“Number of nursing staff experiencing suicidal thoughts rises”… “Youth mental health in decline”…
Many news stories have outlined the deterioration of mental health in the UK and internationally. But the ongoing debate is: has mental health really deteriorated over the past three decades, or have our definitions and understanding of what mental health entails simply changed in the general population?
It is extremely challenging to definitively answer this question, maybe even impossible. But one of the best places to start is with repeated surveys of the same population using standardised mental health questionnaires. The study being reviewed here by Zhang et al. (2023) does just this in three nationally representative surveys across Great Britain.
Another key question is: if the mental health of the population really has deteriorated, when did this begin? If we can pinpoint the time when things began to change, we may be closer to identifying causal factors.
A researcher at San Diego State University postulates that the shift in mental health among young people happened between 2010 and 2015, around the time of rising social media use in this age group (Twenge, 2020, 2023).
Another historical event that may have affected population mental health was the global recession of 2007, followed by UK austerity measures from 2010.
The paper reviewed here (Zhang et al, 2023) tracked changes in psychological distress from 1990 to 2019, spanning a time of many major economic, social and technological changes, and there is a lot we can learn.
“Those who cannot remember the past are condemned to repeat it.” Great insight and value can be gained from exploring 30-year trends in psychological distress, and the UK has several nationally representative surveys that can be used for this purpose.
Methods
Data from three repeated cross-sectional surveys of adults (aged 16 and older) were used for this analysis:
- Understanding Society (1991–2019) – 106,417 participants from England, Scotland, and Wales
- Health Survey for England (2003–2018) – 8,000 participants from England
- Scottish Health Survey (1995–2019) – 5,000 participants from Scotland
Data was weighted to be representative of each country.
Psychological distress was measured by the widely used 12-item General Health Questionnaire (GHQ-12). The authors classified scores of 4 or more as indicating psychological distress, with higher scores representing greater psychological distress. Data was analysed using regression models.
Results
How have rates of psychological distress in adults across Great Britain changed over the past 30 years? Each dataset paints a slightly different picture (see Figure 1).
Understanding Society has the most observations, going back to 1990. It shows year-to-year fluctuation in psychological distress levels in the British population, but largely within the 17-21% range. That is, this nationally representative study suggests that roughly 1 in 5 adults in Great Britain were psychologically distressed, and that this remained true from 1990 to 2017. There was an upturn in 2015, and it is difficult to say whether it reflects a sustained increase or more fluctuation within the pre-observed range (17-21%). Future waves of this survey will tell us more.
The Health Survey for England had the shortest coverage of time, starting only in 2003. If we added a line-of-best-fit to this data, we would probably observe a gradual increase in rates of distress from 2003 to 2018. Unlike the other studies, the Health Survey for England shows a noticeable peak in distress around 2009, possibly reflecting the economic recession and austerity measures. It also showed a peak in 2016, with partial recovery in 2018, reasons for which are unclear.
Like Understanding Society, the Scottish Healthy Survey shows relative stability in levels of adult distress up until 2015/6, with a prevalence around 15%. Echoing patterns in Health Survey for England around the late 2010s, distress in this Scottish sample peaked in 2018, recovering somewhat in 2019.
Without confidence intervals or some other measure of variation in the figures provided in the study, it is difficult to say how reliable these year-to-year changes are. That said, a breakpoint analysis suggested that things changed for the worse for the Great Britain sample in 2016 and for the Scotland-only sample in 2011.
Figure 1.
Adapted from Zhang et al.’s (2023) figures, with the hatched lines showing prevalences of 15% and 20% for reference
Click to view bigger version of this image.
High-risk groups
The authors split trends by key demographic variables such as area deprivation, sex and age. Most readers will not be surprised to read that those living in more economically deprived areas reported higher levels of psychological distress. Within Scotland, it appeared that the disparity in distress between the most deprived groups and everyone else widened from ~2016 onwards. This is also evident (albeit to a lesser extent) in the Great British and England-only samples.
When split by age-group, a long-term increase in psychological distress could be observed for the youngest group (aged 16-34) from about 2010 onwards. Convergently, a decrease in distress was observed in the oldest age group (aged 65+) in the same time period. This was supported by breakpoint analyses (segmented regression).
Findings from Zhang et al. (2023) suggest that 1 in 5 adults in Great Britain were experiencing psychological distress, and that this was relatively stable from 1990 to 2017.
Conclusions
This paper by Zhang et al. (2023) gives quite a nuanced answer to the question of whether mental health has really deteriorated over the past three decades, with the take-home message being that it depends on the area of deprivation in question, the gender and age of the individual, and how you conduct the surveys. The authors concluded that:
The prevalence of psychological distress increased after 2010 in young adults and after 2015 in working- age adults across GB […] The trends in psychological distress are similar to those for overall population life expectancy and healthy life expectancy, indicating a population health challenge that predates the COVID- 19 pandemic and which demands a radical and coherent governmental response.
Has mental health really deteriorated over the past three decades? This study suggests that it depends on the sample and demographic group in question.
Strengths and limitations
The main strength of this study is its inclusion of three nationally representative surveys, which provides an opportunity to triangulate findings across different studies with different methodologies. The time range covered by these studies is also impressive, with the longest-ranging study spanning 1992 to 2015, which gives us more data points from which to establish a “norm”.
However, I have several concerns about the reliability of these findings:
- There are no error bars on trend lines in the included graphs. As such, it could be that, for example, the apparent increase in distress among young people is well within the normal range of error for each year. Relatedly, the authors don’t conduct any inferential analyses on these trends, they only describe the trends narratively, so we cannot say whether year-to-year variation exceeds within-year variation. Inferential statistics would help us say whether changes from year are over and above the “normal” fluctuations due to random error.
- There are quite notable differences across the different surveys and countries, raising important questions about how each study was conducted (e.g., sampling biases, data collection procedures, weighting, etc.), and how comparable they are. For instance, why was there greater year-to-year fluctuation in distress levels in Understanding Society compared to the other two surveys? The authors do acknowledge this in the limitations and suggest that Understanding Society would be more likely to experience attrition over time, it being a household panel survey with a longitudinal aspect.
- Another methodological limitation of the study, not mentioned by the authors but worth considering, is the utility and construct validity of the outcome: a score of 4 or more on the GHQ-12. This cut-off is arguably fairly arbitrary (why not 5? Or 6?) and the authors do not defend this choice. It would have been helpful, in my view, to see changes in the mean distress level on a continuum. The GHQ is a measure of state rather than trait – it asks individuals about changes from their norm (e.g., “Have you recently lost much sleep over worry? More than usual?”). While this could be seen as a strength of the measure, aligning it with functional impairments rather than trait-level emotionality, this may be a reason why these trends don’t align with other prevalence trends (e.g., mental disorder, chronic symptomatology).
What’s in a GHQ score of 4 or more? The authors do not defend their decision to use 4 as a cut-off on the GHQ-12, meaning that those who are categorised as experiencing psychological distress could be arbitrary.
Implications for practice
One key clinical implication from this paper is that (despite some possible recent changes) there was remarkable stability of psychological distress levels among adults over a 30-year period. Levels of significant distress have been largely between 15% and 20% since 2010 across all three samples. While this may differ for specific demographic groups, this provides a baseline target for primary and public services to be prepared for. It seems unlikely that GP surgeries and third level institutions are currently set up to cope with 15-20% of their populations presenting with psychological distress, but this paper provides evidence that this should be the case (pending the weight one places on a GHQ-12 score of 4 or more).
When writing about study implications, the authors focus on the finding that, for young people of working age (16-34), rates of psychological distress increased from the year 2010 onwards. They suggest that this may have been a reaction to austerity measures at national and local levels, and highlight that,
Governments should act to protect population health and reduce health inequalities by addressing the economic causes of these trends as well as the wider range of social determinants of health.
For me, there is another key implication from this paper for young people, for the provision of better youth mental health services. The group aged 16-34 is quite broad, including both working age adults but also adolescents and young people still in education. There has been recent concern about rising levels of depression, anxiety, self-harm and eating disorders among young people (aged ~15-25; McGorry et al., 2024). While economic factors (e.g., cost of living, housing issues) may well be contributing to this phenomenon, there is also an implication to improve preventative programmes and mental health services for this age group.
Mental healthcare splits at age 18 into child versus adult care, despite the increasing likelihood of mental illness in the late teens and early 20s, and the major shifts in societal expectations and responsibilities around this age (McGorry et al., 2022; Cannon et al., 2022). A discontinuity in care can therefore be highly disruptive at this point. Another related recommendation is the need to train more psychiatrists and other mental health professionals who specialise in diagnosing and treating those in late adolescence and their early 20s. Psychiatrists typically train in either child and adolescent care, or adult care; however, it has been proposed that there is now a need for a “youth psychiatry” subspeciality, providing clinicians with the tools to navigate the cultural, biological and educational pressures of this particular group (Cannon et al., 2022).
This paper highlights the need for improved support for young people’s mental health, and begs the question: should mental health care provision be split at the age of 18?
Statement of interests
No conflicts of interest to disclose.
Links
Primary paper
Zhang, A., Gagné, T., Walsh, D., Ciancio, A., Proto, E., & McCartney, G. (2023). Trends in psychological distress in Great Britain, 1991–2019: evidence from three representative surveys. Journal of Epidemiology & Community Health, 77(7), 468-473.
Other references
Cannon, M., Power, E., Cotter, D., & Hill, M. (2022). Youth psychiatry: time for a new sub‐specialty within psychiatry. World Psychiatry, 21(1), 2.
McGorry, P. D., Mei, C., Dalal, N., Alvarez-Jimenez, M., Blakemore, S. J., Browne, V., … & Killackey, E. (2024). The Lancet Psychiatry Commission on youth mental health. The Lancet Psychiatry, 11(9), 731-774.
McGorry, P. D., Mei, C., Chanen, A., Hodges, C., Alvarez‐Jimenez, M., & Killackey, E. (2022). Designing and scaling up integrated youth mental health care. World Psychiatry, 21(1), 61-76.
Twenge, J. M. (2023). Generations: The Real Differences Between Gen Z, Millennials, Gen X, Boomers, and Silents—and What They Mean for America’s Future. Simon and Schuster.
Twenge, J. M. (2020). Increases in depression, self‐harm, and suicide among US adolescents after 2012 and links to technology use: possible mechanisms. Psychiatric Research and Clinical Practice, 2(1), 19-25.
