Who would want a robot caregiver?
This post went out to subscribers to my Substack newsletter on August 23. Check out all the posts, including ones from the archives not reposted here, and subscribe here.
I recently spent time visiting my aging parents—my father will be 88 in November and my mother is 86—so I have elder care on the brain. My parents have a diversified portfolio of children to help them out. I, on the other hand, have none, which makes futuristic elder care predictions all the more salient.
What happens when the population of old people gets too big for the supply of paid and unpaid caregivers? Raising wages is the obvious answer but, as I discussed in a March 2020 Bloomberg column, the economics are terrible.1
To save you a click, I’m going to repeat the whole thing, with the critical facts highlighted, before I get to my new point:
They are the most numerous and lowest-paid U.S. health-care workers: the 4.5 million caregivers who assist elderly and disabled people with such daily activities as bathing, dressing, feeding and toileting. Compared to other medical workers, they have the most frequent and personal contact with the elderly. That puts them on the front lines of the coronavirus pandemic.
Known in the business as direct-care workers, these paraprofessionals include certified nursing assistants, who work in nursing homes, and home health aides and personal-care assistants, who often travel between multiple clients. Their median wage is $12.27 an hour, with home care workers making the least and nursing assistants the most. As the pandemic spreads, these workers are especially vulnerable and increasingly difficult to replace.
Don’t they deserve more money? The National Domestic Workers Alliance certainly thinks so:
Home care workers take action when health crises hit to protect the most vulnerable, and $11/hour isn’t fair pay. We need to invest in care workers — key first responders in this crisis. https://t.co/j9Mnmgh3FY
— Domestic Workers (@domesticworkers) March 14, 2020
But the pay isn’t low because the people who hire caregivers are greedy and mean-spirited. Neither is it because the work is easy or unimportant. It’s a much tougher problem. Caregiving is vital, but so labor-intensive that at higher wages, hardly anyone could afford it.
Consider some basic math. There are 8,760 hours in a year (8,784 this leap year). At $11 an hour, that’s $96,360. At $12.27, it’s $107,485. At $15 an hour, it’s $131,100.2
And that’s before the client pays for room, board, and other medical care. About 16% of caregiver payments come from family budgets, while private insurance covers only 11%. The rest comes from government programs, primarily Medicaid.
Not everybody requires 24/7 care, of course, but many people do. “Need” is as much a matter of what people can afford as it is an objective criterion. For families deciding how to spend dwindling resources, it can be a matter of supplementing paid with unpaid labor, which can require cutting back on a family member’s own work hours. For state legislatures allocating the Medicaid money that covers 52 percent of long-term care, it means tradeoffs between elder care and hospital reimbursements or maternity care, not to mention schools and highways.
Direct-care workers are in short supply, with skyrocketing turnover rates. In 2018, the latest year for which data is available, the turnover rate among home health aides hit a new high of 82 percent, according to the annual survey by Home Care Pulse. The agencies and institutions that employ these workers aren’t just competing with each other. They’re competing with businesses such as Amazon.com that have been raising entry-level wages.
“Turnover was so severe in 2018 that more than half of the participants had to turn away new clients because they didn’t have enough caregivers,” reported the trade publication Home Health Care News. Tight Medicaid reimbursement limits make it hard to raise pay or expand service. Higher reimbursements would permit higher pay but, again, the money would have to come from somewhere else.
We’ve seen this dilemma before: absolutely essential work that takes so long that wages remain low even as it consumes a high proportion of total expenses. Pre-industrial spinners worked for incredibly low wages, yet their pay was often the biggest expense in making cloth. And they were always in demand. “The spinners never stand still for want of work; they always have it if they please; but weavers sometimes are idle for want of yarn,” wrote the agronomist and travel author Arthur Young, who toured northern England in 1768.
Before the Industrial Revolution, Indian hand-spinners, the world’s best, took about 100 hours to produce enough cotton thread to weave the fabric for a modern pair of jeans — not including the time cleaning and preparing the fiber beforehand. Spinning the equivalent amount of wool on a European spinning wheel took about 110 hours. At the low modern wage of $11 an hour just the thread in a pair of trousers would have cost well over $1,000, not including the time spent dyeing, weaving or sewing. (For sources and an in-depth discussion, see my forthcoming book, "The Fabric of Civilization: How Textiles Made the World.") Only by finding mechanical ways to get much more thread per hour did people finally make cloth abundant, leading to the takeoff in worldwide living standards that economic historian Deirdre McCloskey calls the Great Enrichment.
Boosting productivity, and the wages it supports, is tougher for in-person services like caregiving. In his 2017 book, “Who Will Care for Us?” economist Paul Osterman advocates giving caregivers more training and allowing them to provide simple medical services like administering medications, thereby reducing the need for more expensive nurses to provide frequent care. Many people get into the field because of an interest in health care only to find themselves ignored by their medical colleagues, treated as little more than baby sitters and stymied when they seek to learn more.
Current regulations, Osterman notes, can lead to such absurd situations as caregivers guiding the hands of dementia patients to “self-administer" eye drops, skirting rules against aides doing that job. But nurses, who are well organized and politically influential, fight like crazy against any incursion on their territory.
To boost productivity more significantly, potentially improving both care and wages, start-ups are experimenting with artificial intelligence. An intriguing example is Cherry Home, which markets an unobtrusive monitoring system that distinguishes normal behavior patterns from abnormal ones, including falls, restless sleep or signs of confusion. When something looks off, the system alerts a monitoring center, which contacts caregivers, family members or emergency services as needed. The system has a privacy mode that displays stick figures rather than images of people, and it can communicate with someone in distress without requiring them to press a button. In theory, such systems could allow individuals to stay in their homes without having aides or family members present all the time.
By calling attention to the important work of aides in eldercare, the coronavirus offers an opportunity for a conversation about how to increase the respect accorded their work within the health-care system, how to improve communication between the people who spend the most time with the frail elderly and other medical professionals, and how to build their skills and widen their responsibilities. But as long as it takes thousands of hours a year to care for a single person, wages can’t go up much.
In the pre-industrial era, everyone knew that spinners were important. Before smokestacks, a spinning woman was the iconic representation of “industry.” But to raise wages, you need more than respect and goodwill. You need new technology.
Anyone looking for A.I. applications ought to be thinking about how to enhance elder care. Cherry Home’s system demonstrates that new technology may not mean anything as complicated and intrusive as a humanoid robot.
Robots for elder care tend to be pitched as substitutes for friends or pets or human caregivers. They dance! They tell jokes! But as the comments on this YouTube video suggest, people tend to find the cheery forced companionship of robot carers more than a little suspect.
Anthropomorphizing robot caregivers misses one of their biggest advantage: their impersonal nature. The typical approach reflects the widespread assumption that old people would prefer human aides. The mental model of the elder is my late mother-in-law, who treated her caregivers as new best friends. Who wants to be cared for by an inanimate object rather than a person?
The same kind of person who, all else being equal, prefers the self-checkout to the cashier or the ATM to the bank teller. For all the chatter about technology and “privacy,” nobody seems to be thinking about what it’s like to have a strange person come into your house, hang around for hours, and see you in your most intimate and vulnerable moments. The privacy that matters in everyday life has nothing to do with databases.
My fiercely independent parents have no interest in a human caregiver, even though they could use some help. And I understand why. Personally, I’d prefer a robot that acts like a helpful appliance.
Odds and Ends
I'll be giving a Zoom lecture on The Fabric of Civilization Saturday. Sign up at Lunatic Fringe Yarns. I’m adding a live demo explaining how spinning works, since even experienced spinners often don’t understand what’s going on. Let’s hope it works.
The Prepared is an excellent newsletter about “engineering, manufacturing, and infrastructure.” Their book club recently featured The Fabric of Civilization, and I answered members’ questions on Zoom. Hilary Predko did a fantastic job editing the conversation into an interesting Q&A, complete with some nice visuals.
Two fun food writers: Tamar Haspel grows her own and Howard Miller (whose bio is a story in itself) plumbs the weirdness of American food history, although his latest post is on the movie Elvis.
On Substack, there are no editors telling you what you can write about. Or, as you may have noticed, when that free post has to be finished.
In my August 6 post, I announced a contest inspired by thoughts from fellow dynamist Substackers Jim Pethokoukis and Anton Howe:
So here’s are the challenges. You can pick one or try any combination.
- Write an updated version of the Victorian paragraph, looking back at 2014.
- Write a speculative version of the Victorian paragraph, looking back at today from 2030.
- Come up with an inspiring illustration of a possible 2040.
I’ll publish a selection of the best here (you’ll retain rights, of course) as I receive them and will accept entries through September 30. I’ll then award the top two in each category a collection of what Jim would call “Up Wing” books. The judging process will depend on how many entries are received, and I reserve the right to award fewer than six prizes. Email them to me at [email protected].
Full background, including the model paragraph, at the original post.