Skip to main content
. Author manuscript; available in PMC: 2009 Jun 8.
Published in final edited form as: Arch Clin Neuropsychol. 2008 Feb 21;23(2):129–137. doi: 10.1016/j.acn.2007.11.005

Table 7.

Diagnostic classification accuracy based on time and/or error scores on Trail Making Test Part B

Sensitivity Specificity PPP NPP Hit Rate
MCI/AD (n=257) vs. Control (n=269)
Errors1 61.9% 70.3% 66.5% 65.9% 66.2%
Time2 45.1% 91.1% 82.9% 63.5% 68.6%
Errors and time3 35.4% 94.4% 85.8% 60.5% 75.5%
Errors and/or time4 71.6% 66.9% 67.4% 71.1% 69.2%
MCI (n=200) vs. Control (n=269)
Errors1 59.0% 70.3% 59.6% 69.7% 65.5%
Time2 43.0% 91.1% 78.2% 68.2% 70.6%
Errors and time3 33.0% 94.4% 81.5% 65.5% 76.2%
Errors and/or time4 69.0% 66.9% 60.8% 74.4% 67.8%
AD (n=57) vs. MCI (n=200)
Errors1 71.9% 41.0% 25.8% 83.7% 47.9%
Time2 52.6% 57.0% 25.9% 80.9% 56.0%
Errors and time3 43.9% 67.0% 27.5% 80.7% 53.0%
Errors and/or time4 80.7% 31.0% 25.0% 84.9% 42.0%

Note: MCI = mild cognitive impairment; AD = Alzheimer’s disease; PPP = positive predictive power; NPP = negative predictive power.

Identification of impairment within each condition was based on the following contrasts:

1

≥1 error vs. 0 errors

2

z≤−1.0 vs. z >−1.0

3

≥1 error and z ≤ −1.0 vs. 0 errors and z>−1.0

4

≥1 error and/or z ≤ −1.0 vs. 0 errors and z>−1.0