NHS Mental Health Artificial Intelligence (AI) Social Media Listening (SML) | Positive and Negative Patient Experience Provider Data

What Can AI and Social Media Tell Us: Novel Approaches to Understanding the Real-World Patient Experience Across the NHS and Mental Health Care Pathways.

Content by Sanius Health

Earlier this year, we partnered with industry to use AI to amplify the voice and experience of patients across the NHS, addressing challenges in health equality, access, flow, outcomes and poor care. Most of the patients we looked at were from disadvantaged and vulnerable groups across the country. This was a hugely successful project, and we wanted to extend this work to improve experiences, CQC ratings, outcomes, and flow for patients across the NHS overall. So, we got to work with some of our AI tools, OpenAI, and some of the smartest researchers, scientists, and analysts we know to look at every hospital, the real patient experience, and outcomes by trust.

We started with Mental Health providers, as the NHS has seen skyrocketing waiting times in this space since 2017, with recent data suggesting over 1.2 million patients are waiting and struggling to access care. The basic maths tells us that if you and I are using X (Twitter), Facebook (our parents and grandparents), Instagram, TikTok, and more, and telling the world how we feel about everything, this would be a true source and voice for the patient experience.

When we explore some of the data from these sources over the past year, we find that the trusts with some of the most challenging patient experiences are Central and North West London NHS Foundation Trust, West London NHS Trust, Pennine Care NHS Foundation Trust, Nottinghamshire Healthcare NHS Foundation Trust, Norfolk and Suffolk NHS Foundation Trust, Oxleas NHS Foundation Trust, and Tavistock and Portman NHS Foundation Trust. At the other end of the spectrum, the ones with the best patient experiences – given as a proportion of posts with a positive sentiment – were Humber Teaching NHS Foundation Trust and North Staffordshire Combined Healthcare NHS Trust.

Looking at specific services, key areas which have come up include inadequate training within children’s Special Educational Needs (SEN) services, Patient Advice and Liaison Services (PALS), and more general joining up of care from provider to commissioning organisation. However, it should be noted that a number of organisations, such as Oxford Health NHS Foundation Trust, Sussex Partnership NHS Foundation Trust and South West Yorkshire Partnership, were excluded due to low numbers of posts, and others due to the potential impact of responses linked to the acute services provided on these mental health-focused insights. Future waves will explore additional enrichment of these data as, fundamentally, we want to use these insights to improve patient access, experiences, care and outcomes.

Year on year, we’re seeing the demand for mental health care continue to soar, with patient experiences all too frequently taking the brunt of struggling NHS capacity and resources. From December 2022 to December 2023 alone, the number of people in contact with mental health services surged by 13% – from 1.65 million to 1.87 million. With this increasing pressure, mental health trusts have seen falling overall ratings from the Care Quality Commission (CQC), dropping from 77% ‘good’ or ‘outstanding’ in 2022 to 74% in 2023.

Core pathways are beginning to emerge as having particular difficulties, including ‘Inpatient acute and intensive care for working-age adults’ services – dropping by -7% in ‘good’ or ‘outstanding’ ratings from 2022 to 2023 – and ‘Acute wards for adults of working age and psychiatric intensive care units’ – 77% of trusts reporting ‘requires improvement’ or ‘inadequate’ ratings.

In trying to tackle these issues, the NHS often looks for firsthand patient, carer, and staff feedback to better understand what’s going wrong. Worryingly, recent insights have suggested that the proportion of mental health staff who would be happy for a friend or family member to receive care at their organisation under the current standards fell to its lowest level in 5 years (62.8%). In line with this, service user respondents reported a drop from 32% (2014) to 29% (2022) in terms of an overall good experience when accessing services.

In the face of current challenges, it is all too clear that ensuring that the future trajectory of mental health services is an upward one requires an approach driven by this form of firsthand, real-world perspective of the needs of patients and their carers. But with care and leadership teams under ongoing and significant capacity strains that can often impact their ability to conduct a full exploration, how do we use technology to delve into the untapped reserves of patient experiences and extract data that drives service improvements?

Novel Approaches to Understanding the Patient Experience

Over recent years, we have seen Artificial Intelligence/Machine Learning (AI/ML)-driven approaches surge in use across sectors such as healthcare, driving transformation through the provision of real-time, actionable insights that capture patient experience, public health trends, and service quality. As mentioned, interpreting the wealth of patient-centric insights shared publicly through social media sites, with advanced technology that enriches what we know about key pathways from this font of information, has been a particular area of focus.

These social media listening (SML) approaches leverage AI algorithms to extract, monitor, analyse, and contextualise social media data. Already, this has been used by those tracking influenza outbreaks through the monitoring of any mentions of flu symptoms or related terms within posts, comments, hashtags, and other sources. Enhancing more traditional approaches with real-time enrichment, a similar use was launched during the Covid-19 pandemic, with SML tools helping to track viral spread, vaccine sentiments, and general misinformation.

From a longer-term, healthcare service perspective, hospitals and clinics have been utilising SML to evaluate their pathways by analysing patient feedback – boosting care quality by identifying the key areas for improvement. At both a healthcare provider and Pharma level, SML has been used to dig into online discussions about medication side effects, and what this could imply in terms of real-world evidence around drug safety and efficacy. Finally, outside of the direct care environment, these approaches have enabled the monitoring of signs of mental health issues in social media posts, helping mental health professionals better understand the prevalence of conditions such as anxiety and depression, as well as potential triggers across patient populations.

These projects and more have shown us just how pivotal the patient experience is in enhancing healthcare services – fostering trust, improving care pathways, and ultimately achieving better outcomes for patients. Empowering patients and their voices is critical in identifying service gaps and informing quality improvements, letting us ensure that care is not only efficient, but ultimately patient-centred. And while highlighting the core challenges is fundamental to identifying targets for change, understanding and amplifying the positive patient experiences is equally key to driving ongoing improvements from powerful use cases.

What We Have Learned to Date Across the Mental Health Landscape

As we continue to support our partner trusts and clinical colleagues through the design of new approaches driven by advanced AI/ML technology, our recent work has applied SML methodologies to undertake key analyses around the recent experiences of patients in contact with mental health trusts. At its core, our aim was to begin uncovering real-world insights that would enable us to build a better understanding of patient and carer experiences, thereby improving quality of care and patient outcomes based on novel SML-generated insights.

Our approach included an initial AI-driven extraction of relevant data from multiple online sources, including Instagram, TikTok, Twitter, YouTube, Google Reviews, articles, and Reddit. With over 1,000 responses generated stretching across a 1-year historical search period, we focused on the most recent service-focused challenges from a patient perspective. From here, all responses were reviewed to curate overarching themes and key challenges at an individual trust and wider mental health service level, complemented by AI-powered tools like Natural Language Processing (NLP) algorithms for context and sentiment analysis.

Already, this output has helped us to identify the core service delivery challenges and biggest areas of importance for patients throughout the mental health care pathway. From the quality of care received, to patient perceptions of staff effort and the ease of accessibility of key services, our SML analysis has enabled us to pinpoint the core themes in which trusts have been both excelling and struggling. These work to guide clinical, operational, and leadership teams in planning and decision-making, creating opportunities to optimise services and truly provide the best experiences and outcomes possible for patients.

By drilling down into areas with potentially the largest impacts for service users, our goal is to support our partners in revolutionising the mental health landscape in a way that is driven by the patient voice, and enhanced by novel approaches that showcase the growing ways in which AI/ML-technologies can help. With additional insights into the potential factors underlying the issues faced by patients, our work offers an important route to explore each layer of the challenge.

For those who would like to learn more about our work in this space, please reach out here or through info@saniushealth.com.