Big Data

NHS Providers urges government to prioritise digital investment to boost productivity

NHS Providers has called for urgent government investment in digital technology as a means to enhance productivity, following the release of a survey involving leaders of NHS trusts. This call comes amid growing concerns about financial pressures and operational challenges facing the health service.

The survey, whose results were published on 8 September 2024, painted a stark picture of the financial health of the NHS. Trust leaders indicated that their finances have been severely impacted by a number of factors, including ongoing industrial action, the effects of inflation, demands for increased efficiency, and the overall operational strain on the NHS. These challenges have exacerbated existing financial pressures, placing trust leaders in a difficult position as they attempt to manage their budgets for the upcoming year.

The survey found that a significant 51% of respondents were “extremely concerned” about their ability to meet operational priorities within their respective organisation’s financial budget for the 2024/25 fiscal year. This concern highlights the precarious financial state of many NHS trusts as they face increasing demand for services with limited resources.

Moreover, 92% of those surveyed expressed the view that the efficiency challenge for 2024/25 is more daunting than in the previous year, with the need to do more with less becoming ever more critical. This sentiment reflects the growing pressure on NHS leaders to find ways to increase efficiency in an already overstretched system.

When it comes to meeting recovery targets, trust leaders were notably divided in their confidence levels. While almost a third of respondents (32%) were optimistic that their system would achieve its targets to reduce waiting times for physical health services in 2024/25, only 8% felt confident that similar targets for mental health services would be met. This disparity underscores the particular difficulties the NHS faces in addressing long waits in mental health care, a sector that continues to experience significant challenges.

Respondents also shared their views on accident and emergency (A&E) services, with 46% of those providing these services indicating that their trust is likely to meet the new target of seeing 78% of A&E attendees within four hours by March 2025. This reflects the ongoing challenges in delivering timely emergency care, as the NHS grapples with rising demand and workforce shortages.

On the financial front, the outlook is similarly divided. A significant proportion of respondents (44%) forecasted a deficit for their trust’s financial position in 2024/25, while 45% predicted a breakeven position, and only 11% anticipated a surplus. These figures highlight the fragile financial footing on which many NHS trusts currently stand.

Sir Julian Hartley, Chief Executive of NHS Providers, praised the resilience and adaptability of trust leaders and their teams in the face of these considerable challenges. He stated: “It’s incredibly impressive how trust leaders and their teams are responding to the productivity challenge. You can see the benefits of collaboration, incentivising staff, and using digital tools to help them work more efficiently.” However, Sir Julian emphasised that these efforts, while commendable, can only achieve so much without substantial strategic investment.

He continued: “That’s all happening, but it can only go so far without significant strategic investment in infrastructure and digital technologies. We need to recognise the value of investment in the NHS – its impact as a catalyst for economic growth, employment, and research, and the importance of funding infrastructure, digital, and other technology to drive productivity.”

The timing of these survey findings is significant, coming just ahead of a much-anticipated review of the NHS by esteemed NHS surgeon and independent peer, Lord Ara Darzi. Commissioned by Health Secretary Wes Streeting, this review is expected to be published on 12 September 2024. Early indications suggest that the report will include sharp criticism of the NHS’s productivity levels, as well as concerns about the state of children’s health.

Prime Minister Sir Keir Starmer also weighed in on the state of the NHS during an interview broadcast by the BBC on 8 September 2024. In his remarks, the Prime Minister described the NHS as having been “broken” by successive Conservative governments. He referenced the upcoming review, underscoring the severity of the challenges facing the NHS.

Responding to the Prime Minister’s comments, Sir Hartley expressed that these concerns mirrored those frequently raised by NHS trust leaders. He said: “The Prime Minister’s comments today echo what trust leaders have been telling us. The shockwaves of the longest and deepest squeeze in NHS financial history, a growing mismatch between capacity and demand, major workforce challenges, and the after-effects of the Covid-19 pandemic are still being felt throughout the health service.”

Sir Hartley went on to highlight the need for sustained, long-term financial investment in the NHS, pointing out that the current approach to funding is unsustainable. He stated: “Time and again, trust leaders tell us they want to see long-term, multi-year investment in the health service which allows them to plan for the future instead of this stop-start approach to NHS funding which leaves them constantly worrying about budget cuts followed by quick-fix, short-term funding announcements.”

Adding further weight to the conversation around NHS productivity, a report published by the Health Foundation in April 2024 examined clinicians’ views on the technologies that could offer the greatest opportunities to enhance productivity within the NHS. The report underscored the potential of digital tools to drive efficiencies, streamline operations, and improve patient outcomes, aligning closely with the calls from NHS Providers for greater investment in this area.

In conclusion, the message from NHS Providers is clear: while trust leaders and their teams are doing all they can to meet the challenges they face, there is a pressing need for the government to step up with significant investment in digital technologies and infrastructure. Such investment is critical not only for improving productivity but also for securing the long-term future of the NHS as it continues to navigate unprecedented operational and financial pressures.

Researchers apply innovative mathematical technique to help combat obesity

Researchers from Katz School and UMass Dartmouth have unveiled a groundbreaking method that employs an advanced mathematical technique to deepen understanding of the factors influencing weight loss. Detailed in their white paper, “A Choquet-Integral Based Approach to Identify Weight Loss Component Subsets,” their research was showcased at the IEEE/ACM International Conference on Connected Health: Applications, Systems, and Engineering Technologies (CHASE) in June.

Traditionally used in theoretical mathematics and economics, the Choquet Integral is now making strides in the biomedical field. This technique provides a sophisticated method for targeting and optimising various metrics and factors affecting weight loss, showing promise in the practical application to health data.

The innovative use of the Choquet Integral has enabled the researchers to identify key health factors critical for weight loss, improving the precision and efficiency of their study methods over traditional approaches. This method not only minimises errors but also enhances the extraction of useful information by concentrating on the most pertinent data. The Choquet Integral utilises a fuzzy measure, which unlike traditional metrics that provide definite values such as size or length, can manage uncertainty and overlaps. This allows for a more adaptable evaluation of complex, interrelated data sets.

Matthew Fried, the leading author of the paper and a Ph.D. student at Katz School under the mentorship of Dr. Honggang Wang, explained, “We believe this methodology could pave the way for more efficient and accurate health data analysis, ultimately contributing to better health outcomes and advancing the fight against obesity.”

The methodology was rigorously tested across four distinct datasets, including random numbers, fabricated data, standard heart data from the UC Irvine Libraries, and National Institutes of Health (NIH) health data. The technique proved effective in differentiating between actual data and noise, underscoring its suitability in modelling interactive features and measuring variables like insulin and glucose levels, LDL (bad cholesterol), HDL (good cholesterol), height, and more.

Dr. Honggang Wang, co-author and chair of the Graduate Department of Computer Science and Engineering at Katz School, stated, “We studied how different health factors affect each other, whether positively or negatively, using this special mathematical approach. This method helped us understand more clearly which biological factors are most important for weight loss.”

This cross-disciplinary application of the Choquet Integral to health data analysis not only enhances the efficiency of machine learning models by selecting reduced versions of power sets but also has the potential to transform weight loss studies and other areas in the biomedical field.

Hua Fang, another co-author and professor of computer & information science at the University of Massachusetts Dartmouth and UMass Chan Medical School, noted, “The benefits of this technique extend beyond weight loss studies. It has broad potential applications in various biomedical fields where analysing complex inter-variable relationships is crucial.” This marks a significant step forward in the use of mathematical models to enhance biomedical research and health care outcomes.

Machine learning enhances tracking and prediction of Parkinson’s disease symptoms

A groundbreaking study published in the IEEE Transactions on Neural Systems and Rehabilitation Engineering has introduced a novel automated system employing machine learning (ML) techniques to assess and predict the progression of Parkinson’s disease (PD). This system offers a promising enhancement in the evaluation of motor symptoms, potentially transforming how this neurodegenerative disorder is managed.

Parkinson’s disease, a condition with no known cure, is primarily managed through symptomatic treatments focusing on alleviating tremors, mood disturbances, bradykinesia (slowness of movement), and postural instability. Traditionally, clinicians have relied on the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) to gauge disease progression through stages such as mild, moderate, or advanced, and to assess patient responses to treatments. While the MDS-UPDRS Part III scoring method is deemed sensitive and reliable, it is not without its shortcomings, including its subjective nature and limited capacity to detect early-stage or prodromal PD symptoms.

Recognising these limitations, researchers have explored a digital method that utilises ML algorithms to extract movement markers from MDS-UPDRS Part III video recordings, such as those from the finger-tapping test, which evaluates limb bradykinesia. This approach has proven to yield a higher diagnostic and prognostic accuracy regarding PD severity.

Existing digital detection models often assume a uniform set of kinematic features across all stages of PD, varying only with disease severity. However, this assumption is increasingly challenged by findings suggesting that motor symptoms evolve non-uniformly throughout the disease’s progression. The recent study hypothesised that a more nuanced consideration of varying kinematic features could enhance the detection and accurate classification of PD severity at different stages.

The study analysed video data from 66 PD patients and 24 age-matched healthy controls, excluding any individuals with a history of brain tumours, strokes, or neurostimulatory implants. All PD diagnoses were validated using the United Kingdom PD Brain Bank criteria. Data collection occurred at baseline and was repeated a year later, with participants abstaining from anti-Parkinsonian medications overnight prior to the recordings. These videos not only captured motor tasks but also cognitive evaluations under the MDS-UPDRS III.

Researchers compared three classification models: a multiclass classification model that uses consistent features across all severity levels; an ordinal binary classification model that accounts for the progressive nature of the disease; and a novel tiered binary classification approach that adjusts the kinematic features considered based on the severity of symptoms.

In total, 180 videos were analysed, including 123 from PD patients. These were categorised by motor symptom severity scores of zero to three. The videos demonstrated significant variations in kinematic features, such as movement amplitude and sequence effect, which is the diminishing amplitude during repeated movements. Notably, the study found that kinematic features associated with lower severity scores differed distinctly from those linked with higher scores, supporting the hypothesis that PD-related motor impairments evolve uniquely over time.

The study identified several innovative kinematic features, including amplitude decay and variations in the speed of opening and closing movements. These features, quantifiable through video analysis, showcased differences between various severity levels with higher precision than traditional methods. The tiered binary classification model, in particular, demonstrated superior effectiveness in predicting PD severity, suggesting that a multi-stage model or a combination of models that consider different features at various disease stages could significantly enhance PD management and treatment efficacy assessments.

This machine learning-based approach to video analysis holds great potential to revolutionise PD management by enabling more accurate monitoring and quantification of motor symptoms, thus paving the way for more tailored and effective treatment strategies.

AI-generated alerts proven to reduce death risk in hospital patients

Recent research highlights the transformative role of artificial intelligence (AI) in improving patient outcomes in hospitals. AI-generated alerts, designed to notify hospital staff of any significant changes in a patient’s condition, have been found to dramatically reduce the risk of mortality.

The innovative system increases the likelihood of timely medical intervention, with patients 43% more likely to have their care escalated, markedly boosting survival rates. According to Matthew Levin, Professor of Anesthesiology, Perioperative and Pain Medicine at The Mount Sinai Hospital, this approach leverages AI and machine learning technologies to proactively address potential declines in patient health. “We wanted to see if quick alerts made by AI and machine learning, trained on many different types of patient data, could help reduce both how often patients need intensive care and their chances of dying in the hospital,” Levin explained.

Traditionally, tools like the Modified Early Warning Score (MEWS) have been used to assess the risk of clinical deterioration. However, Levin points out that the automated machine learning algorithms used in this study surpass these older methods in both accuracy and timeliness, facilitating earlier and potentially life-saving interventions.

The study evaluated the effectiveness of these AI alerts across four surgical units at The Mount Sinai Hospital in New York, encompassing a total of 2,740 patients. These individuals were divided into two groups: the first group received real-time alerts about health deterioration directly to their doctors or rapid response teams, while the second group had alerts generated but not immediately transmitted to their healthcare providers.

David Reich, President of The Mount Sinai Hospital and Mount Sinai Queens and a leading figure in both Anesthesiology and Artificial Intelligence and Human Health at Icahn Mount Sinai, highlighted the substantial impact of real-time, AI-driven alerts on patient care. “Our research shows that real-time alerts using machine learning can substantially improve patient outcomes,” Reich stated. He described these tools as “augmented intelligence” that enhance clinical decisions, ensuring that timely and appropriate interventions are made to improve patient safety and outcomes, aligning with the goals of a learning health system.

Further benefits of the AI alert system include a higher likelihood of patients receiving necessary medications for heart and circulation issues, alongside a decreased mortality rate within 30 days of care. The system is under continual development, with a dedicated team of intensive care doctors assessing high-risk patients daily, providing targeted recommendations to the treating physicians. This ongoing refinement of the algorithm, driven by an increasing amount of patient data, is enhancing its accuracy and reliability.

This study not only reaffirms the critical role of AI in modern healthcare but also points towards a future where technology and human expertise converge to deliver superior patient care.

Advancements in infant precision medicine through creation of ‘digital babies’

In an ambitious collaboration between the University of Galway and Heidelberg University, along with the Heidelberg Institute for Theoretical Studies and Heidelberg University Hospital, a breakthrough has been made in the realm of infant healthcare through the development of ‘digital babies’. These are sophisticated computational models that replicate the metabolic processes of newborn infants, aiming to enhance the precision in diagnosing and treating their medical conditions.

At the Digital Metabolic Twin Centre at the University of Galway, the team utilised real-world data gathered from a significant cohort of 10,000 newborns. This data encompassed various parameters including sex, birth weight, and metabolite concentrations. From this, the researchers successfully constructed 360 detailed whole-body computational models. These models are specifically designed to emulate the metabolic systems of infants over their first six months of life.

Published on 3 June 2024 in the journal Cell Metabolism, the findings of this study underscore the efficacy of these models. Not only were they capable of predicting known biomarkers for inherited metabolic diseases, but they also accurately forecasted the metabolic responses infants would have to different treatment methodologies.

Elaine Zaunseder, the lead author from Heidelberg University, highlighted the unique metabolic characteristics of babies, which are pivotal for their growth and health. She explained, “Babies require more energy for body temperature regulation due to their high surface-area-to-mass ratio and the fact that they cannot shiver during the first six months. Our task was to decode these metabolic activities and integrate them into mathematical frameworks within our computational models.”

Zaunseder also emphasised that this research marks a significant initial stride towards creating digital twins for infants. Such digital twins could potentially transform paediatric healthcare by offering customised disease management that aligns with the distinct metabolic needs of each infant.

Professor Ines Thiele, who led the project, stressed the importance of newborn screening programmes which are vital for the early detection of metabolic diseases, thus improving survival rates and health outcomes for infants. “However,” she noted, “the observed variability in disease manifestation among babies highlights the critical need for personalised treatment plans. Our models enable in-depth studies into the metabolism of both healthy and diseased infants, including those conditions screened for in newborns.”
Parallel to these developments, Imperial College London announced in May 2024 that their researchers are also working on digital twin models, specifically heart models for NHS patients suffering from pulmonary arterial hypertension. This indicates a broader trend towards adopting digital twin technology across various areas of healthcare, promising more targeted and effective treatment strategies.

How Finland’s Digital Health Village is redefining healthcare delivery

Launched in 2016, Finland’s Digital Health Village represents a pioneering digital service platform within the healthcare and social welfare sectors. This platform integrates digital care pathways with expert information online, designed to alleviate pressure on traditional healthcare infrastructures and improve service delivery. The initiative has not only met but surpassed its expectations, notably by enabling healthcare workers to allocate time to additional patients each day, thus enhancing operational efficiency.

Globally, healthcare systems face challenges such as ageing populations, increasing demand for services, workforce shortages, lengthening treatment queues, and financial constraints. In 2023, the Finnish government allocated over EUR 20 billion to its 21 wellbeing services counties, a substantial portion of the national budget, highlighting the growing financial demands on healthcare and social welfare. The Southern Finland healthcare system alone employs around 60,000 professionals, with the Helsinki and Uusimaa Hospital District (HUS) managing a significant workforce of 27,500. In 2023, HUS provided specialist medical care and emergency services to more than 690,000 patients, underlining the critical need for efficient service delivery mechanisms.

The Digital Health Village, developed through the national Virtual Hospital 2.0 project coordinated by HUS, serves as a central piece in Finland’s strategy to modernise healthcare. This cloud-based service was designed from the start to consolidate expertise, offer updated information, and improve patient care both digitally and locally. It includes over 30 specific hubs catering to various patient groups, featuring dependable expert information on different health conditions. The platform’s My Path service offers digital care pathways on referral, open-access self-care programs, and remote consultation capabilities. HealthVillagePRO, another integral component, provides clinical guidelines and online courses for healthcare professionals to foster skill development and the adoption of innovative work practices.

Matti Bergendahl, CEO of HUS, highlighted the transformative effect of this digital overhaul, noting that rethinking patient needs and service delivery has been crucial. The establishment of the Digital Health Village as a cloud-based platform from its inception has been particularly advantageous, allowing for cost-effective, secure, and scalable development.

The impact of this digital transition is significant, saving healthcare professionals up to 26 minutes per day, thereby freeing up time to treat additional one to two patients daily. This efficiency gain, derived from expert interviews and external impact analysis, translates to annual savings of more than EUR 42 million for Southern Finland’s healthcare system. Moreover, the broader societal benefits, such as enhanced treatment quality, reduced wait times, and economic savings due to decreased patient travel and income loss, contribute to an estimated EUR 689 million in annual savings for the region.

The platform’s user satisfaction rates are also notable, with up to 69% of users reporting improved quality of life. The frequency of service usage correlates with greater perceived benefits, underscoring the value of regular interaction with the digital platform.

Looking forward, the integration of AI and large language models is set to further personalise and enhance the Digital Health Village. Current applications include AI-driven chatbot features, and future possibilities involve leveraging AI for multilingual translation capabilities. Sirpa Arvonen, Digital Manager at HUS, is enthusiastic about the potential to improve information flow by enabling both patients and professionals to access digital resources in their preferred languages, thereby making communication more effective and inclusive.

Pioneering AI tool developed by NHS teams to transform patient care

A ground-breaking artificial intelligence (AI) tool, designed to forecast patient health trajectories, has been developed by a collaboration of healthcare researchers within the NHS. Dubbed “Foresight,” this innovative tool is poised to revolutionise clinical decision-making processes, enhance monitoring in healthcare environments, and bolster clinical trials.

The development team is composed of specialists from two NHS foundation trusts in London—King’s College Hospital and Guy’s and St Thomas’—as well as academic experts from King’s College and University College London. Utilising the Cogstack platform, which is renowned for its capability in information retrieval and extraction, Foresight leverages natural language processing to efficiently mine data from NHS electronic health records. This allows the tool to be trained on vast amounts of healthcare data, employing a deep learning methodology to identify intricate patterns within both structured and unstructured data sources.

A critical evaluation published in The Lancet Digital Health illustrates Foresight’s effectiveness: the tool successfully predicted the next ten possible health disorders in a patient’s timeline with impressive accuracy rates—68% at King’s College Hospital, 76% at Maudsley NHS Foundation Trust, and an outstanding 88% with the US-based MIMIC-III dataset.

The tool’s utility was further underscored through a practical test where five clinicians created 34 hypothetical patient timelines based on simulated scenarios. An impressive 93% of the predictions made by Foresight were deemed clinically relevant, affirming their practical applicability in real-world settings.

According to a report from the NIHR Maudsley Biomedical Research Centre, Foresight’s capabilities are not limited to forecasting; it can also emulate clinical trials, facilitate longitudinal research, generate synthetic datasets, and simulate interventions to study disease progression. Professor Richard Dobson, a leading figure in medical informatics at King’s College London and the senior author of the study, expressed enthusiasm about the multitude of applications for Foresight. He highlighted its potential in creating digital health twins and advancing medical education, among other uses.

Professor Dobson emphasised the importance of employing high-quality data to refine AI models and expressed a vision for expanded collaboration. His aim is to involve more hospitals in the development of “Foresight 2,” an iteration that promises even greater accuracy through enhanced language models.

This initiative has garnered substantial support, receiving funding from the NHS AI Lab, the National Institute for Health and Care Research, and Health Data Research UK (HDRUK). Professor Andrew Morris, director of HDRUK, noted that the success of such innovations hinges on the quality and representativeness of the data used. He advocated for continued investment in the UK’s data infrastructure to ensure these advancements can be realised in a manner that is both secure and respectful of patient privacy.

NHS adopts AI to combat absenteeism and expedite elective care waiting times

The National Health Service (NHS) is on the brink of an innovative leap, deploying artificial intelligence (AI) across an additional ten trusts with the aim of curtailing missed appointments and thus liberating valuable staff hours. This strategic move is anticipated to significantly dent the backlog in elective care waiting lists.

This initiative’s expansion comes in the wake of a triumphant pilot programme at the Mid and South Essex NHS Foundation Trust. This particular programme witnessed a near one-third reduction in patient no-shows within a mere six-month span.

The pioneering software, a collaborative creation between Deep Medical and contributions from both a frontline worker and an NHS clinical fellow, utilises algorithms alongside anonymised patient data to foresee potential appointment absences. It ingeniously taps into various external factors, such as weather conditions, traffic situations, and employment statuses, to deduce possible reasons behind a patient’s failure to attend. To counteract these challenges, it proposes alternative booking options that align more closely with the patient’s availability, including after-hours and weekend slots for those unable to take daytime leave.

Moreover, the software cleverly incorporates a system of intelligent backup bookings, ensuring that no clinical time goes to waste and optimising overall operational efficiency.

The six-month trial at the Mid and South Essex NHS Foundation Trust produced remarkable results: a 30% decrease in no-shows, 377 prevented missed appointments, and an additional 1,910 patients seen. Given these outcomes, projections suggest the trust could save approximately £27.5 million annually by persisting with this programme, benefiting a population of 1.2 million.

Deep Medical, under the co-founder duo Dr Benyamin Deldar and AI connoisseur David Hanbury, is at the forefront of this technological advancement. Dr Deldar highlights the software’s dual benefit: drastically reducing missed appointments and repurposing these slots for other patients, thereby enhancing both financial savings and public healthcare delivery.

The imminent roll-out to an additional ten trusts across England marks a significant step forward in this AI-powered journey.

In a concerted effort to recuperate elective care services post-pandemic and address prolonged waits for routine procedures, the NHS is embracing cutting-edge technologies and innovations, including AI. This approach aims to tackle the prevalent issue of missed hospital appointments, which amounts to hundreds of thousands each month, thus ensuring more judicious use of clinical time and expedited access to care for waiting list patients.

Annual statistics reveal a startling 6.4% no-show rate among the 124.5 million outpatient appointments across NHS England, translating into a financial strain of £1.2 billion.

The analysis also uncovers that physiotherapy appointments bear the brunt of absenteeism, with an 11% no-show rate, followed closely by cardiology, ophthalmology, and trauma and orthopaedics.

Dr Vin Diwakar, NHS England’s National Director for Transformation, praises the NHS’s innovative spirit and its openness to novel operational methods that ensure timely patient care. He underscores the AI initiative’s capacity to not only refine patient services but also to foster a more economical utilisation of taxpayer funds.

Moreover, he emphasises the role of such AI pilots in empowering patients to manage their healthcare more effectively and in addressing health inequalities.

The University Hospitals Coventry and Warwickshire (UHCW) NHS Trust exemplifies another facet of AI application through ‘process mining’. This technique offers insights into the efficacy of existing processes, spotlighting bottlenecks and areas ripe for improvement.

Notably, during its pilot, the Trust identified a correlation between high deprivation scores and increased DNAs, with a marked surge in last-minute cancellations post two SMS reminders. Adapting their strategy to send reminders 14 days and then four days prior to an appointment significantly reduced no-show rates from 10% to 4% among a targeted patient group.

Encouraged by these results, the Trust is now exploring the application of process mining to theatre scheduling, aiming for further efficiency gains and enhancements

Mayo Clinic teams up with Cerebras Systems to advance healthcare AI

The Mayo Clinic, a renowned non-profit medical institution based in Rochester, Minnesota, announced a strategic partnership with the Silicon Valley-based startup Cerebras Systems on Monday (15th of January, 2024). This collaboration aims to harness the power of artificial intelligence (AI) in enhancing healthcare services.

With significant presences across three major campuses in the United States and additional facilities in the United Kingdom and the United Arab Emirates, the Mayo Clinic is set to utilise cutting-edge computing chips and systems supplied by Cerebras. The collaboration will delve into the Mayo Clinic’s extensive archive of anonymised medical records and data, laying the groundwork for the development of bespoke AI models.

According to Matthew Callstrom, Mayo’s Medical Director for Strategy and Chair of the Radiology Department, these AI models are poised to revolutionise various aspects of medical record management and diagnostics. Some models are being designed to interpret and summarise extensive medical records for new patients, streamlining the patient onboarding process. Additionally, other models will focus on identifying intricate patterns in medical imagery and genome data, patterns that might elude even the most experienced medical professionals. However, Callstrom emphasised that these systems are intended to augment medical decision-making, not replace it. The human expertise of doctors remains paramount in the clinical decision-making process.

“The integration of AI is about enhancing the decision-making process for each patient, considering the multitude of factors and drawing upon extensive experience,” Callstrom explained.

The Mayo Clinic’s collaboration with Cerebras is anticipated to yield results that will be accessible via the Mayo Clinic Platform. This platform is a comprehensive data network already utilised by various healthcare systems such as Mercy in the U.S., the University Health Network in Canada, and other systems in Brazil and Israel.

While the pricing model for the AI technology developed through this partnership has not been finalised, Callstrom indicated that the Mayo Clinic intends to disclose further details about this venture during a presentation at JPMorgan Chase’s healthcare conference in San Francisco.

Andrew Feldman, CEO of Cerebras, described the deal as a multi-million-dollar agreement spanning several years, though he refrained from disclosing specific financial details. Cerebras, aspiring to compete with industry leaders like Nvidia, will provide both the necessary hardware and software development expertise as part of the agreement. This collaboration marks a significant step in integrating advanced AI capabilities into the healthcare sector, aiming to enhance patient care through technological innovation.

New machine learning approach transforms behavioural health medication practices

At the recent AMCP Nexus 2023 conference in Orlando, Florida (16th-19th of October), presenters showcased a groundbreaking machine learning program designed to address medication-related issues in individuals with behavioural health conditions. This program has shown promising results in reducing polypharmacy, enhancing medication adherence, and decreasing healthcare costs.

Behavioural health conditions pose a significant challenge to healthcare systems. A 2020 Milliman study examining commercial healthcare claims data from 2017, which encompassed 21 million people, revealed that although only 27% had a behavioural health condition, they accounted for over half of the total healthcare expenditures. In this context, machine learning offers a potential solution for better managing these conditions.

The presenters highlighted the issue of polypharmacy, a common concern in behavioural health where 60% of adults with a condition are prescribed two or more psychotropic medications. Polypharmacy not only increases the risk of drug interactions and adverse events but also contributes to soaring healthcare costs. Dr. Caroline Carney, Chief Medical Officer at Magellan Health, underscored the tendency for medication overlap and overprescription in treating conditions like depression and anxiety, often leading to unnecessary medication layers.

Another issue in managing behavioural health medications is the multiple prescribers involved, including primary care doctors, specialists, and inpatient clinicians, often resulting in uncoordinated treatment. This lack of coordination can leave patients confused and overwhelmed with their medication regimens.

To combat these challenges, Magellan Health collaborated with Arine, a medication management tech startup, to create inforMED (formerly known as Navigate Whole Health). This AI-driven program identifies prescribers who can potentially optimise patient care, generating comprehensive care plans with treatment recommendations and patient education. The program’s effectiveness is continuously improved by incorporating new clinical outcome data.

Dr. Carney elaborated on the program’s approach, which considers hundreds of parameters, providing prescribers not just with medication change suggestions but also with reasons, implications, and evidence-based support for the recommended changes.

Yoona Kim, PharmD, PhD, co-founder and CEO of Arine, explained that machine learning algorithms are utilised to target prescribers based on their prescribing patterns and the presence of prescribing outliers in their patient panels. The program also considers social determinants of health, using ZIP code data to assess potential barriers to healthcare access, such as low income or lack of vehicle access.

The results of this program have been significant. Dr. Kim reported a reduction in behavioural health polypharmacy by 45% to 55%, a 20% increase in medication adherence, a 20% reduction in average daily morphine milligram equivalents, and a savings of $360 to $840 in pharmaceutical costs per enrolled member annually.

Dr. Carney emphasised the program’s success in providing actionable data and guidance to healthcare providers, leading to improved patient outcomes and stronger, longer-lasting professional relationships. This innovative approach signifies a major step forward in the management of medication for behavioural health conditions.

NHS England’s controversial £330m data deal with Palantir draws mixed reactions

NHS England’s recent decision to award the £330 million Federated Data Platform (FDP) contract to US data analytics firm Palantir, in collaboration with Accenture, PwC, NECS, and Carnall Farrar, has elicited a spectrum of reactions. This announcement marks a significant step in NHS’s digital transformation efforts, with the FDP designed to connect and streamline access to healthcare data across the NHS.

Matthew Taylor, CEO of the NHS Confederation, acknowledged the FDP’s potential in enhancing care delivery by freeing up clinical time and fostering efficient, safer patient care. However, he emphasised the need for substantial efforts to garner public support for the initiative.

Contrasting opinions emerged from within the NHS. Paul Jones, Chair of the Digital Health Networks CIO Advisory Panel, expressed disappointment at NHS England’s decision to proceed despite reservations from trust digital teams and the financial constraints on other NHS digital budgets.

Nick Wilson, CEO at System C, recognised the wealth of experience within the NHS and technology companies in digital transformation but voiced disappointment over the contract not being awarded to a British consortium. He also highlighted concerns about the exclusion of GP data from the FDP, stressing the complexities of interoperability in health and social care and urging Palantir and Accenture to learn from past challenges.

The decision faced criticism from those who had campaigned against Palantir’s involvement. The Good Law Project, a not-for-profit campaign organisation, is preparing legal challenges to ensure proper handling of sensitive NHS data, focusing on maintaining privacy.

Cori Crider, Director of Foxglove, a legal advocacy organisation, raised questions about the FDP’s effectiveness, citing unsuccessful hospital trials of Palantir’s technology. Dr David Nicholl, spokesperson for Doctors’ Association UK, echoed these concerns, questioning the scrutiny around the deal and the preliminary results of NHS trials with Palantir’s technology.

David Davis, MP for Haltemprice and Howden, expressed reservations about Palantir’s suitability for handling sensitive data, citing their background in espionage and concerns about data protection.

In response to these varied perspectives, NHS England has garnered support from several organisations, including the NHS Confederation, National Voices, and the Academy of Royal Medical Colleges. National Voices CEO Jacob Lant and Dr Jeanette Dickson, Chair of the Academy of Medical Royal Colleges, acknowledged the FDP’s potential in driving digital innovation and enhancing data connectivity in the NHS.

To address public concerns, NHS England is developing an engagement portal for the public to learn more about the FDP and submit queries. Additionally, Dr Nicola Byrne, National Data Guardian, and Dr Nicola Perrin of the Association of Medical Research Charities have joined the independent Check and Challenge Group for the FDP, overseen by NHS England.

Furthermore, NHS England has assured that data sharing under the FDP will not commence until new ‘Privacy Enhancing Technologies’ (PET) are developed and implemented, expected by April 2024. Details on these technologies, being developed by a separate supplier, are anticipated to be released later in the year.

Personalised healthcare revolution: The rise of digital twins in medicine

The trajectory of medicine is being redefined by pioneering research into computational models, advancing towards a future where medical treatments are tailored not to the average patient, but to each individual. Envision possessing a ‘digital twin’—a virtual counterpart that can undergo trials and treatments, sparing you the need for direct medication or surgical intervention. Scientists project that within the next decade, we could witness the routine use of ‘in silico’ trials, utilising virtual organs to evaluate drug safety and effectiveness, while bespoke organ models might be employed to customise patient care and avert medical complications.

Digital twins represent sophisticated computer-generated replicas of physical entities or processes, continuously refined with data from their actual counterparts. In the medical realm, this entails the fusion of extensive biological data—including genetic, proteomic, cellular, and systemic information—with individual patient data to craft detailed virtual models of their organs, and potentially, in time, their entire body.

Professor Peter Coveney, Director of the Centre for Computational Science at University College London and co-author of ‘Virtual You’, suggests that much of current medical practice lacks a scientific underpinning. He compares it to navigating by looking in the rear-view mirror—basing treatment for the patient at hand on historical cases. “A digital twin utilises your own data within a model that encapsulates your unique physiology and pathology. It’s a move away from decisions based on potentially unrepresentative population data to truly personalised medicine,” explains Prof. Coveney.

Cardiology is at the forefront of this cutting-edge model. Companies are already harnessing patient-specific heart models to aid in the design of medical devices. Meanwhile, the Barcelona-based enterprise ELEM BioTech is at the forefront, granting companies the capability to test drugs and devices on simulated human hearts. “We’ve conducted numerous virtual human trials on several compounds and are on the cusp of launching a new phase, with our cloud-ready product accessible to pharmaceutical clients,” shares Chris Morton, co-founder and CEO of ELEM.

At the recent Digital Twins conference hosted by the Royal Society of Medicine in London, Dr. Caroline Roney from Queen Mary University of London detailed the development of tailored heart models which could significantly aid surgeons in planning interventions for atrial fibrillation patients. “Surgeons typically resort to average-based approaches, but crafting patient-specific predictions that forecast long-term outcomes remains a formidable challenge,” Dr. Roney stated. She foresees widespread application of this technology in cardiovascular treatments, including decisions on valve selection and placement during replacements.

The field of oncology is also poised to benefit from digital twins. Teams from GSK and King’s College London are joining forces to construct virtual duplicates of patient tumours, amalgamating imaging, genetic, and molecular data with 3D cultures of cancer cells, and observing their drug responses. Leveraging machine learning, researchers can foresee how individual patients may react to various treatments, drug combinations, and dosages. “Conducting repetitive trials on a real patient with multiple treatments isn’t viable. Our aim is to devise a strategy while the patient is still with us, preparing us for any recurrence of cancer,” said Professor Tony Ng from King’s College.

The advent of digital twins extends even to the realm of pregnancy, offering the potential to develop treatments for conditions such as placental insufficiency or pre-eclampsia, and deepening our grasp of pregnancy and labour physiology. Professor Michelle Oyen, Director of the Center for Women’s Health Engineering at Washington University in St Louis, is crafting placenta models from ultrasound scans and post-birth high-resolution imagery to predict complications during pregnancy. “We’re striving to identify measures in a live person that could forewarn us of placental issues, aiming to preempt adverse outcomes like stillbirth,” Prof. Oyen elucidates.

In collaboration, Professor Kristin Myers from Columbia University is modelling the cervix, uterus, and foetal membranes, with the overarching goal to merge these into a comprehensive individual model to predict pregnancy outcomes. “We hope to analyse a simple ultrasound scan to understand how the uterus will adapt and when labour might occur,” Prof. Myers aspires, potentially guiding decisions on interventions like caesarean sections.

Moreover, the concept of digital twins is being expanded to model entire hospitals to enhance patient flow and healthcare system efficiency. Dr. Jacob Koris, a trauma and orthopaedic surgeon and digital lead at Getting It Right First Time, describes how tracking digital footprints left by patient interactions—from X-rays to outpatient appointments—can provide a granular, real-time view of patient treatment pathways. “Such insights could pinpoint areas for improvement and exemplary practices that could revolutionise patient care,” Dr. Koris believes.

This ambitious step forward in computational medicine promises a leap from the traditional, one-size-fits-all model to a future where every treatment is as unique as the patient it serves.