How to evaluate a remote pain program: a clinician's checklist.

Why does the evaluation framework matter more than the technology?

Because the technology is the part most likely to change and the least likely to decide the outcome. A headset, an app, and a video platform are delivery channels. Whether a patient with eight months of low back pain gets back to lifting a grandchild depends on what is delivered through them: which clinical model, selected for whom, progressed by whom, and checked against what.

Remote programs deserve to be taken seriously. The 2020 AHRQ systematic review update found that several noninvasive, nonpharmacological approaches, among them exercise, multidisciplinary rehabilitation, and psychological therapies, were associated with improvements in function and pain for specific chronic pain conditions, with most effects small to moderate and some persisting for months after treatment ended. CDC guidance likewise places nonopioid and nonpharmacologic therapies at the front of chronic pain care. Remote delivery is one way to get those approaches to a patient who lives two hours from the nearest pain clinic.

The difficulty is that remote delivery also lowers the cost of building something that looks like rehabilitation and is not. A fixed framework protects against that. It lets you compare a VR embodiment program, an app-based cognitive behavioral course, and a video physical therapy service on the same terms, and it keeps the conversation on the questions a vendor would rather skip.

What are the seven areas to check?

1. Clinical model

What the program is doing to the patient, in terms a clinician would recognize. Graded exposure, pain neuroscience education, cognitive behavioral therapy, acceptance and commitment therapy, exercise progression, and VR embodiment can all be described. 'Proprietary neuro-retraining' cannot.

• Which named therapeutic approaches does the program apply, and what does a typical week look like in minutes and activities?
• Does the model treat the pain as real and the patient as capable, or does it imply the pain is imagined?
• Who designed the protocol, and is it written down somewhere I can read?

2. Patient selection

Who the program is for, and more importantly who it is not for. A remote program cannot examine a patient. It depends on exclusion criteria and a route back to in-person care. Our page on who is a candidate for virtual pain rehabilitation describes what good selection looks like from the patient's side.

• What are the inclusion and exclusion criteria, in writing?
• Does the program require a prior medical evaluation, and what does it do if a patient reports new weakness or numbness, loss of bowel or bladder control, unexplained weight loss, fever, or thoughts of self-harm?
• Who makes the final decision to enroll, and is that person licensed?

3. Evidence

What is known about this program, and separately what is known about the method it applies. Those are different questions, and vendors often answer the second when asked the first. The table in the next section sets out what each type of evidence can support.

• Has the program itself been studied, with what design, in what population, and where can I read it?
• What is the evidence for the underlying method, independent of this vendor?
• What share of patients did not improve, and what does the program do for them?

4. Human oversight

How many qualified people stand between the patient and the software. Automated content can teach; it cannot notice that a patient has stopped sleeping or that a flare has a new quality.

• Does a licensed clinician review each patient before the program starts, and is that review a conversation or a form?
• Who does the patient speak to each week, what are their qualifications, and is it the same person throughout?
• What happens when a patient reports worsening symptoms, and who has the authority to pause or stop the program?

5. Progression and safety

How the program increases load, exposure, or difficulty over time, and what it does when that goes wrong. In chronic pain, flares are expected. A program with no plan for them will either push patients into large flares or let them stall at the first one.

• What decides when a patient moves to the next level: a person, a schedule, or an algorithm?
• How are flares handled, and what counts as a reason to step back or stop?
• What are the known adverse effects, including motion discomfort if VR is involved, and how often do they occur?

6. Outcome reporting

What the program measures, how, and what it tells you afterward. The IMMPACT consensus recommends six core domains for chronic pain trials, and they are a fair standard for a clinical program too: pain intensity, physical function, emotional function, the patient's global impression of change, symptoms and adverse events, and adherence or dropout. See measuring functional progress for the instruments that fit each domain.

• Which validated instruments are used at baseline and discharge, and will I receive the scores?
• Are results reported for every enrolled patient, only for those who completed, or only for those who improved?
• What is the completion rate, and what happens to the data of patients who drop out?

7. Data and privacy

What is collected, where it goes, and who can see it. Remote programs may collect more than a clinic does, including movement data from a headset and records of coaching conversations.

• Will the vendor sign a business associate agreement, and which data flows does it cover?
• What is collected from the device, the app, and the sessions, and how long is it kept?
• Is any patient data shared with advertisers or data brokers, or used to train software, and can the patient opt out?

What evidence should a vendor be able to show?

A vendor can show several kinds of evidence, and each supports a different kind of claim. The table runs from the strongest to the weakest. Nothing on it is worthless, and nothing on it proves more than it proves.

Uncontrolled real-world data deserves a fair reading, because it is the type most programs, Karuna included, can offer first. Reported in full, it answers a real question: what happened to the patients who enrolled. It cannot separate the program's effect from the passage of time, from other care the patient was receiving, or from the tendency of people to seek help when their pain is at its worst and then improve regardless. The tells are in the reporting: honest real-world data states the sample size and who was in it, reports every measure including the ones that moved least, says what share of patients did not improve, and avoids the word 'proven'. Dishonest real-world data takes a single headline percentage from the responders and calls it a success rate.

How should the program fit with the patient's existing care?

As one component, with the treating physician still in charge of diagnosis and medication. A program that asks the patient to stop other treatment, or that advises on medication without the prescriber, has stepped outside its role.

• Diagnosis stays with the treating clinician. The program should require a prior medical evaluation and defer to it. A remote team cannot order imaging or examine a joint.
• Medication decisions stay with the prescriber. Coaching that touches on medication should route the question back, not answer it. Any tapering plan belongs to the prescriber.
• Physical therapy can run in parallel. Ask how the program coordinates dosing with an active PT plan, so the patient is not following two progressions that fight each other.
• Mental health care continues. Pain programs that use cognitive or acceptance-based methods are not a substitute for treatment of depression, PTSD, or substance use disorder, and they should screen for each and refer out.
• There is a route back. Worsening symptoms, new neurological signs, or a crisis disclosure should trigger a defined handoff to the physician or to emergency services, with the referring provider told the same day.

For a patient who lives far from in-person services, a remote program may be the only structured rehabilitation on offer, and the fit question becomes how to hold the medical home together around it. Our guide to at-home chronic pain rehabilitation covers what that looks like from the patient's side.

What should a referring provider get back?

At minimum, three written reports: a baseline, a mid-point, and a discharge summary. Each should carry validated scores rather than adjectives.

• Baseline. The enrollment decision and the clinician's reasoning, the exclusions checked, starting scores on pain, function, fear-avoidance, and mood, and the patient's own goals in their own words.
• Mid-point. Adherence (sessions completed against sessions planned), the current progression stage, flares and how they were handled, any adverse events, and interim scores.
• Discharge. Change on every baseline measure with the instrument named, the patient's global impression of change, a flare plan, recommendations for ongoing care, and a plain statement of what did not improve.

The instruments matter because they let you compare across programs and over time. The Oswestry Disability Index for low back function, the FABQ or Tampa Scale for fear of movement, the Pain Catastrophizing Scale, and a standard pain rating scale are widely used and well understood. A program that invents its own 'recovery score' and reports nothing else has made its results impossible to check.

For workers' compensation cases, ask also for functional statements that map to the job: tolerances for sitting, standing, lifting, and reaching, where the program is in a position to observe them. Our page on workers' compensation and chronic pain rehabilitation goes into what adjusters and nurse case managers need from these reports.

The checklist: fourteen yes-or-no questions

Copy this into the referral file. A 'no' is not automatically disqualifying, but every 'no' should come with a reason you can write down.

1. The program names the therapeutic approaches it uses, in standard clinical terms.
2. Written inclusion and exclusion criteria exist, and the vendor will share them.
3. A prior medical evaluation is required before enrollment.
4. A licensed clinician reviews each patient before the start.
5. A named, qualified person meets with the patient at least weekly, and it is the same person throughout.
6. There is a defined procedure for red-flag symptoms and crisis disclosures, with a same-day route to a physician or emergency care.
7. Progression decisions involve a person, not just a schedule or an algorithm.
8. Flares and adverse events are tracked, and the vendor can state how often they occur.
9. The vendor can point to published evidence for the underlying method, independent of itself.
10. The vendor's own outcome data is labeled by design (trial, cohort, or uncontrolled), with the sample described.
11. The outcome data states the share of patients who did not improve.
12. Validated instruments are used at baseline and discharge, and the scores are shared with the referring provider.
13. Baseline, mid-point, and discharge reports are provided in writing.
14. A business associate agreement is available, and the vendor states what data is collected, kept, and shared.

How does Karuna answer these questions?

Karuna runs a remote program, so it is fair to hold it to the same list. The answers below use only what Karuna publishes on how the program works.

• Clinical model. Karuna Virtual Embodiment Training (KVET)™ is a 12-week, at-home program built on pain neuroscience education, graded exposure, and pain reprocessing principles, delivered through short daily VR embodiment exercises on a provided headset. It is designed for a sensitized pain system, it treats the pain as real, and Karuna is founded on the biopsychosocial model.
• Patient selection. The program is for adults with chronic pain that has already been medically evaluated. It is not for acute injury, for unexamined pain, or for anyone with red-flag symptoms, who are sent for medical evaluation first. A free consultation and a clinician review decide enrollment, and that review can say no.
• Evidence. Karuna's own outcome data is uncontrolled real-world evidence: a cohort of 50 patients, all workers' compensation cases, with self-reported measures, no control group, and a share of patients who did not improve on any given measure. It shows what happened to that group and cannot establish cause. The full table and its caveats are on how it works. The underlying methods are covered in our guides to VR for pain management and pain reprocessing therapy. Karuna has also run institutional review board approved studies on the tool itself, including validation of range-of-motion measurement in VR.
• Human oversight. A clinician reviews each patient before the start and provides oversight throughout. A dedicated pain coach meets the patient one-on-one every week for the full 12 weeks, the same coach throughout, and adjusts the plan.
• Progression and safety. Exposure expands in graded steps, in the headset first and then in real-world movement, with the coach adjusting the plan weekly and the final weeks spent on a flare plan and on continuing after discharge.
• Outcome reporting. The measures Karuna tracks are fear-avoidance (FABQ-PA), pain, activity level, confidence, and sleep, taken before and after the program.
• Data, privacy, and reports to referrers. These are questions to put to the care team directly at the free consultation, along with how reports reach the referring provider. Karuna takes workers' compensation referrals, and the care team confirms coverage case by case.

What this means for you

If you are a prescriber or rehabilitation clinician

Run the seven areas before the demo, and ask for the written exclusion criteria and a sample discharge report first. If the vendor cannot produce both, the rest of the conversation is marketing. Keep diagnosis and medication with you, and tell the patient plainly that the program is an addition to their care, not a replacement for it.

If you are a nurse case manager or adjuster

The checklist doubles as a utilization review file. Ask for outcomes in the IMMPACT domains and in functional terms that map to the job, and ask for the non-response rate up front so a later plateau is not a surprise. Coverage is a separate question from quality. A program can pass every item here and still need case-by-case confirmation.

If you are a patient or family member

You can ask the same questions. The ones that matter most: who will I talk to each week, what happens if I get worse, and what will my doctor receive at the end. Your pain is real, and a good program will say so and show you the measures it uses to track your progress. Keep your own physician in charge of diagnosis and medication throughout, and if you develop new weakness or numbness, loss of bowel or bladder control, fever, unexplained weight loss, chest pain, or thoughts of harming yourself, seek medical care first.