Skip to main content
. 2014 Dec 19;4(12):e005811. doi: 10.1136/bmjopen-2014-005811

Table 2.

Diagnostic accuracy of self-reported alcohol consumption compared with reference police breathalyser

Threshold of self-reported alcohol consumption (UK units) Total participants Correctly positive by reference standard Correctly negative by reference standard Sensitivity (%) Specificity (%)
1 207 38/38 9/169 100.0 (90.8 to 100.0) 5.3 (2.8 to 9.8)
2 207 38/38 23/169 100.0 (90.8 to 100.0) 13.6 (9.2 to 19.6)
3 207 38/38 52/169 100.0 (90.8 to 100.0) 30.8 (24.3 to 38.1)
4 207 38/38 71/169 100.0 (90.8 to 100.0) 42.0 (34.8 to 49.5)
5 207 37/38 84/169 97.4 (86.5 to 99.5) 49.7 (42.3 to 57.2)
6 207 35/38 105/169 92.1 (79.2 to 97.3) 62.1 (54.6 to 69.1)
7 207 34/38 117/169 89.5 (75.9 to 95.8) 69.2 (61.9 to 75.7)
8 207 29/38 130/169 76.3 (60.8 to 87.0) 76.9 (70.0 to 82.6)
9 207 26/38 142/169 68.4 (52.5 to 80.9) 84.0 (77.7 to 88.8)
10 207 20/38 148/169 52.6 (37.3 to 67.5) 87.6 (81.8 to 91.7)

1 UK unit is equivalent to 10 mL or 8 g alcohol (total 207 participants due to missing data on alcohol consumption for 1 participant). 95% CIs are shown in parentheses.