Patient Engagement Isn't a Metric. It's the Treatment.

I built a mental health app once. Beautiful thing, honestly. Clean interface, evidence-based modules, a mood tracker my clinical brain was genuinely proud of. We shipped it to people living with bipolar disorder and then we watched 15% of them use it in any given week.

Fifteen percent.

I'm a clinical psychologist. I knew the adherence literature cold. I knew that a treatment nobody opens is not a treatment, and I still built a thing whose entire design depended on a person in a depressive episode remembering to tap an icon. Then we tried something that felt almost too dumb to say out loud in a product meeting. We called them. With AI. Weekly engagement went to 85%.

That's the origin story of HANA. It's also the only thing I'd defend in a knife fight: engagement isn't a number you report at the bottom of the quarterly deck. It's the intervention.

Why do most patient engagement programs fail?

They fail because they put the first move on the patient. The portal, the app, the printed discharge packet, the "call us if anything changes" at the end of a visit. Every one of those designs hands the hardest job in the system to the person least equipped to do it that week.

There's a quality improvement study out of Cardiac Solutions, published in JACC Advances, that makes this painfully concrete. Heart failure patients enrolled in an automated daily outreach program readmitted at 15.5%, against a national historical baseline of 24%. Good result. Publishable result.

But the real finding is buried inside the finding.

What actually separated the patients who came back from the ones who didn't?

Engagement. Nothing else. In that same digital outreach study, patients who opened more than half their messages readmitted at 7.7%. Patients who opened less than half readmitted at 21.6%, which is basically the national average with extra steps.

Same program. Same content. Same clinical team. Same disease. The only variable that moved was whether the patient actually showed up in the conversation.

So when a vendor tells you their content is clinically superior, ask them what percentage of patients read it. That's the whole product.

Does the channel matter more than the content?

Yes, and I learned this from a circus, not a clinic.

I spent a stretch of my twenties crossing Australia with a travelling circus (feels like another lifetime, honestly). The aerial silk act was technically astonishing. Years of training, real artistry, genuinely the hardest thing anyone in that tent could do. It drew polite applause. The guy spinning fire chains in the dirt out front drew a crowd three deep.

Not because fire chains are better. Because fire chains are reachable. People stopped, watched, understood, stayed.

Clinical content works the same way. A voice call in someone's own language at 7pm, asking about their swelling and their meds, outperforms the most rigorously designed patient education portal ever built. Not because it's smarter. Because someone picks up. That's most of what our patient outreach use cases come down to in practice.

How fast does a clinic actually need to reach out?

Fast, and the window is narrower than most teams think. A Kaiser Permanente analysis of nearly 12,000 heart failure patients found follow-up contact within 7 days of discharge was associated with roughly 19% lower odds of 30-day readmission. Contact after day 7 showed no significant benefit. None.

That's a brutal sentence if you run a care management team. It means the call your coordinator makes on day nine, the one she stayed late to get through, statistically did nothing.

It's not a staffing problem you can hire your way out of. A clinic discharging fifty patients a day cannot reliably reach all of them inside 48 hours, confirm the follow-up, screen for red flags, and document it back. Something slips. The patients who slip are the sick ones.

What does closing that gap actually return?

Across our deployments, HANA runs at 85% weekly engagement against an industry baseline of 15 to 20%, with over a million patient interactions and zero critical adverse events. The clinic-side ROI works out around 31:1, which sounds like a made-up marketing number until you price a single avoided readmission at the AHRQ average of $15,200 and do the arithmetic yourself.

But I'd rather you ignore my numbers and look at the Cardiac Solutions gap again. 7.7% versus 21.6%. That's the size of the prize sitting inside a program most systems already own. They just can't get anyone to answer.

Key Takeaways

Engagement is not a reporting metric, it's the active ingredient, and the heart failure data makes that hard to argue with. The same program produced a 7.7% readmission rate in engaged patients and a 21.6% rate in disengaged ones, which means the delivery mechanism is doing more work than the clinical content. Reach matters more than sophistication, speed matters more than thoroughness, and the seven-day window is real. If your program can't reach people inside it, you're running an expensive documentation exercise. Build for the patient who won't call you. That's the one who comes back.

FAQ

Does AI voice outreach replace care coordinators?

No, and any vendor who says otherwise is selling you something. It handles the volume layer, the calls that don't need clinical judgment, and it surfaces the ones that do. Your coordinators stop dialling and start intervening.

Is automated patient outreach safe for clinical conversations?

Within clear limits, yes. HANA has run over a million patient interactions with zero critical adverse events, with escalation to a human clinician on any red flag. The agent doesn't diagnose and doesn't prescribe. It listens, documents, and hands off.

How long does it take a clinic to go live?

Days, not quarters, if your systems are reachable. HANA is open-source and self-hosted with no OpenAI dependency, so your data stays where your compliance team wants it. The integration docs cover the technical setup.

If you're running a program that looks great on paper and can't get anyone on the phone, book a call with me and we'll look at your actual reach numbers together.