Last week I had the pleasure of attending NHS Digital’s Industry Briefing, held in Leeds. The interactive event featured a handful of speakers informing the audience of what NHS Digital is doing to support healthcare providers across the country. However, one presentation on Natural Language Processing (NLP) stood out for me. Delivered by Joshua Symons, Associate Director of Data Optimisation at NHS Digital, the presentation alluded to the need for reliable NLP due to the explosion of consumer technologies such as smart speakers.
The reason why I was so eager to write about the application of smart speakers in healthcare is rather personal. I had previously bought a smart speaker for my father, who is blind, and the additional independence it has given him at home is remarkable. There are numerous products on the market from the likes of Amazon, Google and Apple and the ecosystem of complementary devices, accessories and skills is growing rapidly, with some devices being relatively affordable (around £20 during the Black Friday sales.)
It’s undeniable that smart speakers can assist healthcare providers, as evidenced by California-based AI company, Suki, which enables automatic voice note recording for GPs. However, in my humble opinion, the real value comes from its use at home as there is a global outcry for technologies that can reduce the number of inpatient visits to alleviate the burden on hospital staff.
There are lots of instances where smart speakers can be used in their basic form to improve health and promote wellbeing. However, there are countless impressive use cases being deployed by organisations, primarily in the U.S., such as:
- Providing basic medical advice for common ailments from the NHS website – NHSX
- Managing health improvement goals and earn wellness incentives – Cigna
- Checking the status of home delivery prescriptions – Express Scripts
- Enabling parents and caregivers to give updates to clinicians post-surgery – Boston Children’s Hospital (Kids MD)
- Searching for a nearby urgent care centre and scheduling same-day appointments – St. Joseph Health
- Querying last blood sugar readings, analysing trends and receiving advice – Livongo
- Tracking blood pressure – Omron Healthcare
- Listening for agonal breathing as an early warning sign for cardiac arrests – University of Washington
- Enabling elderly or frail users to connect with caregivers, set up reminders about medication, schedule appointments and report weight and blood pressure – Libertana
On top of all these specialised skills, Alexa also became HIPAA-compliant in the U.S. this year, which means that patient records can be transmitted via the device (and approved health record providers) for use at home, which opens up the possibilities even further.
One of the biggest reasons why these devices could significantly help vulnerable demographics is because they use the simplest user interface possible; the voice. Admittedly, they would require some basic knowledge to install but once up and running, they can be operated by almost anyone.
However, some concerns around their use persist, particularly around data privacy. Many continue to worry about devices recording conversations when not in use as well as the use of personal and confidential data by large, commercial enterprises. Another issue is the consumer’s reliance on technology and the impact of failure. For example, as mentioned above, a user may rely on their smart speaker to remind them to take medication. If the device were to then fail for whatever reason, there is an obvious health impact and potentially even legal implications.
If smart speakers were distributed by health and social care organisations, in a similar way to how Fitbits were distributed to health insurance customers, what would this mean for IT suppliers? First of all, smart speaker manufacturers would obviously benefit tremendously, especially if they could then apply analytics to the captured and stored data to improve population health. Secondly, if this information was made available to healthcare providers then support would be needed to make the data/information interoperable and actionable. Thirdly, but not finally, third party app developers would be able to contribute to the ever-evolving app and skill ecosystem.
It goes without saying that a smart speaker may not be a priority for many who would actually benefit from it and so relying on patients’ own devices may not be the answer. To properly achieve widespread distribution across the vulnerable demographics, the NHS would probably need to subsidise these types of devices although in this challenging economy, this may be easier said than done. Whilst it is certainly not a silver bullet to the NHS’s challenges, the smart speaker may just be a much needed lifeline that it could leverage to help reduce the burden on healthcare providers across the country.