Table 3:
BiobankConnect reduces the amount of data elements that need to be checked
| Biobank (number of elements) |
R1 (via BiobankConnect) | R2 (string matching) | R3 (random search) |
|---|---|---|---|
| KORA (75) | 1.5 | 1.8 | 36 |
| MICROS (119) | 2.0 | 1.3 | 59 |
| FINRISK (223) | 1.5 | 1.9 | 111 |
| Hunt (353) | 2.5 | 4.1 | 174 |
| NCDS (516) | 1.2 | 1.8 | 260 |
| Prevend (6174) | 2.2 | 4.3 | 3109 |
| Average | 1.8 | 2.7 | 3730 |
| Missed elements | 3 | 20 | 0 |
R1,2,3 shows the average rank of the ‘best’ match when searching using BiobankConnect, using Lucene string matching only, and random iteration, respectively.