Table 1.

Localized meningitis epidemics identified through analysis of surveillance data in health centre level resolution and comparison of the epidemic signals with a surveillance at the district level, by period and by threshold. 14 health centres in Burkina Faso, 2004–2014.

	Years 2004–2009 « NmA »	Years 2010–2014 « other Nm »
	Cut-off 100	Cut-off 50	Cut-off 75	Cut-off 100
N district years with a localized epidemic in at least one health centre	22	34	15	10
N (%) district years with a localized epidemic but without any signal at the district level	4 (18%)	23 (68%)	7 (47%)	3 (30%)
N (%) district years with localized epidemics and a signal at the district level	18 (82%)	11 (32%)	8 (53%)	7 (70%)
N district years where the signal at the district level was preceded by the signal at the HC level	5	8	3	1
Delay between health centre and district level signal*:
Mean	− 0.4 weeks	− 2.1 weeks	− 1.4 weeks	0.6 weeks
Median (range)	0 week (− 5–4)	− 1 week (− 9–1)	0 week (− 9–2)	0 week (− 2–4)
N (%) district years with localized epidemic with gain in signal detection using surveillance at the HC level	9 (41%)	31 (91%)	10 (67%)	4 (40%)
Median (range) number of cases that occurred in the HC during the four weeks following the epidemic signal	17 (0–269)	6 (0–219)	2.5 (0–54)	1.5 (0–54)

Cut-offs are expressed as N weekly cases per 100,000 inhabitants.

*negative values mean that the health centre signal preceded the district signal. Positive values usually mean that HC level data were missing for the weeks when the district epidemic was detected.