Table 2 Note characteristics according to patient level-corpus character thresholds

From: Machine learning to predict penumbra core mismatch in acute ischemic stroke using clinical note data

Character threshold

Number of unique notes included

Time delay between note and CTP scan time (minutes)

Word count

Character count

500

2 [1,2]

29.8 [15.6–92.9]

138 [67–481]

841 [488]

1000

2 [1–3]

31.3 [16.1–112.9]

212 [67–591]

1584 [173]

5000

3 [2–4]

42.4 [21.9–165.5]

464 [67–1098]

3187 [935]

  1. Note: All figures are reported as median [IQR] and correspond to all patient notes that were added to patient-level corpora until minimum character threshold was reached.
Back to article page