How can data privacy ensure AgeTech is both effective and trustworthy?

In March, I wrote about building technology worthy of older adults’ trust.

I wrote about varied preferences for technological transparency. Older adults often find reassurance in tangible proof of security, like handwritten passwords. Younger generations embrace abstract proofs of security, like Google single sign-on. Unfortunately, most software engineering teams exclusively comprise younger adults, who build products to meet their preferences and expectations - not those of older adults.

In the months since writing about trust in AgeTech, I’ve become more and more convinced that rendering technology in the physical world is the key to fostering trust with older generations. The importance of trust helps explain why most technology interfacing with older adults has a physical component, dating back to the LifeAlert alerting button (“help I’ve fallen and I can’t get up”).

LifeAlert is a case study in effective yet ageist tech - it’s a physical marker that connotes infirmity. Who wants that? When it comes to alerting, however, newer health monitoring systems are taking a more compassionate and technically sophisticated approach to leveraging physical devices to build trust with older adults - and their families.

Can we trust health monitoring systems?

Health monitoring systems include remote patient monitoring (RPM) and ambient patient monitoring (APM).

Remote patient monitoring systems are most reminiscent of LifeAlert, using a piece of wearable hardware to track health metrics. Today’s RPM systems like CarePredict require active patient participation, whether by charging a device or inputting data. The folks who’d benefit most from this technology, however, may lack the drive or ability to remember to charge and wear the device. As a case in point, Apple Watch may serve as an age-agnostic and user-friendly RPM for fall detection. But an Apple Watch can’t alert the circle of care to a fall if the user has forgotten to put it on. Ultimately, low compliance may result in unreliable data and may limit the efficacy of an RPM.

Anecdotally, one of the caregiving clients I work with shared that she got an Apple Watch after a terrible fall that landed her in the hospital - but she removes the device to charge overnight anytime there is a system update. Nighttime presents great risk of falling, whether due to balance challenges changing position¹ ; medication side effects² ; or nighttime confusion associated with dementia (“sundowning”)³ - so removing an RPM at night for a system update may mean the device isn’t available when it’s most needed.

Can we trust remote patient monitoring given the risk of low user compliance?

Ambient monitoring systems like SafelyYou embed sensors in the user’s environment to monitor health and safety without requiring user participation. SafelyYou partners closely with clinicians, boasting 100K+ clinician-reviewed falls (“on-the-ground events”). That’s a pretty impressive level of clinician partnership, even among the broader landscape of ambient monitoring systems like Tellus and Care.ai that partner closely with clinicians. Ultimately, transmitting sensitive data like fall event videos to clinicians raises a question of data privacy protections for APMs available today¹.

Can we trust ambient monitoring given data privacy concerns?

What does data privacy look like in care?

My experience building highly regulated financial technology products convinced me that the biggest threat to data privacy is the period in which the data are in transit between systems. It’s pretty hard to steal data while it’s stored in a server (which acts as a vault), but sending data to another server cracks open the vault door for bad actors.

As I think about data security for both remote and ambient monitoring, I’m increasingly excited about the opportunity presented by edge computing. Edge computing sounds very technical, but is actually simple: a smart device intakes and processes data itself (“locally”) rather than sending the data over WiFi to be processed in an external server far away. The smart device may send an alert over WiFi but the actual data that triggered the alert isn’t sent in most cases.

In practical terms: I think of my Grandpa Gene who had a very smart personal assistant come to his home to help with filing and taxes - none of his data/files left the premises, but she occasionally called a colleague to verify or escalate a question if needed.

Taking edge computing one step further, “inference on the edge” makes a smart device even smarter, by pre-downloading artificial intelligence models onto the smart device. This is a one-way download: your data aren’t feeding back into the artificial intelligence model, but the device can use the models to process your data very intelligently. Additionally, inference on the edge means the device can process data and react very quickly (“low latency”), because there’s no lag time in sending data across the world wide web.

So, a smart device using inference on the edge is effectively a self-contained box of your data held within your own home that’s smart enough to send an alert to a loved one if something is wrong.

What does a secure monitoring system look like?

A couple of considerations strike me as key in a secure monitoring system to address the range of needs of adults aging in place or in a facility:

1. Many older adults don’t have internet⁴ .

   1. A successful monitoring device should not require a WiFi connection to send alerts.

2. Most families I surveyed cared most about the security of health and wellness data - asking questions about storage and data retention policies.

   1. A successful monitoring device needs to process data locally on the device rather than on an external server

3. All families I surveyed preferred DIY installation and maintenance.

   1. A successful monitoring device needs to have simple installation and a very long battery life so that families can “set it and forget it.”

4. The top risks leading families to escalate an older adult on the continuum of care are fever, fire, fall, and forgetfulness.

   1. A successful monitoring device needs to have adequate data to alert (maybe even predict!) fever, fire, fall, and forgetfulness.

Building Trust Through Transparency

The promise of health monitoring systems is immense, but their success hinges on addressing data privacy and security concerns. Transparency in data handling practices is crucial. All stakeholders (facility operators, families, older adults) need to know exactly how data is collected, processed, stored, and deleted.

As I think about the potential of monitoring systems to support care for older adults, I get equally excited about the potential of these systems to support care for all of my loved ones and even myself. I’m hopeful that we can build toward a future in which our lived environment cares for our loved ones just as much as we do.

Enjoying Golden Age of Technology?

Subscribe to Golden Age of Technology here.

Access Golden Age of Technology archive here.

¹  ReddyCare. “Why Are Older Adults Prone to Falling at Night?”

²  Appeadu MK, Bordoni B. Falls and Fall Prevention in Older Adults. [Updated 2023 Jun 4]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan.

³  SlumberYard, United Zion Retirement Community‘s Therapy Department“Tips for Avoiding Nighttime Falls.” 2020 Sept 24. 

⁴ According to a report published by Older Adults Technology Services (OATS) from AARP's Aging Connected initiative, more than 21 million people age 65 or older in the U.S. lack broadband access at home.