SUD Screening Tools and How They Work in Practice

Substance use disorder (SUD) screening uses validated questionnaires and, sometimes, biological tests to identify people who may have a problematic relationship with alcohol or drugs. Most screening happens through brief, standardized tools administered in a doctor’s office, emergency department, or behavioral health setting, and the process can be as short as a single question. The challenge is not just picking the right tool but getting honest answers, reaching the right people, and connecting a positive screen to actual care.

What SUD Screening Actually Looks Like in Practice

At its simplest, SUD screening is a short set of questions asked during a routine health visit. The dominant framework is called SBIRT, which stands for Screening, Brief Intervention, and Referral to Treatment. A clinician or intake worker administers a quick questionnaire. If the score crosses a threshold, the patient gets a brief motivational conversation on the spot, and those with more serious results are referred to specialized treatment. In a large U.S. implementation of SBIRT, reported alcohol use dropped by roughly a third and illicit drug use dropped by about three-quarters at a six-month follow-up, though the researchers cautioned that these decreases cannot be attributed solely to the intervention itself because of confounders and natural patterns in substance use.1PubMed. Substance use outcomes of patients served by a large US implementation of Screening, Brief Intervention and Referral to Treatment (SBIRT) A separate evaluation in New Mexico found similar reductions, and participants who received more intensive follow-up showed sharper drops in drinking frequency compared to those who received only a brief intervention.2PubMed Central. The relationship between services delivered and substance use outcomes in New Mexico’s Screening, Brief Intervention, Referral and Treatment (SBIRT) Initiative

The screening step itself is fast. Many primary care clinics now embed standard questionnaires directly into the electronic health record so that a patient answers a few questions on a tablet or screen before the physician even walks in.3PubMed Central. Electronic health records: essential tools in integrating substance abuse treatment with primary care Emergency departments have taken a similar approach, using nurse-driven triage screening questions built into the record that trigger automated prompts for next steps when a patient screens positive for opioid use disorder.4PubMed Central. Impact of Universal Screening and Automated Clinical Decision Support for the Treatment of Opioid Use Disorder in Emergency Departments

The Most Common Screening Tools

A systematic review of validated SUD screening instruments identified sixteen tools suitable for primary care. All of them screened for substance-related disorders; none covered behavioral addictions like gambling.5Preventive Medicine Reports. Identifying available substance use disorder screening tests feasible for use in primary care: A systematic review The tools vary in length and focus, but a few names come up constantly:

When a Single Question Is Enough

One of the more surprising findings in this field is that ultra-short screening, sometimes just one question, can perform nearly as well as longer instruments. A single screening question asking about drug use in the past year was 100% sensitive for detecting a drug use disorder and about 74% specific, meaning it caught every case but flagged some people who did not actually have a disorder. Its performance was comparable to the full ten-item DAST.10PubMed Central. A single-question screening test for drug use in primary care A separate study tested single screening questions for both alcohol and other drugs and found that at the best cutoff point, the single drug question caught 97% of people with substance dependence, with about 79% specificity. For alcohol, the single question captured 88% with 84% specificity.11PubMed Central. The ability of single screening questions for unhealthy alcohol and other drug use to identify substance dependence in primary care

Self-administered single-item screening questions have also been validated. In a primary care population, a self-administered single question for drug use had about 71% sensitivity and 94% specificity for detecting unhealthy drug use, and stronger numbers for detecting an actual drug use disorder: roughly 85% sensitivity and 89% specificity.12PubMed Central. Validation of Self-Administered Single-Item Screening Questions (SISQs) for Unhealthy Alcohol and Drug Use in Primary Care Patients These single-question approaches matter because the most common reason screening does not happen is lack of time. A question that takes five seconds removes that excuse.

Self-Report Versus Biological Testing

Most SUD screening relies on what patients say about their own use. This raises the obvious question of accuracy. The answer is complicated and depends heavily on the substance, the setting, and how much the patient trusts the person asking.

A study comparing self-reported drug use to urine testing in both research and primary care settings found that self-report dramatically underestimates actual use. Sensitivity of self-report for detecting opium use within the past 72 hours was only around 39-50%, and for cannabis it was roughly 10-13%. For amphetamines, no one who tested positive by urine admitted to use at all.13PubMed Central. Validity of self-reported substance use: research setting versus primary health care setting That is a sobering picture of how unreliable self-report can be when stigma or legal consequences are in play. However, the picture is not uniformly bleak. In a study of Black American women, agreement between self-report and urinalysis was classified as “good” for most substances and time points, with moderate agreement being the exception rather than the rule.14PubMed Central. Concordance between self-report drug use and urinalysis in a sample of Black American women And in adolescents in an opioid treatment trial, self-report generally held up well against urine screens, though patients in one treatment arm were more likely to underreport cocaine and opioid use, which would have exaggerated the apparent benefit of that treatment.15PubMed Central. Concordance between self-report and urine drug screen data in adolescent opioid dependent clinical trial participants

For alcohol, biological markers can fill gaps that self-report misses. Phosphatidylethanol (PEth), a direct biomarker of alcohol consumption, was found to be the most frequently positive and most sensitive marker for excessive alcohol consumption in a study of homeless individuals, detecting it with 100% sensitivity at a lower cutoff and 94% at a higher one. It outperformed carbohydrate-deficient transferrin (CDT), a more traditional marker.16PubMed. Investigating the use of PEth, CDT and MCV to evaluate alcohol consumption in a cohort of homeless individuals Biological markers like PEth are especially useful when a patient’s self-report may not be reliable, such as in forensic or child-welfare contexts, but they measure consumption over weeks or months, not whether someone meets the criteria for a disorder. A screening questionnaire and a blood test are answering different questions.

Why People Underreport and How Technology Helps

The gap between what patients say and what tests find is not random. People underreport substance use because they fear judgment, legal consequences, or losing access to care. In a prospective study of people who use drugs in Vancouver, fewer than half disclosed their drug use to a healthcare provider. Indigenous participants were significantly less likely to disclose, and those who did disclose actually rated their quality of care lower than those who kept quiet.17PubMed Central. Non-disclosure of drug use in outpatient health care settings: Findings from a prospective cohort study in Vancouver, Canada That last finding is the opposite of what you would hope: disclosure is supposed to lead to better care, but patients experienced it as leading to worse treatment.

One consistent finding across multiple studies is that removing the human from the screening process increases disclosure. When patients in an emergency department were screened via a kiosk rather than by a person, they were far more likely to report high-risk drug use. The odds ratio was striking: people were over twenty times more likely to disclose high-risk drug use to a kiosk than to a clinician doing an in-person screen.18PubMed Central. Kiosk versus In-person Screening for Alcohol and Drug Use in the Emergency Department: Patient Preferences and Disclosure A study of audio computer-assisted self-interviews for the ASSIST tool in primary care found similar themes: participants said self-administration made them more willing to be honest, especially active drug users.19PubMed Central. Feasibility and Acceptability of an Audio Computer-Assisted Self-Interview Version of the Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) in Primary Care Patients A large repeated cross-sectional study in pediatric primary care confirmed this pattern across age groups: youth screened via computer self-administration were significantly more likely to report any substance use compared with those screened by a clinician, with the effect strongest among 12- to 13-year-olds.20Frontiers in Adolescent Medicine. Does screening mode matter? A repeated cross-sectional study of computer self-administered vs. clinician-administered screening of youth substance use in pediatric primary care

Screening in Pregnancy and the Mandatory Reporting Problem

Prenatal care is one of the settings where SUD screening is most consequential and most fraught. Many states require healthcare providers to report prenatal substance use to child protective services, and that legal landscape shapes everything about how screening actually plays out. Obstetric and pediatric providers have described implementation of mandatory reporting policies as “often targeted” and potentially biased, with providers unsure what happens to patients after a report is made.21PubMed Central. Obstetric and pediatric provider perspectives on mandatory reporting of prenatal substance use

From the patient’s side, the consequences of a positive screen in pregnancy are not abstract. Qualitative research has found that pregnant women expected psychological, social, and legal consequences from being identified, including judgment by providers and reports to child protective services. Rather than disclosing, women took active countermeasures: avoiding prenatal care, emotionally disengaging from it, timing substance use to evade urine tests before visits, and sharing strategies within their social networks for getting the medical benefits of prenatal care while dodging its surveillance aspects.22PubMed Central. Women’s perspectives on screening for alcohol and drug use in prenatal care Even women who were doing the medically recommended thing, taking prescribed medication for opioid use disorder during pregnancy, described mandated reporting as unjust and stigmatizing. They felt caught between competing demands: their doctor’s recommendation to stay on medication and a child welfare system that treated that same medication as grounds for investigation.23PubMed Central. Prescribed and Penalized: The Detrimental Impact of Mandated Reporting for Prenatal Utilization of Medication for Opioid Use Disorder

Universal screening, where every patient is asked the same questions regardless of appearance or demographics, is the main response to concerns about bias in targeted screening. The NIDA Quick Screen has been integrated into prenatal electronic health records in at least one large public healthcare system for exactly this reason.24AJOG Global Reports. Implementation of universal screening for substance use in pregnancy in a public healthcare system But universal screening does not resolve the underlying tension: if the legal consequence of a truthful answer is a child welfare investigation, screening becomes a deterrent to honesty no matter how it is administered.

Making Screening Work Across Languages and Cultures

A screening tool validated in English on a mostly White population does not automatically work for everyone. Several efforts have tested whether common instruments hold up across languages and cultural contexts. The AC-OK, a combined mental health and substance use screener, was evaluated with Latino immigrants in both the United States and Spain. Its factor structure held up in both samples, and it performed well as an approximation of more specific substance abuse measures.25PubMed Central. Evaluation of the AC-OK mental health and substance abuse screening measure in an international sample of Latino immigrants The TAPS tool, which screens for tobacco, alcohol, prescription drug misuse, and illicit substance use, has been developed in a Spanish-language version with strong performance: sensitivity and specificity above 90% for tobacco and above 80% for alcohol at recommended cutoffs, under both self-administered and interviewer-administered formats.26PubMed Central. Development and Feasibility of a Spanish Language Version of the Tobacco, Alcohol, Prescription Drug, and Illicit Substance Use (TAPS) Tool

Race and sex also shape who screens positive and how. In a community-based SBIRT implementation in Allegheny County, Pennsylvania, men were more likely than women to screen positive on both the prescreen and full screen, and Black patients were more likely to screen positive than White patients.27PubMed Central. Project Lifeline-II: Feasibility of Implementing Screening, Brief Intervention, and Referral to Treatment (SBIRT) in Allegheny County, Pennsylvania These disparities may reflect genuine differences in substance use patterns, differences in willingness to disclose, or differences in the populations presenting for care. Screening alone cannot distinguish among these explanations, which is why linking a positive screen to culturally competent follow-up care matters as much as the screening itself.

Screening Outside the Doctor’s Office

SUD screening does not have to happen in a traditional clinic. Community health centers, pharmacies, and even jail-based health programs have experimented with SBIRT. A pharmacy-based pilot study confirmed that screening was feasible in a retail pharmacy setting, with pharmacists able to administer validated tools and compare results to prescription drug monitoring data to flag patients who might benefit from a referral.28PubMed Central. Screening, Brief Intervention, and Referral to Treatment (SBIRT) in a retail pharmacy setting Emergency departments are another high-yield venue. A data-linkage study of ED patients with substance-related visits identified predictors of subsequent overdose death, and the authors argued that routine data already in patient records could be used to flag high-risk individuals for connection to lifesaving care.29PubMed Central. Predictors of Overdose Death Among High-Risk Emergency Department Patients With Substance-Related Encounters

When a primary care clinic implemented a low-burden universal screening program called Helping HAND, roughly 47% of the nearly 1,900 patients screened over a year endorsed some level of substance use. Only about 4% declined to participate. In the year before the screening program existed, the same clinic had referred just nine people for substance abuse treatment. In the first year after implementation, 44 individuals were enrolled in counseling, evidence-based treatment, and case management, and nearly 800 received informational brochures.30PubMed Central. Low-Burden Universal Substance Use Screening in a Primary Care Clinic to Lower Implementation Barriers That kind of before-and-after picture illustrates the basic premise of universal screening: people with substance use problems are already in the healthcare system, but unless you ask, you will not find them.

Screening for Co-Occurring Mental Health Disorders

Substance use disorders rarely exist in isolation. Anxiety, depression, PTSD, and other psychiatric conditions overlap heavily with problematic substance use, and screening that ignores this misses a large part of the clinical picture. The Centre for Addiction and Mental Health Concurrent Disorders Screener was developed for use by addiction counselors who may have limited mental health training. In validation testing, it was highly sensitive to identifying psychiatric disturbances, had a low rate of false negatives, and showed excellent test-retest reliability, though its ability to distinguish among specific psychiatric diagnoses was more limited.31PubMed. The Centre for Addiction and Mental Health Concurrent Disorders Screener The AC-OK screener used with Latino immigrants similarly covers both mental health and substance abuse on a single instrument.32PubMed Central. Evaluation of the AC-OK mental health and substance abuse screening measure in an international sample of Latino immigrants Catching both conditions at once matters because treating one without addressing the other tends to produce poor outcomes for both.

Artificial Intelligence and Clinical Notes

The newest frontier in SUD screening does not involve asking patients anything at all. Researchers have been training large language models to scan free-text clinical notes, the unstructured narratives that clinicians write after patient encounters, and extract evidence of substance use. A fine-tuned model called Llama-DrugDetector achieved near-perfect accuracy for detecting mentions of most individual substances and strong performance on harder tasks like identifying prescription opioid misuse and polysubstance use.33PubMed Central. Enhancing substance use detection in clinical notes with large language models A separate effort using a different model family showed that larger language models excelled at extracting SUD severity information from clinical notes without needing any task-specific training.34npj Mental Health Research. Decoding substance use disorder severity from clinical notes using a large language model

These tools are not replacing questionnaires yet. They are best understood as a complementary layer: a way to catch signals of substance use that are documented in the record but never formally coded or acted on. A clinician might write “patient smells of alcohol” or “reports using friend’s Percocet” in a progress note without entering a formal SUD diagnosis. Language models can surface those mentions systematically across thousands of records, potentially identifying patients who were never screened or whose screening results were never followed up. The technology is still in early validation, but it represents a genuinely different approach to detection, one that works passively on data that already exists rather than requiring active participation from the patient or time from the provider.