CRIS blog: Loneliness – so obviously important but very under-reported

Robert Stewart, Professor of Psychiatric Epidemiology at King’s College London and Clinical and Population Informatics Lead at the NIHR Maudsley Biomedical Research Centre writes about the work of his team to identify loneliness in healthcare records using natural language processing, and how this can help improve our understanding of its relationship to mental health.

Loneliness is clearly important in mental healthcare, particularly in large cities where so many people are living in quite isolated circumstances. There’s no argument, really, about this. Loneliness, of course, isn’t the same as social isolation – you can be lonely with lots of friends and family around you, or you can live in quite isolated circumstances and be perfectly happy that way. Loneliness describes a state of being unhappy with your social environment and clearly has important implications for mental health.

The relationship between loneliness and mental health is probably quite complicated. Being lonely can clearly cause someone’s mental health to get worse. On the other hand, having a mental health condition may result in loss of social contacts and loneliness as a result.

The important thing is that loneliness matters but isn’t generally available as information for health services. If you asked the head of a mental health service how many patients receiving care were lonely, they would struggle to provide an answer.

A new way to identify loneliness in medical records

The NIHR Maudsley Biomedical Research Centre Clinical Record Interactive Search (CRIS) team have been working for over ten years to ‘unlock’ this sort of information for mental health services.

A lot of detailed information is routinely collected in mental healthcare on patients’ symptoms and other relevant circumstances. This tends to be recorded in text (e.g., case notes or letters) rather than as tick-boxes or other structured fields. However, we have been able to develop simple automatic mechanisms for recognising pieces of information using a technique called ‘natural language processing’ and can now capture a very wide range of different features from the symptoms that people are experiencing, to the treatments that they are receiving.

The natural language processing application for loneliness (see page 205 in the NLP catalogue) was developed in CRIS by an MSc student, Mayur Parmar, working with Gayan Perera, Jyoti Sanyal, and other experts in the CRIS team. This work was begun in April 2020 as one of the first of our responses to the COVID-19 pandemic and lockdown. It seemed fairly clear that social distancing might have consequences on loneliness for people receiving mental health services; however, at the time, we were not able to capture this. We were able to finish the development by early June 2020 and research projects using the app were being completed by the end of the summer.

Discovering links between certain mental health conditions and loneliness

Initial work, carried out by Mayur Parmar for his dissertation and later published, looked at loneliness across mental healthcare and its relationship with general hospital admissions.

The research found that loneliness recorded in case notes was most common in people with eating disorders and severe mental illnesses such as schizophrenia and bipolar disorder. This could be due to the stigma surrounding these conditions, but more research is needed to understand the relationships.

People recorded as being lonely were more likely to have a general hospital admission and, if they were admitted, had longer periods in hospital. This illustrates the importance of identifying loneliness, as those affected may have unmet healthcare needs.

Older adults and the impact of the pandemic

Further work on loneliness - led by our colleague Flora Greig, a Consultant Old Age Psychiatrist at South London and Maudsley NHS Foundation Trust - was published more recently. This study showed that loneliness in older people receiving mental healthcare was more commonly recorded during the first UK lockdown than the same months in 2019 and that it had become more strongly linked with risk-taking behaviour such as non-accidental self-injury and alcohol or drug use.

The research concluded that addressing and preventing loneliness should be a priority and as we start to think about how we ‘live with COVID’, we need to consider the impact of further social distancing measures on this already vulnerable population, with an emphasis on investigating ways to possibly mitigate against these.

Recognising the importance of loneliness

As well as their specific findings, both these studies emphasise the value of being able to capture loneliness from mental healthcare records in the first place.

Real experiences of loneliness will probably always be underestimated because they won’t necessarily be picked up or recorded in health services. By its very nature loneliness is an experience that disconnects us from society and can prevent us from seeking help.  However, the more that we recognise its importance and make the information about the impact of loneliness available, the more likely it is to receive attention.

Mental Health Awareness Week is taking place 9 - 15 May 2022 and the theme is loneliness. Follow us on Twitter to find out more about our research into loneliness.