Is healthtech triggering a rise in self-diagnosis?
Join Our Newsletter - Get important industry news and analysis sent to your inbox – sign up to our e-Newsletter here
X

Is home health tech fuelling self-diagnosis?

By Chloe Kent 29 Jun 2021

From AI symptom-checkers to diagnostic devices, more and more patients are taking their health into their own hands.

Is home health tech fuelling self-diagnosis?
Not everybody believes that increased at-home understanding of health always a positive thing. Credit: Shutterstock

Artificial intelligence (AI) diagnostic apps have soared in popularity in recent years. These platforms often come with prominent disclaimers that they are there to provide health information only, rather than a final diagnosis, but have increasing appeal among those wishing to have a better understanding of their health before deciding to speak to a clinician.

The prevalence of home-use health tech is also growing, with many consumer smart products like the Apple watch now able to collect comprehensive data about the wearer’s health. Numerous home monitoring solutions are emerging for clinical use too.

But not everybody believes that this increased at-home understanding of health always a positive thing. The developers of such tools consistently assert that they are not to be used for outright self-diagnosis, but in learning more about their health without the supervision of a clinician there is always a chance that a patient will take the management of their condition into their own hands.

With this newfound freedom to monitor and track one’s own health and wellbeing comes the risk that patients may forgo seeing a trained medical professional in favour of self-diagnosis.

Home health monitoring

Researchers at the University of York recently received a £1m grant to develop a device to enable patients to self-assess their immune systems at home.

The portable instrument will allow patients to examine their own immune systems using a single drop of blood and is initially intended for use by people with rheumatoid arthritis (RA). Currently, RA patients have to undergo monthly blood tests in hospital to monitor their condition, which can be both stressful for patients and costly for the healthcare system.

The device is based on photonic biosensor technology, which will be used to detect infection-related protein biomarkers.

University of York professor Thomas Krauss says: “We’re aiming for the same thing as a glucose sensor. Just take a pinprick of blood, and it goes onto a sensor and gives you a readout.”

Krauss says that, if a single drop of blood isn’t sufficient for the sensor, then the team will look into combining the blood with a diluting solution, similar to the use-process of a lateral flow test for Covid-19, which many people now administer at home.

The device also supports the NHS long-term strategy of moving to patient-reported methods with less reliance on hospitals and clinics. While patient-reporting may lead to some sacrifices in accuracy due to test administration errors, Krauss says the benefits far outweigh any risks.

When asked how the device developers can ensure that patient-reported readings from this device are as reliable as those recorded in hospitals and clinics, Krauss says: “We can’t. But you make up for it with quantity. If a patient reports once a week, rather than once a month in-clinic you get much more data. I would argue that from the quantity of data you get better quality. You can see trends better, so you can afford a higher error rate.”

Self-diagnosis: a thorny issue

But what of non-clinical solutions? Devices like the one being developed at the University of York are designed to be used in tandem with the guidance of a medical team, who will use the self-reported data to inform patient care. When it comes to symptom-checking AIs and biomarker data from smartwatches, the issue of self-diagnosis becomes far more complex.

A study recently published in the BMJ found that Covid-19 symptom checkers in both the US and UK failed to identify the symptoms of severe Covid-19, bacterial pneumonia and sepsis, advising these cases to stay at home and not seek urgent treatment. The researchers concluded that patient-led assessment tools had the potential to worsen outcomes by delaying appropriate clinical assessment.

Another study, published in BMJ Open, found significant variance in coverage, accuracy and safety of the eight most popular online symptom assessment apps, raising concerns about how appropriate their everyday use is.

The developers of medical AI firm Infermedica, which recently added paediatric care to its suite of triaging apps, have a different take. Instead of providing patients with a suite of potential diagnoses, like popular symptom-checker Ada, Infermedica aims to triage them and tell them whether they need to seek medical care or can administer self-care at home.

Infermedica founder and CEO Piotr Orzechowski says: “I wouldn’t refer to [Infermedica] as self-diagnosis, we don’t use such terms, but self-care is important if used appropriately. We never replace medical-grade diagnosis, we’re an educational tool to guide you to the right kind of doctor.”

Orzechowski says that, when appropriately administered, the self-care AI apps may prompt patients into can relieve the burden placed on overstretched healthcare systems. They could also be useful in resource-limited settings where medical care may not be readily available.

Wearables still subject to scepticism

According to GlobalData, the wearable tech market is likely to be worth $64bn by 2024, but many medical professionals remain sceptical about the practical potential of the technology. While these lifestyle-oriented products have the capacity to record a wide range of user data, the usefulness of this information beyond satisfying personal curiosity is still very much up for debate.

The Covid-19 pandemic has expedited the wearables trend, with a recent survey carried out by Stanford Medicine finding that wearable usage increased by 33% between 2019 and 2020. But studies have delivered differing results on how well wearables can compare to clinical devices – and whether, as Krauss says, the sheer volume of data available can make up for any disparities.

Orzechowski says: “I think we’re getting to a point where AI algorithms can draw data from so many sources, not just what you type on your keyboard but also taking data from your smartwatch or from medical devices.

“If you combine all of it together, you still won’t replace a GP, but you can safely take some self-care measures if you can treat yourself at home. I think that’s a part of our developing health education, which is very important.”