LOCUS or ASAM? Choosing the right level-of-care framework

A patient arrives at intake with an active alcohol use disorder and a bipolar I diagnosis. The clinician’s substance use assessment points toward a residential ASAM placement. The same clinician’s psychiatric assessment, scored on LOCUS, comes back lower, because the patient is currently housed, employed, and taking medication as prescribed. One authorization request has to satisfy both pictures, often at two different payers, on the same intake day.

ASAM (the ASAM Criteria) and LOCUS (the Level of Care Utilization System) are the two leading frameworks behind behavioral health placement decisions: ASAM is required by most Medicaid programs and commercial payers for substance use disorder placement, and many Medicaid managed care organizations and commercial behavioral health carve-outs require LOCUS for psychiatric placement. Both use a six-dimension assessment model, but different clinical bodies built them for different presentations, and they do not always agree on the same patient.

This is rarely a clinician training gap. Experienced behavioral health clinicians generally understand both frameworks well. Usually, it comes down to one platform built around one framework, with the other framework bolted on wherever it happens to fit: an attachment, a workaround, a form the clinical record never really sees. This piece gets into what each framework measures, where they clash on the same patient, and where payer review tools fit into all of it. Then it gets into what an EHR has to get right, so documentation, authorization, and billing hold together no matter which framework a program is working under.

ASAM’s six dimensions determine which of its six levels of care a patient with a substance use disorder needs, and that level-of-care designation is what drives the billing code on the claim.

The American Society of Addiction Medicine has published national placement criteria for substance use disorder care for decades, and today it functions as the operating framework, not a reference document, for admissions, treatment planning, authorization, and revenue cycle at most SUD programs.^[1] SAMHSA describes it as the most widely used set of guidelines for placement, continued stay, and transfer or discharge decisions in addiction treatment nationally,^[2] and most major commercial payers and Medicaid programs apply it as the basis for prior authorization and concurrent review.

The current Fourth Edition, released in 2023, organizes assessment across six dimensions: intoxication, withdrawal, and addiction medications; biomedical conditions; psychiatric and cognitive conditions; substance use-related risks; recovery environment; and person-centered considerations, a new dimension in this edition covering social determinants of health, patient preference, and the need for motivational enhancement.^[3]

Those six dimensions map to six levels of care. Early intervention, formerly a separate Level 0.5, was removed from ASAM’s specialty addiction continuum in the Fourth Edition and reframed as “Early Intervention and Secondary Prevention”, a service outside the specialty continuum, though it remains authorized under a Level 1 license in at least some states, such as Illinois.^[4] The remaining continuum:

• Level 1 covers outpatient care: long-term remission monitoring (1.0), outpatient therapy (1.5), and medically managed outpatient (1.7).
• Level 2: intensive outpatient (2.1), high-intensity outpatient (2.5, the level most state licensure rules and payer contracts still call partial hospitalization), and medically managed intensive outpatient (2.7).^[5]
• Level 3: clinically managed residential services (3.1 through 3.7).
• Level 4: medically managed inpatient services.

For a full dimension-by-dimension breakdown of the Fourth Edition and what each level of care requires, see blueBriX’s ASAM levels-of-care guide. The level-of-care designation is also the reason a documentation gap on the ASAM side becomes a billing gap and not only a clinical one. For SUD-focused intensive outpatient, the code is H0015. S9480 covers psychiatric IOP instead, so it isn’t something to swap in for H0015 on an SUD claim. At the high-intensity outpatient or partial hospitalization level, H0035 and S0201 are the commonly billed codes, but which one applies can depend on diagnosis and on whether the payer is Medicaid or commercial, so that’s worth confirming per contract. Residential care maps to H0018 or H0019.

LOCUS measures psychiatric functional impairment, not substance use severity, and it is the framework most Medicaid behavioral health carve-outs and community mental health programs require for general psychiatric placement.

The Level of Care Utilization System was developed by the American Association for Community Psychiatry (AACP) for psychiatric and general mental health placement decisions.^[6] Community mental health centers, Medicaid managed care organizations with behavioral health carve-outs, and Assertive Community Treatment programs most commonly require it. Unlike ASAM, LOCUS is a proprietary instrument owned by AACP, which has practical licensing implications for programs adopting it.

LOCUS’s core design principle is the reason it can diverge from ASAM on the same patient: it weighs observable functional impairment more heavily than diagnostic severity or history. Its six dimensions, each scored 1 to 5, are:

• Risk of harm asks whether the patient is a danger to themselves or others or can’t safely care for themselves.
• Functional status looks at how well they handle daily activities, work, relationships, and medication.
• Medical, addictive, and psychiatric co-morbidity weighs physical health, substance use, and psychiatric conditions together, not separately.
• Recovery environment: level of stress and level of support in the patient’s living situation.
• Treatment and recovery history: response to prior treatment episodes.
• Engagement and recovery status: motivation, insight, and willingness to accept recommended care.

The six scores sum to a composite ranging from 6 to 30, and the composite maps to one of six recommended service intensities:

• Recovery maintenance and health management: ongoing monitoring and support for patients in stable recovery.
• Low-intensity community-based services: routine outpatient-level care.
• High-intensity community-based services: more frequent or intensive outpatient support, such as IOP.
• Medically monitored non-residential services: structured day treatment, such as PHP.
• Medically monitored residential services: around the clock residential care with psychiatric monitoring.
• Medically managed residential services: the high-intensity residential or inpatient psychiatric care.

That licensing model is stricter than it looks: AACP does not sanction paper or PDF scoring for clinical, service-planning, or utilization-management use.^[7] The only sanctioned path is an electronic scoring algorithm licensed through Deerfield Solutions, the company AACP made its sole authorized partner for building and distributing the LOCUS family of tools,^[8] either as a standalone platform or through a certified EHR integration. For a program running LOCUS alongside ASAM, that makes EHR integration one of only two compliant ways to generate an actual LOCUS score, the other being a separate standalone Deerfield license.

Clinical judgment can deviate from the composite’s recommendation when the deviation is documented. LOCUS doesn’t come with its own per-diem billing taxonomy the way ASAM does; instead it rides on the outpatient, IOP, PHP, and inpatient psychiatric codes a program already bills under for standard behavioral health services.

They share the same six-dimension structure, but they are scoring different things: ASAM answers for substance use severity, LOCUS for psychiatric functional impairment. That structural overlap is also where the two frameworks do the most good together: strong risk and safety documentation reads directly into ASAM’s Dimension 1 (intoxication, withdrawal, and addiction medications) and LOCUS’s Risk of Harm dimension at the same time. Document it once, with real detail and a number attached, and it covers both. Two separate write-ups aren’t necessary.

There’s one specific, recurring way they disagree. A patient with a strong ASAM case for residential or PHP care, built on relapse history or a high-risk living situation, can still score low on LOCUS if they’re currently functioning okay day to day, since LOCUS weighs current function over history or diagnosis. Same episode, same patient: authorization clears on the substance use side and gets denied on the psychiatric side.

Because a patient with bipolar disorder and alcohol use disorder needs the substance use side of the picture documented in ASAM terms and the psychiatric side documented in LOCUS terms, often for two different utilization reviewers, sometimes at two different payers, within a single treatment episode. For integrated and dual-diagnosis programs, this is not an edge case. It is the default intake scenario.

Arizona’s Medicaid program shows the pattern at scale: AHCCCS provides LOCUS access and AACP-delivered training to providers at no cost,^[9] and separately requires the ASAM CONTINUUM assessment for adults with substance use conditions, explicitly including those with co-occurring mental health conditions.^[10] A single AHCCCS-contracted program treating co-occurring patients is expected to run both instruments as a matter of course, not as an exception.

Doing both without doubling documentation time isn’t complicated in concept. It’s just hard to build well. An assessment tool has to run both instruments from a single encounter, without making the clinician retype demographics, presenting problem, or the safety screen a second time. The scoring logic can’t be shared either. ASAM’s dimensional-severity model and LOCUS’s 1-to-5 composite are different math, and treating them as interchangeable breaks both. Then there’s where the scores end up. They need to sit somewhere the clinician can see both, tied into the same treatment plan, instead of stuck in two forms that never reference each other.

When an EHR can’t support this, staff duplicate documentation by hand, a spreadsheet, a second system, whatever gets both frameworks covered. Or a program just goes with whichever framework its EHR already supports and makes the other diagnosis fit around that. Either way, the denials show up on the underdocumented side of the chart.

ASAM and LOCUS are professional clinical standards, developed by clinical bodies for clinical placement decisions. Many payers layer a different kind of tool on top of both: a proprietary utilization management product, licensed to the payer rather than the provider, used by commercial payers, some Medicare Advantage plans, and government agencies to make medical necessity determinations across medical and behavioral health. It is the payer’s framework, not the clinician’s, and providers do not typically choose whether to use it, or which one a given payer has adopted.

These payer-side tools do not map cleanly onto either clinical framework, which is why documentation built around one framework alone can still fail payer review. They tend to weight current acute symptoms and failed lower levels of care more heavily than ASAM’s relapse history or LOCUS’s engagement dimension. The practical takeaway: thorough, functionally specific documentation, not framework selection by itself, is what protects an authorization, regardless of which criteria set the payer applies on the other side of the request.

At minimum, it needs to run more than one structured, weighted-scoring instrument from a single encounter, keep that scoring logic configurable without a development cycle, and link every assessment result to the treatment plan and the authorization record. Platforms built around a single clinical framework push programs treating both populations into workarounds. That’s true whether the tooling is ASAM-only, common in SUD-focused EHRs, or a general mental health EHR with no SUD-specific logic anywhere. It’s a software problem more than a clinical one. Clinical staff generally understand both frameworks fine; the software is what was never built to hold both at once.

Three configuration requirements turn that gap into evaluation criteria you can bring to any vendor conversation, blueBriX included:

• Can the intake workflow run more than one structured, weighted-scoring instrument from a single encounter without duplicate data entry?
• Can assessment scoring logic be reconfigured without a development ticket when a program adds a second framework or a state changes its required instrument?
• Are assessment results linked to the treatment plan and the authorization record, rather than living as a standalone form disconnected from billing?

In a demo, ask the vendor to build a second assessment live rather than describe it on a slide, ask how authorization records are linked when a patient’s level of care changes, and ask whether dual-framework documentation requires two separate encounters.

blueBriX’s ASAM support is built into the encounter and program management workflow: ASAM score assignment happens at intake, the referral engine can be configured to route patients to the appropriate program (detox, residential, IOP, or PHP) based on the documented score rather than a manual handoff, and each ASAM level of care is configured as its own program with its own enrollment criteria, authorization requirements, and billing strategy, as described in blueBriX’s ASAM levels-of-care guide.

Underneath that ASAM workflow sits a broader assessment architecture: a library of 200+ pre-built, validated instruments, including PHQ-9, GAD-7, and C-SSRS, alongside a no-code, drag-and-drop form builder that supports weighted items, subscale calculations, and branching logic. That same engine is designed to support a second structured, weighted-scoring instrument, such as a LOCUS-style assessment, from the same encounter, since new or updated assessments are configured in an admin console rather than requested as a development ticket. For a program treating both SUD and general psychiatric populations, this is what a configuration-first assessment engine looks like, as opposed to a single ASAM-shaped form.

The platform supports outpatient, IOP, PHP, residential, SUD, crisis, and group therapy programs within one system instance, and patient transitions between levels of care preserve treatment plan continuity and documentation history rather than resetting in a new silo. Authorization records are tracked per service line and per payer combination, not only per patient, which is what prevents a PHP authorization from incorrectly auto-applying to an IOP claim after a step-down, a common denial trigger in mixed SUD and mental health caseloads when authorization tracking isn’t service-line specific.

For the Clinical/Compliance Owner, documentation gaps and incomplete dimensional records surface through the AI orchestration layer’s human-in-the-loop review, where AI suggests, the platform validates, and the clinician decides, before a gap becomes an audit finding. For the Econ/Ops Owner, prior authorization status is linked to the treatment plan rather than tracked separately, and blueBriX’s RCM services team supports tracking authorization windows through the step-up and step-down cycles specific to IOP, PHP, and residential care.