Inpatient mental health: a digital desert – a blog by Karen West and Professor Dan Joyce

Summary

In October, the Healthcare Safety Investigation Branch will tweak its name, become independent from NHS England and the UK government, and gain new powers to strengthen investigations. With the announcement of the change in status, Health Secretary, Steven Barclay, reported it would be leading an investigation into inpatient mental health. This follows swiftly on the heels of the Strathdee rapid review into data on mental health inpatient settings, which itself was launched in response to well-documented failures in these settings. The aim of this new investigation into mental health is simple: to improve safety.

In this blog, Karen West, Head of Transformation (Mental health) at Oxehealth, and Professor Dan Joyce from the University of Liverpool, discuss the importance of data in patient safety improvement and explain why inpatient mental health data is so difficult to collect and what can be done to improve this.

Content

The importance of data on patient safety

So where will the investigation start? As with all reviews, it will start with data collection – gathering evidence from national partners, staff, patients and families – to understand the risks to mental health patients and learn what and how to change to drive safety.

Good data is fundamental to safety. The recommendations in Strathdee’s report made this link clear:

“The aim of improving mental health information and information systems is ultimately about keeping patients safe and providing consistently high quality, evidence-based therapeutic treatments that enable patients to achieve the outcomes they need to have the quality of life they want back in their communities.”[1]

However, the type of data you collect is important. Strathdee found:

“patients and families and clinical teams told us that the [healthcare] system is not measuring what matters. They do not consider we [NHS England] are measuring what will truly have an impact on patient safety and outcomes.”[1]

Inpatient mental healthcare: a digital desert?

But what if this data is sparse. In mental healthcare, objective, longitudinal data about “patient and clinician reported progress” is limited, especially in inpatient settings. Professor Dan Joyce explains: “As a psychiatrist working on inpatient wards, I often found that I would only have limited time per week with each patient, so the information I had about how a patient was doing was formed from those short contacts, which may not be indicative of their overall progress. Information that would help me understand how a patient was doing – which would be automatically available in a general hospital – was really limited.” 

Without technology to support them, time-strapped staff may find data collection a burden. And data is likely to be unstructured and incomplete, particularly if it is collected as handwritten notes on paper forms. Dan provides an example of this challenge: “I was responsible for a patient with bipolar affective disorder. We knew him well and had often admitted him in a manic state. We knew that he would often stop taking his regular medication that maintained his wellness, so we put him back on his usual stabilising medication as this had been proven to work previously. But after 10 days, we were still struggling to contain symptoms of mania. It was clear to me that he was not sleeping and resting. So I asked the nursing team to complete a sleep inventory. However, after two weeks, it was clear that staff were simply too busy to observe and collect this detailed information. So on a night shift, I sat with him for some time, observing his activity around sleeping. I was able to confirm that his sleep was significantly dysregulated around sleep initiation, so with that data we were able to adjust his care plan to 'reset' his sleep and he showed a marked recovery. The data was essential in being able to provide the right care”.

In general inpatient healthcare, staff will have a plethora of technology to support data collection, which not only reduces the burden on them, but provides a rich source of data to track patient improvement or deterioration, responses to care and to support learning.

In contrast, inpatient mental healthcare appears to be a digital desert; however, at the same time, the Strathdee report found some clinicians were spending half of their time entering data. This paradoxical situation results in what Strathdee summarised as being:

“Too much data collection is about activity and processes and too little about patient experience, what therapeutic treatments are provided, and the ‘real time’ patient and clinician reported progress and outcomes.”[1]

Limited patient-focused data restricts the ability to gain actionable insights that are fundamental to providing better care. For example, less information about how a person is doing will:

• Significantly limit the ability to inform personalised care plans, such as how someone reacts to specific interventions.
• Impact the ability to understand specific risks – relevant to the patient – and put in place appropriate risk management, such as a falls risk management plan for frail patients.
• Impact the ability to understand if the patient is responding to care or deteriorating, and whether that necessitates changes to care as noted in our example above.
• Limit the ability to characterise a patient’s condition and understand what does and does not work to learn how to improve the care for other patients with similar presentations or conditions.

With high staff vacancy rates (around 20%) on inpatient mental health wards[2] and increasing demand, it’s not surprising that in many cases where safety has been compromised, the CQC's reports point to insufficient staffing. But increased staffing is not a quick fix, especially if these staff must rely on existing processes and data for understanding how the care they provide impacts on a patient's wellbeing and recovery. Investing in staff training, recruitment and retention is essential to improve safety but it will take a long time. Which is no doubt why cross-governmental parties have called on technology to help.[3]

“Embracing innovation will keep the NHS fit for the future…. That's why smart use of tech is a key part of our NHS Long Term Workforce Plan.” Steve Barclay, Secretary of State for Health.

Entering the era of AI

No one can have failed to notice the flood of interest in AI, particularly in healthcare. Its use in physical healthcare is already providing significant advances; for example, using image analysis to diagnose strokes[4] or assess mammograms; using virtual assistants to book appointments or use in remote monitoring supporting people in virtual wards. It also has tremendous potential as a tool for improving safety.[5]

But contemporary AI is fundamentally a data-driven technology, requiring large, purposefully-curated and representative data. So if we are not collecting the right data, we cannot even begin to apply recent technological advances to develop algorithms that will either assist with detailed analysis of complex data or help automate prone-to-human-error and mundane tasks that free up specialist healthcare professionals to do what they do best.

Investment in technology to support and improve data collection that enables the advances that data-driven technology could bring would, therefore, seem like a no brainer. But you can’t just “lift and shift” tech from the general healthcare setting into mental healthcare. Unlike the familiar investigations of general medicine, such as an ECG, MRI imaging or vital sign measurements, improving our understanding of how to help people with mental illness may involve new technology that exposes tensions between the needs of professionals and the rights of individuals. Emerging technology – and how best to deploy it in sensitive healthcare settings such as mental health – provides new challenges for healthcare that are only now being realised and addressed. But the opportunity to significantly improve safety and how patients are treated effectively is worth it and we should not shy away from the challenge if we are to bridge the gap between physical and mental healthcare.

Technology is not a panacea but it is a tool that can, when implemented carefully, support safer care. As former CNIO for NHS England, Professor Natasha Phillips wrote recently in a very personal blog:

“The reality is we cannot be there all the time with our patients and so we must harness all the technology available to support and keep them safe.  I truly believe that access to timely information and the use of data as clinical decision support, artificial intelligence and robotic process automation can enable us to target care more effectively and be more purposeful and person centred in our time with patients.”[6]

References

1. Department of Health and Social Care: Independent report. Rapid review into data on mental health inpatient settings: final report and recommendations. 28 June 2023.
2. Committee of Public Accounts. Progress in improving NHS mental health services. Sixty-Fifth Report of Session 2022–23. UK Parliament. 21 July 2023.
3. Policy Mogul. Wes Streeting speech to NHS Confederation. 14 June 2023.
4. Department of Health and Social Care and The Rt Hon Steve Barclay MP. Press release. Artificial intelligence revolutionising NHS stroke care. 27 December 2022.
5. Bates DW, Levine D, Syrowatka A, et al. The potential of artificial intelligence to improve patient safety: a scoping review. Nature 2021; 4(54).
6. Natasha Philips. Summer of change. LinkdedIn. 4 September 2023.

About the Author

Karen West is Head of Transformation for mental health at Oxehealth.  Oxehealth works alongside patients, doctors and nurses to develop technology to transform the diagnosis, care and experience of mental health patients. In contrast to physical health, mental health has not had the benefit of significant technology investment and it was the opportunity to address this gap that attracted Karen to join Oxehealth in April 2022. Prior to her role at Oxehealth, Karen was part of the founding team at start up healthcare quality app, Tendable (Perfect Ward). Karen led on business development for several years and developed a passion for supporting frontline staff in utilising technology to improve and empower them to drive quality improvements.

Dan Joyce is a consultant psychiatrist and professor of connected mental health, researching how data science and machine learning can be applied to improving the detection and treatment of mental illness. Before moving to the University of Liverpool in 2022, he worked at Oxford Health NHS Foundation Trust as a consultant psychiatrist in Early Intervention for Psychosis, clinical trial delivery and as a Senior Clinical Research Fellow at the University of Oxford's Department of Psychiatry working on applications of statistical and machine learning to develop deployable technology for clinical decision support.