eBAC — Estimated Blood Alcohol Concentration

Part of Statistical Methods — MORIE’s statistical-methods reference.

eBAC is a continuous outcome derived from self-reported alcohol consumption data in CPADS. MORIE computes two eBAC variants.

Widmark formula

MORIE uses Widmark’s formula in US units (Matthews & Miller, 1979):

\[\text{eBAC} = \frac{A \times 5.14}{W \times r} - 0.015\, t\]

where

  • \(A\) = fluid ounces of ethanol consumed; a standard drink (14 g of ethanol) is 0.6 fl oz, so \(A = 0.6 \times\) drinks

  • \(W\) = body weight (lb)

  • \(r\) = Widmark distribution factor (0.73 for men, 0.66 for women)

  • \(0.015\) = elimination rate (percent BAC per hour)

  • \(t\) = hours since drinking began

The result is a percentage (g/dL); values below zero are floored at zero. Five drinks for a 150 lb man over two hours give \(5 \times 0.6 \times 5.14 / (150 \times 0.73) - 0.03 = 0.111\).

MORIE variants

ebac_tot

Total eBAC from the full CPADS drinking episode as reported.

ebac_legal

Binary indicator: \(\mathbb{1}[\text{eBAC} \geq 0.08\text{ g/dL}]\) (Canadian legal driving limit).

Both are available as canonical CPADS variables and participate in the CPADS_REQUIRED_VARIABLES contract.

Python API

from morie import calculate_ebac, is_over_legal_limit

# weight in pounds; the Widmark constant is 0.73 for men, 0.66 for women
ebac = calculate_ebac(drinks=5, weight_lbs=154, hours=2.0, gender_constant=0.73)
over = is_over_legal_limit(ebac)

eBAC-IPW module

The ebac-selection-adjustment-ipw module uses eBAC strata as a selection-correction mechanism. See Causal Inference for the statistical framework.

References

  • Matthews DB, Miller WR (1979). Estimating blood alcohol concentration: two computer programs and their applications in therapy and research. Addictive Behaviors 4(1):55-60.

  • Widmark EMP (1932). Die theoretischen Grundlagen und die praktische Verwendbarkeit der gerichtlich-medizinischen Alkoholbestimmung. Urban & Schwarzenberg.

  • Brick J (2006). Standardization of alcohol calculations in research. Alcoholism: Clinical and Experimental Research, 30(8):1276–1287. https://doi.org/10.1111/j.1530-0277.2006.00155.x