The Causal Attribution of Injury to Alcohol
Consumption: A Cross-National Meta-Analysis From
the Emergency Room Collaborative Alcohol
Analysis Project
Cheryl J. Cherpitel, Yu Ye, Jason Bond, and Guilherme Borges
Background: Whereas a substantial literature exists documenting the association of alcohol and injuries,
causal associations are less well established.
Methods: The relationship of drinking-in-the-event variables with attributing a causal association of
alcohol consumption and the injury event was examined by using meta-analysis across 13 emergency room
studies from 8 countries included in the Emergency Room Collaborative Alcohol Analysis Project.
Results: Pooled odds ratios for both log-transformed blood alcohol concentration at the time of the
emergency room visit and the amount of alcohol consumed in the 6 hr before injury were positively
predictive (1.19 and 1.80, respectively) and heterogeneous across studies. Effect size changed little when
age and gender were controlled. When stratifying on reporting five or more drinks on an occasion during
the last year (5+ yearly drinkers), the amount consumed was positively predictive of reporting a casual
association of drinking and injury only for 5+ yearly drinkers. The effect size of feeling drunk at the time
of injury, controlling for the amount of alcohol consumed, was positively predictive (2.04) but heteroge-
neous across studies. Meta-analysis regression found the level to which alcohol is consumed in a detrimental
pattern to be a significant predictor of blood alcohol concentration, and of the amount consumed and
feeling drunk at the time of injury, on causal attribution, with a lower detrimental pattern level with a larger
effect size.
Conclusions: The association of acute use of alcohol on causal attribution may be affected by chronic use
to some extent, but this association is negatively affected by the degree to which a society exhibits harmful
drinking patterns.
Key Words: Causal Attribution, Alcohol Injury, Emergency Room, Drinking Pattern.
A
GREAT DEAL OF literature now exists documenting
the association of alcohol with injuries on the basis of
studies of emergency room (ER) samples (Cherpitel, 1993;
Romelsjö, 1995). Although larger proportions of injured
patients compared with noninjured patients have been
found to be positive for blood alcohol concentration (BAC)
at the time of arrival in the ER and although larger pro-
portions also report drinking before the event that brought
them to the ER, causal associations are less well estab-
lished. Drinking-in-the-injury event variables have been
analyzed in some of these ER studies, including the number
of drinks consumed before injury, the time lapsed between
drinking and the event, and whether the patient believed
that the event would have happened had he or she not been
drinking (causal attribution; Cherpitel, 1996, 1998; Cherpi-
tel et al., 1993). These studies suggest that drinking occurs
in closer proximity to an injury event than to a noninjury
event (with more than half of the injured patients reporting
less than an hour’s time lapse between the last drink and
the event), and substantial proportions of injured patients
have been found to report feeling drunk at the time of the
event and to attribute a causal association of their drinking
with the event (Cherpitel, 1996). It is important to note that
causal attribution as defined here, or the patient’s attribut-
ing a causal association of drinking with the injury event, is
not the same as the epidemiological definition of causality,
in which a cause/effect association between a risk factor
and an outcome is established through the application of
formal criteria, including the strength, consistency, tempo-
rality, and specificity of the association (Susser, 1973).
The associations of drinking-in-the-event variables and
From the Public Health Institute (CJC, YY, JB), Alcohol Research Group,
Berkeley, California; and Instituto Nacional de Psiquiatria & Universidad
Autonoma Metropolitana-Xochimilco (GB), Mexico City, Mexico.
Received for publication April 28, 2003; accepted August 28, 2003.
This study was supported by National Alcohol Research Center Grant AA
05595-21 from the NIAAA.
Presented at the Annual Meeting of the Research Society on Alcoholism,
June 21–26, 2003, Ft. Lauderdale, FL.
Reprint requests: Cheryl J. Cherpitel, DrPH, Public Health Institute, Al-
cohol Research Group, 2000 Hearst Ave., Berkeley, CA 94709; Fax: 510-642-
7175; E-mail: ccherpitel@arg.org.
Copyright © 2003 by the Research Society on Alcoholism.
DOI: 10.1097/01.ALC.0000095863.78842.F0
0145-6008/03/2711-1805$03.00/0
ALCOHOLISM:CLINICAL AND EXPERIMENTAL RESEARCH
Vol. 27, No. 11
November 2003
Alcohol Clin Exp Res, Vol 27, No 11, 2003: pp 1805–1812 1805