Concurrent validity of the CAGE questionnaire. The Nord-Trøndelag Health Study

Abstract

The aim of this study was to examine the psychometric properties of the CAGE questionnaire, and the questionnaire's concurrent validity with current and previous alcohol consumption. This study employed data from the Nord-Trøndelag Health Survey wave 1 (HUNT-1 in 1984–86: N = 24,900) and wave 2 (HUNT-2 in 1995–97: N = 36,350). The concurrent validity of the CAGE questionnaire was examined both as a dichotomous variable with the recommended cut-off (≥2 affirmative answers) for alcohol problems, and as a categorical scale. The categorical scale was constructed by counting responses from 0 to 4, and a separate category for current abstainers in HUNT-2. Current self-reported consumption above the gender specific 80th percentile was defined as “current excessive consumption”. “Previous excessive consumers” were defined by meeting at least one of the two following criteria at the time of HUNT-1: reporting drinking too much alcohol in any period of their life, or reporting a high level of alcohol consumption. The internal reliability of CAGE was adequate, and in relation to alcohol consumption, there was a linear relationship between the CAGE score and both the current and previous excessive consumption. In conclusion, this study indicates good concurrent validity and adequate psychometric properties of the CAGE questionnaire. The dose–response pattern seen between the CAGE score and alcohol consumption, suggests that it can be used as an ordinal measure, rather than with a cut-off of two or more. The concurrent validity of the CAGE is better in women than in men.

Introduction

Alcohol use disorders (AUDs) are associated with increased risk for a host of detrimental outcomes (Cargiulo, 2007, Rehm et al., 2004). Like other conditions, early identification and treatment improve prognosis, and detection of AUD is therefore relevant in primary care settings (O'Connor, 2005). It is therefore unfortunate that this issue seems insufficiently addressed in clinical practice (CASA, 2000, Cheeta et al., 2008, Friedman et al., 2000), with time constraints as one of the important reasons (Barry et al., 2004, Harris et al., 2007, Murphy, 2009). The CAGE questionnaire was first presented in 1970 as a screening instrument for efficient and precise identification of potential alcohol-related problems (Ewing, 1984, O'Brien, 2008). Since then, CAGE has become one of the most frequently used screening tools for alcohol problems, in clinical practice, clinical studies and in general population studies (Alvarez and Rio, 1994, Amaral and Malbergier, 2004, O'Brien, 2008, Steinbauer et al., 1998). Its advantages include brevity, ease of use, and scoring without the need for pen-and-pencil (O'Brien, 2008).

The CAGE contains four questions and each contribute towards the acronym CAGE to make the questions easy to remember: 1) Have you ever tried to cut down on your drinking, 2) do you get annoyed when people talk about your drinking, 3) do you feel guilty about your drinking, and 4) have you ever had an eye-opener (a drink first thing in the morning) to steady your nerves or get rid of a hangover? The response categories are binary (affirmative or negative), and the questions are framed in a life-time perspective (Ewing, 1984). Most studies suggest that two or more affirmative answers should increase suspicion of alcohol-related problems (Aeertgeerts, Buntinx, & Kester, 2004), and prompt further investigation. There has been numerous studies looking at the psychometric properties of CAGE, both in clinical settings (Aeertgeerts et al., 2004, Allen et al., 1995, Berks and McCormick, 2008, Bradley et al., 1998, Bradley et al., 2001, Dhalla and Kopec, 2007, Malet et al., 2005, McIntosh et al., 1994, O'Brien, 2008), and in non-clinical population-based studies (Alvarez and Rio, 1994, Amaral and Malbergier, 2004, Bühler et al., 2004, Crowe et al., 1997, Koppes et al., 2004, Messiah et al., 2008, Nyström et al., 1993, Saunders et al., 2006, Smart et al., 1991).

Although most studies have found CAGE to be both practical and useful (Allen et al., 1995, Alvarez and Rio, 1994, Amaral and Malbergier, 2004, Berks and McCormick, 2008, Crowe et al., 1997, Dhalla and Kopec, 2007, Malet et al., 2005, McIntosh et al., 1994, O'Brien, 2008, Saunders et al., 2006), some dissents still exist (Aeertgeerts et al., 2004, Bisson et al., 1999, Bradley et al., 1998, Bühler et al., 2004, Koppes et al., 2004, Messiah et al., 2008, Nyström et al., 1993, O'Hare and Tran, 1997). In relation to general population-based studies, the main controversy regards indications of a questionable validity (Bisson et al., 1999, Koppes et al., 2004, Messiah et al., 2008), cross-cultural differences (Bloomfield et al., 2002, Messiah et al., 2008) and gender differences (O'Hare & Tran, 1997). For instance, Bisson et al. (1999) concluded that the use of the CAGE as a screening tool for heavy drinking and drinking problems in a general population study should not be advised due to poor discrimination between heavy drinkers and non-heavy drinkers. Additionally, another general population-based study found that the CAGE performed worst of three screening questionnaires for alcohol problems (Koppes et al., 2004). The uncertainties regarding CAGE's applicability in a general population-based setting prompted the current study.

In the present study we aimed to examine the psychometric properties of CAGE in a general population sample, by investigating the internal reliability of the questionnaire, as well as its concurrent validity against current and previous excessive consumption.