Emergency departments (EDs) are special places for primary care, where frequent heavy and problem drinking are prominent among admissions (1–5). Identification of patients with alcohol use disorders (AUDs) is important in EDs, since these patients may be motivated to change drinking behaviors and to accept treatment or referral for problem drinking, particularly if their admission was related to their alcohol use (6). Indeed, in EDs, patients with alcohol-related disorders may be more prone to initiating a course of treatment following intervention by a health professional who can demonstrate their alcohol abuse and suggest an appropriate therapeutic strategy (7, 8). The optimal aim of alcohol interventions in ED is to assist the patient in reducing consumption (in the case of risky drinking and abuse) or to enter into specialized alcohol treatment for those who are alcohol dependent (9–11). Different types of Interventions [brief alcohol interventions (BAIs), brief negotiation interviewing (BNI), or motivational intervention (MI)], which require different type of people (ED workers, ED workers trained to alcohol interventions, or trained alcohol health workers) can be offered.

Brief alcohol interventions have been found to be effective in a number of clinical settings in primary care practice (12–14). They are comprised of short counseling sessions, which can be provided by emergency staff (nurses, physicians) and have a very good cost-time-effectiveness ratio for those with unhealthy alcohol use (15). However, there is little evidence for efficacy of BAIs among patients with very heavy alcohol use or dependence, particularly those identified by screening tests who are not actively seeking help or advice and who therefore are less likely to be amenable to change (16). Particularly, a majority of patients admitted to the ED for acute alcohol intoxication (AAI) fit this profile (17). In order to assist these patients in changing drinking behavior, BAIs can incorporate BNI in time limited health care settings (9, 18). BNI can be provided by ED workers trained to alcohol interventions or ideally trained alcohol health workers who are aware of the psychopathologic underpinnings of dependence and are able to address the patient’s ambivalence and perceived difficulty in behavioral change related to his alcohol problem (19). Nevertheless, BNI interventions may still result in failure (16). Thus, severely dependent patients who often exhibit significant denial of their alcohol problems (20) may require a more lengthy MI (60 min compared to the usual 15 or 20 min), which may be more effective in addressing excessive resistance (21). MI requires a high level of understanding of the psychological mechanisms of dependence and must be understood as adopting an empathic and non-confrontational style. This psychological intervention (MI) is not classically a part of the ED culture where important barriers to translating alcohol interventions to clinical practice exist, and requires participation of trained alcohol health workers (22).

To maximize their interest, these different types of Interventions, which require different type of people (ED workers, ED workers trained to alcohol interventions, or trained alcohol health workers) should be adjusted with severity of Alcohol Use Disorder (AUD). It is therefore important to identify as precisely as possible the severity of the drinking problem (at-risk, abuse, dependence, severe dependence) by using screening tools capable of distinguishing these conditions, so that each patient may be offered the most suitable intervention. Thus, ED admission for drunkenness is an opportune time for initiating interventions for AUDs but the specific screening test used to identify severity of the disorder and the resulting intervention must be determined (23, 24). Hungerford and Pollock have recommended that screening instruments have high sensitivity and specificity, and are not time consuming, expensive, or difficult to use (25). The use of the CAGE (26–28), RAPS4-QF (29), and AUDIT (30–32) tests have all been recognized as effective for detecting alcohol-related disorders in the ED setting (1, 2, 4, 33, 34). Comparisons, taking into account gender, of the performance of these screening instruments among intoxicated patients in a French emergency service site have been reported, which included the optimal cut points for detecting different levels of severity of AUDs for each of these scales (35). As reported in the earlier paper, for detecting alcohol abuse the optimal cut point for the RAPS4-QF was ≥2 for men and ≥4 for women. In the case of CAGE, the optimal cut point was ≥3 for men and ≥2 for women. The optimal cut point for the AUDIT was ≥12 for men and ≥7 for women. For detecting alcohol dependence, the optimal cut point for the RAPS4-QF was ≥3 for men and ≥4 for women. For the CAGE, a cut point ≥3 was found for both men and women. For AUDIT, an optimal cut point ≥18 was found for the total sample, ≥14 for men and ≥11 for women (35). However, to our knowledge, there are no studies, which propose to graduate levels of severity of dependence (moderate or severe) by these screening scales in ED for patient admitted for IAA.

The aim of this paper was to help clinicians to identify the severity of AUD from optimal thresholds found for recommended scales. Especially, taking account of the high prevalence of alcohol dependence among patients admitted to the ED for AAI, we have tried to define thresholds of severity of dependence based on the AUDIT score.

The CAGE, RAPS4-QF, and AUDIT screening tests are those most frequently used in the ED to identify problems associated with alcohol use (29, 39, 43–45). However, these screening tests are typically used for general populations admitted to ED and there are no recommendations for interpreting results related to sensitivity and specificity by gender and severity of alcohol disorders for these instruments (46, 47). The results of these instruments could provide guidance for optimizing interventions in the ED because patient’s level of engagement with treatment will depend on the strategy adopted at the initial consultation and this strategy could depend on the severity of the patient’s problem with alcohol (16, 48–50). Tests used to screen for and distinguish different levels of alcohol-related disorders should have good sensitivity at the suggested cut points for each level of severity.

Earlier, we reported results showing that RAPS-QF test (incremental score) seem to possess better psychometric properties than the CAGE in men while the CAGE seems more adapted to female populations (35). This finding was also reported by Cherpitel and Bazargan (51). Furthermore, different thresholds are evident depending on the severity of alcohol-related disorders (abuse or dependence). The AUDIT demonstrates good performance for detecting alcohol abuse and alcohol dependence in male patients, but at a higher cut point [12] than the cut point traditionally used [8] (52). In women, the recommended cut point of seven (46) is confirmed here for identifying AUDs. For the purpose of distinguishing alcohol dependence from other conditions, the AUDIT displays good performance at the cut point of 14 in men, 11 in women, and 18 for the total sample. In the study population here of those admitted to the ED for drunkenness, compared with the general population of those admitted to the ED, the higher cut point for the AUDIT is in agreement with that reported by Conigrave et al. (52) who advised a cut point of 15 (Se: 0.73, Spe: 0.84) for patients whose admission to emergency care was associated with acute alcoholism. Further, analysis of optimal thresholds based on the indices of quality for the AUDIT suggest the possibility of defining different cut points depending on the seriousness of the alcohol-related disorder beyond the classical distinction between abuse and dependence.

Thus our results demonstrate three severity intervals, depending on the two optimal points found for the RAPS4-QF for men and the two optimal cut points found for the CAGE for women. Similarly, our study demonstrate severity intervals depending on the cut point on the AUDIT (scores 0–11, 12–13, 14–40 for men and 0–6, 7–10, 11–40 for women). These cut points are compatible with risk intervals recently suggested by Rubinsky et al. (47) (0–4, 5–10, 11–14, 15–40 for men and 0–1, 2, 3–4, 5–8, 9–12, 13–40 for women), which could encompass, respectively, zones of occasional use (I), risky use and abuse (II), and dependence (III, scores 16–19; and IV scores 20–40) according to the World Health Organization (WHO) guidelines for the AUDIT (53). In this paper, risk intervals III (16–19) and IV (20–40) suggest a possible approach of the severity of dependence with the AUDIT scores.

The different thresholds proposed for CAGE, RAPS4-QF, and AUDIT could be used to guide the choices of intervention for the patient. Two “diagnoses tracking and interventions” can be proposed: a short diagnosis track using CAGE and RAPS4-QF, or a long diagnosis track using AUDIT. Thus, if one considers AUDIT too time consuming for persons admitted to the ED, the severity intervals could be designated in men and in women using, respectively, the RAPS-QF and the CAGE scales, because of their good performance in these populations. The score could help practitioners choose appropriate interventions, depending on the severity of alcohol-related disorders. Bazargan-Hejazi et al. (54) in a study testing the effect of brief interventions in the ED have underlined the efficacy of brief intervention for patients screened positive for at-risk drinking as defined by AUDIT scores of 7–18. Brief interventions were not effective for patients with scores in the 19–40 range, which could refer in this study to dependence. These results are compatible with a recent review published by Saitz (16) who highlighted the absence of evidence for efficacy of brief interventions in primary care in people with dependence or very heavy drinking. For these patients more lengthy MI, should be proposed (50). It requires being able to distinguish mild and moderate dependence (which requires BNI) to severe dependence (which requires MI) taking account gender. Consequently, it is necessary to screen patients in order to provide the most efficient intervention. Thus for patients admitted to ED for AAI we proposed, tacking account gender, the use of the AUDIT or CAGE, and the RAPS4-QF for screening with cut points orienting the choice of interventions. Interventions would be provided by ED workers (nurses or physicians) sensitized to brief interventions when screening tests guide toward alcohol abuse (12 ≤ AUDIT < 14 for men or 7 ≤ AUDIT < 11 for women or 2 ≤ RAPS4-QF < 3 for men or 2 ≤ CAGE < 3 for women) and by trained staff when screening tests guide toward very heavy drinking or dependence (14 ≤ AUDIT for men or 11 ≤ AUDIT for women, or 3 ≤ RAPS4-QF for men or 3 ≤ CAGE for women).

Moreover literature and experience suggest it would be important to propose more adapted interventions appropriate to the severity of dependence, since the more severe the disorder the stronger the denial and resistance to treatment. Additional analysis of thresholds for the AUDIT from the sum of items of the MINI allowed us to distinguish three supplementary severity intervals: a zone of mild dependence (AUDIT score 14–17 for men, 11–17 for women), a zone for moderate dependence (AUDIT score 18–25), and a zone of severe dependence (AUDIT score ≥26). For these thresholds, we propose adapted interventions that can use the brief negotiation approach for mild and moderate dependence and longer more intensive MI (more time consuming, requiring trained alcohol health workers) for severe dependence. These types of interventions require a thorough understanding of problematic alcohol use and solid training in this approach for nurses or practitioners who provide it (55).

Several limitations apply, however to our results. First, only the AUDIT provides threshold for levels of dependence severity. Short questionnaires like the CAGE and RAPS4-QF are not sufficient for this distinction in order to adapt accordingly the type of intervention. Indeed the AUDIT, for patients admitted for AAI, seems the more efficient screening test. However, the CAGE and RAPS4-QF can be used in ED setting where time required for AUDIT cannot be given. Secondly, we should have correlated our results with the recent definition of alcohol use disorder [DSM5, Ref. (56)]. Nevertheless, patients were not initially evaluated with this classification, which was not the gold standard diagnosis when we performed our work. However, we can assume that the levels of current severity proposed by the DSM5 for Moderate (presence of four to five symptoms) and Severe (presence of six or more symptoms) could correspond to a higher AUDIT score to 18. Further studies may confirm this. In the same way, the use of the MINI according to a uni-dimensional perspective can be criticized, however, this approach is the one retained in the new version of the DSM5. Finally, it is difficult to propose intervention according to a criterion of severity of AUDs defined by this only study. In fact it should have done a follow-up study to allow recommendations for different interventions and a control population. Today, review of the literature on brief interventions is challenged and our ambition was to offer more intervention strategies that determine the best intervention. Naturally, we followed the clinical intuition that leads us to think that interventions must be more intense when clinicians faced with a severe disorder, but this is not demonstrated with this study. It is also not known whether the level of denial and resistance to treatment depends only on severity of dependence or whether it may be associated with other factors such as readiness to change, independent of the level of severity of AUD (57). Undoubtedly, studies should be conducted in order to validate the relevance of the guidelines proposed here.

Patients with alcohol-related disorders are frequent in the ED. Among them those admitted following acute intoxication represent a specific population. They are frequently alcohol dependent. During their presence in the ED they can benefit from intervention for an alcohol-related disorder. ED practitioners have to decide quickly who can benefit most from an intervention and the nature of the intervention. The choice is guided by knowledge of the severity of the disorder and by reserving lengthy and specialized interventions applied by highly trained caregivers when the severity of AUD suggests a graduated motivational approach. The systematic use of the CAGE, RAPS4-QF, or AUDIT in order to screen for the severity of alcohol-related disorders should be recommended in this setting.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.