“…Twenty studies used the first interpretation of the term in their assessment of surveillance systems [ 12 , 13 , 14 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 48 , 49 , 50 , 52 , 53 , 54 ]. They used retrospective analysis of routinely collected data [ 12 , 13 , 14 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 30 , 49 , 54 ] and retrospective analysis of routinely collected data combined with capture-recapture statistical methods [ 29 , 50 , 52 , 53 ], with only one study [ 48 ] opting to collect collected data prospectively. The proportion of cases captured by the different surveillance systems was compared with a variety of alternative sources of data, or estimated using the capture-recapture method, and ranged from 45% of cases in a TB surveillance system in Saudi Arabia [ 20 ] to 99.9% for a salmonellosis surveillance system in Sweden [ 52 ].…”