From my Facebook profile, today:
Me:
Thinking back to CareWheels and spin-offs. This was to be a home monitoring service whereby elders, seniors, I'd add unsupervised teens and children, any age really, with caveats (not infants), would stay in their homes but send off lots of signals just by moving around and doing stuff. It's not about spycams which no one has time to look at or space to store.
It's about medical cabinet open close, bathtub use, shower on/off etc.
Ron Braithwaite could tell you how Machine Learning informs a human staffed set of dashboards (humans working from home also, they monitor each other), with dispatch teams at the ready for both scheduled and unscheduled home visits.
Like if the pill dispenser isn't activated within a time frame, or the bathtub gets turned on but not off... or a stove burner. IoT. Protocol might be to call first, contact a signed up neighbor if no response, only escalating as makes sense.
You don't dispatch a team at the drop of a hat. Patients at especially high risk for some medical condition will presumably have sensors for that, e.g. blood pressure or whatever. It's up to you and your doctor what level of monitoring is indicated. Some people opt out of some options.
Machine Learning has only gotten better since this infrastructure was first envisioned.
Ron Braithwaite:
Yeah, CareWheels was a fantastic idea. In fact, it was one of the driving reasons why we were going to move to Canada.
Honeywell owned (owns) a bunch of patents in this area, which they have never done anything with (TTBOMK), so we thought we would make it happen in Vancouver, BC with the help of a couple of non-profits.
The basic idea was to use machine learning (https://en.wikipedia.org/wiki/Bayesian_statistics) to determine when things usually happened and how long they would take, using simple sensors all over the place, wiring up an elder’s home to allow them to age in place.
I worked out a deal with Canada’s largest non-profits working with elders and another working with those who were disabled to provide a computer and modem on each end, targeting those who were primarily house-bound both as clients and as care providers.
When our machine learning algorithms detected something unusual, they would contact the assigned care provider, who would attempt to contact the client and, failing that, then contact designated family members, and (worst case) emergency services. In addition, the care provider was tasked with making contact with each of their assigned clients every day.
Just as an aside, did you know that 90% of all calls to emergency services by elders are for socialization? People get lonely living at home, so having someone - anyone - calling them daily takes a huge load off emergency services.
At any rate, the way it worked is that each sensor for each person had a specific typical time value associated with when the sensor detected activity. If the elder typically uses the restroom for 45 minutes at 9am, but doesn’t exit the restroom after 90 minutes, somebody needs to check up on them, just to make sure they haven’t slipped in the tub and broken their hip. The Bayesian algorithm learned the pattern of each client, so we were looking for exception conditions for that specific person and avoiding the problems rule-based expert systems.
Perhaps the most heart-breaking tragedy of getting turned away by a Canadian Immigration agent (I have always assumed it was because I’m an obvious hippie, but I really don’t know) was having to call the first person I had hired as a caregiver. I had interviewed several and picked him to lead the caregiver cohort, for a variety of reasons. Not the least being that he had suffered a traumatic brain injury about 12 years previously and couldn’t get a job, even though he was not cognitively impaired. He sobbed into the phone as I told him that we were turned away and it wasn’t going to happen.
I still think it is a viable service and I would love to contribute to the project if someone else were to resurrect it.
Thanks for reminding me of this, Kirby. This was fun to dredge up old memories of tilting at windmills.
Yeah, CareWheels was a fantastic idea. In fact, it was one of the driving reasons why we were going to move to Canada.
Honeywell owned (owns) a bunch of patents in this area, which they have never done anything with (TTBOMK), so we thought we would make it happen in Vancouver, BC with the help of a couple of non-profits.
The basic idea was to use machine learning (https://en.wikipedia.org/wiki/Bayesian_statistics) to determine when things usually happened and how long they would take, using simple sensors all over the place, wiring up an elder’s home to allow them to age in place.
I worked out a deal with Canada’s largest non-profits working with elders and another working with those who were disabled to provide a computer and modem on each end, targeting those who were primarily house-bound both as clients and as care providers.
When our machine learning algorithms detected something unusual, they would contact the assigned care provider, who would attempt to contact the client and, failing that, then contact designated family members, and (worst case) emergency services. In addition, the care provider was tasked with making contact with each of their assigned clients every day.
Just as an aside, did you know that 90% of all calls to emergency services by elders are for socialization? People get lonely living at home, so having someone - anyone - calling them daily takes a huge load off emergency services.
At any rate, the way it worked is that each sensor for each person had a specific typical time value associated with when the sensor detected activity. If the elder typically uses the restroom for 45 minutes at 9am, but doesn’t exit the restroom after 90 minutes, somebody needs to check up on them, just to make sure they haven’t slipped in the tub and broken their hip. The Bayesian algorithm learned the pattern of each client, so we were looking for exception conditions for that specific person and avoiding the problems rule-based expert systems.
Perhaps the most heart-breaking tragedy of getting turned away by a Canadian Immigration agent (I have always assumed it was because I’m an obvious hippie, but I really don’t know) was having to call the first person I had hired as a caregiver. I had interviewed several and picked him to lead the caregiver cohort, for a variety of reasons. Not the least being that he had suffered a traumatic brain injury about 12 years previously and couldn’t get a job, even though he was not cognitively impaired. He sobbed into the phone as I told him that we were turned away and it wasn’t going to happen.
I still think it is a viable service and I would love to contribute to the project if someone else were to resurrect it.
Thanks for reminding me of this, Kirby. This was fun to dredge up old memories of tilting at windmills.
