The bioengineer Takuo Aoyagi died in Tokyo in April, around the same time that people began clamouring to panic buy his 1970s invention, the pulse oximeter.

Blood oxygen saturation, which pulse oximeters measure by transmitting red and infrared light into, say, your middle finger and comparing how much of each is absorbed by the blood, is considered one of five valuable vital signs, alongside temperature, heart rate, blood pressure and respiratory rate. As the NHS used clip-on pulse ox devices as part of its at-home coronavirus screening and GPs began loaning units to patients with shortness of breath, two things happened.

The £40 pulse ox devices from China sold out on Amazon and the consumer tech world remembered that SpO2 (peripheral capillary oxygen saturation) sensors were part of almost every hardware company's roadmap. The only problem? Neither your phones nor your Fitbit can diagnose you.

In other words, don't rely on an iphoness app or a Samsung Galaxy S10 to tell you what percentage of oxygen-carrying haemoglobin is in your blood. And think twice about headlines such as 'The Apple Watch 6 could combat coronavirus with this new feature' or 'The next Apple Watch could be a Covid-19 early-warning system'.

The SpO2 sensors in Fitbit fitness trackers, Garmin watches and the new Huawei Watch GT2 are interesting for their potential to enhance features like high-altitude training and sleep tracking. And 2020 will likely be a big year for pulse ox measurement in mainstream wearables as references to SpO2 have actually been spotted in snippets of code for ioses 14, suggesting we will see it on an Apple Watch Series 6 in September. So if a £40 clip-on can do the job, why can't a £400 smartwatch?

"Typically it's because these sensors are not precise, that's the main limitation, even the heart-rate monitors are not very precise. So the ones that you wear are only for the consumer level, not for the clinical level," says João Paulo Cunha, associate professor, department of electrical and computer engineering, University of Porto. "Pulse oximetry is another level of detail and precision that you need so that you have a usable value. Some of the companies, for example Fitbit knows it would be lying to customers if it puts out the values and says 'use these values' so that's why they're not doing it."

So just as a standard PPG (photoplethysmography) heart-rate sensor on a smartwatch is not medical-grade accuracy, the sensors on wearables aren't designed or tested for medical use. It's also the case that medical pulse oximeters tend to be used on parts of the body, such as the finger, earlobe and toes, in which the LED light passes through to a photodetector on the other side, not the place where most wearables sit: the wrist. It's unlikely that even Apple will be able to solve the limitations to the extent that regular people would be able to spot check for coronavirus with an Apple Watch.

Pulse ox readings from both wearables and finger pulse ox devices are less accurate when the wearer is moving making them much less useful for continuous monitoring of healthy, mobiles users. This has been shown in a number of studies including a 2019 observational study of COPD patients which saw a huge 25 per cent drop in the percentage of valid data with moderate-to-vigorous activity. One theory as to why is that motion causes movement of venous blood and other things like tissue fluid, along with arterial blood.

Rich Robinson, Garmin's UK head of product, says the SpO2 sensors on its Fenix adventure watches are precise enough to help users to train and make decisions at high altitude in mountainous regions, where oxygen saturation can drop, but they are categorically not designed for medical use: "To a certain extent, it is accurate but there's a whole host of variables, whether you've got the device fitted correctly or are you moving while you're trying to take the reading?" (Another block to health-related SpO2 monitoring on smartwatch and fitness trackers is that it's another sensor to drain the battery.)

The movement issue is likely one of the reasons why health and fitness companies, such as Fitbit in particular, have focused on sleep tracking use cases, specifically the potential to monitor for signs of sleep apnea, a disorder where your breathing stops and starts while you sleep and the most likely focus for Apple. Fitbit, which is in the process of being acquired by Google, took years to 'turn on' the SpO2 sensors for its fitness devices. A Fitbit rep says that the Estimated Oxygen Variation chart seen in its app is "not intended to track slow fluctuations in relative SpO2 or sustained hypoxia as might occur with acute or chronic respiratory problems."

Instead it was designed to track shorter term fluctuations in blood oxygen level over time scales of 30 to 60 seconds, as is seen in breathing disturbances during sleep. Garmin has seen the SpO2 sensor trickle down to its Vivoactive range, where it's one of the metrics that can laid over the data for a bad night's sleep in its Garmin Connect platform. "During sleep, when you are not moving, you can get very good quality data," says Cunha. "This is where I think these wearables really have a chance."

There have been some attempts to counter the movement factor, including promising algorithms which use the device's accelerometer to control for the confidence of the measurement. Researchers at Loughborough University have also experimented with using green and orange light, instead of red, from a multi-wavelength optoelectronic patch sensor to reduce motion artefacts. The results from 31 participants who were instructed to either move their hands freely or take the oxygen saturation readings while walking and cycling, suggests that this could make the readings more robust "without restricting physical movement".

It's unlikely to be enough to close the gap, though. Is Apple likely to make a breakthrough with the next Watch in 2020? "No, I don't think so. This type of noise will always be there," says Cunha. "If they do it, they will do it slowly, with controlled studies. Even with ECG (electrocardiogram) they are very careful to stick to atrial fibrillation, which they are certified for, because you're providing health information so people may use it in the wrong way."

So far Huawei has been the most inclined to put out information on how the SpO2 sensor on its new Huawei Watch GT2e smartwatch, launched in April, can be used to spot test for acute or chronic hypoxemia – low levels of oxygen in the blood. That might be because US companies such as Fitbit are more wary of litigious users than Chinese tech brands.

That said, even though Huawei states the smartwatch can "help you perform health management actively and comprehensively" and outlines that groups such as the elderly and people living at high altitude might want to check their estimated SpO2 levels regularly, it stops short of serving users with health advice. Its guidance says that if you get a reading below normal levels – which Huawei deems to be 90 to 100 per cent, lower than the standard medical advice in the UK – "this condition requires immediate treatment from the doctor".

The question of who should be looking at and interpreting this data – you or a medical professional – is both the final reason we're not all tracking our blood-oxygen levels with personal devices at home right now and a cause for hope. Expect to hear a lot more about SpO2 and blood-oxygen levels from Silicon Valley and beyond this year. In San Francisco, sleep-tracking ring company Oura is sponsoring research at the University of California, to see if its hardware can detect Covid-19 symptoms via SpO2 and temperature monitoring, with 30,000 users signed up so far. Just don't expect your phones or your fitness tracker to medically diagnose you anytime soon, at least not without a doctor monitoring it.

Everything from shellac nails to poor circulation and low blood pressure can affect the readings of finger-based pulse ox monitors and we know even less about wrist-based sensors and inaccurate smartphones camera readings. It's also likely that by the time your SpO2 levels have dropped below say, 95 or even 92 per cent, you will already be experiencing the symptoms – such as not being able to say a whole sentence without taking breaths or having shortness of breath for over 24 hours – that indicate you or someone with you should call 111 immediately for a phones consultation. There have been anecdotal stories of cheap pulse oximeters acting as an early warning to genuine cases so a consumer-grade spot check reading in conjunction with 111's symptom checker may have some use. But there's also a chance that inaccurate results could overwhelm the system.

"It's overkill because if you don't know how to use it and don't know how to interpret it, it will raise more calls to the NHS and it will be even more difficult to manage," says Cunha. "But if it's prescribed by your doctor, when you are a suspected case or having your Covid tests, I think that would be very useful."

While Garmin execs cannot stress enough that their devices are not medical devices, it has partnered with digital health company Aparito on a trial in North Wales to monitor at-risk cancer patients during the coronavirus outbreak. With funding from NHSX and SMART Cymru, Elin Haf Davies, Aparito CEO and a former paediatric nurse, says her team has signed up 65 patients discharged from the Betsi Cadwaladr Health Board for at-home monitoring so far. Over 40 patients have been set up in the first nine days and a long term target of 1,200 has been set for the coming months.

People in the scheme get alerts twice a day to fill out the NHS 111 Covid-19 symptom checker on their phones and carry out a spot check for vital signs, including heart rate and blood oxygen saturation, via Vivoactive 3 and 4 devices with the capability to add further data from medical grade devices such as thermometers and blood pressure cuffs.

"The doctors know that we're using Garmin as a consumer-grade device, but we're still getting good trends of data of how things are changing over time," says Davies. Crucially, Aparito has access to the raw data from the Garmin devices, unlike regular users. Saturation levels from 97 to 100 per cent don't get captured, from 95 to 97, they're asked to check again in a few hours with follow up questions on breathlessness and between 92 and 95, they are prompted to call the hospital.

Davies is also planning two new 'deployments' for post-Covid and other groups of immunocompromised patients: "We're anticipating a lot of patients that have been in hospital with Covid will have quite long-term lung damage and quite long-term rehabilitation." She says the general population doesn't need this kind of monitoring and is "definitely not for every single person".

There are similar at-home monitoring projects in very early stages across Europe, including Porto, Portugal. "I think having an infrastructure that can easily upload x times a day temperature and oxygen saturation would be a very, very good tool, even using non-clinical grade measurements as long as there is some medical guidance," says Cunha. "If we have a next wave, I hope that this type of infrastructure is already there."

😓 How did coronavirus start and what happens next?

❓ The UK's job retention furlough scheme, explained

💲 Can Universal Basic Income help fight coronavirus?

🎲 Best video and board games for self-isolating couples

👉 Follow WIRED on Twitter, Instagram, Facebook and LinkedIn

This article was originally published by WIRED UK