With more than 350,000 digital health apps on the market, and an average of 250 new solutions released every day, Ross O’Brien, managing director for the UK and Europe at Wysa, looks at how can commissioners and clinical leads can uphold safety standards while putting the best tools into the hands of clinicians and patients
The case for digital clinical tools in the NHS is increasingly clear.
It has been shown that mobile and digital health apps can improve access for patients, offer support during what are often-long waiting times, provide better data insight in order to deliver the right treatment and better treatment, and give higher patient engagement.
An acceleration in the use of digital tools as a result of the COVID-19 pandemic means patients are more accepting of online and technological treatments, and even favour it – with 22 million people now usinh the NHS App.
But technology is constantly evolving and there is a wealth of choice available.
With over 350,000 digital health apps on the market, and an average of 250 new health apps released every day, how can commissioners and clinical leads uphold safety standards while putting the best tools into the hands of clinicians and patients?
For this to happen, it is essential to make informed decisions that do not stand still, but evolve and develop as needs of clinicians, patients and organisations do.
There has historically been no common consistent standard across the NHS when it comes to measuring the ability and efficacy of digital health tools
There has historically been no common consistent standard across the NHS when it comes to measuring the ability and efficacy of digital health tools.
Unlike other industries, where there is an ISO or equivalent benchmark, due to its evolution at pace, the NHS has a number of different standards that do not always align, making both monitoring standards and integration across systems challenging.
An initial consultation by the Innovation Department at NHS Transformation Directorate set out to define the standards for clinical safety, data protection, cyber security, and technical assurance alongside a view of accessibility and usability.
Their findings were developed into a standard and absolute baseline, Digital Technology Assessment Criteria (DTAC) advisory assessment criteria that digital health technologies need to meet to operate safely within health and social care in England.
Rather than a one-off exercise, this standard must be consistently being met.
Each use case, each interaction, and each client record gets updated to make sure the clinical safety risk profile is dependent on a daily update.
Having a deep level of rigor means that the creators are not having to duplicate every time they go to a new market, also expanding the pool of data available to help the learning system, which is essential when using AI and machine learning to pick up on user intent or keywords that are used for flagging potential risks.
The NHS is a national body, but it is committed to delivering the best care to patients as individuals whose symptoms can differ from person to person.
To help minimise risk we must assess and analyse products through systems, national programmes, and patient focus.
Having a blanket approach just doesn’t deliver for patients and is where gaps in digital systems arise, as even the most sophisticated of all technologies cannot do everything for everyone
Each requires something a little different, and it is essential to make sure that the needs of patients are being met at the same as designing for population and organisational-wide systems.
This requires recognising which segments of the population a specific technology can work for, and targeting interventions appropriately depending on needs, and then distributing high-quality digital technologies to people who need them.
Having a blanket approach just doesn’t deliver for patients and is where gaps in digital systems arise, as even the most sophisticated of all technologies cannot do everything for everyone.
At the same time as providing a tool for self management, technology offers an opportunity to use data to make clinical inference through an integrated system, which also contributes to real-world evidence for continuous improvement.
But none of this works unless we ensure clinical teams are invested to ensure patient buy-in and better relationships, and upskilled to use digital tools.
The best technology in the world will not work if it slows down progress, or the interface is difficult to use.
Excellent deployment of technology in healthcare systems requires training that has human knowledge front and centre.
The fundamental original premise of the internet was about making sure that the right data was in the right place, and that people had access to the best information at the time.
Without good data, healthcare professionals will lose interest and patient trust will be damaged.
Great technology does not only look great, but makes the best information accessible and available to make the best health decisions.
This kind of information is essential to commissioners, who want to make evidence-based assessments for the tools they choose to deploy.
In a landscape where technology and social media are posing risks, as well as benefits, there can sometimes be a nervous attitude towards the use of technology within healthcare settings.One concern is the way AI can replicate bias. This can be addressed by really targeting the communities that the intervention is meant for to ensure that their specific needs are being met, and always seeking to understand the data and its provenance in its entirety.
In a landscape where technology and social media are posing risks, as well as benefits, there can sometimes be a nervous attitude towards the use of technology within healthcare settings
By not thinking that a tool can fix everything for everyone, we move towards much-more-clinically-robust and targeted applications that can actually work.
Wysa, for example (pictured right), directly addresses bias by using scripted clinical interventions which are supported by natural language processing (AI chat) rather than allowing the AI therapist to make a best guess in terms of the quality of care.
The move towards developing robust clinical regulations and benchmarking do not only help address the vast choice architecture, but ensure that standards are being met on a consistent basis, delivering the best in patient outcomes.
It’s an ongoing evolution of real-world evidence gathering, understanding patient attitudes and responses, and investing in the development of systems.
That continuous improvement is the only way that we can ensure that rewards trump any risks.