The influence of the COVID-19 pandemic and the first UK lockdown on older age presentations to a regional psychiatric liaison service: A retrospective observational study.

Background: Loneliness and social isolation may have increased during the coronavirus disease 2019 (COVID-19) pandemic, possibly increasing mental health problems. However, due to fears of contracting COVID-19, patients may not have presented to hospitals. Aims: We assessed the impact of the COVID-19 pandemic measures and lockdown on elderly (≥65 years) presentations to a psychiatric liasion service. Settings and Design: A UK regional liasion psychiatry service. Materials and
Methods: Mixed methods methodologies were used to assess data between 01/01/20 and 31/08/20, which were compared to data from the same time period in 2019. Statistical and thematic analyses were conducted to understand drivers of presentations. Statistical Analysis: Nonparametric testing and analysis of variance (ANOVA).
Results: Presentations to liaison psychiatry reduced by 13% for 01-08/2020 compared to 01-08/2019, with a 42% reduction during lockdown. Average weekly presentations significantly decreased during the main lockdown months only (April 2019 vs. 2020: 213.0 [12.3] vs. 110.3 [22.9] [adjusted P = 0.006], May 2019 vs. 2020: 209.5 [14.6] vs. 148.8 [12.3] [adjusted P = 0.006]). There was only a trend toward reduced elderly presentations from January to August 2020 and during lockdown. Lockdown pressures did not drive significantly more new elderly presentations. However, we still found a highly significant difference in the distribution of causal factors for the elderly presentations affected by the pressures of lockdown compared to those who were not. Conclusions: A trend toward reduced elderly presentations during lockdown and the pandemic was found. Thematic analysis supported by further statistical analysis of the drivers of elderly presentations showed that the pressures of lockdown clearly affected older age liaison psychiatry presentations. Copyright:

INTRODUCTION

The global population is aging, and the prevalence of mental disorders in the population over 60, the risk factors for which include life stressors impacting on loneliness and social isolation, is around 15%.[1] The coronavirus disease 2019 (COVID-19) pandemic changed the pattern of patient presentations to mental health services, with governments initiating “lockdowns” and “stay-at-home” instructions globally and in the UK.[23] Moreover, other factors such as concerns of becoming afflicted with COVID-19 or misconceptions that the National Health Service (NHS) hospitals were only available for COVID-19 patients might have influenced the pattern of patient presentations to UK hospitals.[4] Psychiatric liaison services often operate in general acute hospitals; thus, they are more likely to see mental health service users with comorbid physical and mental illnesses.[5] Older adults in mental health crisis are more likely to present to psychiatric liaison service than other psychiatric emergency services due to having a comorbid physical illness and mental or neurological condition. Considering that the risk factors for elderly mental health disorders may have been increased by the COVID-19 pandemic, but, on the other hand, the elderly might have presented less frequently to hospitals due to changes in health-seeking behaviors, it is imperative to examine how the COVID-19 pandemic and UK national lockdown may have influenced the pattern of elderly patient presentations to psychiatric liaison services. We, therefore, did a retrospective service evaluation of older patient presentations to Birmingham and Solihull Mental Health Foundation Trust liaison psychiatry department during the COVID-19 pandemic period from January to August 2020 and compared this with data from January to August 2019, focusing on the particular dates of the lockdown period between 23 March 2020 and 28 May 2020. This service evaluation aimed to investigate the influence of COVID-19 pandemic on older adults’ presentation to psychiatry liaison services.

MATERIALS AND METHODS

This study was approved by the Research and Innovation Department of Birmingham and Solihull Mental Health Foundation Trust (BSMHFT), which serves a diverse population of approximately 1.3 million people over 172 square miles. This is a retrospective service evaluation of older patient presentations to the liaison psychiatry services of BSMHFT during the COVID-19 pandemic period from January to August 2020 (NHS Research Ethics Committee approval were deemed to be unnecessary when the service evaluation’s team of BSMHFT approved this service evaluation (SE0227 - on 29/06/2020). Mixed method methodology was employed in this study. This choice was based on the ability of this technique to generate holistic evidence by combining the use of both qualitative and quantitative analytic methods.[6]

Cohort Identification

An informatics request was submitted to the information governance team of the trust. Detailed information was requested about referrals to liaison psychiatry services between January 2019 and August 2020. Date-based filtering allowed specific data to be extracted for all liaison psychiatry presentations from January 2019 to August 2020. Considering that the first reported case of COVID-19 in the UK occurred at the end of January 2020, this corresponded to the 12 months before the approximate beginning of the impact of the COVID-19 pandemic on the UK.[7]

From this data set, we extracted presentations from the pandemic period that we defined as 1 January 2020 to 31 August 2020 and compared them with that of the same period in the preceding year. We further focused our analysis on the first national lockdown period in the UK between 23 March and 28 May 2020 (the date when groups of up to six persons or more were first permitted to meet outside) and specifically looked at older patient (defined as ≥65 years) presentations during that period. Electronic records of patients were then accessed for elderly presentations from the lockdown period via the data repository software of the trust to permit qualitative analysis.

Qualitative Analysis

Qualitative methods were used to explore themes emerging from COVID-19–related presentations among older adults during the first national pandemic lockdown period. Data was thematically analyzed inductively to generate in-depth and trustworthy themes from the content of noting and not from preconceived ideas or theory.[8] The six-step approach of Braun and Clarke utilizing familiarization, coding, theme development, revision, naming, and writing up was used to identify themes and causes of presentations. Trustworthiness and rigor were achieved using Lincoln and Guba’s criteria: credibility, transferability, dependability, and confirmability.[9] Member checking and triangulation were used to attain credibility. Member checking was carried out by the authors, who validated the themes generated. The main author ensured triangulation was attained by discussing the findings with other authors. If there were disagreements, they were resolved by discussions. Dependability was achieved by clearly documenting the research process. Confirmability was attained by allowing the interpretations and findings derived from the data to be supported by quotes. Transferability was achieved by providing a detailed description of the research in a manner that it could be easily applied in other contexts.

Quantitative Analysis and Statistics

Data fields extracted from the informatics request included date of presentations, patient demographics (age, gender, ethnicity), psychiatric history, presenting complaint, and history of prior encounters with our psychiatric services. Presenting complaints were classified according to broad diagnostic categories (anxiety/depression, suicidal ideation, acute psychosis, acute stress). Patients were also classified according to whether or not there was an involvement of alcohol abuse in their presentation. Additional information of interest included whether patients tested positive for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) or not, any infection symptomatology, and whether patients were affected by the lockdown measures or not. Categorical data was presented as absolute numbers and percentages and was analyzed using the Chi-squared test. Analysis of categorical data, which included categories with a null return, was done by merging the groups together until the categories no longer contained null results; this was done with maintenance of a logically constructed new group (e.g., grouping all Black, Asian and minority ethnic [BAME] categories together). Patient presentations were grouped according to the week of presentation (week commencing Monday), and this was used to compare trends in patient presentations over time. For comparison, a lockdown equivalent period in 2019 was selected based on day equivalence rather than date equivalent (i.e., Monday of the last full week of March to 9 weeks and 4 days later). The Shapiro–Wilk test was used to assess if weekly presentations were normally distributed or not. Normally distributed data was expressed as mean and standard deviation (SD). Weekly presentations per month were analyzed using multiple t-tests with Welch’s correction and a Holm–Sidak correction for multiple comparisons. Significance was accepted at P < 0.05, or in the case of multiple t-tests, when the adjusted P was <0.05. Data was presented and analyzed using GraphPad Prism 8.4.3.

RESULTS

The total number of patients presenting to the Trust Liaison Psychiatry Services for the whole of the period between 1 January 2019 and 31 August 2020 was 16,621 [see Table 1]. The number of presentations during the studied pandemic period of 2020 (from 1 January to 31 August) and the same period of the preceding year was 13,251, and more patients were seen in 2019 than 2020 (7088 and 6163, respectively – 13.1% reduction, absolute). During the lockdown period (defined as beginning on Monday 23 March 2020), there were 1145 patient presentations, whereas the same 67-day period (beginning on Monday 25 March 2019) saw 1982 patient presentations – 42% reduction, absolute. Table 1 provides a breakdown of demographic features for the full time period including 1 September–31 December 2019, each of the 8-month period, and the national lockdown period. Although the sex distribution of patients was not significantly altered across these times, statistically significant, though small differences were noted in the distribution of ethnicity among the presenting patients [Table 1].

Table 1

Demographic data for patients presenting to BSMHFT for the different periods

	01/01/19-31/08/20	01/01/19-31/08/19	01/01/20-31/08/20	Lockdown-equivalent period 25/03/19-30/05/19	Lockdown period 23/03/20-28/05/20	P
Number, n (% of all presentations)	16,621 (100)	7088 (42.6)	6163 (37.1)	1982 (11.9)	1145 (6.9)	n/a
Male gender, n (% of epoch)	8513 (51.2)	3591 (50.7)	3207 (52.0)	978 (49.3)	614 (53.6)	0.093
Ethnicity, n (% of epoch)
White	11,235 (67.6)	4796 (67.7)	4195 (68.1)	1350 (68.1)	802 (70.0)	0.003
Asian	2135 (12.8)	915 (12.9)	776 (12.6)	255 (12.9)	130 (11.4)
Black	893 (5.4)	403 (5.7)	293 (4.8)	102 (5.1)	50 (4.4)
Mixed	730 (4.4)	333 (4.7)	231 (3.7)	99 (5.0)	54 (4.7)
Arab	37 (0.2)	19 (0.3)	6 (0.1)	4 (0.2)	1 (0.1)
Any other	386 (2.3)	157 (2.2)	135 (2.2)	34 (1.7)	24 (2.1)
Unknown	1205 (7.3)	465 (6.6)	527 (8.6)	138 (7.0)	84 (7.3)

BSMHFT = Birmingham and Solihull Mental Health Foundation Trust. Significant differences were not found in the gender distribution of patients. However, significant differences were found in the proportions of presenting ethnicities, with fewer non-White ethnicities presenting during 2020 and the lockdown. Data analyzed using Chi-squared test

Analysis of weekly patient presentations by month for all age groups showed significant differences between 2019 and 2020 [Figure 1]. Also, 1 April till 28 May accounted for 58/67 (87%) of the days of lockdown. There was a prominent and significant decrease in weekly patient presentations when comparing 2020 with 2019, though only for these 2 months (April 2019 vs. 2020: 213.0 [12.3] vs. 110.3 [22.9] [adjusted P = 0.006], May 2019 vs. 2020: 209.5 [14.6] vs. 148.8 [12.3] [adjusted P = 0.006]).

Figure 1

Weekly patient presentations (mean, SD) per month for all age groups from January to August for 2019 and 2020. Adjusted t-tests indicated that only April and May 2020 had a significantly reduced mean number of weekly presentations, compared to the same months in 2019 (star indicates P < 0.005). Analysis by t-tests with Welch’s correction and a Holm–Sidak correction for multiple comparisons. SD = standard deviation

Elderly Presentations

The total number of elderly presentations from January 2019 to 31 August 2020 was 715. Elderly presentations made a relatively constant proportion of all presentations across the 2 years, including during lockdown. As was the case with the whole cohort, fewer elderly presentations were seen in the period January–August 2020 than for the same period in 2019 (265 vs. 321, respectively – 17.4% reduction, absolute). When comparing the spread of presentations across January–August 2019 and 2020, the proportion of elderly presentations across the 8 months appeared to be comparatively decreased during the lockdown months, but this was exaggerated by a relative increase in presentations over the same months in 2019. Indeed, on statistical analysis, the mean number of weekly presentations of elderly patients during these two periods was seen to be not significantly decreased during lockdown months or at any other time (April 2019 vs. 2020: 11.8 [5.2] vs. 4.8 [1.7] [not significant], May 2019 vs. 2020: 10.3 [3.9] vs. 6.0 [2.4] [not significant]) [Figure 2]. As with the entire cohort, we found that the gender distribution of elderly presentations was not significantly altered across these epochs, but unlike the entire cohort, no significant differences in the distribution of ethnicities was found either (data not shown).

Figure 2

Weekly presentations per month (mean, SD) in 2019 and 2020 for those of age 65 years and above. There were no significant differences between any months for the number of weekly presentations to liaison psychiatry services. Analysis by t-tests with Welch’s correction and a Holm–Sidak correction for multiple comparisons. SD = standard deviation

We found a decrease in the proportion of elderly patients presenting to liaison psychiatry services for the period January–August 2020, who were not previously known to services. This was also the case for the lockdown period. However, this was not significantly different to that for the same period in the preceding year. No elderly patients were referred to our services with neuropsychiatric manifestations of COVID-19, and, in fact, only one patient tested positive for SARS-CoV-2 at the time of presentation.

During the lockdown period, 51 elderly patients were referred to our services, though one patient self-discharged prior to medical and psychiatric assessment and was therefore not included in the analysis. Analysis of the causes of presentation when dichotomized according to whether they were related to COVID-19 or not (i.e., due to the presence of the pressures of lockdown in the history) showed that 15/50 (30.0%) patients presented purely due to the pressures of lockdown [Figure 3]. A greater proportion of new referrals than those already known to services was affected by lockdown measures (unknown 39.3% [11/28] vs. already known 18.2% [4/22]), though this difference was not statistically significant. However, analysis of causes of referral showed a highly significant difference in the distribution of causes between those patients affected and unaffected by lockdown restrictions [see Figure 4].

Figure 3

Patient presentations during lockdown previously known versus unknown to psych liaison services. This is broken down according to whether they are related or unrelated to lockdown pressures (P = 0.1305). Data analyzed using Chi-squared test

Figure 4

Percentage of elderly presentations during lockdown according to different causes dichotomized according to whether or not patients felt they were affected or unaffected by the pressures of lockdown. Patients were grouped into multiple categories (e.g., alcohol-related causes and suicide attempt/ideation). Data analyzed using Chi-squared test

Thematic Analysis

Thematic analysis revealed two main themes as contributing to presentations in those affected by lockdown. These were social isolation and acute stress. In fact, all but one elderly presentation was related to social isolation. Other psychosocial factors such as the closure of community facilities and the need for shielding and isolation further compounding symptoms in already vulnerable individuals with a past psychiatric history were noted. We provide two quotes that typify these themes:

“Y is finding it more difficult recently as a result of social distancing due to COVID-19 pandemic, usually social and now can go out and saying ‘can’t see the end.’” (social isolation, Patient Y).

“X continues to think that he is responsible for the current situation relating to COVID. X spoke about his home being dirty, and how this may have contributed to the spread of COVID.” (acute stress, Patient X).

DISCUSSION

This was a retrospective study analyzing the effects of COVID-19 pandemic measures and lockdown on elderly presentations to a liaison psychiatry service of a large metropolitan mental health trust through comparison with the same period of the preceding year. We found a substantial variation in weekly presentations by month for all age groups, which was significantly decreased for the lockdown months of April and May when compared to the same time period in 2019. We further concentrated on elderly patients (≥65 years old) and noticed a nonsignificant trend for decreased numbers of elderly patients presenting to our services. The proportion of previously unknown elderly patients during this time was not significantly altered either. We noticed that despite an absolute reduction in elderly presentations during lockdown, lockdown pressures themselves contributed many elderly presentations, and the majority of these lockdown-related presentations were in patients who were not previously known to our services. A significant difference in the distribution of causes between those patients affected and unaffected by lockdown restrictions was seen. In keeping with these findings, the main themes emerging from the qualitative analysis of lockdown-related presentations were social isolation and acute stress related to the pandemic. On the other hand, the risk of presentation to our services with acute psychosis was much higher in those who were not affected by lockdown measures. Interestingly, no patients were referred to our services with neuropsychiatric manifestations of COVID-19, though this may be because of both the high mortality in older age patients with COVID-19 and also because any problems that occurred were likely being dealt with by hospital specialists in the context of acute illness.

We suggest that significantly fewer overall patient presentations to mental health services and the trend toward fewer elderly presentations during the COVID-19 pandemic, especially during the lockdown period, mirrored the generally decreased number of hospital presentations and admissions that was seen across the country.[1011] However, the profile of elderly presentations (whether previously known or unknown to services) was relatively stable for the 2 years we studied and apparently was not significantly affected by the lockdown or pandemic. Although it was difficult to show many significant differences in assessed metrics in the elderly, the psychological pressures of lockdown in the elderly could, nevertheless, be shown to have measurable effects on their mental health, as patients with suicidal ideation, depression, alcohol-related and acute stress were significantly more likely to report being affected by lockdown. This is in keeping with data obtained from a large group of elderly persons living through the pandemic in the UK.[12]

Strengths and Novel Findings of This Study

The COVID-19 pandemic has generated literature in relation to outbreaks of the disease in psychiatric institutions, the neuropsychiatric manifestations of the infection, and also the effect of the pandemic on health-care workers’ mental health after having to deal with the massive influx of patients.[131415] Several large studies have also been done, which have looked at the effects of COVID-19 pandemic on mental health in the UK general population.[161718192021] They have mainly concluded that the pandemic measures have adversely impacted mental health, particularly in the younger population, women, and people of low socioeconomic groups.[22] The majority of these studies with primary data from the UK are based on surveys. However, unlike our study, none of these have focused on the impact of pandemic measures and lockdown on the elderly population, possibly due to concerns for the younger population who have been found to be more affected than the elderly by the pandemic in these studies.[22] Our study clearly showed that the findings of these previous studies could produce a false impression of the effect of the pandemic on the elderly. Recent data suggests that elderly mental health has been affected by the pandemic, and that they do not behave uniformly in terms of psychological response to lockdown stresses.[12] Those patients with impairments or disabilities have increased odds of elevated loneliness, sleep disturbance, depressive symptoms, and anxiety.[23] In this regard, one of the major strengths of our study is the inclusion of data which is based on actual referrals and presentations of all age groups, particularly the elderly, to our liaison psychiatric services. Furthermore, we provide a profiling of causes and factors within that subgroup. Our findings corroborate those of another study from a large UK mental health service where adult acute and elderly acute admissions were noted to have significantly decreased during lockdown.[3] However, that study did not provide granular data on the nature of admissions and associated causal factors. So, this study adds crucial insight into the nature of presenting factors, which has been lacking to date.

Limitations

Our study, nevertheless, does have some limitations. We focused on presentations to BSMHFT, which covers a socially diverse population of 1.3 million people. Notwithstanding this fact and that other studies support our findings, presentations at this trust may not be fully representative of other regions in the UK, or the UK in general. Furthermore, as this was a study of presentations to liaison psychiatric services, it may, therefore, underestimate the effect of pandemic measures and lockdown on the mental health of the general population, as it reflects the clinical tip of the population iceberg. While it is unlikely that severely disruptive mental health conditions such as acute psychosis and suicide attempts would not present to mental health services for long, it is quite possible that anxiety, depression, and alcohol abuse would remain the problems that were managed by many people in their own households. Alternatively, patients may have deliberately avoided seeking medical attention due to fears of contracting the virus and/or government advice to avoid hospital emergency departments if possible. Finally, our comparison of cases for 2020 was with cases for 2019 and was of course subject to the inherent weaknesses of retrospective observational studies. Access to data for 2018 and other previous years may have improved the strength of our statistical comparisons, but unfortunately this was unavailable.

In summary, this study shows that pandemic measures and lockdown resulted in decreased patient presentations across all age groups to mental health services in our trust, though only significantly in the non-elderly. When focusing on the elderly, we found that lockdown measures resulted in social isolation and acute stress, which contributed to many new presentations. These findings highlight the effects of pandemics, and especially lockdown measures, on the mental health of the elderly population and indicate that despite a greater focus on the mental health of under-65-year-old people, we clearly need to better cater for the mental health needs of the elderly during such times too.[12]

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.