One in four people are forced to endure “life or death” delays of over a 12-hour wait in England’s worst performing NHS trusts, significantly higher than the 4-hour target and only 57% of patients were treated, discharged or admitted to a bed the previous month at any large NHS trust in November 2024 [1]. In the United Kingdom the NHS is the largest employer, the fifth worldwide after the US Department of Defence [2] and consequently, with a budget of £192 billion for 2025/26 [3], has the resources for economies of scale by reshaping using innovation and initiatives of digital transformation to make changes across the health system and to provide a patient centric model. The growth of social media has dramatically expanded in recent years, with over 5.17 billion users worldwide [4], providing positive and critical comments which could be harnessed to provide solutions in using transformation to make positive changes in performance targets. There are some good examples of digital solutions in the NHS but they are in different regions with moderate coordination or repository of solutions to use “off the shelf” solutions. Utilising “off the shelf” software provides faster benefits than developing a bespoke solution including quicker implementation, reliability, user support, and better scalability reducing the risk of failure [5]. This paper aims to answer the research question: Can sentiment analysis provide transformational change to improve healthcare targets? Section 2 discusses the challenges that the National Health Service (NHS) trusts are facing in England around waiting times and the action being taken by the government to increase performance against targets. Examples are provided where trusts are undertaking innovative ways to make improvements. Big Data Analytics and social media are discussed, with examples of how they have been successfully used in healthcare. Section 3 illustrates how social media analytics can be adopted by healthcare trusts to undertake sentiment analysis to identify hotspots. Section 4 discusses the findings of research and proposes a number of recommendations for digital transformation initiatives. Section 5 concludes the paper and proposes future research.

Recent surveys have shown that satisfaction with the UK’s National Health Service is at a record low. The 2023 Kings Fund survey recorded the lowest levels of public satisfaction since the survey began in 1983 with only 24% of the public being satisfied with the NHS and 13% being satisfied with social care [6]. One of the biggest issues is hospital waiting times. A quarter of patients waiting for accidents and emergencies (A&E) treatment have been left waiting up to 12 hours in the busiest hospitals, with almost 14,000 people dying due to excessively long waiting times [7]. The Guardian ([8], p. 1) reported that: “dissatisfaction with the service has more than doubled in just two years, from 25% in 2020 to 51% in 2022, a period marked by COVID-19’s arrival, the NHS suffering a worsening shortage of staff and growing numbers of patients—now 9 million across the UK—facing long waits for treatment”. As of the summer of 2024, the number of people visiting accidents and emergencies was above pre-pandemic patients, and the number of patients waiting over 4 hours for emergency treatment rose consistently between 2015 and 2020 with a new record being set in December 2022 [9]. Figure 1 shows Public Satisfaction with the NHS between 1983 to 2023 reproduced from The King’s Fund [6].

Figure 1. Public satisfaction with the NHS 1983 to 2023 reproduced from The King’s Fund (2024).

Lord Darzi was recently commissioned by the UK government to undertake an investigation into the state of the NHS and produced a series of recommendations, one of being investment and utilisation of technology to transform UK healthcare [10]. The cover letter to the report stated: “Over the past 15 years, many sectors of the economy have beenradicallyreshaped by digital technologies. Yet the NHS is in the foothills of digital transformation. The last decade was a missed opportunity to prepare the NHS for the future and to embrace the technologies that would enable a shift in the model from‘diagnose and treat’to‘predict and prevent’—a shift I called for in High Quality Care for All, more than 15 years ago” ([10], p. 1). Similarly, a report published in 2023 identified a number of reasons why digital transformation in the NHS is challenging including: insufficient digital infrastructure and issues with legacy systems; the lack of integration between the different information systems used; a lack of digital leadership; the lack of specialist digital skills and digital skills in the wider workforce; and the risk of excluding patients from care who cannot or prefer to not use digital services [11]. Another issue faced by the NHS is an ageing population in the UK, with over 900,000 older people being admitted to hospital annually due to the NHS failing to keep them healthy at home [12]. A number of other reasons why NHS England is failing have been suggested including: healthcare buildings and estates being old and crumbling; population health is worsening; continuing staff burnout and work pressures; continuing lack of access to adult social care; an inadequate workforce plan; lack of access to urgent and emergency care; and continued industrial action [13][14]. As part of the UK governments plan to reform the NHS, hospital league tables are being introduced to monitor performance and managers at failing hospitals could be dismissed if they are unable to make adequate improvements while top performing hospitals will have more freedom over their spending [15].

To demonstrate how healthcare providers can analyse social media data to compare their performance against targets, a trial of Brand24 [49] was utilised to perform an analysis of social media data of the NHS Trusts displayed in Table 1. A 14-day trial of Brand24 was utilised which commenced on 22^nd November 2024. Alerts were configured to monitor posts containing positive or sentiment from the 10 NHS trusts displayed. The software utilised determined the sentiment of posts, however the posts analysed in this study were sense checked. The majority of social media analytics tools reviewed had similar features as shown in Table 2. One of the limitations of undertaking a trial is that the software could only analyse the last 30 days of social media data, to analyse data beyond 30 days a full software licence is required. This was found to be similar to Hootsuite [50] which in the trial acquired only allows 7 days of data to be reviewed. This suggested that the limitation would likely apply to other social media analytics tools. Therefore, the examples shown in this paper cover the period 25 October 2024 to 24 November 2024.

For the purpose of this research, alerts were setup on the NHS Trusts displayed in Table 1 to compare the perception on social media between trusts with quicker waiting times against trusts with higher waiting times. It was expected that patients would be posting on social media their views on the treatment they received, potentially providing insight as to what the trusts are doing well or could be doing better. Brand24 and other social media analytics allow the user to compare themselves against their competitors, in this case other healthcare providers. Figure 4 shows a comparison between 2 of the NHS Trusts based on a number of metrics such as the number of times mentioned on social media, positive and negative mentions. These metrics will give a healthcare provider an indication of how they are perceived by the public. There are filters which can be applied, such as a date filter to measure social media performance over a time period. There is a rounding error in the software as Bradford Teaching Hospitals shows 78 positive of 125 social-media mentions = 63% (62.4%); 9 negative of 125 = 8% (7.2%).

Figure 4. Comparison between two healthcare providers social media performance (online data in the public domain).

Brand24 and other social media analytics packages allow users to view more in-depth analytics such as positive and negative sentiment over time as shown in Figure 5.

Figure 5. Summary of positive and negative sentiment in social media posts (online data in the public domain).

Social media analytics software have the ability to identify common keywords in social media analytics posts based on the sentiment. For example, posts thanking NHS staff often contain words such as “thank”, “great”, “fantastic” or “support”. By clicking on the words, the user can view all of the posts where these words are mentioned. This would be helpful for a healthcare provider to view what their patients believe they are doing well. Similarly, they can look at where other healthcare providers are performing well and drill down into the specific social media posts for more insight. This also works with negative sentiment, for example words such as “negligence”, “death” and “fatal” were identified as negative in the research undertaken across multiple NHS Trusts. By analysing the posts which contain these words, healthcare providers can get an understanding of what patients are unhappy about. Figure 6 shows keywords identified in social media posts displayed in a word cloud format, with posts with positive sentiment in green, negative sentiment in red and neutral sentiment in grey. Similarly, the larger the size of the word, the higher frequency it appears. The social media analytics package and/ or integrated Computer Assisted Qualitative Data Analysis Software (CAQDAS) allows qualitative user feedback to be quantified, allowing trends to be monitored. For example the number of times a positive or negative word appears allowing hotspots, which need to be addressed, to be identified.

Figure 6. Example of a Word Cloud of key words found on social media and the sentiment.

Figure 7 shows a sample of two posts with positive sentiment. Both posts appear to be praising the healthcare trusts. The second post is from a family member praising the care their father received, specifically praising the speed of care. Multiple posts praising specific services could allow healthcare providers to understand where they are performing well.

Figure 7. Posts with positive sentiment on X.

Figure 8 shows a sample of three posts showing negative sentiment. The first post (formerly referred to as Tweets) appears to be a general comment that patients are going to hospital as they are unable to see their GP. The second post provides a specific example of long waiting times in the accident and emergencies department, which the NHS trust could investigate or respond to explain the waiting time. The third post is a more general post around NHS abandonment and negligence.

Figure 8. Posts with negative sentiment on X.

A common trend in the social media posts analysed during this study is that many of the posts do not give specific information. They appear to be generalised posts which do not give sufficient information to investigate as identified in Figure 8. It would be helpful if the posters provided suggestions how the problems they identified could be rectified. Similarly, whilst many of the posts provided examples where NHS trusts had provided excellent care, there were also posts which did not provide helpful information to analyse (absurd comments).

There is an opportunity for healthcare providers to learn from other providers who are performing well. For example, in England there is the Care Quality Commission (CQC) [73] who are a national body to regulate and inspect health and social care providers. They use a rating that works on a scale from one to four: inadequate, requires improvement, good and outstanding. Leaders and subject matter experts from healthcare providers who are achieving the Outstanding rating could be utilised to provide advice and guidance to trusts who are struggling. A catalogue of accredited off the shelf products and services could be developed from tried and tested solutions by healthcare providers who are exceeding their targets. This would allow other providers to replicate what is being undertaken elsewhere, but specialists from the outstanding healthcare providers could lead the implementation and tailor them to the healthcare provider where they are being implemented as it is likely that they will have requirements unique to their organisation. Change management would also need to be considered as staff would need to be trained or upskilled, therefore key personnel would be required to oversee the transition. This would also speed up the implementation of these systems. The Public Accounts Committee [74] highlighted repeated failures in NHS digital transformation, noting that the Department’s previous decade-long attempt at IT reform had been ‘both expensive and largely unsuccessful’, and citing the NHS COVID-19 contact-tracing app. An estimated £11 million had already been spent on a delayed first phase and this is a further example of poor financial transparency and accountability. Therefore, seconding experts into advisory roles to spearhead and help other healthcare providers adopt best practice could increase the success rate of digital transformation projects.

Additionally, healthcare providers could share best practice through a Community of Practice (CoP). A CoP is defined as: “groups of people who share a concern or a passion for something they do and learn how to do it better as they interact regularly” ([75], p. 1). Key personnel from outstanding rated healthcare providers could join community of practices to share ideas about what they have implemented successfully and how they overcame challenges with other healthcare providers. Regular meetings would allow new developments to be shared between providers and shared solutions to new problems to be created. Similarly, they can discuss the cost of digital transformation vs the value for money in terms of efficiency and patient experience.

NHS trusts could offer incentives for the public/stakeholders to suggest ideas to problems. This could identify innovations which other NHS trusts and healthcare providers could adopt if they are successful. The NHS launched the Change NHS online portal on 21 October 2024 to allow the public to share their views, experiences and ideas [76][77]. Whilst there are no rewards for the public to share their views, the success of this scheme could further justify why healthcare providers could setup a system for encouraging the public to suggest ideas to solve problems. Therefore, there is an opportunity for healthcare providers to develop a web portal to allow patients, staff and stakeholders to document the challenges they have encountered and suggest their solutions. Social media data could be streamed into the web portal and social media analytics could be used to identify trends, utilising techniques such as sentiment analysis. Healthcare providers could respond to the public. A reward and incentive system could be implemented to encourage suggestions, which could help healthcare providers to improve the quality of their care and potentially reduce costs. Given the increased pressure on healthcare providers with penalties such as management “being sacked” for poor performance, implementing these improvements could help them to move up the league tables [15]. Patient feedback could also be recorded at the time they are discharged, as this may provide a higher response rate than surveys and may also provide more timely feedback. This could be compared against social media feedback identifying problems to triangulate with patient journey and wait time.

Implementing the recommendations outlined above could enable healthcare providers to move to a Patient Centric (PC) model of healthcare delivery. Patient centric care is defined: “as Putting the patient first in an open and sustained engagement of the patient to respectfully and compassionately achieve the best experience and outcome for that person and their family” ([78], p. 1). Providers who have adopted a patient centric care approach have achieved benefits including greater patient satisfaction with care received, increased job satisfaction for staff, increased quality of care and safety of care, greater quality of life and well-being of patients [79].

One of the limitations of the study is the limited access to social media analytics tools. Whilst there are many social media analytics tools publicly available, the trials available from the software reviewed appear to be limited both in terms of the number of days the software can be used and how far back in time social media posts can be analysed. Future studies will require a full licence to a product such as Hootsuite or Brand24 to undertake a detailed analysis of social media posts over a time period of a year or greater.

This paper has provided a literature review of the causes of public dissatisfaction with the health and social care provided by the NHS which is evidenced by surveys undertaken. Patient waiting times for emergency care is one of the key targets of the NHS which most healthcare providers have struggled to meet. Examples have been provided of healthcare trusts who have made improvements through the use of innovation and digital transformation. An approach to monitoring patient feedback and the performance of other healthcare providers has been outlined through the use of social media analytics. It is a relatively low-cost solution for healthcare providers, for monitoring patient feedback and identifying problems to triangulate with patient journey and staff perceptions which if used effectively could allow healthcare providers to be proactive in finding the cause of problems and could implement innovative solutions. A number of recommendations have been provided which could allow healthcare trusts to move to a patient centric model of care including implementing rewards and incentives for patients, staff and stakeholders, implementing a community of practice and for the development of a catalogue of off the shelf solutions which could be easily adopted by healthcare providers from a central repository and will be investigated in future research.

Conceptualization, M.W. and A.S.A.; methodology, M.W. and A.S.A.; software, M.W. and A.S.A.; validation, M.W. and A.S.A.; formal analysis, M.W. and A.S.A.; investigation, M.W. and A.S.A.; resources, M.W. and A.S.A.; data curation, M.W. and A.S.A.; writing—original draft preparation, M.W. and A.S.A.; writing—review and editing, M.W. and A.S.A.; visualization, M.W. and A.S.A.; supervision, A.S.A.; project administration, M.W. and A.S.A. All authors have read and agreed to the published version of the manuscript.