Screening and Following Up Harmful Alcohol Use “… is Not Necessarily Your Primary Focus”: A Qualitative Study Exploring Health Professionals’ Experiences Addressing Harmful Alcohol Use in a Norwegian Hospital

Introduction

Alcohol use remains a leading cause of excess mortality and morbidity worldwide. ^1,2 A recent study on the alcohol attributable burden of disease implicated alcohol in an estimated 3 million deaths and 131.4 million disability adjusted life years (DALYs) globally in 2016.² Global per-capita consumption has been forecast to increase from 6.5 liters in 2017 to 7.6 liters by 2030, with rates of current and heavy episodic drinking following a parallel trajectory and the prevalence of lifetime abstinence declining.³ Progress towards the 10% relative reduction targets outlined in the World Health Organization (WHO) Global strategy to reduce the harmful use of alcohol 2010 has therefore been limited.^1,4 Only the European region has exceeded this target, albeit from a very high baseline.⁴ Despite this, per capita consumption, current drinking, heavy episodic drinking and alcohol use disorders remain higher/more prevalent in Europe than in any other WHO region.⁴

The WHO’s (2023) draft action plan to implement a global alcohol strategy defines harmful alcohol use as:

drinking that causes detrimental health and social consequences for the drinker, the people around the drinker and society at large, as well as patterns of drinking that are associated with increased risk of adverse health outcomes.¹

This clearly encompasses a broader population than the International Classification of Diseases (ICD) ‘harmful pattern of use’ included in the diagnostic criteria for alcohol use disorder, and the action plan is correspondingly focused on a range of interventions to reduce the acceptability, availability, and affordability of alcohol, including through screening, brief intervention and referral to treatment (SBIRT).¹ The action plan builds on The Safer Technical package, which highlights the efficacy of SBIRT both in terms of harm reduction and the primary prevention of alcohol-related harm.⁵

Studies in Sweden, Finland and Denmark have demonstrated the significant impact of alcohol and other substance use on excess mortality among patients with psychiatric disorders.^6–8 Psychiatric comorbidities in alcohol use disorder are well established,⁹ and a Scandinavian cohort study implicates substance use disorders in a threefold increase in mortality from diseases and medical conditions and a ten to fifteenfold increase in death from other external causes, primarily accidents.⁷ Despite this, screening rates in hospital emergency departments remain very low.¹⁰

The Norwegian Alcohol strategy¹¹ is correspondingly focused on developing robust alcohol treatment pathways, including screening all patients admitted to hospital emergency departments in Norway. A recent study found that 1 in 5 patients admitted to a hospital emergency department in Oslo reported harmful alcohol use,¹² and this proportion is likely to be higher among injury patients.^13,14 A study from the United States found that 50% of patients with high healthcare utilization costs met the criteria for a substance use disorder, most commonly alcohol, with only 10% receiving treatment.¹⁵ Emergency departments therefore present clear opportunities both for secondary (identification and lifestyle interventions) and tertiary prevention.^12,14

A recent review reports substantial evidence for the effectiveness of alcohol interventions in urgent and emergency care,¹⁶ with SBIRT associated with reductions in alcohol use, in the negative consequences of use and reduced hospital emergency department visits.¹⁷ However, heterogeneity in study designs, populations, interventions, assessment tools, implementation, content and delivery problematize generalization from results from clinical trials.¹⁶ Control conditions (information and advice) may be sufficient to impact on consumption in some studies,^16,18 and a meta-analysis of the effectiveness of screening and brief interventions (SBI) in emergency care found that more intensive intervention showed no clear benefits over shorter interventions.¹⁹

Recent studies have highlighted a need for an increased focus on alcohol at all levels of the Norwegian healthcare system.^20–22 However, while a recent review highlights the efficacy of SBIRT in emergency care,¹⁶ few qualitative studies have explored health personnel’s experiences identifying and following up harmful alcohol use in these settings.

Aim

To explore health professionals’ views and experiences of systematic screening and tailored follow-up of harmful and hazardous alcohol use in a Norwegian hospital.

Methods

Design

The study was carried out as part of a multi-site study evaluating the impact and implementation of systematic screening and tailored follow-up of harmful and hazardous alcohol use in emergency care in Oslo, Norway. The new clinical procedures included screening all patients admitted to the emergency department using the Alcohol Use Disorders Identification Test – Consumption (AUDIT-C), a validated short form of the full AUDIT test incorporating four questions pertaining to alcohol consumption.²³ Patients with AUDIT-C scores between 3–5 and 4–6 for women and men, respectively, were given structured feedback. Patients with AUDIT-C scores >5 or 6 were assessed using the Clinical Institute Withdrawal Assessment Scale revised (CIWA-Ar)²⁴ following additional risk assessment, and received appropriate preventive treatment and referral to specialist alcohol treatment services.

Participants

Qualitative semi-structured interviews were carried out with nine specialty registrars and four nurses with experience administering AUDIT-C and CIWA-Ar between May and December 2022. Participants were recruited by the consultant responsible for delivering training on the new clinical procedures and a research nurse on the observation ward of the first hospital to implement systematic screening and tailored follow-up of harmful and hazardous alcohol use. Participants represent a convenience sample of health professionals with experience applying the new clinical procedures and available for interview on the scheduled interview dates. They were aged between 25 and 35 years, comprised four men and nine women, and most had been employed between 2 months and 3 years in their current position except two nurses who had been employed for >5 years.

Data Collection

A topic guide based on three thematic questions pertaining to the phenomena of interest was applied (see Box 1). Follow-up questions were formulated depending on participants’ responses. Additional questions were identified based on these responses, recorded in field notes and incorporated in the topic guide in subsequent interviews to enrich and expand on the interview data.²⁵ All thirteen interviews were conducted by the first author TT and were carried out face to face in a visitors’ room in the emergency department and a staff room on the observation ward at the hospital.

Box 1 Primary Topics Addressed in Interviews

Interviews were transcribed verbatim by TT. Quality assurance was achieved by listening to the interviews while reading through the transcripts. A total of 450 minutes of interviews were recorded and transcribed, with interviews averaging 35 minutes and varying between 15 and 60 minutes in duration.

Analysis

Data were analyzed using a deductive approach extrapolating from the topics in the interview guide.²⁶ Braun and Clarke’s²⁶ semantic thematic analysis was used to identify and code coincident and divergent patterns in the data as these pertained to the phenomena of interest. Interviews were coded using NVivo v.14.²⁷ Descriptive themes were identified through open coding, with analytical themes developed through an iterative process of coding, categorization and merging identified themes into overarching analytical themes pertaining to participants’ experiences screening and following up harmful alcohol use. The original transcripts were revisited to check for consistency. The first author TT performed the initial coding to generate an overview of semantic and latent content in the data, using NVivo v.14²⁷ to create and merge nodes to “segment” the data before sorting identified codes into potential themes. The process of reviewing and refining themes was performed by the first author TT in collaboration with the last author AJ. This involved assessing themes for coherence and ensuring candidate themes captured the contours of the data.²⁶ Semantic content was given precedence over cross-cutting latent themes in this collaborative review process to maintain a primary focus on participants’ experiences with the new clinical procedures in their own words. The themes are presented in the results section. Quotes foregrounding health personnel’s experiences screening and following up patients’ alcohol use, and their views and experiences of the implementation process, are used to illustrate the themes.

Ethics

The study followed the ethical principles outlined in the revised Declaration of Helsinki²⁸ and was approved by The Norwegian Data Protection Authority (number: 816616). The Regional Committee for Ethics in Medical Research, South-Eastern Norway (number: 279073) determined that the study did not fall within their remit. Participants received oral and written information on the study including assurances regarding the anonymization of all personal information in any associated publications and their right to withdraw consent at any time. Written consent was obtained before the interviews were undertaken. Participants’ informed consent included publication of anonymized responses/direct quotes.

Results

Results are presented in terms of the following descriptive themes: (i) Standardized and clinical assessment; (ii) Formal and informal treatment guidelines, and; (iii) Training delivery and barriers to implementation.

Standardized and Clinical Assessment

Participants reported positive experiences using AUDIT-C to screen patients, with only a few highlighting the additional associated workload or specific difficulties inherent in screening critically ill patients. Patients’ alcohol use had previously been recorded in writing in compiling patient histories as part of the initial patient assessments. Discussion pertaining to participants’ overall assessment of the new procedures tended to focus on the perceived utility of the assessment tool as a mnemonic aid, and notions of “standardization”, more “systematic” assessment and more objective measurement and categorization as opposed to the subjective clinical “characterization” of patients’ alcohol use. One participant suggests:

We’re generating more reproducible results now. If you look in the journals there’s all sorts of information that isn’t necessarily very useful in a clinical or research context… drinks a glass of wine most weekends or drinks a half bottle of wine every other week… some kind of approximation. It’s far easier to identify and follow up patients with harmful or hazardous use using AUDIT-C (Participant 2).

However, while routinely using AUDIT-C to screen patients was represented as constituting a reasonable and minor adjustment to established clinical practice, not all patients were screened. Aside from critically ill patients who may be “purple”, “unconscious” or otherwise unresponsive, screening varied by participants’ clinical assessment of the relevance of alcohol consumption to the primary reason for admission. The participants mentioned forgetting or not prioritizing screening patients who were not visibly inebriated or admitted with alcohol-related injuries, or curtailing assessments where patients reported drinking only infrequently:

It’s great when you already suspect alcohol is directly relevant to the presenting reason for admission. But when you have patients where you feel it’s not relevant to the overall picture… if you’ve got a swollen foot, you’ve fallen off your bike… it can feel a little artificial…it’s not that it’s a waste of time, but it’s not necessarily your primary focus. If I forgot I wouldn’t necessarily go back and ask (Participant 7).

Participants described three broad patient groups: “our patients”, ie repeat admissions with known alcohol use disorders, “high risk” patients who may be intoxicated or admitted with alcohol-related illness or injury, and “low risk” patients, where alcohol was not directly implicated in the primary reason for admission. While screening was not represented as problematic, assessments were not routinely completed with low-risk patients, and existing procedures for managing patients with acute alcohol intoxication were well established. Screening was therefore considered superfluous also with this patient group, with “our patients” routinely assessed for alcohol withdrawal and receiving preventive treatment directly on admission.

The perceived utility of screening was therefore primarily associated with the notion of patients “flying under the radar”. Participants represented AUDIT-C as a useful tool for structuring alcohol-related assessment in these contexts, and pre-empting patients “suffering acute alcohol withdrawal at three in the morning” in the words of one. However, despite clear concerns surrounding the number of patients “flying under the radar” and the perceived utility of screening in identifying these patients, assessments were also frequently curtailed where participants sensed patients becoming resistant to consumption questions. Patients’ alcohol use was considered difficult to “quantify” or convert readily into units of alcohol, and the specificity and repetitive nature of follow up questions this necessitated was considered potentially invasive, in particular on open wards where other patients might overhear:

Generally, when you ask patients they’ll say they don’t really drink very much, but if you keep asking they might say “A couple of glasses at the weekend”, which can mean anything really. Is every day the weekend? How big are the glasses? So, you try to get a more reliable estimate. Any spirits? Do they drink beer? Try to boil it down to units, ideally. Is this a typical week? Any periods without drinking? You have to unpack it, break it down (Participant 2).

Participants posited varying iterations of the assessment that patients’ alcohol use required “unpacking”, that “patients don’t think in units”, and that generating robust scores required time and repetition. While clinicians’ perceptions regarding the veracity or accuracy of patients’ self-reported alcohol consumption varied, all those posing follow-up questions in attempts to establish a more reliable estimate of a patients’ alcohol use were alert to patients feeling “pressured”, “uncomfortable” or “judged”: As such, not all patients were screened, and not all remaining patients were screened sufficiently robustly to generate a score:

It’s fairly full-on isn’t it. It can be confronting. And other patients can generally overhear. If the answers are vague, you have to repeat the question, and it can feel like you’re pressuring patients or trying to catch them out somehow. If I have to ask more than twice I’ll generally just avoid recording a score. If it’s not in the journals it’s far more likely because the patients haven’t answered than because we haven’t asked (Participant 1).

Participants’ assessment of the level of risk implied by patients’ AUDIT-C scores also varied depending on a range of individual-level characteristics and the perceived acceptability of drinking in various social contexts. However, most shared a perception that the risk thresholds implied by AUDIT-C were very conservative. Several participants reflected on their own or their colleagues’ drinking, equating “risky” drinking with alcohol-related illness or injury and the breakdown of social and familial relationships. Patterns of alcohol consumption consistent with harmful or hazardous use were therefore not problematized by participants: problematic use was associated with the long-term physiological and psychological health impact of harmful or hazardous consumption over time. One elaborated on the association between alcohol risk assessment and alcohol-related norms as follows:

Most people drink, a glass of wine or two maybe, which’ll generate a certain score without this necessarily being a concern. Some obviously drink more, and… I mean, I drink myself, so when I’m speaking with the patient I’ll be thinking: “This isn’t excessive”, but then it’ll generate a surprisingly high score. So, you definitely get a different perspective on your own drinking by ‘scoring’ it (Participant 8).

Formal and Informal Treatment Guidelines

Nurses following up patients on the observation ward reported positive experiences using CIWA-Ar with patients at risk of acute alcohol withdrawal and had been using the scale for a number of years following clear internal guidelines on the assessment and management of delirium tremens. However, none had observed any changes in clinical practice or in the frequency of CIWA-Ar assessments following the introduction of routine screening in the Emergency Department. The AUDIT-C scores were not experienced to be routinely recorded in the patient journals, and the nurses were either unfamiliar with AUDIT-C or somewhat sceptical in terms of its utility in identifying patients who may be “flying under the radar”:

I’ve not really seen AUDIT scores recorded in the journals, and if I do it’s like: “OK. Sure”. We’ve had seminars about it… and it’s maybe not the most accurate assessment tool?… there’s plenty of healthy, normal, functional people who score 4 or 5, and…patients with similar scores who clearly have a problem, so it’s difficult to know what the score is telling you sometimes (Participant 11).

Consistent with participants’ accounts of screening patients using AUDIT-C, robust CIWA-Ar assessment was represented as contingent on repeating and reframing questions and “unpacking” patients’ responses to generate a score. This required the allocation of sufficient time with individual patients. Where sufficient time was unavailable – in particular with non-Norwegian speakers who were represented as comprising a significant proportion of this patient group – interaction with patients was limited. Even under optimal conditions, scoring was experienced to be variable and subjective, depending both on the vagaries of subjective judgement and the veracity or otherwise of patients’ assessment of the severity of their symptoms:

It’s hard to quantify sweating. Some things are obviously easier, tremors for example, but there’s always a few numbers between the various descriptors, so we try to go: “I’ll do CIWA this shift” to minimise discrepancies. We do talk about how we score it, because we obviously don’t ask patients: “How sweaty do you feel?” But it’s very subjective. Patients can minimise their symptoms, or play them up to get more medication. You ultimately have to take them at their word (Participant 12).

The account is fairly typical, with participants foregrounding the subjective nature of CIWA-Ar observations and describing the follow-up of patients’ alcohol use as variable, individualized and characterized by what one participant referred to as: “gefühl”. Participants highlighted a need for clearer alcohol guidelines and more systematic patient follow-up. They also suggested that while repeat admissions with established alcohol use disorders or patients admitted with acute alcohol intoxication were followed up routinely with CIWA-Ar, benzodiazepines, supportive care and referral to relevant outpatient treatment services, there was clear scope for working more proactively. As one participant states:

The classic is being asked to just keep an eye on patients, not using CIWA, just seeing if they develop symptoms. There’s no clear guidelines in these cases. It really depends on the individual doctor (Participant 11).

Patients treated for alcohol withdrawal were referred to specialist alcohol treatment services, and procedures for onward referral were well established. Transport arrangements could be made, hospital social workers involved as appropriate, and several participants described “informal” procedures: for example delaying discharge to give patients time to consider treatment options. As one participant suggests:

We get all sorts here but acute intoxication, alcohol withdrawal…substance misuse and mental health generally. These are “our” patients. So, unlike other patients - there’s always a pressure on beds - we tend to let them stay an extra night, so they don’t necessarily need to decide immediately whether to accept treatment or go home. It’s safe here, familiar. So, if they want to stay we always try to accommodate them (Participant 12).

Training Delivery and Barriers to Implementation

Participants recalled receiving emails or micro-training sessions on the use of AUDIT-C during morning meetings, and some had attended internal courses to support the implementation of the new alcohol guidelines. However, none had received training or reported experiences following up patients with positive AUDIT-C scores through structured feedback or brief interventions at the time the interviews were undertaken. While procedures for following up hazardous alcohol use were well established, including the informal procedures outlined above, none had observed any changes in clinical practice other than the use of AUDIT-C to screen patients and a general “increased focus” on alcohol in the emergency department and on the wards. As one participant suggests:

It’s a great start, because alcohol is something we always talk about, in terms of what do we do about it. Maybe it’s a lack of appropriate prioritization, maybe it’s a lack of effort, which is obviously partly about a lack of clear guidelines. We have all sorts of clinical guidelines to follow, how long to continue treatment, when to review treatment, and alcohol is obvious implicated in a lot of this, medically, psychologically, functionally, and if you manage to address it earlier, to intervene, …I think screening obviously helps. But what comes next? (Participant 1).

This account is typical, with most participants recognizing the scope for upstream interventions while highlighting the challenges associated with addressing lifestyle behaviours in emergency care and a perceived lack of clarity around the relevant procedures. Some mentioned having encouraged patients to contact their general practitioners, recording relevant information in the discharge summaries, and most identified a need for more readily available information surrounding appropriate community-based alcohol treatment services and available referral options and mechanisms. Participants also highlighted a need for information they could give to patients, and standardized procedures for following up on AUDIT-C scores.

The difficulties inherent in identifying and engaging with the number of patients “flying under the radar” was a recurrent theme in interviews. Aside from the challenges associated with “unpacking” alcohol use described above, participants highlighted the fragmented nature of patient care and expressed concerns that patients may not be followed up routinely on the wards:

If they have an acknowledged problem it’s easier to deal with. The difficulty is when you have a patient who drinks heavily but is fully functional. Working people, people who’ve recently retired, people who drink heavily but don’t necessarily recognize they have a problem. You’re not necessarily sure what to do, where to refer them, and I think…if I identify excess drinking as an issue then do nothing…is that ethical? If I record it in the journal, send them off for an x-ray, they end up going home. I’ve basically achieved nothing. Maybe I’ve offended them, eroded their confidence in the health service even. I’m not sure there’s a point (Participant 7).

As such, participants both welcomed the standardized assessment implied by screening using AUDIT-C and highlighted a need for an equivalent standardization of follow-up care. As one participant observes: “That’s where we have some work to do. Screening is not a problem, does not take too long, and it’s already standard operating procedure. But what comes next?”. While most participants anticipated the implementation of some form of follow-up on the wards, procedures had not been fully implemented at the time the interviews were undertaken, and none had read the procedural manual, with most positing iterations of the following assessment:

We have all kinds of things we’re supposed to follow up in a very structured way. In practice there simply isn’t time. You have to prioritize. Ideally, we need something that’s very easy to learn, something clear and simple (Participant 12).

Discussion

Our results suggest clinicians welcomed the use of AUDIT-C to screen patients and the “increased focus” on alcohol in the emergency department this implied. Clinicians experienced minimal disruption to existing clinical procedures and identified hospital emergency departments as an appropriate locus for more upstream interventions. This ostensibly high level of emergency department staff participation contrasts with results from other recent qualitative and mixed method studies that have highlighted cultures of practice not amenable to SBIRT delivery.^29,30 Grønkjær et al’s²⁹ mixed method study of health professionals’ experiences screening patients in four hospital emergency departments in Denmark, for example, highlights clinicians’ ambivalence both in terms of the appropriateness and utility of routine screening and the use of standardized screening instruments in emergency care.

This impression is compounded by studies examining implementation-related outcomes, which report clear variation in participation, retention and adherence rates across studies.^10,16,31 In Drummond et al’s¹⁸ multi-center randomized controlled trial, for example, low levels of emergency department staff participation resulted in the research team delivering interventions across six of nine sites. Interventions delivered by research teams have also been associated with lower rates of refusals and higher follow-up rates than those delivered by clinicians,³¹ with a lack of time, appropriate screening tools and the effect of screening and brief interventions on patient relationships the most frequently cited barriers to implementation.^10,16

However, while results in our study suggest clinicians reported positive experiences using AUDIT -C, relatively few assessments were completed. Screening did not represent a priority where alcohol was not implicated in the primary reason for admission. Further, while clinicians did not explicitly foreground a lack of time or concerns surrounding the potential negative impact of SBIRT on patient relationship as barriers to implementation,^10,16 these concerns are clearly implied both in the curtailment of assessments requiring repeated efforts to reframe consumption questions and in participants’ alertness to patients feeling “uncomfortable” or “judged” in these contexts. Concerns in relation to the lack of privacy on open wards have also been addressed in other studies.^10,29 Completing as opposed merely to initiating assessments as clinical guidelines are embedded more robustly in clinical practice is also clearly likely to impact on clinicians’ overall assessment of the appropriateness and utility of standardized screening.

Typologies of patients and alcohol-related norms have also been foregrounded in other qualitative studies.^29,30 Grønkjær et al²⁹ found that the social acceptability or “normalization” of potentially harmful or hazardous alcohol use, including among emergency department staff, precipitated a “nervousness” or reluctance to challenge prevailing norms, with health professionals “protecting” patients who were not obviously inebriated from the “stigma” associated with hazardous or harmful use. Sivertsen et al³⁰ similarly evoke the notion of “protection” in describing the practice of separating intoxicated patients from others, with health professionals reluctant to challenge prevailing norms or appear “judgmental” or “paternalistic”. Both these studies identify dichotomized typologies of patients, with clinicians distinguishing primarily between patients identified as “ravaged”, “worn out” or “untidily dressed” and others.^29,30

While results from this study suggest clinicians distinguished primarily between patients admitted with alcohol-related illness or injury and others and were similarly alert to the latter feeling “pressured”, “uncomfortable” or “judged”, most also highlighted the number of patients “flying under the radar” and the perceived utility of standardized assessments in identifying these patients. Studies have shown that even where routine screening has been implemented, assessments are more likely to be curtailed with patients at higher risk of hazardous alcohol use, in particular where a lack of patient cooperation precludes assessment.³² Patient surveys suggest risky drinkers may be more likely to experience alcohol-related conversations to be “awkward” or “judgmental” relative to low-risk drinkers or abstainers, although high risk drinkers are also more likely to reevaluate their drinking following these conversations.³³ A recent survey undertaken in a hospital emergency department in Norway suggests most patients respond positively to alcohol assessment on admission.³⁴ Results from this study suggest clinicians may also reevaluate their own alcohol use and question social norms around alcohol while carrying out AUDIT C assessments.

Participants’ concerns in relation to the lack of clear alcohol procedures and the variable nature of follow-up care have also been described elsewhere, with Næss et al’s³⁵ study on the implementation of SBIRT in Norwegian hospitals suggesting few have implemented robust systems for alcohol screening despite this representing a core tenet of Norwegian alcohol policies since 2013. Næss et al³⁵ suggest a combination of personal involvement or individual “champions” and adequate project support is critical for establishing robust procedures for identifying and following up harmful alcohol use. This clearly resonates with results from this study in which nurses in particular identified clear variation in terms of patient follow-up and highlighted a need for clear, unambiguous procedures in order to translate these into routine clinical practice. While clear clinical guidelines had been introduced, these had not been clearly integrated into routine clinical practice at the time the interviews were undertaken, and individual champions are likely to be instrumental in integrating these more fully into the “informal” procedures and practices enacted in emergency care settings.

Strengths and Limitations

A strength in this study was that interviews were carried out with both men and women, both medical and nursing staff, and a broad range in terms of age, professional experience and medical specialization. Another strength was the use of open, thematic questions to facilitate a narrative interviewing style, and the generation of interview data reflecting a correspondingly broad range of experiences and perspectives. This study was also undertaken during the early stages of implementation following the introduction of routine screening in the emergency department, but before procedures for following up on AUDIT C scores had been fully implemented on the wards. Results may therefore be useful in informing the implementation process moving forwards.

The latter is clearly also a limitation. Participants’ experiences were primarily limited to using AUDIT-C to screen patients: none had experience delivering structured feedback or brief interventions or detailed knowledge of the new alcohol treatment guidelines. Convenience sampling and the recruitment of patients by individuals responsible for training delivery also present clear limitations, including selection bias and limited transferability. Further, while the first author TT and last author AJ collaborated in interpreting findings, the primary reliance on a single analyst (TT) represents a further limitation, as does the number of study participants and being based in a single Norwegian hospital. While data saturation was achieved to the extent that no new information was generated in interviews, additional interviews, or a more reflexive application of thematic analysis may have generated more nuanced results.

Conclusion

Identifying and following up harmful alcohol use represents a key component of alcohol harm reduction policies. This study highlights tensions between alcohol-related harm and alcohol-related norms as these pertain to standardized and clinical assessment in screening and following up harmful and hazardous alcohol use in a Norwegian hospital. Results suggest training should focus on zero alcohol recommendations, the use of assessment tools, the acceptability of screening to patients and introducing “clear and simple” patient follow-up procedures.