Compassion and compliance are not opposites

IntellaRx is the Compassionate Compliance Program, a nonprofit division of the Sevadar Foundation. Its customer is the public agency: health departments, Medicaid programs, corrections and reentry health services, county and municipal public-health authorities, and the task forces they convene.

It is not a consumer service. A patient looking for pain treatment should see the practice's opioid stewardship page instead. Unsure where you fit? Tap the directory assistant, the light in the corner of the page.

Why did overdose deaths keep rising after opioid prescriptions fell?

Because policy aimed at the wrong number.

For two decades, the target was the count of opioid prescriptions. The count came down. Deaths did not follow. They climbed for six more years, to nearly double the 2016 figure. By the time prescribing rules took hold, the lethal supply was illicit.

IntellaRx lays the chain out in eight steps:

1. Chronic pain is a mass problem. New cases arise faster than new cases of diabetes, depression or high blood pressure.
2. Compassionate prescribing reform created ground for pill mills. By 2010 that industry had a geographic center.
3. The crackdown worked on prescribing. The demand it displaced moved to heroin, then to illicit fentanyl.
4. Prescriptions fell. Deaths kept climbing.
5. Dose arithmetic ignored the individual. Bioavailability, metabolism and genetics produce large differences in the dose a person needs.
6. Enforcement analytics counted without context: prescriptions and doses, no diagnosis, no treatment intent.
7. A stranded population formed.
8. Diversion stayed unaddressed, because nobody positively identifies who receives the medication.

Under all eight sits a substrate: the social and economic conditions that make a population vulnerable to misuse, addiction and diversion.

Who are the "pain refugees"?

Patients cut off from established pain treatment. They do not stop existing. They cycle through primary care, pharmacies, urgent care and emergency departments with no stable answer.

When researchers called Michigan primary care clinics posing as such a patient, 40.7% of eligible clinics said they would not prescribe. Tapering or stopping therapy is linked to more overdoses, suicides and mental health crises.

The enforcement analytics that created this group cannot tell a pill mill from a specialist who treats the most complex patients in a region. Both generate the same numbers. So the specialist's patients lose access, and diversion, the thing everyone set out to stop, stays possible.

Why does the framing matter to an agency?

Call the problem "too many prescriptions" and the fix is a limit. You get fewer prescriptions.

Call it "decisions made without the information to make them well" and the fix is better information. You get fewer overdoses, fewer people cut off from care, less diversion.

IntellaRx is built on the second framing. Prescribing volume is explicitly not one of its outcome measures. That is the metric two decades optimized, while opioid-involved mortality rose 95% to its 2022 peak. An agency that needs prescribing reduction as its headline measure is, by IntellaRx's own account, probably looking at the wrong program.

What does IntellaRx consist of?

Sixteen components in five groups. One rule governs all of them: they exist to prevent diversion, identity theft, misidentification and preventable overdose, and to give clinicians and agencies the context counts lack. They do not exist to police people in pain.

Getting the medication to the person

Pharmaceutical delivery. Two failures meet at the pharmacy counter. A patient who cannot reach a pharmacy, or whose pharmacy is out of stock, stops a medication they are physically dependent on. Withdrawal, uncontrolled pain, an emergency visit and sometimes the illicit supply follow. And when a patient cannot collect medication, someone else does. That handoff is an unrecorded link in the chain.

Licensed pharmacy partners dispense. Delivery is timed ahead of run-out, signed for by the patient or an authorized adult with identity confirmed at the door, carried by drivers trained on confidentiality, and updated the same day a prescriber changes a dose. Delivery is never a condition of treatment.

Safe storage and rescue medication. The largest measured source of prescription opioid misuse is not theft. In the 2024 National Survey on Drug Use and Health, 31.3% of people who misused a prescription pain reliever got it free from a friend or relative, and 4.0% took it without asking. More than a third of misuse starts in somebody else's medicine cabinet. Enrollment includes a lockable container, naloxone with household training, and a disposal route. Naloxone is no more an accusation than a fire extinguisher is a comment on a household's cooking.

Knowing who receives it

Biometric cross-validation. Palm-vein imaging or 3D facial recognition at the point of transfer. IntellaRx states the limit up front: there is no US estimate of how much diversion comes from identity fraud, so it claims no diversion-prevented figure. The case is patient protection. A misattributed result hands the wrong person a positive toxicology or a lost prescription. The system answers one question at one moment. It does not track location, search face databases or feed law enforcement. Refusing enrollment costs nobody their medication.

DNALock and custodial toxicology. A buccal-swab identity anchor and chain-of-custody collection, so a result can never attach to the wrong person.

Understanding the clinical picture

ProviderSynch. A pull API that returns a scoped health-history summary when a prescription is transmitted. The pharmacist sees diagnosis and treatment context, not a bare number.

EMR and laboratory integration. Diagnosis coding, lab results, chronic care management and remote therapeutic monitoring in one view.

Metabolic determination. Dose requirements vary roughly thirteen-fold between people. Genetic differences in liver metabolism alone account for three-fold or more. A CYP2D6 poor metabolizer on an apparently high dose and a person acquiring more than they need produce the same number, and the first gets treated like the second. Metabolic characterization and co-medication review put an explanation in the record before anyone has to defend it.

The limits are published beside it. No genotype tells you the right dose, and trials of pharmacogenomic-guided opioid dosing found no better pain outcomes. What the evidence supports is narrower: explaining an outlier, and avoiding codeine and tramadol in CYP2D6 poor and ultrarapid metabolizers.

MedimetRx psychometrics. Validated behavioral health screening, completed by the patient in their own time, scored against normative data and their own history. Critical results, suicidality first, go to a clinician immediately. A positive screen means more support, never less treatment. No individual result is reported to an agency.

Remote patient monitoring. Consented, revocable between-visit signal, paired with remote therapeutic monitoring of function, usually the more useful signal in chronic pain. Almost everything that goes wrong happens between visits. Monitoring is scoped to what someone will act on and gives the patient a route to raise a problem. Monitoring that only flows toward the clinic is surveillance. Monitoring the patient can use is care.

Assessing risk honestly

Risk assessment from medical, social and family history, per the federal Pain Management Best Practices Inter-Agency Task Force. Social determinants verification, so support matches need. AI integrated rule discovery using the simplest model that explains the data, and never a per-person suspicion score.

Watching the supply, not the patient

Drug-supply signal analysis, emerging threat analysis and a real-time overdose registry. CDC surveillance tied nearly seven in ten 2023 overdose deaths to illegally manufactured fentanyls. The supply keeps moving: carfentanil went from 29 deaths in the first half of 2023 to 238 in the first half of 2024. The gap between a change entering a local supply and a jurisdiction knowing is measured in months. People die in that gap who did not have to.

IntellaRx corroborates toxicology, overdose events and street-vocabulary signals and issues warnings for a county, a corridor or a set of ZIP codes. Never for an individual. It is a public-health instrument, not an interdiction tool.

What will IntellaRx not do?

• Prescribe, dispense, diagnose or treat.
• Build enforcement cases. It declines deployments structured that way.
• Score patients by suspicion or monitor anyone's social media.
• Report individual behavioral health, social or monitoring data to an agency. Agencies get aggregates.
• Make treatment conditional on participation.
• Tell an agency whether HIPAA, 42 CFR Part 2, state confidentiality law, corrections rules or public records law apply to it. That belongs to the agency's counsel. IntellaRx describes the data precisely enough for counsel to decide, and warns agencies not to conclude Part 2 is irrelevant just because something else is.

The enforcement refusal is practical as well as ethical. Every clinical component runs on honest answers. A program believed to feed enforcement stops getting them almost immediately.

What principles govern the program?

Seven commitments, each one a constraint:

1. Assume positive intent. Start by naming a villain and you build instruments that mostly catch other people.
2. Break the cycle of shame. Judged patients stop telling the truth, and every component depends on the truth.
3. Feedback and education over optics. If nothing changes when a measurement changes, stop collecting it.
4. Publish the weak evidence. A federally funded review found no evidence that urine drug testing mitigates risk in opioid prescribing, and that results can be misread in ways that harm patients. IntellaRx publishes that at full prominence.
5. Never make treatment conditional on the program.
6. Never tell an agency its own legal status.
7. Identify outliers instead of marginalizing them. A patient whose dose does not match the arithmetic is a person to understand, not a number to correct.

The frame is the federal Task Force created under the Comprehensive Addiction and Recovery Act of 2016: a biopsychosocial model, individualized care, risk assessment grounded in history, and behavioral health treated alongside physical pain.

How does an engagement work?

Four phases, each with an output and an exit.

1. Scoping. The population, the problem, what has been tried, what data exists. Output: a written problem statement and an honest verdict on whether IntellaRx helps. Sometimes it does not, and saying so costs the agency nothing.
2. Design. Components, population, data flows and outcome measures, agreed before launch, including measures that could make the program look bad.
3. Pilot. Defined population, geography and duration, with pre-registered measures.
4. Evaluation and decision. Results reported whatever they say, analyzed by the agency's own epidemiologists. A pilot that fails should stop.

The agency brings a named clinical or public-health owner, a defined population ("the state" is not a pilot), a route for participants to be heard, and counsel's written determination before any data moves.

A defensible pilot enrolls by rule, not convenience, including the hard-to-reach. It uses a comparison, such as a matched county or stepped-wedge rollout, because national opioid deaths fell from 82,851 in 2022 to 55,007 in 2024 with no help from anybody's pilot. It follows patients at least 12 months, because overdose is too rare for a three-month pilot to detect. It carries a pre-agreed stopping rule and an independent evaluator. And IntellaRx will agree in writing to publish results whatever they show.

What should IntellaRx be judged on?

• Harm: fatal and non-fatal overdose, emergency visits for pain, withdrawal or overdose, mental health crises, deaths from any cause.
• Continuity of care: patients who lose a prescriber, therapy gaps, patients who leave care, time to a new clinician. One cohort study found tapering associated with termination of care at an adjusted odds ratio of 4.3.
• Delivery: component reach, screening completion and response time, naloxone and storage in place, on-schedule deliveries.
• Population: time from a supply change to a warning.

And the signals of failure: more patients leaving care, more therapy gaps, screening used to restrict treatment, identity or toxicology results producing dismissals, differential outcomes by race, insurance or geography, and enrollment skewed toward the easy-to-serve.

How is the data governed?

Collect only elements that change a decision. Keep clinical detail with treating clinicians and send agencies aggregates, on separate paths rather than one path with a permission switch. Write down the purpose of every flow. End every retention period with actual deletion. Log all access. Make consent refusable without consequence.

Why a nonprofit?

The harm in this history came from incentives. Distributors had a volume incentive. Pill mills had a cash incentive. Enforcement had a caseload incentive. Insurers had a cost incentive. Almost nobody's incentive pointed at the person in pain.

A 501(c)(3) makes nobody virtuous. It removes one class of incentive and puts the finances on the public record. No prices are published; cost is set in writing during scoping.

The program's argument came out of the interventional pain and metabolic medicine practice of Dr. Gurpreet Singh Padda, MD, MBA, MHP: patients dropped by someone else, pharmacies that could not fill, doses that did not match anyone's arithmetic. His biography.

How do I contact IntellaRx?

IntellaRx 4477 Woodson Rd, Suite 203 St. Louis, MO 63134 Call (314) 250-5100

The first conversation is scoping, not selling. Send no patient names, diagnoses or identifying health information through a web form.

Frequently asked questions

Does IntellaRx replace a state prescription drug monitoring program?

No. It is complementary to a PDMP, and its published evidence on PDMPs includes their limits.

Can IntellaRx tell us whether HIPAA applies to our agency?

No, and any vendor who answers that for you deserves caution. IntellaRx describes what data the program touches so your counsel can decide.

Do patients have to join any component to keep their treatment?

No. Declining biometric enrollment, delivery, monitoring or screening removes nobody from care. A consent that cannot be refused is not consent.

Why is prescribing volume not an outcome measure?

Because optimizing it produced the stranded-patient harm the program exists to fix. IntellaRx is judged on overdose, continuity of care and diversion.

Does IntellaRx make claims about how much diversion it prevents?

No. That figure has never been measured, and a vendor offering a percentage is offering a number nobody has.