I’m sure you’ve heard of the recent tragic death of Jason, a Team Valhalla rider in our community during a race the other evening. There has been a lot of discussion recently around emergency contact details and how to flag help should a similar incident happen again. Each one of the cycling clubs seems to be taking their own approach on how this can be done. I was just wondering if Zwift have any plans on adding something to the software that could trigger an emergency call if somebody got any trouble while riding.
I think a singular approach would work better than each club doing their own thing. Just wondering your thoughts on this issue and I’m happy to help with ideas if you like.
We have looked at a solution for our small race team within our club.
The tragic events earlier this week have lead a lot of the racing community to try and figure out a good solution to needing to contact an ICE should something happen.
It would be great to see Zwift implement something in the game itself, I’m sure it’s a GDPR minefield but something to look at.
What are some examples things Zwift can do here? Warnings on extreme HR spikes or risky HR patterns, some emergency contact info in your profile, some kind of “you ok” dialog if you drop out randomly without touching anything?
What is really hard about a lot of this is it’s hard to tell the difference between equipment failure (like the tickr just stopped working), someone rage quitting, etc. vs an actual emergency, but hopefully there’s something that can be useful for Zwift to make it safer etc.
Totally agree in that I’m not sure this is something that can be detected and triggered from server side necessarily (as you say, tech failures are way too common) but maybe something that can be initiated from the user side if things go south and they are in need of help?
One of the biggest issues is that, unlike a lot of sports/strenuous exercise, there are often no other eyes on the individual at the time, so no way of knowing whether it is a technical or a medical issue. Having some way to alert others if you are isolated so that there is a chance help can be summoned in time is, I believe, the ideal end goal here.
How that’s achieved I don’t know, but then this is what we want to explore.
It should be relatively straightforward to have a personal “I don’t feel well” button implemented in the Companion app, as well as a Facebook-esque “I’m concerned about this person” button. Both could then lead to Zwift attempting to contact the user directly, and if not able to potentially dispatching emergency services, but relies on having a customer service person 24/7 available to make the call. Allowing users to optionally put NOK contact details into the app would also be easy enough to do, given we already give Zwift a load of data under GDPR.
Automated cardiac risk screening is something I’ve been trying to engage with Zwift about for ages. It would be a highly non-trivial task (lots of literature about the issues of using HRMs for such a purpose) and fraught with potential liability, but I think would be very beneficial for research into user health from a population basis if nothing else.
Probably not some sort of safeguard as that’s too difficult due to the examples you give.
Maybe somewhere central to store emergency contact details that you can opt to share with other zwifters like how you can follow etc. So you can share with a very select group who you are comfortable with having ICE details.
Then if you’re online with them and they think something is wrong then they have easy access to a number to call?
Many smart watches do have features for HR spikes and emergency contact calling.
I highly recommend setting this up.
I personally don’t think Zwift needs to nor should undertake what indeed could be a hideous red tape battle of personal information; whether anyone is willing to let their contacts information go public or not. I can only imagine it would be a privacy hay-day for someone wanting to poke and prod about additional personal information; no matter the amount of warning labels slapped on.
Regardless; those who are using or have smartwatches, DO enable these features. They’re designed to function in case of emergency 24/7; and that’s what matters.
I guess there are two aspects, one is preventative, and the other is reactive.
In terms of preventative (which I think is ideal, because re-active means someone is already having an incident), there are things Zwift can probably do.
Preventative:
Zwift could partner with some fancy group that can provide logic to identify abnormal HR patterns that might be a symptom of a bigger “cardiac risk” issue, then Zwift can piggy-back off that, and use your HR data to scan for issues you might need to watch out for. What they could do is send an email to you after a ride saying "Hey, we noticed “x” which could be a symptom of “y”, if you have no reason to believe your HR monitor was mis-reading it would be good for you to go check with your doctor on this because here are the risks … ".
In this way nobody else is getting this email, it’s up to you to act on it to minimize your risk in the future.
Then there is re-active:
Having a spot for emergency contact info on your profile might help, but my guess is most cases folks won’t be paying attention to people dropping out, so it’s hard to guess how often that would be used - not to say it’s a bad idea or anything. I understand the idea of an “I’m not well” button in the companion app, but in order for Zwift to make that button more effective in an emergency than simply dialing 911 the UX would have to be so prominent that I imagine a lot of people would complain. So there’s issues there. It sounds like other watches etc, have provided some similar functionality, so I suppose a scan to check for best practices there would be helpful.
@nickopotamus great suggestions I like the idea of the ‘help button’ in the companion app. Guessing that 911, 999 or any emergency dial would be more difficult. My thoughts where for the rider to add their emergency contact (relative or friend) for the ‘help button’ to message - that way the emergency contact can physically visit or call and assess weather 911, 999 need to be called etc. Just a thought feature-request
Some of this falls into an area I’ve much experience in.
In terms of generating a response from emergency services in the UK it won’t happen. There is no way a call taker would schedule an ambulance attendance from a Zwift representative who THINKS there MIGHT be a possibility of somebody having a medical emergency.
And let’s be honest 99.9% of the time it’ll turn out to be a false alarm causing the end user embarrassment & wasted time for the emergency response.
We currently have significant waiting time for ambulances for actual confirmed emergencies. For example it’s not uncommon for somebody having a stroke to wait for over 3 hours.
The problem with preventative measures is two fold.
Firstly, it’s really not a simple problem to solve. HRM based arrhythmia detection is not easy (there’s a good body of literature out there comparing HRM to ECG and it’s mostly false negatives), and not all events are arrhythmia related.
There may potentially be something predictive of cardiac risk in terms of power/HR decoupling, time spent at threshold, cumulative work, etc but there isn’t a decent enough dataset out there as yet to do this work in this sort of population using real world (i.e. imperfect) data. Myself and a cardiology colleague approached Zwift nearly a year ago to question if their dataset - probably the largest and richest in the world - could be used for such research but haven’t heard anything for a while. Zwift data science team, if you read this, please get in touch!
Secondly, there’s the liability issue. Zwift’s EULA places the burden of risk on the user. Tragedies like Sausage’s sadly do happen, as they do road cyclists, runners, and people using the gym. If Zwift monitors people and suggests they seek help, who is liable if they suffer an ill effect of over investigation? Or if Zwift miss an event?
It’s not insurmountable, and I’d love to see users being warned if their cardiac health is considered to be potentially at risk, but until we understand the scale of the problem, have the data to be able to make these predictions, and have a robust system for warning people, reactive measures are probably the way forward (as well as a lot quicker and easier to implement).
These features in things like the Apple Watch have required significant investment in regulatory approval, taking years and having to be repeated in different countries.
I’m not 100% sure what is meant by suffering an ill effect over investigation. In what situations is seeing a doctor in advance of a cardio issue a bad thing? Isn’t the advice almost always to see a doctor ahead of doing any new exercise regiment to begin with?
I guess my thought is if the logic just lets people know they are in a risk category because of ‘x’ or ‘y’ data from their HR, and suggests talking to a medical professional about it if you happen to think your HR was not misreading, then it’s not like they are saying “yes you have this issue”, more that “we noticed something in the data that could be an issue worth discussing with your doctor, perhaps tread lightly in terms of intense sessions”.
So it’s not like there is an expectation that Zwift catches every issue, this is just in the case where the data shows there’s elevated risk.
Every test and investigation carries risks. Those vary from the trivial (infection from a blood test) to the serious (MI during an angiogram). Over-investigation is a real risk that needs to be taken into account when any screening program is introduced, which this in effect would be (for example, see potential harms from PSA based screening for prostate cancer).
HRMs alone are not a reliable source of information regarding arrhythmia (eg PMID 28543790, sorry I can’t include links). Zwift could flag every abnormally high HR to the user for example, and maybe refer them to the algorithm in the above paper, but we have no evidence (yet) that would have any impact on cardiac events at all, let alone events such as befell Sausage.
There’s research to be done in this field, with the potential for real benefit, for sure. I can’t believe that there won’t be some gold hidden in Zwift’s data that would be able to go some way to achieving what you’re thinking of. But high HR alerts almost certainly aren’t the way forward.
My Garmin chest strap would do that nearly every ride if it wasn’t sufficiently moist.
I have an expensive BP monitor owing to a previous issue which i now have to closely monitor.
Every now and again it will warn me that it’s detected Arrhythmia.
However having had consultant led cardiac tests i know this is not the case.
I’m afraid relying on data from relatively inexpensive equipment is not the way forward. I can’t think of any other organisation that does this.
We don’t know the circumstances of the sad loss of Jason so it’s impossible to say whether there would have been any advanced warning for him.
Great idea Rick, we’ve been discussing this as a team before and again recently. I think it’s pretty straight forward as far as implementing a button on screen or the CA that’s simply a notification we don’t need to over complicate it with triggers based on HR or pedalling stopping etc as long as it’s customisable so we can add whatever note to it and then select how it’s sent which is the part that’s going to require a bit of thought.
Have a simple warning in case of accidental pressing that says please press stop within 10sec or this message will be sent to your contacts.
Much as it would be great to have a more standard option/function in-game, I can’t realistically see that happening any time soon given the legal, technical, privacy, etc issues.
Jason’s death has made me think about how we as a team try to mitigate the risk of riders alone. So far, I’m thinking of an optional ICE register for my group that only a very small number of people have access to but which could be used if we have a concern about a rider (I don’t know any details about Jason’s death so no idea if this would help in that situation but I can see that it’s at least a mitigation for some risks).
The biggest issue I still see are that losing wifi may look very similar to a rider having a health problem.
