EPM Behavioral Mathematics  ·  SUD Clinical Decision Support

See what other tools can't.

The patient who looks stable but isn't. The risk that standard assessments miss entirely. EPM surfaces what's hidden — so your team can act before the crisis, not after.

Request a Live Demo See the Evidence
EPM patient detail
EPM why this read

Why EPM Exists

The patient who looked fine.

"Standard measures said stable. EPM said otherwise. Fifteen days later, we understood why."

A patient in an IOP program — attendance perfect, MAT compliant, engaged in group. Every standard measure said success story in progress. EPM said otherwise.

Stability was high, but Risk was climbing. Sober social connections quietly fraying. Cravings increasing. A declining trajectory the standard picture had no mechanism to detect.

The clinical team was flagged. The overdose happened fifteen days later. It was survived. Without that window, it might not have been.

EPM surfaces information. Your clinicians make the calls.

Q2 flag — high stability high risk
Why this read breakdown
⚡ Arizona Providers — Time Sensitive

AHCCCS & The New AI Billing Review

Arizona just changed the rules.

Starting July 2026, Arizona deploys the nation's first AI-informed Medicaid prepayment review system. Before AHCCCS pays a behavioral health claim, AI screens it for fraud risk. The question for legitimate providers is now simple: does your documentation prove the work was actually done?

  • 01EPM generates timestamped, role-attributed clinical records across every staff role — house manager, case manager, facilitator, UA. A sham facility cannot fabricate this data shape.
  • 02Every UA result is correlated with the prior week's behavioral risk score. A positive screen matching a documented rising-risk pattern is proof of real clinical work.
  • 03Facilities that can't demonstrate clinical reality in their data are the target. Facilities that can are protected. EPM is the difference between those two categories.
Audit Readiness — EPM vs Standard EHR
Timestamped per-event clinical logEPM ✓
Multi-role corroboration (3+ staff)EPM ✓
UA result tied to prior risk scoreEPM ✓
Longitudinal behavioral trajectoryEPM ✓
Standard EHR check-box documentationBasic only
Prediction matched to outcomeNot available

Behind the Math

A clinical summary of what most tools don't know.

Standard assessment tools are built from research datasets. EPM was built from that — and from hundreds of clinical and real-world observations that never make it into a published study. What follows is a small sample of what's in the engine.

Critical
Powdered fentanyl is not the same risk as heroin
Pharmacologically harder to stabilize. Standard buprenorphine dosing is often insufficient for fentanyl-primary patients. The math weights them differently. Most tools use a single opioid category. Powdered fentanyl — the deadliest form — is invisible to the eye but fluoresces under UV light, a harm reduction fact most clinicians have never been taught.
Critical
Counterfeit pills are a separate risk category
Fake Xanax. Fake Adderall. Fake oxycodone. The patient believes they took a benzo. It was fentanyl. Zero tolerance. EPM flags counterfeit pill exposure independently — because intent and actual exposure are clinically different, and the patient genuinely didn't know what they were taking.
Emerging
Xylazine skin lesions are being misread in the field
Patients presenting with severe necrotic skin wounds — wounds that look like leprosy — are often being assessed as homeless with bacterial exposure. In many cases it's xylazine coming out through the skin from the inside. Naloxone doesn't reverse it. EPM accounts for xylazine exposure. Most EHRs don't have a field for it.
Clinical Flag
Methadone on fentanyl is a different clinical picture
A fentanyl-primary patient selecting methadone over buprenorphine carries a different risk profile. Methadone produces its own euphoric effect. EPM separates MAT type from MAT compliance and weights them differently based on primary substance. That distinction is in the patent.
Clinical Flag
Day 87 of a 90-day program is a danger zone
Highest attendance. Best behavior. Most compliant. And statistically one of the most vulnerable windows in recovery. EPM detects patients approaching program end without locked aftercare and flags the cliff before discharge. No other tool measures this.
Clinical Flag
Polysubstance is a multiplier, not a checklist
Standard tools add substances together. EPM treats polysubstance use as a compounding risk multiplier — because cocaine and alcohol together isn't cocaine plus alcohol neurobiologically. It's a different patient with a different trajectory.
Built In
Getting a job in the first 30 days can be a relapse trigger
Employment early in recovery looks like progress. The stress of a new job in the first month is a documented relapse risk. EPM fires an automatic clinical alert when employment is logged before 30 days sober — a non-intuitive signal that no standard EHR captures.
Built In
Sleep disruption, trauma, and social contagion are weighted variables
Not checkboxes. Weighted mathematical inputs that change the entire risk profile. A patient whose entire social network is actively using isn't at elevated risk — they're at maximum contagion risk. The math treats them that way.
These are eight examples. There are hundreds of observations like these built into the engine — drawn from clinical literature, validated datasets, and direct field experience with the population this tool was built to serve. The depth is the point.

The Evidence

How EPM performs against the tools you already use.

EPM's scoring engine was validated against four consecutive years of national substance use treatment data — the Treatment Episode Data Set published by SAMHSA. Same engine, same math, four independent years. It held. The validation cohort was specific: opioid-primary patients in IOP programs, community-dwelling, measured on real treatment outcomes. Not a simulation. Not a mixed population.

Existing Tool Their Score EPM
OQ-45
SUD treatment outcome
0.62 AUC 0.80–0.84 AUC
ASI Composite
SUD relapse prediction
0.75 AUC 0.80–0.84 AUC
COMPAS
Recidivism · 137 variables
0.71 AUC 0.716 AUC
12 vars · zero racial inputs
C-SSRS
Suicide risk sensitivity
13% 93%

SUD: four years TEDS-D federal data, N=103,168, opioid IOP clean cohort, 2020–2023. Recidivism: NIJ Challenge, cold — no training on this data. C-SSRS: 15-case head-to-head. EPM outperforms on every domain measured.

108/100
When the math breaks 100 — that's not a bug. That's a signal.
EPM scores can exceed 100 or go negative. A stability score of 108 means a patient is performing beyond the standard ceiling — exceptional engagement, exceptional support, exceptional trajectory. A negative trajectory score means the decline is accelerating faster than the baseline model predicts. The math doesn't cap reality to fit a scale. It reflects what's actually happening. That's the point.
Patent Pending · US App 64/063,351 · 108 claims
EPM Clinical is clinical decision support software. Scores surface information for clinician review. Clinicians retain all clinical decision authority. Not an FDA medical device.

Built By

R
Robert S. Entis
PRSS · Principal Inventor · EPM Behavioral Mathematics
Not a researcher. Not a theorist. Just another casualty of the opioid war — a 26-year legitimate pain patient who unexpectedly became a homeless fentanyl addict, flatlined three times in a detox facility, and happened to have 40 years of programming and enterprise technology experience to turn the pain and regrets into a human behavioral prediction engine.

To be honest — EPM doesn't exist to repair what I've destroyed in my life or to heal the close family relations I miss every day. It cannot bring my son back to life. It doesn't exist to profit from addiction.

It exists to help end the daily fentanyl and drug overdoses. To change the meaning of R·I·P — Recovery Is Possible. Because Relapse Is Predictable.
R
Dr. Rodney Pearson, MS, D.Min, LPC
Co-Inventor · Licensed Professional Counselor · Phoenix, AZ
Dr. Rodney Pearson has worked in counseling and behavioral health for over 25 years. Founder of RodTree Behavioral Health and Wellness, he holds a Master's in Professional Counseling and a Doctorate of Ministry, and is a licensed professional counselor specializing in trauma, substance use, and co-occurring disorders. His community-centered, whole-person approach directly shaped EPM's treatment engagement architecture.
K
Dr. Cameron Kenne, PhD, DBA, MST
Co-Inventor · Clinical Psychologist · Behavioral Health · Gilbert, AZ
Dr. Cameron Kenne brings a rare combination of clinical depth and technical sophistication to EPM. A doctoral-level clinical psychologist specializing in addiction recovery, trauma, and co-occurring disorders, Dr. Kenne also holds advanced degrees in business administration and technology — uniquely positioned to bridge behavioral science and clinical software. Evidence-based, recovery-focused, and fluent in English and French, he maintains an active commitment to community recovery advocacy.
C
Connie Entis
Co-Inventor · Disclosed Relapse Claim Group
Proud mother of three. Thirty-six years of marriage through riches, rags, three years on the hot concrete streets of Phoenix, the loss of a son, and almost her husband in medical detox. She never left. Not once.

"She is the strongest person I have ever met." — Robert, every day.

Connie knows recovery from the inside — 120-day inpatient, Crossroads for Women, a 90-day IOP, and two years of hard-won sobriety. Today she co-manages recovery housing and walks residents through housing crises and the first fragile steps back into normal life. Her insight into the difference between disclosed and detected relapse shaped one of EPM's core patent claim groups.

The person who stays when everything falls apart sees things no researcher ever will.

RIP — Recovery Is Possible

EPM runs on top of your existing system.

Your EHR stays. Your billing system stays. Your workflows stay. EPM adds a behavioral scoring and monitoring layer on top of whatever you're already using — without replacing it.

Think of it as a scoring engine and longitudinal monitor that your existing system doesn't have. One assessment at intake. Automatic rescore when anything changes. A compliance-ready audit trail that runs alongside your current documentation.

For Arizona providers: EPM's timestamped multi-role clinical record is designed to be audit-ready for the new AHCCCS prepayment review system. It works alongside your EHR — not instead of it.

Kipu BestNotes AZZLY Welligent Qualifacts + Any FHIR-compatible EHR
1
Keep your EHR
Demographics, billing, scheduling, treatment plans — stay exactly where they are.
2
Add EPM scoring
5-minute intake assessment. Behavioral scores calculated server-side. No math exposed to the browser.
3
Monitor automatically
Log a UA, a relapse, a housing change — EPM rescores instantly. Clinical flags fire when patterns emerge.
4
Audit-ready by default
Every event is timestamped and role-attributed. Your documentation proves the work was done.

Get In Touch

Let's talk about your program.

Every demo is live, not recorded. We walk through real scoring, a blind-spot case, and what your documentation looks like under the new AHCCCS review — using fictional patients in a protected environment.

If you're an Arizona IOP or residential program, the AHCCCS conversation alone is worth 15 minutes.

Or email directly: robert@epmclinical.com

Response within 1 business day. Your information is never sold or shared.

EPM Clinical is clinical decision support software. Scores surface information for clinician review — not to replace clinical judgment. Clinicians retain all clinical decision authority. EPM is not an FDA medical device.