How the AvoidRehab Decision Walkthrough works.

What the walkthrough is

The AvoidRehab Decision Walkthrough is a self-paced set of plain-language questions that surface the factors a thoughtful clinician would consider before recommending residential treatment. It produces one of six "things to consider" outcomes, each tied to the specific answers a user gave. The walkthrough does not assume that residential treatment is wrong: for some people it is empirically the right starting point. It assumes that the decision should be driven by clinical factors, that residential length of stay should be tied to clinical progress rather than a fixed calendar, and that the system tends to default to residential when other options would fit better.

It is built for two readers: a person trying to figure out their own next move, and a family member trying to think through someone else's situation. Both are stressed, both have short attention spans, and both deserve plain language.

What it is not

• Not a diagnosis. The walkthrough never tells you that you have a substance use disorder. Diagnosis is the work of a licensed clinician using DSM-5-TR criteria.
• Not a level-of-care prescription. The walkthrough never tells you "you need IOP" or "you do not need rehab." It surfaces factors and frames them as things to discuss.
• Not a validated screening instrument. The walkthrough is educational. It has not been studied, validated, or normed against any clinical population.
• Not a substitute for a clinician. The most important thing a user can do with the walkthrough is bring its output to a licensed provider and have a real conversation.

Source basis

Every clinical claim in the walkthrough traces back to one of the following publicly available sources. Most are free to download from federal agencies. The two clinical instruments referenced (CIWA-Ar and the C-SSRS framing) are public-domain or freely licensed.

SAMHSA Treatment Improvement Protocols (TIPs)

• TIP 45: Detoxification and Substance Abuse Treatment. The federal reference for which substances require medical detox and which do not. Drives the alcohol and benzodiazepine pathways in the walkthrough, and the explicit note that stimulant and cannabis withdrawal do not require medical detox.
• TIP 42: Substance Use Disorder Treatment for Persons With Co-Occurring Disorders. Covers the psychiatric crisis flags, environmental factors, and treatment-planning considerations.
• TIP 63: Medications for Opioid Use Disorder. Drives the medication-first framing for the opioid pathway. Buprenorphine and methadone are first-line treatment.
• TIP 35: Enhancing Motivation for Change in Substance Use Disorder Treatment. Drives the readiness-to-change question, the goal question, and the treatment of motivational interviewing as evidence-based.

NIDA

• Principles of Effective Treatment (research-based guide). The framing that duration of engagement, not residential intensity, is the strongest predictor of outcome comes directly from NIDA's published principles.

NIAAA

• Rethinking Drinking. Drinking-level thresholds (low-risk, heavy) come from NIAAA's public Rethinking Drinking materials.

Public-domain or freely licensed clinical instruments

• CIWA-Ar (Sullivan et al., 1989, Br J Addict). The walkthrough does not score users on CIWA-Ar. It uses CIWA-Ar's symptom domains (prior seizure, prior DTs, tremor) to flag for the user that these symptoms matter to a prescriber.
• COWS (Wesson & Ling, 2003, J Psychoactive Drugs). Used as background framing for the opioid pathway, not as a scored instrument.
• AUDIT (WHO). Referenced for self-assessment in the minimal-treatment outcome.
• DAST-10 (Skinner, 1982). Background reference for non-alcohol substance screening.
• PHQ-9 (Pfizer / Spitzer). Not directly used in the walkthrough; cited in the source library for clinicians who want to follow up on co-occurring mood concerns.
• C-SSRS (Columbia Lighthouse Project). The walkthrough uses only the core ideation framing (with vs without plan or intent) for the crisis question. It is not scored.
• WHO mhGAP Intervention Guide v2.0. Background reference for crisis routing.

The five questions, in our voice

The walkthrough is organized around five plain-language questions, named by AvoidRehab. They map loosely to factors the source documents discuss, but the structure, naming, and flow are ours.

1. What is actually happening physically? Substance, frequency, duration, withdrawal history. Drives the detox-or-not question. Drawn from TIP 45 and (for opioids) TIP 63.
2. Is there an immediate crisis? Suicidality with plan or intent, active psychosis or mania, pregnancy with active use, acute medical instability. Drawn from TIP 42 and the C-SSRS ideation framing.
3. What is home and daily life like right now? Stable housing, active use in home, domestic violence, support system, plus a separate question on whether daily-life functioning and any subacute psychiatric symptoms would let outpatient appointments actually be attended. Drawn from TIP 42 (treatment-planning factors and co-occurring disorders) and TIP 35.
4. What has been tried, what is wanted? Prior treatment, current goal, current readiness. Drawn from TIP 35 and NIDA Principles.
5. What does the math say? Insurance, time off work, caregiving obligations. Drawn from TIP 42's treatment-planning framework, applied to real-world feasibility.

The six outcomes

Based on a user's answers, the result page surfaces one of six "things to consider" frames. Each is tied to specific input flags, not a calculated score:

1. Right now, this looks like a medical or psychiatric emergency. Triggered by suicidality with plan or intent, active psychosis or mania, or acute medical instability. Routes to 988, 911, and SAMHSA's National Helpline. Acute hospital-level needs are deliberately separated from residential treatment needs.
2. Withdrawal management is the right first step. Triggered by alcohol with prior withdrawal seizure or DT history (or current severe pattern), long-term benzodiazepine use, or daily illicit opioid use without current medication. Result page describes three settings (at home with daily check-ins, a short residential withdrawal program, or a hospital), not a binary detox-vs-outpatient framing. For the opioid pathway, leads with medication induction (buprenorphine, methadone) followed by IOP or PHP, per TIP 63.
3. A 24-hour structured setting may be a better fit than outpatient. Triggered when nothing rises to an emergency, but daily-life functioning, subacute psychiatric symptoms, or recovery environment makes outpatient unrealistic. Distinguishes residential treatment from a recovery residence (sober-living house) paired with outpatient care, since the two are often confused and have very different cost and structure.
4. IOP or PHP deserves a hard look first. Triggered by significant use plus environmental complexity, or by a history of outpatient that was not enough. Frames intensive outpatient (9 to 19 hours per week) or partial hospital (20 or more hours per week) as the level of care to evaluate before residential.
5. Outpatient plus medication may be enough. Triggered by mild-to-moderate use with a workable home environment and no acute instability. Names the FDA-approved medications by pathway: buprenorphine, methadone, or naltrexone for opioids; naltrexone, acamprosate, or disulfiram for alcohol.
6. This may not require formal treatment at all. Triggered by use patterns that do not meet typical clinical thresholds, or by low readiness where formal treatment rarely sticks. Surfaces brief intervention, harm reduction, peer support, and individual therapy as legitimate first steps.

For pregnancy with active substance use, the walkthrough routes to a separate perinatal-specific outcome. Methadone or buprenorphine is the standard of care for opioid use disorder in pregnancy, and abrupt cessation of opioids, alcohol, or benzodiazepines can cause fetal distress. This routing is drawn from TIP 63.

Decision logic, in plain terms

The result a user sees is determined by a cascade, not a sum. Each layer is checked in order, and the first one that matches is the outcome shown:

1. If any emergency flag is set (suicidality with plan or intent, active psychosis or mania, acute medical instability): emergency outcome.
2. If alcohol use is combined with a prior withdrawal seizure, prior DTs, or a current severe pattern: withdrawal-management outcome.
3. If benzodiazepine use is long-term or high-dose: withdrawal-management outcome (medical taper).
4. If opioid use is daily and illicit, and the user is not currently on medication for opioid use disorder: withdrawal-management outcome that leads with buprenorphine or methadone induction.
5. If the user flagged pregnancy with active use: perinatal outcome.
6. If daily-life functioning is severely impaired, subacute psychiatric symptoms are present, there has been a recent psychiatric hospitalization, the home is unsafe, or significant use is combined with active use in the home and no support: 24-hour structured outcome.
7. If there is significant use, active use in the home, or a prior failed outpatient or intensive-outpatient episode: IOP or PHP outcome.
8. If use is moderate and stable, or the user is currently on medication for opioid use disorder: outpatient plus medication outcome.
9. Otherwise: minimal-treatment outcome.

Every cascade decision is documented in the page source. The walkthrough does not call out to any server; the logic runs entirely in the browser. Nothing is hidden behind an API.

Privacy and data handling

The walkthrough runs entirely in the user's browser. No answer set is sent to a server, logged, attributed, or retained.

• State persists only in the URL hash on the user's own device, which is how a result page can be reproduced or shared without storing data anywhere central.
• The page loads fonts from Google Fonts and uses no third-party analytics on individual answers. Aggregate page-view analytics may be collected by the hosting platform at the page level, never tied to walkthrough responses.
• If a user shares their result URL, the hash carries the answers needed to recreate the page. Anyone with that URL can see the same result. Treat shared URLs like any other personal information.
• The provider prep PDF is generated client-side from a static template; it does not transmit anything.

Accessibility

The walkthrough is built to meet WCAG 2.1 AA where it can. Specifically:

• All controls are keyboard-operable. Question options work with Enter and Space.
• Live regions (aria-live="polite") announce question and result changes to screen readers.
• Color is not the only signal. Each result tag has a label, each option has a visible mark, and progress is reported as text in addition to the visual bar.
• A no-JavaScript fallback message and the static SEO section keep the page usable for users without JS, search engines, and assistive technology that pre-renders.
• A skip-link is present on the landing page.

If you find an accessibility issue, the fastest fix path is to email the editorial team at the contact address in the footer.

What this walkthrough deliberately will not do

• It will not produce a numeric score.
• It will not produce a single "right answer" that bypasses a clinician.
• It will not collect contact information as a gate to the result.
• It will not refer users to a specific treatment program, brand, or facility.
• It will not assume that residential rehabilitation is always wrong. For some people it is empirically the right starting point. The walkthrough flags those cases.
• It will not pretend to be a clinical instrument it is not.

Updates and review

The walkthrough and this methodology page are reviewed at least every six months and whenever a referenced federal source publishes a substantive update. The current version is dated at the top of this page. Substantive changes are logged in the public changelog (linked from the footer when material updates ship).

Feedback

If you are a clinician, person in recovery, family member, researcher, or harm-reduction worker who sees something wrong, vague, or misleading, the editorial team wants to hear about it. Use the contact link in the footer.