[Updated 21/07/19: the blog previously incorrectly stated the information sourced by Alexa was part of the agreement the NHS. This information is from the NHS which is openly syndicated through open APIs. Thanks to Anne Marie Cunnigham (@amcunningham) for the info.]
You may have seen in the news last week Amazon and the NHS have partnered to improve access to health information through Alexa, Amazon's voice assitant. There’s been a lot of opinions shared and news in the press with a range of views communicated ranging from this is a disaster waiting to happen to its the solution we’ve all been looking for. The truth as usual is somewhere in the middle. But what are the ramifications to normalising voice access to healthcare information?
Hey voice assistant!
The Amazon-NHS partnership enables people to make simple health related queries to Alexa and get answers from trusted sources of information - so are then able to make informed decisions on their health. This sounds great in principle, the concern however is this only supports those who are typically already aware of (and engaged in) their health and wellbeing.
Socio-economically deprived populations remain unsupported. They may be unable to get a voice assistant device also information accessed isn’t tailored to their specific need which can increase anxiety. If information is incorrect for their specific circumstances it can drive patients to seek medical guidance. As well articulated by Dr Sam Finnikan, improved information improves awareness and helps people to have a more informed discussion on their health and wellbeing. It enables shared decision making to support the person being in charge of their health and care. But the person needs to already have a certain level of engagement in their health already to use voices assitants in this way as demonstrated by active signposting. So in essence voice assistants are most likely to support already well equiped members of the public. But the challenges around regulation, equitable access, evidence and misconceptions that the device can provide diagnosis not just information are significant. Additionally with recent publications on how staff have access to snippets of recordings from voice assistants from Amazon and Google what are the privacy approaches and are they appropriate?
From reactive to proactive
How might using voice assistants access to health information actually shape healthcare systems.
One of the challenges facing health systems is the move from a reactive system treating disease to a proactive system that intervenes earlier to stop the decline of health or even prevent onset of illness. Currently healthcare systems are set up to engage with most people when there’s something wrong. It is well established life expectancy is increasing and so are the number of people with complex multiple conditions. As a result the demand on healthcare is increasing with associated costs escalating, this is unsustainable and a more proactive and preventative approach could change this. However with the current reactive model of healthcare it means people visit the doctor usually with declined health. This means health care data has gaps on the healthiness of individuals and initial decline in wellbeing.
A number of different approaches are being tried to address this. For example wearables are being brought to market with a suite of sensors to detect specific clinical indicators. The most well publicised example is the Apple Watch which was updated in 2018 with ECG capability. The Apple Watch is now capable of screening for atrial fibrilation. We have national screening programmes, such as breast screen to detect early onset of cancer. Private provider organisations offer health checks and “pre-scans” to enable prevention scans. Additionally in July 2019, the Mayo Clinic indicated it will be offering head to toe scans and investigations as part of its drive to provide personalised preventative care. These are all screening based approaches to early detection of deteriating health. The benefits of screening are limited with significant potential for increasing costs and demand on health systems due to false positives.
Voice assistants offer another approach to looking for onset of decline in health but at the risk of putting a middleman between the patient and the NHS.
Will voice assistants shape healthcare systems
The potential of voice assistants is accessing trusted healthcare information to improve knowledge. But, it’s important to note that this also provides the company information on the wellbeing of the person and their dependents. Information that doesn’t reach the healthcare system but information that can be used to create personal health profiles for each individual. This then means its possible to gather ongoing information on a persons health, specific concerns and possible deterioration in health. This can then be expanded, if there’s been continued query into certain symptoms or a query flags up as potentially serious then this might then be a good intervention point. Perhaps for an asynchronous communication method (e.g. chat or messaging) with clinicians to enable light touch support and prevent possible deterioration. It also enables other services, voice has been identified as a possible health indicator. This develops an ongoing health support relationship between the individual and the voice assistant provider.
What we are potentially seeing here are voice assistants avoiding overloading the system by putting themselves in the middle between the individual and the healthcare system. Also it normalises accessing NHS information through these devices. This gives a unique point in the exchange to generate ongoing information on the individuals health and wellbeing.
Of course the NHS symptoms information should be used to improve individual and public health. It's amazing that the NHS supports open syndication of informatino using open APIs. There is the additional reputational support from the NHS adds reputational value to voice assistants.
What is not understood is: if the NHS is supporting the use of voice assistants and sharing copyrighted material, then how is the patient information and knowledge being shared to help the NHS develop into a proactive care system? Or is the knowledge being used to develop products that ultimately circumvent established healthcare systems?
Are voice assistants setting up a new interface between individuals and the healthcare system which means a new point of control which in the longer term disadvantages healthcare systems?
Ideally the use of voice assistants would mean a long term partnership between industry, NHS and place based care as equals for health system transformation. Such a partnership would research if the information from voice queries can be used for flagging potential early intervention points and if so the appropriate mechanisms for doing so without overloading care providers or increasing costs to the healthcare system. This would enable relational and human aspects of care to be supported while introducing technology to provide improvements where beneficial.