Hospital Doctors Miss Most Chances to Spot Risky Drinking and Drug Use

Every year, millions of patients pass through the doors of internal medicine wards and emergency departments carrying a hidden burden that has nothing to do with the illness that brought them there. At-risk alcohol consumption and illicit drug use silently raise the odds of disease, complications, readmission, and even death, yet they often go unrecorded by the very clinicians best placed to intervene. A new study from Oslo, Norway, has now put hard numbers on this blind spot, and the results are striking: when researchers compared verified substance use data with what physicians actually wrote in patients’ electronic records, the majority of cases were either incompletely documented or missed entirely.

The research, published in the journal Advances in Drug and Alcohol Research, drew on a prospective sample of 2,872 acutely hospitalized internal medicine patients recruited at Lovisenberg Diaconal Hospital in Oslo between November 2016 and December 2017. During the original study, patients self-reported their drinking using the four-item Alcohol Use Disorder Identification Test, known as AUDIT-4, and provided blood samples analyzed with liquid chromatography-mass spectrometry, or LC-MS, to detect illicit drugs such as tetrahydrocannabinol, amphetamine, methamphetamine, cocaine, MDMA, and heroin. Crucially, none of this information was ever shared with the hospital staff treating those patients. That design choice transformed the study data into a gold-standard benchmark against which the physicians’ own documentation could be measured.

From the original cohort, the researchers identified 688 patients whose drinking exceeded low-risk guidelines, defined as AUDIT-4 scores of at least 5 for women and 7 for men, and/or who tested positive for one or more illicit drugs. Electronic records were available for 653 of them, comprising 548 alcohol-positive patients and 157 illicit drug-positive patients, with 48 individuals appearing in both groups. The team then combed through admission records and discharge summaries, scoring an alcohol assessment as complete only if it described both the quantity and frequency of consumption, for example units of alcohol per week, and a drug assessment as complete only if it named the substance and described how often it was used. Vague phrases such as “drinks moderately” or “recreational drug use” were classified as incomplete.

The findings were sobering. Among the 548 patients with study-verified drinking above low-risk thresholds, physicians documented quantity and frequency in only 43.2 percent of records, equivalent to 237 patients. An alcohol-related intervention, whether counseling, planned follow-up with a family physician, or referral to a treatment institution, was recorded in just 22.0 percent, or 121 patients. Perhaps most telling, only 9.1 percent of alcohol-positive patients received an alcohol-related ICD-10 diagnosis code at discharge, and even among those with scores indicating harmful drinking or possible dependence, coding rose to just 20.7 percent. For illicit drug users, the picture was similar: a complete assessment appeared in 34.4 percent of records and an intervention in 26.0 percent.

The timing of documentation mattered as much as its presence. In admission records of alcohol-positive patients, 41.8 percent contained complete quantity and frequency assessments and 15.2 percent contained no mention of alcohol at all. By discharge, the situation had deteriorated sharply: only 18.6 percent of discharge summaries included a complete assessment, and 56.9 percent made no reference to alcohol whatsoever. A parallel pattern emerged for drug-positive patients, with complete assessments falling from 31.2 percent at admission to 28.5 percent at discharge, while 44.6 percent of discharge records omitted illicit drug use entirely. Because discharge summaries are the primary channel of information to the general practitioners who manage post-hospital care, this attrition represents a systematic loss of clinically vital information at exactly the moment it matters most for continuity of treatment.

Statistical modeling revealed which patients were most likely to receive thorough attention. Complete quantity and frequency assessments of alcohol use were independently associated with harmful drinking levels, with each additional year of age, and were significantly less likely during outpatient emergency department evaluations compared with ward admissions. Interventions were strongly linked to harmful drinking, with an adjusted odds ratio of 4.87, and to the presence of a complete quantity and frequency assessment itself, which tripled the odds of an intervention at an adjusted odds ratio of 3.66. Psychological distress, measured by the Symptom Checklist-5, and unemployment also independently predicted intervention. The relationship between assessment and intervention is the study’s central mechanistic insight: physicians cannot counsel patients about drinking patterns they have never quantified, making documentation the gateway through which all subsequent preventive action must pass.

Why does this gap between verified substance use and recorded substance use persist? The authors point to the unforgiving tempo of emergency care, where time pressure and crowding squeeze the space available for detailed history taking, a hypothesis supported by the lower documentation rates among outpatients evaluated and discharged within hours. But the stakes of omission are high even at consumption levels well below overtly harmful thresholds. Recent population-level analyses have demonstrated a dose-response relationship between alcohol intake and all-cause mortality beginning around 100 grams per week, and at-risk drinking worsens outcomes across a remarkable range of common hospital conditions: it increases recurrence of atrial fibrillation, raises the risk of heart failure readmission, undermines glycemic control in diabetes, contributes to falls in older adults, reduces medication adherence, complicates anticoagulant therapy, and interacts dangerously with psychoactive prescriptions. Unrecognized heavy drinking also leaves patients vulnerable to withdrawal syndromes, whose most severe manifestation, delirium, can be fatal.

The study’s authors argue that the tools to close this gap already exist and are neither burdensome nor experimental. AUDIT-4 takes minutes to administer and can be simplified further to just the quantity and frequency questions, and patients in the study readily disclosed both at-risk and harmful drinking when asked. Brief interventions delivered in emergency departments have been shown to reduce at-risk drinking in the short term, and alcohol interventions in primary care produce measurable reductions in harmful drinking at twelve months. The most promising structural solution may be technological: embedding screening questions directly into electronic triage forms has achieved screening rates approaching 97 percent in prior research, whereas physician-directed training alone has yielded more modest results. The authors also suggest that clinicians could make fuller use of ICD-10 codes that denote alcohol use without implying harm, since more consistent coding would improve the registries used to estimate the true scale of alcohol-related harm in hospital populations.

Limitations remain, and the authors acknowledge them candidly. The data come from a single mid-sized urban hospital, so local documentation practices may not generalize everywhere; self-reported alcohol consumption is subject to recall bias, though typically in the direction of under-reporting, which would make the true prevalence of at-risk use even higher; and some counseling may have occurred without ever being written down. The illicit drug analyses were under-powered for regression modeling, and laboratory detection of a drug says nothing about dosage or frequency. Yet the core conclusion stands with unusual clarity. Hospital admission is one of the few moments when a person with risky drinking or drug use is guaranteed contact with the health system, and this study shows that in most cases that moment slips past unused. Treating substance use documentation with the same rigor as tobacco status, allergy lists, and medication records, and verifying that discharge summaries carry it forward, would convert a routine hospital stay from a missed opportunity into a genuine public health intervention delivered at scale.

Subject of Research: Documentation and management of at-risk alcohol and illicit drug use among acutely hospitalized internal medicine patients

Article Title: Missed opportunities: the detection and management of at-risk drinking and illicit drug use in acutely hospitalized patients

Article References: Gamboa, D., Kabashi, S., Jørgenrud, B., Lerdal, A., Nordby, G., & Bogstrand, S. T. (2025). Missed opportunities: the detection and management of at-risk drinking and illicit drug use in acutely hospitalized patients. Advances in Drug and Alcohol Research, 5, Article 14149. https://doi.org/10.3389/adar.2025.14149

Image Credits: AI Generated

DOI: 10.3389/adar.2025.14149

Keywords: alcohol screening, illicit drug use, emergency department, electronic health records, AUDIT-4, LC-MS, brief intervention, internal medicine, discharge summaries, ICD-10 coding, substance use assessment, hospital care