Skip to main content
. 2021 Mar 13;28(7):1468–1479. doi: 10.1093/jamia/ocab027

Table 1.

Average query execution times in seconds for 100 randomly generated queries over electronic health records for ∼2.8 million patients from the Stanford Medicine CDW, using ACE or BigQuery; and over health insurance claims records for ∼65 million patients from the Optum Clinformatics Datamart, using ACE

Command Average [min-max] query response time (seconds)
Stanford Medicine CDW
Claims
ACE BigQuery ACE
OR of 4 features 0.015 [0.005–0.211] 198.7 [121.0–642.0] 1.224 [0.017–1.813]
UNION of 4 features 0.018 [0.005–0.306] 167.7 [73.0–227.0] 0.205 [0.026–1.618]
AND of 2 features 0.026 [0.005–0.681] 221.5 [124.0–940.0] 0.0684 [0.026–1.530]
INTERSECT of 2 features 0.024 [0.005–0.314] 233.6 [152.0–748.0] 1.018 [0.023–0.545]
SEQUENCE of 2 features 0.017 [0.005–0.214] 202.3 [148.0–266.0] 0.0602 [0.018–0.237]

Abbreviations: ACE, advanced cohort engine; CDW, clinical data warehouse; OR, odds ratio.