Technology Could Help us Age in Place…Here’s How | Deloitte US has been saved
By Edgar Kalns, Ph.D., specialist leader, Applied AI, Deloitte Consulting, LLP
In 2009, researchers in England created a software program to compare the language used in one of Agatha Christie’s final novels to the vocabulary she used in 16 other books. The analysis revealed the author likely suffered from Alzheimer’s disease or dementia toward the end of her career. Along with a 31% drop in vocabulary between her earliest works and her 73rd novel (ironically titled, Elephants Can Remember), the author also repeated several phrases and used more indefinite words, such as something, nothing, and anything.1
There is often a correlation between language-usage patterns and the early stages of cognitive decline. People who have dementia tend to repeat words more frequently. They also tend to pause often when speaking and might use more ‘ahs’ and ‘ums’ than usual.2 What if we caught these signs early? In Deloitte’s vision for the Future of Health, smart devices might be equipped with artificial intelligence (AI) and machine learning to do just that.
Consider this: An always-on smart speaker could be programed to continuously analyze a user’s speech patterns to identify the earliest stages of dementia. Speech analysis could also help a clinician determine if a new medication is making a patient feel confused. Similarly, a smart speaker could be set up to listen to breathing patterns to determine when someone is out of breath, or to distinguish between an occasional dry cough and an increasingly frequent wet cough, which could indicate a more serious condition. A smart toilet might be able to generate valuable medical information about the user’s health and diet. It might also be able to determine when the user is chronically constipated or dehydrated.
How can tech address the health needs of boomers?
In the US, the growing health needs of the baby-boom generation represents a potentially massive social problem. As this generation ages and requires increasingly more care, the number of professionals who can provide that support is dwindling. In 2010, there was an average of seven caregivers for every person age 80 or older. By 2030, this “caregiver-support ratio” is expected to fall sharply to 4:1, according to AARP.3 One reason is that an estimated 660,000 nurses are baby boomers, and many are likely to retire by 2030. Moreover, overworked caregivers might not be trained to diagnose complex health care needs or might not be able to recognize subtle changes in health or behavior. Our Future of Aging research explored a key question around how much analogue (i.e., personal touch, face-to-face interactions) will be needed in the digital age. This question has become increasingly important as social isolation and loneliness became more common during the pandemic.
The ubiquity of low-cost microphones and cameras—coupled with advances in AI (deep learning)—could make it possible to analyze large amounts of longitudinal audio and visual data and use it to detect subtle changes in health. Armed with such insights, caregivers could intervene earlier to provide medical attention or prevent injuries, hospitalizations, or even death. Case in point: I worked on a project several years ago where we tried to use subtle changes in gait and balance to identify older adults who were at risk of falling. We analyzed videos of people walking and used applied machine learning to detect tiny changes that might merit an alert to a caregiver. Our solution was able to spot changes that are nearly undetectable to the human eye.
We are just beginning to tap into data
The longevity paradigm shift has led to the development of an interdisciplinary ecosystem that looks different from traditional health care models. It comprises a growing community of life sciences and health care, health tech, and companies focused on solutions that address underlying drivers of disease and aging. However, many important health metrics aren’t yet well defined or easy to measure. Monitoring changes to body temperature, heart-rate, and weight is relatively easy because the metrics are well-defined, commonly understood, and tools to measure these vital signs are widely available. However, the metrics needed to detect changes in coughs, bowel movements, or gait and balance, are not well established. Analyzing this data will require the expertise of a clinician. Moreover, while deep learning could help identify health issues in the earliest stages, vast amounts of data are needed to train the algorithms. For example, teaching a deep-learning model to distinguish between a dry cough and a wet cough requires a large number of audio samples coupled with a medical expert to categorize the coughs correctly.
The ability to identify the root causes of aging and disease could make it possible for all of us to age in place as we get older. Technology could make it easier for people to manage chronic illnesses at home and automatically alert clinicians to subtle changes, such as a shortness of breath. Targeting the root causes of aging and disease has the potential to revolutionize the world of health care as we know it, according to our new report on the Future of Aging and Longevity.
From innovative startups, to multibillion-dollar pharmaceutical organizations, to regulatory agencies, a rich ecosystem of new and emerging stakeholders is defining the future of longevity. There is a tremendous opportunity for companies to develop devices that can collect unique health data from older adults and generate insights that can help them manage existing conditions or detect illnesses or other health risks at the earliest stages. Such technology could provide caregivers with essential tools and allow more people to age in place.
Endnotes:
1. Agatha Christie and nuns tell a tale of Alzheimer’s, National Public Radio, June 1, 2010
2. How dementia/Alzheimer’s affects communication and tips to help caregivers, Dementia Care Central, September 27, 2019
3. The aging of the Baby Boom and the growing care gap, AARP Public Policy Institute, August 2013
