.9-99.7)	99.8 (99.6-100)	.03
Rennes	Se, BIF: 2000 Se, BOOF: 500 Sp, BIF: 2000 Sp, BOOF: 2000	97.8 (97.1-98.4)	87.6 (84.7-90.5)	<.001	98.2 (97.6-98.7)	99.8 (99.5-100)	<.001
Lille	Se, BIF: 2000 Se, BOOF: 1700 Sp, BIF: 2000 Sp, BOOF: 2000	98.3 (97.7-98.9)	86.8 (85.2-88.4)	<.001	99.4 (99-99.7)	99.4 (90-99.7)	>.99

^aSe: sensitivity.

^bMax total distance used: 2

^cBIF: patient born in France.

^dBOOF: patient born out of France.

^eSp: specificity.