Table 2.

Reasons for true and false classifications identified by manual chart review

Reasons for true negative classification	n=31, 21% (15% to 28%)
Examinations ordered but not completed	15, 48% (32% to 65%)
Colonoscopy ordered but not completed	8
Faecal immunochemical test ordered but not completed	6
CT colonography ordered but not completed	1
Lack of documentation or incorrect documentation	9, 29% (16% to 47%)
Declined screening	4, 13% (5% to 29%)
Insufficient time to discuss	3, 10% (3% to 26%)
Reasons for false-negative misclassification	n=104, 69% (62% to 76%)
Screening outside of UCSF	53, 51% (41% to 60%)
Screening prior to Epic EHR implementation	29, 28% (20% to 37%)
Database and query errors	22, 21% (14% to 30%)
Misclassified as eligible for screening	n=15, 10% (6% to 16%)
Poor life expectancy, or risks outweighing benefits	8, 53% (30% to 75%)
Above risk (personal or family history of polyps)	6, 40% (20% to 64%)
Not primary care empanelled	1, 7% (0% to 32%)
Reasons for false-positive misclassification	n=2, 4% (0% to 14%))
Ordered but incomplete faecal immunochemical test	1, 50% (9% to 91%)
Performed colonoscopy revealed inadequate bowel preparation	1, 50% (9% to 91%)

The second column lists the number of charts and associated percentage of the group with 95% CIs.

EHR, electronic health records; UCSF, University of California, San Francisco.