From Patterns to Prescriptions: A Smarter Way Forward for Long COVID Care

May 6, 2025

The Problem With Clusters

From Labels to Action

Our 5 Recovery-Driven Subtypes

Long COVID research has come a long way—but clinical care hasn’t caught up.

And while studies like the Nature Medicine machine-learning analysis (Zhang et al., 2023) identified four major subtypes using 30,000 electronic records, we’re still missing a critical step:

Turning patterns into prescriptions.

At AHA, we’ve been working on a recovery-informed framework that doesn’t just label the condition—but guides people out of it.

In the research, long COVID was grouped into four phenotypes:

Cardiac & Renal

Respiratory, Sleep & Anxiety

Musculoskeletal & Nervous System

Digestive & Respiratory

This validated what many of us already knew—long COVID isn’t a single illness. But in clinical practice, these clusters don’t give us direction.

That’s why we built something different.

We used anonymised case data across hundreds of recovery journeys and asked not just what symptoms were present—but:

What was driving dysfunction?

How did the body respond to pacing, movement, or stress?

What actually helped?

The result? A recovery-informed model built on real-world outcomes, not just co-occurring symptoms.

Anosmia/Dysgeusia-Dominant

Respiratory-Dominant

Neurological-Dominant

Cardiovascular/GI-Dominant

Multisystemic & Complex

Each of these subtypes responds differently to medications, supplements, breathwork, pacing strategies, and rehab programming.

And each demands more than “just rest.”