We are days away from the general election – where patient care and experiences remain a key challenge for any new administration. Sanius Health’s recent release of the Mental Health (MH) provider AI-powered social media data insights was well received by NHS leadership colleagues. As a company, we therefore agreed to bring forward our schedule for a long-awaited Acute provider release.

We looked at 50,000 data points extracted from a multitude of public social media posts, with the data showing that 35% of providers have a negative feedback proportion of over 25% – patient experience scores showing levels of sensitivity and variation outside of more structured measures, such as the Friends and Family Test. Critically, the value of AI-supported social media listening (SML) is that it gives us the unedited experience of many patients and families while receiving care as opposed to after, where they may have forgotten the true experience of the service. 

The rising demand for acute care is evident, with a notable 10% increase in A&E attendances alone from December 2022 to December 2023. Placing additional strain on inpatient services, this surge underscores the need for effective feedback mechanisms to better understand and address patient concerns. Now trialled and tested across multiple NHS pathways, and with acute services facing their own unique challenges, this initial phase within the acute health landscape has helped us to derive valuable insights for healthcare colleagues. 

Turning once again to Twitter, TikTok, YouTube, and Reddit, we wanted to see what could be unlocked from the vast volumes of firsthand insights captured on these platforms. From our total extract of 50,000 posts, we manually validated 25% – the figure above highlighting a number of key patient or carer comments from the 5 trusts with the highest proportion of negative datapoints. 

Tapping into the rich resources social media provides to gauge real patient sentiments, our goal is ultimately to help improve patient experiences, outcomes, and overall service delivery from an acute healthcare perspective. Indeed, from a snapshot taken over the past year, our analysis revealed that several acute trusts faced significant challenges in delivering optimal patient experiences. Among these were Stockport NHS Foundation Trust, James Paget University Hospitals NHS Foundation Trust, and Mid Cheshire Hospitals NHS Foundation Trust. 

Conversely, trusts like Coventry and Warwickshire Partnership NHS Trust, Royal Papworth Hospital NHS Foundation Trust, and Chelsea and Westminster Hospital NHS Foundation Trust stood out with the highest proportions of positive patient sentiment. While the current phase provides an overview of an initial selection of the NHS’s acute trusts, due to insufficient data points to make a balanced assessment of negative vs. positive sentiments for others, our upcoming second wave of insights will provide further enrichment of the sources, keywords, and resulting data to capture the full spread of organisations.

Current Pathways to Assessing Patient Experiences in the NHS

In the ever-evolving landscape of healthcare, understanding patient experiences has been a cornerstone for improvement. The Friends and Family Test (FFT) has long served as a feedback tool to give patients a voice when evaluating the quality of their NHS service experiences. Since its inception in 2013, the FFT has provided invaluable insights by asking users if they would recommend the services they received, capturing over 75 million pieces of feedback to date.

However, as with many legacy systems, the FFT has its limitations. While revisions in 2020 marked an improvement to the original single-question gauge of general satisfaction, the FFT remains reliant on broad, qualitative data and varied collection methods that can prove challenging when attempting to compare insights across each organisation. 

Comparison of the FFT’s acute care domains with sentiment analysis captured through our AI-driven SML approach did indeed show an overall correlation in regard to the proportion of positive patient experiences. However, the level of variation seen between organisations within FFT feedback highlighted a potentially lower sensitivity, with all trusts reporting positive experiences at 80% or above for inpatient care, and an average of 70% or above across acute service-linked domains.

A core goal as we move forward is to ensure that we continue to build on the strong foundation laid by tools like the FFT, all while embracing emerging innovative approaches. Leveraging technologies such as advanced analytics, machine learning, and real-time feedback systems like SML can offer deeper, more actionable insights. These new methodologies provide us with a much more comprehensive and precise understanding of patient experiences. 

Indeed, by integrating cutting-edge technologies with patient feedback mechanisms, we can advance beyond the limitations of traditional methods and foster a more dynamic, responsive healthcare system that better meets the needs of patients and providers alike.

Public Perceptions and Their Role in Reshaping NHS Pathways

For those of you unfamiliar with SML or our previous work in targeted disease and mental health SML extraction, the core concept is this: through AI-driven methodologies, data is extracted against specific keywords from various online platforms and analysed for the core themes and sentiments expressed by posters. Natural Language Processing (NLP) algorithms allow us to perform context and sentiment analysis on thousands of responses within a proprietary AI platform. 

As we and others have shown, this approach is one that lends itself to a multitude of use cases and sources of information. Indeed, alongside our more detailed trust-level analysis, we were also interested in exploring public and service user views on the NHS at a wider level. To do this, we scraped comments from an online health-focused journal, with several key themes emerging from reader feedback on recent publications.

From both a patient experiences perspective and a more general NHS operational, transformation, and flow perspective, these comments highlighted key themes within healthcare pathways. For patients, key areas of challenge were linked to the quality of care, potentially underlined by ongoing capacity constraints at the frontline. This seemed to be compounded by service inefficiency, specifically around the discharge process and follow-up, as well as poor patient-staff interactions with noted communication failures.

At a broader level, many comments were focused on employee experience and satisfaction – 75% of responses displaying a negative sentiment – as well as staffing levels or compensation (71% negative), financial management or funding allocation (68% negative), and the leadership team (61% negative). Moreover, issues were identified regarding ongoing systemic changes and strategic planning (59% negative), with underlying technology that could support data management and data flow (58% negative) also highlighted as critical challenges.

Transforming the Healthcare Landscape Across Service Types

Our insights have already been hard at work pinpointing critical areas needing improvement for NHS colleagues, from systemic capacity constraints to inefficiencies across pathways, and with a particular focus on discharge and follow-up. Building upon our prior work across other NHS pathways, we have uncovered some key themes that strike a particular chord with service users. Critically, these SML analyses offer actionable insights for NHS teams seeking new ways to guide decision-making and planning – enhancing services, patient experiences, and quality of care through novel insights into what’s happening on the ground.

By focusing on the most impactful areas for service users, we aim to support our partner trusts in transforming the acute health landscape. Leveraging patient voices and advanced AI/ML technologies, we are uncovering the underlying factors affecting patient care and identifying solutions to address them. We are constantly enriching the depth and breadth of our SML databank and insights, and for those interested in learning more about our work and its implications for acute health services, please reach out here or through [email protected].