TABLE 1.
Description of MGI and UKB cohorts
| MGI (Discovery) | UKB (Validation) | |||
|---|---|---|---|---|
| Analytic cohort size | 38,359 | 393,092 | ||
| Phecodes | 1,859 | 1,686 | ||
| Pancreatic Cancer Cases | Controls | Pancreatic Cancer Cases | Controls | |
| n (%) | 429 (1.12) | 37,930 (98.88) | 659 (0.17) | 392,640 (99.83) |
| Age* | ||||
| Min | 32 | 18 | 52 | 49 |
| Mean | 66.36 | 56.70 | 72.67 | 67.70 |
| Median | 68 | 59 | 74 | 69 |
| Max | 92 | 103 | 82 | 85 |
| Sex | ||||
| Female | 181 (42.19) | 19,960 (52.62) | 307 (46.59) | 212,809 (54.23) |
| Male | 248 (57.81) | 17,970 (47.38) | 352 (53.41) | 179,624 (45.77) |
| Length of follow-up (in years)** | ||||
| Min | 0.05 | 0 | - | |
| Mean | 6.86 | 7.74 | - | |
| Median | 4.44 | 5.42 | 7† | |
| Max | 20.86 | 40.19 | - | |
| Number of unique phecodes | ||||
| Min | 11 | 1 | - | - |
| Mean | 99.59 | 66.74 | - | - |
| Median | 79 | 48 | - | - |
| Max | 495 | 536 | - | - |
*
Age represents age at last electronic health record encounter.
**
Follow-up data on UK Biobank participants comes from 413,591 individuals with time-stamped ICD code data. 86,717 individuals have only a single encounter. These data are not stratified by pancreatic cancer case status. This data was requested for the purposes of this comparison and was unavailable to us for our analysis.
†
Jani BD, Hanlon P, Nicholl BI, et al. Relationship between multimorbidity, demographic factors and mortality: findings from the UK Biobank cohort. BMC Med. 2019;17(1):74. doi:10.1186/s12916-019-1305-x