Welcome back to Healthy Innovations! 👋

The BBC published two articles in late March about people living with long Covid five years on. Darren Parkinson, a former NHS nurse, can't shower more than once a week. Gabby Rudge-Cox says she's existing, not living.

Neither has a diagnosis a blood test can confirm. Neither has a treatment.

This week's Deep Dive looks at new research that may finally be giving science something to work with.

Let’s dive in!

When Darren Parkinson contracted Covid in 2021, it felt like a bad flu – a few days in bed, then back on his feet. Five years later, he cannot work, socialise, or shower more than once a week. Cooking a meal takes three days: chopping one ingredient, resting, then the next. The last time he left the house was for a dental appointment.

Before long Covid, Parkinson was an NHS learning disability nurse and a local councillor in Bradford. Now he spends most of his time lying on his bed or sofa.

"I never ever feel well," he told BBC News in March 2026. "It's like gravity working overtime – that draining never-ending fatigue."

Gabby Rudge-Cox, 42, from near York, caught Covid on New Year's Day 2021 and has since been diagnosed with fibromyalgia, ME, and postural tachycardia syndrome. Having a shower or getting dressed can be the main event of her day.

"Obviously, the world came out of lockdowns," she said, "but for me it feels like I'm still in a lockdown. I'm existing and not living."

Darren and Gabby are two faces of a condition that medicine still cannot reliably diagnose or treat. A US study published in Communications Medicine (Nature Portfolio) in March 2026 – one of the first population-level analyses of long Covid’s disability burden – found its impact on daily functioning rivals that of Alzheimer's disease and asthma. Half of US adults with disabling long Covid are under 50. Around 64% are women.

Five years in, no biomarker has been validated for routine clinical diagnosis and there is no approved disease-specific treatment. That gap is where some of the most consequential research in medicine is now underway.

In October 2025, researchers from Montpellier University in France and Stellenbosch University in South Africa published a study in the Journal of Medical Virology with one of the clearest biological pictures of long Covid yet.

They analyzed blood samples from 50 long Covid patients and 38 healthy volunteers across cohorts in France and South Africa.

This is what they found:

The structural link between microclots and NETs had never been reported before. In long Covid patients the interaction was far more pronounced – and the researchers believe it may make microclots resistant to the body's normal breakdown process.

The result: dense, stubborn structures that may be blocking capillaries and limiting oxygen delivery to tissues – which could help explain the fatigue, brain fog, and wide-ranging symptoms that look different from person to person.

When the research team anonymized the samples and ran them through an AI model, it identified long Covid patients with 91% accuracy – promising, but requiring replication in larger, more diverse cohorts before it can be considered clinically meaningful.

Important caveat: The study did not include disease-matched controls such as people with diabetes or autoimmune conditions – more validation is needed before microclots and NETs can function as clinical diagnostic tools.

The consequences of having no diagnostic test go well beyond the clinic.

The CDC is explicit: no laboratory test can currently diagnose long Covid. Standard blood panels often return normal results even when patients have clear physiological dysfunction – producing years of disbelief, delayed referrals, and missed diagnoses. A 2025 review in Frontiers in Medicine found clinical awareness among physicians remains surprisingly low.

"Often we are told, 'Everything's testing normal. We tested all the standard panels, everything's come back normal.' The blood panels we test for only capture a very narrow scope of our physiology," said Dr. David Putrino, director of rehabilitation innovation at Mount Sinai Health System, who has collaborated with Pretorius on microclot research.

Without a validated biomarker, clinical trials cannot classify patients into biologically distinct subtypes – and continue testing interventions against a group whose underlying biology may differ substantially.

The picture on treatments is mixed – some signals of progress, a series of setbacks – and the absence of a diagnostic test runs through both.

Antivirals – tested on the hypothesis that persistent viral replication drives long Covid:

New targets: Harvard Medical School researchers published findings in late 2025 suggesting long Covid may be driven by persistent chronic inflammatory pathways – distinct from viral persistence. Ongoing trials are testing immunomodulators, JAK inhibitors, and B-cell depletion therapies against these pathways.

Trial access: Scripps Research launched the LoCITT trial in October 2025 – a fully remote study enrolling a thousand participants to test repurposed drugs. Many long Covid patients are housebound; standard trial recruitment excludes them entirely.

Gap to watch – funding equity. The March 2026 Communications Medicine analysis found that conditions predominantly affecting women receive 5.2 times less NIH research funding per year lived with disability than those predominantly affecting men. Long Covid received just 14% of the funding its disability burden would justify. ME/CFS, which frequently co-occurs with long Covid, received less than 1% of its proportionate share.

In spring 2025, the US government compounded this by cutting long Covid research grants and closing the dedicated federal office coordinating the national response.

The microclot and NETs research points toward a potentially tractable diagnostic approach – a test that quantifies these structures in blood and uses AI to classify results. Prof. Pretorius and collaborators are working toward a clinical-grade version, though timelines depend on validation studies with broader, disease-matched populations.

The path is complicated by how much biology remains unexplained. Researchers have identified multiple plausible drivers of long Covid:

These mechanisms may operate differently in different patients – which is why Dr. Iwasaki and others argue that long Covid is likely several diseases under one name, each requiring interventions matched to its underlying biology.

A mass disabling event without a diagnostic test, without approved treatments, and with research funding under pressure on both sides of the Atlantic. Five years in, the patient community defined this condition before medicine did – in support groups, on social media, by insisting something was wrong when every test said otherwise.

The science is finally catching up. But for the patients still searching for answers – for Darren, for Gabby, and for tens of millions like them – the pace of progress is not an academic question.

Most cancer detection waits for tumours to give themselves away. Earli flips that logic entirely. Its Synthetic Biopsy platform uses programmable genetic constructs to enter cancer cell nuclei and force them to produce a synthetic, non-human biomarker — detectable via blood test, PET imaging, or other readouts. Rather than hunting for naturally occurring markers that early-stage tumours may not yet produce, Earli creates the signal itself.

Founded in 2018 by Stanford cancer researcher Dr. Sam Gambhir, CEO Cyriac Roeding, and CSO Dr. David Suhy, the company carries a deeply personal origin story. Gambhir lost his son Milan to glioblastoma at 16, and his wife Aruna to cancer in 2023. Gambhir himself was diagnosed in 2020 and died before seeing his research reach the clinic. His co-founders are continuing the work.

The platform has shown preclinical signal detection in lung, breast, liver, colon, and ovarian cancer models, with regulatory conversations with the FDA underway. The scientific advisory board includes Nobel Laureates Jim Allison and Lee Hartwell, alongside MIT's Bob Langer and Sangeeta Bhatia.

Earli was cited in the White House Cancer Moonshot Fact Sheet as a key initiative for future cancer treatment technologies.

🧢 The beanie that reads your mind. California startup Sabi has emerged from stealth with a brain-reading beanie that decodes your internal speech into text on screen — no typing, no talking. The hat packs between 70,000 and 100,000 miniature EEG sensors (most clinical devices have a few hundred), feeding signals into a brain foundation model trained on 100,000 hours of neural data. Target speed: around 30 words per minute to start, improving with use. A baseball cap version is also in development.

Investor Vinod Khosla is backing it, calling thought-to-text "the biggest and baddest application of BCI [brain computer interface]." The company says neural data is encrypted end to end. The cyborg future, it turns out, will be machine washable.

Thank you for reading the Healthy Innovations newsletter!

Keep an eye out for next week’s issue, where I will highlight the healthcare innovations you need to know about.

Have a great week!

Alison ✨

P.S. If you enjoyed reading the Healthy Innovations newsletter, please subscribe so I know the content is valuable to you!