Table 3

Multiple matching OR estimates of substance use before injury, by type of substance and measurement and type of road traffic injury

Pedestrian: multiple matchingDriver: multiple matchingPassenger: multiple matching
obsOR95% CI obsOR95% CI obsOR95% CI
Alcohol
Self-report1232.67(0.62 to 11.50)6575.84*(2.77 to 12.33)1877.26*(2.04 to 25.86)
Alcosensor (BAC≥0.001)914.48(0.47 to 42.70)5193.79*(1.79 to 8.04)1413.34*(1.11 to 10.06)
Self-report or alcosensor1233.31(0.81 to 13.54)6576.49*(3.10 to 13.60)1877.92*(2.25 to 27.92)
Alcosensor or SR if missing1231.76(0.45 to 6.86)6573.72*(1.87 to 7.40)1874.13*(1.43 to 11.98)
Cannabis
Self-report1231.82(0.28 to 11.95)6520.96(0.36 to 2.55)1871.20(0.16 to 9.20)
Oral fluid4921.04(0.41 to 2.60)1380.59(0.05 to 7.43)
Self-report or oral fluid1231.82(0.28 to 11.95)6522.26(1.00 to 5.10)1875.16(0.53 to 50.41)
Oral fluid or SR if missing6520.94(0.42 to 2.13)1872.06(0.33 to 13.05)
  • The multiple matching estimate for pedestrian’s cannabis detection with oral fluid was not computed due to cells with zero counts.

  • ORs were estimated using conditional logistic regression.

  • *P<0.05.

  • obs are the number of case and control periods used in each model (ie, each person contributed with a maximum of three observed case and control periods for the multiple matching models).

  • BAC, breath alcohol concentration.