SNOMED-CT

Mental health presentations to Christchurch Hospital Emergency Department during COVID‐19 lockdown – Joyce – – Emergency Medicine Australasia


Introduction

The New Zealand government‐mandated ‘lockdown’ because of the novel coronavirus (SARS‐CoV‐2), which causes coronavirus disease (COVID‐19), has had a significant impact on health‐seeking behaviours including ED presentations. Quantifying these changes allows greater understanding of the impact of the virus and associated restrictions.

The first case of COVID‐19 was reported in New Zealand on 28 February 2020, followed by a steady increase in cases over the next 2 weeks. The New Zealand government announced that they would ‘go hard and go early’, introducing public health measures to contain, and ideally eliminate, the virus.1 On 19 March 2020, New Zealand’s borders were closed to non‐residents, and gatherings limited to 100 people. Within 6 days, the country moved through increasing levels of restriction, until the Prime Minister issued an Epidemic Notice under section 5 of the Epidemic Preparedness Act 2006 on 23 March, and the ‘Lockdown’, at the most restrictive COVID alert level 4 in New Zealand, was announced to start on 26 March and would end 33 days later on 28 April.2

This nationwide lockdown significantly restricted domestic travel, banned public gatherings, closed public venues, educational facilities and non‐essential businesses, with the population asked to ‘Stay Home, Save Lives’. The lockdown had a significant impact on the health system with a 70% reduction in face‐to‐face primary care consultations, and all non‐essential surgeries and outpatient appointments postponed. It has been argued that New Zealand had one of the toughest lockdown restrictions in the world,3 giving rise to concern for the potential impact on mental health of the population.4, 5

In addition to the threat of an unseen life‐threatening virus possibly being transmitted within the community and the financial and societal ramifications of effectively closing the economy for lockdown are the additional impacts of isolation under a mandatory quarantine. The National Telehealth Service in New Zealand reported that calls to their helplines increased by almost a third during the first week of lockdown.6 Throughout the world, the COVID‐19 pandemic has been reported to have significant effects on mental health as well as physical health.7, 8 It is important that significant changes in presentations to the ED are recognised, enabling effective alignment of workforce and service distribution. The objective of the present study was to review mental health presentations to a tertiary ED in New Zealand during a national COVID‐19 lockdown.

Methods

Study design

A retrospective, comparative cohort study was conducted using data from patients seen at the Christchurch Hospital ED, with an acute mental health issue. Data were captured for two cohorts, across a similar length of time and range of diagnostic codes, both before and during the national lockdown.

Setting

Christchurch Hospital is a tertiary level teaching hospital, the largest in the South Island of New Zealand, with a catchment of approximately 550 000. It is the only major acute referral centre in the region with over 100 000 attendances each year. Mental health‐related presentations account for almost 6% of these attendances.

Access to mental health services in Christchurch can be achieved in several ways. There is a Single Point of Entry phone number, used predominantly by individuals already known to mental health services, who can arrange appointments with teams or direct patients to the ED for review after hours. Alternately, access is via ED or primary care (GP or after hours clinic referral). There is a single psychiatric hospital in the city with clinic rooms, and a Mental Health Liason Team (nursing) based in the ED who can refer to the Crisis Resolution Team, based elsewhere in the Christchurch Hospital. Christchurch has no 24‐h private psychiatric services, but does have a 24‐h medical clinic and two after‐hours medical clinics that operate in the evenings and weekends until 10 pm. These do not typically treat mental health patients and have a fee for service, whereas the ED has no charge.

Participants

The two cohorts consisted of patients who presented over the 33 days before the declaration of alert level 2 on 19 March 2020 (15 February–18 March 2020) – cohort 1: ‘Pre‐Lockdown’, and 33 days of the level 4 (26 March–28 April 2020), cohort 2: ‘Lockdown’. The week of 19 March to 25 March 2020 was excluded from the data, as this was a period of partial lockdown across levels 2 and 3, with rapidly changing levels of national quarantine. A comparison of total ED presentations versus mental health presentations during similar periods in 2019 was also made. A new data entry system was introduced in 2018 so earlier data sets were not directly comparable.

Inclusion criteria

Patient presentations were included based on date of attendance, and presence of a mental health code entered into their medical record for that presentation. The specific codes sought are those entered using the SNOMED CT clinical terminology classification system.9 The codes represent patient diagnoses as entered into the emergency patient record and are listed in Appendix S1. Any presentation to the ED with a subsequent referral to mental health services was also included.

Exclusion criteria

Data for patient presentations with diagnostic codes for ‘Dementia’ and ‘Delirium’ were excluded, as were presentations solely for alcohol or recreational drug intoxication without suicidal ideation.

Data sources

Data were extracted from routinely collected administrative data extracted from Christchurch Hospital’s electronic medical record system.

Analysis

Simple descriptive statistical analysis was undertaken to enable comparison of the demographics across the two cohorts. Patients within each cohort were grouped according to age, gender and ethnicity. Following this, more in‐depth analysis occurred for the diagnostic categories of overdose and self‐harm. This included analysis of triage code and admission to hospital versus direct discharge from ED. In order to gather more detailed information relating to the specific diagnoses, the patient’s clinical record was then reviewed by a member of the research team. The subgroup analysis of overdose presentations involved identification of the type of drugs taken, and coding of these under the categories of: prescription medications, medications available ‘over the counter’ (paracetamol and ibuprofen), recreational drugs or other toxins.

Statistical methods

Descriptive statistics were calculated using IBM spss Statistics for Windows, version 24.0 (IBM, Armonk, NY, USA). Proportions of presentations were performed using MedCalc for Windows, version 15.0 (MedCalc Software, Ostend, Belgium).

Ethical approval

The present study has been granted ethical approval under the ‘Minimal Risk Health Research’ University of Otago Human Ethics Application HD20/047.

Results

A cumulative total of 15 410 patient presentations occurred during the two study periods in 2020; 9460 during the 33‐day pre‐lockdown period and 5950 during the 33‐day lockdown period. Compared to the 1.2% decrease between the two periods in 2019, the 37% decrease in all ED presentations between the 2020 pre‐lockdown and lockdown periods is significant (P < 0.00001). Mental health‐related presentations decreased by 34.1% during the study period, from 564 presentations during pre‐lockdown to 371 presentations during lockdown. This is significantly more than 2019 where there was a 9.8% decrease (P < 0.001) (Table 1).

TABLE 1.
Total ED presentations versus mental health presentations 2019–2020
Pre‐lockdown

Lockdown

Change (%)P‐value of X2
All ED presentations
202094605950−37.1<0.00001
201990748963−1.2
Mental health presentations
2020564371−34.1<0.001
2019540491−9.8

  • ‘Pre‐lockdown’ refers to 2020 from 15 February 2020 to 18 March 2020, with the same period in 2019 shown as a comparison.

  • ‘Lockdown’ refers to 2020 from 26 March 2020 to 28 April 2020, with the same period in 2019 shown as a comparison.

The number of ED presentations related to mental health issues decreased in a similar proportion to the decrease in all other non‐mental health‐related presentations (P = 0.48) (Table 2).

TABLE 2.
Mental health presentations as a proportion of total ED presentations 2020
Pre‐lockdown, n (%)Lockdown, n (%)Change in proportion of presentations (%)(95% CI), P‐value

All ED presentations 20209460 (100)5950 (100)
Mental health presentations 2020564 (6.0)371 (6.2)0.3(−0.5%, 1.1%), P = 0.48

  • P‐values are based on difference using n−1 χ2 for difference in proportions.

Demographics

A similar pattern of presentations is evident across the two cohorts, in relation to gender, ethnicity and triage code. The pattern of patients presenting with mental health diagnoses across both cohorts included a proportionally higher percentage of females than males (cohort 1: 58% [n = 327]:42% [n = 237]; cohort 2: 57% [n = 212]:43% [n = 159]), both with a majority who identified as New Zealand European (60% [n = 340], 58% [n = 217]), followed by those who identify as NZ Maori (21% [n = 117], 22% [n = 83]).

Two points of variance reached statistical significance in regards to the age bands for presenting patients – the first of these was those aged between 15–19 years, which showed a 5% decrease in the proportion of presenting patients between the two cohorts (P = 0.018). The second group was the older adults, aged 60 and above, which showed a proportional increase of 4.7% (P = 0.016) (Table 3).

TABLE 3.
Demographics of mental health presentations 2020
Pre‐lockdown presentations, n (%)Lockdown presentations, n (%)Change in proportion (%)(95% CI), P‐value

Sex
Female327 (58.0)212 (57.1)−0.8(−5.6%, 7.3%), P = 0.80
Ethnicity
NZ European340 (60.3)217 (58.5)−1.8(−4.6%, 8.2%), P = 0.59
Maori117 (20.7)83 (22.4)1.6(−3.7%, 7.1%), P = 0.55
Pacific16 (2.8)9 (2.4)−0.4(−2.0%, 2.5%), P = 0.70
Asian16 (2.8)11 (3.0)0.1(−2.0%, 2.6%), P = 0.91
Other75 (13.3)51 (13.8)0.5(−3.9%, 5.1%), P = 0.84
Age (years)
0–1426 (4.6)20 (5.4)0.8(−2.0%, 3.9%), P = 0.59
15–1981 (14.4)34 (9.2)−5.2(−9.2%, −0.9%), P = 0.018
20–29180 (31.9)120 (32.4)0.4(−5.6%, 6.6%), P = 0.89
30–39117 (20.7)64 (17.3)−3.5(−1.7%, 8.5%), P = 0.19
40–4975 (13.3)53 (14.3)1.0(−3.4%, 5.7%), P = 0.67
50–5943 (7.6)35 (9.4)1.8(−1.8%, 5.7%), P = 0.33
60+42 (7.5)45 (12.1)4.7(0.8%, 8.9%), P = 0.016
Triage
14 (0.7)2 (0.5)−0.2(−1.3%, 1.3%), P = 0.75
231 (5.5)17 (4.6)−0.9(−2.2%, 3.7%), P = 0.53
3367 (65.0)258 (69.5)4.5(−1.7%, 10.5%), P = 0.15
4143 (25.4)83 (22.4)−3.0(−2.7%, 8.4%), P = 0.30
519 (3.4)11 (3.0)−0.4(−2.2%, 2.7%), P = 0.73

  • P‐values are based on difference using n−1 χ2 for difference in proportions.

Although the absolute number of presentations due to overdose and self‐harm stayed the same between the two study periods, there were small but significant increases in the proportion of individuals presenting with overdose (6.5% increase, P = 0.035) and self‐harm (3.5% increase, P < 0.049), whereas the proportion of patients attending with other mental health issues that did not involve overdose or self‐harm decreased (10% decrease, P = 0.002) (Table 4).

TABLE 4.
Mental health presentations by diagnostic category
Pre‐lockdown mental health presentations, n (%)Lockdown mental health presentations, n (%)Change in proportion (%)(95% CI), P‐value
Overdose158 (28.0)128 (34.5)6.5(0.5%, 12.6%), P = 0.035
Self‐harm35 (6.2)36 (9.7)3.5(0.01%, 7.3%), P = 0.049
Other371 (65.8)207 (55.8)−10.0(−16.3%, −3.6%), P = 0.002

Among the subgroup of those patients who presented after overdose of medications or other toxins, there was an increased admission rate to general medicine inpatient wards or intensive care during lockdown (11.8% increase, P = 0.003). This is primarily due to the increase in significant paracetamol overdoses that required N‐Acetylcysteine treatment, which is managed as an inpatient in Christchurch Hospital. There were four admissions to general medicine for N‐Acetylcysteine treatment during the pre‐lockdown period compared to 10 admissions during lockdown. There were no significant differences between the pre‐lockdown and lockdown periods in demographic data including age, gender and triage code. The proportion of patients who took paracetamol and/or ibuprofen in overdose increased by 13.4% (P = 0.005) during lockdown, whereas there were no significant changes in presentations because of other overdoses (Table 5).

TABLE 5.
Analysis of overdose presentations
Pre‐lockdown overdose presentations, n (%)Lockdown overdose presentations, n (%)Change in proportion (%)(95% CI), P‐value
Total overdose presentations158 (28.0)128 (34.5)6.5(0.5%, 12.6%), P = 0.035
Admitted for overdose
Admitted to ward/ICU11 (7.0)24 (18.8)11.8(4.1%, 20.0%), P = 0.003
Age (years)
0–143 (1.9)7 (5.5)3.6(−0.9%, 9.1%), P = 0.10
15–1934 (21.5)17 (13.3)8.2(−0.8%, 16.8%), P = 0.07
20–2959 (37.3)45 (35.2)2.2(−9.0%, 13.1%), P = 0.70
30–3926 (16.5)18 (14.1)2.4(−6.3%, 10.6%), P = 0.58
40–4917 (10.8)22 (17.2)6.4(−1.6%, 14.9%), P = 0.16
50–5911 (7.0)9 (7.0)0.1(−6.0%, 6.6%), P = 0.98
60+8 (5.1)10 (7.8)2.8(−3.0%, 9.2%), P = 0.34
Sex
Female105 (66.5)84 (65.6)−0.8(−10.0%, 11.8%), P = 0.88
Triage
14 (2.5)2 (1.6)−1.0(−3.3%, 4.9%), P = 0.57
216 (10.1)11 (8.6)−1.5(−5.7%, 8.3%), P = 0.66
3123 (77.9)100 (78.1)0.3(−9.5%, 9.8%), P = 0.95
415 (9.5)15 (11.7)2.2(−4.9%, 9.9%), P = 0.54
50 (0.0)0 (0.0)0.0
Drug(s) taken in overdose
Prescription/mixed123 (77.9)90 (70.3)−7.5(−2.6%, 17.7%), P = 0.15
Paracetamol +/− ibuprofen22 (13.9)35 (27.3)13.4(4.1%, 22.9%), P = 0.005
Recreational

8 (5.1)2 (1.6)−3.5(−1.2%, 8.2%), P = 0.11
Other toxins (vitamins, rat poison, plants)4 (2.5)1 (0.8)−1.8(−2.1%, 5.6%), P = 0.26
Unknown1 (0.6)0 (0.0)−0.6(−2.3%, 3.5%), P = 0.37

  • Recreational drugs taken as a self‐harm attempt, as opposed to for recreational purposes.
  • ICU, intensive care unit.

Discussion

The findings in the present study are suggestive of a change in pattern of presentation that may be associated with periods of lockdown or quarantine. While the New Zealand experience of the COVID‐19 outbreak has resulted in a more limited period of quarantine than has occurred in other countries, the sense of fear, anxiety and distress associated with the unknown is the same. In this way, while each country will experience the pandemic in a unique manner, there is the potential that these findings may represent wider trends and as such offer insight into possible necessary responses.

Anxiety, depression and self‐reported stress are commonly described psychological reactions to the COVID‐19 pandemic.7 Studies of the 2009 H1N1 pandemic reported high levels of anxiety in the community,10 and the negative psychological effects of quarantine were also seen during the SARS‐2003 outbreak.11 Anxiety as an emotional response in relation to COVID‐19 can increase following the first death reported in a country, increasing numbers of new cases, and increased media reporting.12

In countries with low numbers of COVID‐19 cases, or those early in the outbreak, ED presentations have been reported to have decreased significantly. The Royal College of Emergency Medicine in England reported a 25% decrease in ED presentations in the week after lockdown compared to the previous week, and a 49% decrease when compared to 1 month prior.13 Christey et al.14 from Waikato Hospital, a tertiary hospital in the North Island of New Zealand, measured a 43% reduction in trauma admissions during the first 2 weeks of lockdown, when compared to the 2‐week period prior to level 2.

Concerns experienced among the general public related to presenting to the ED (resulting in decreased attendances) may be exacerbated in those with existing mental health conditions. Of interest, fewer patients were seen to present with a mental health concern that did not involve deliberate self‐harm or self‐poisoning. It may be that COVID‐19 messaging common at this time had an unanticipated impact of deterring people from seeking care unless they felt they had a visible/physical reason for doing so.

There are several factors present in the data worth further consideration. These include recognition that despite heightened anxiety, depression and the potentially limited access to usual mental health services, there remains a decrease in these presentations to the ED. The increased focus in the media around telephone support for mental health and community follow up for existing clients may have contributed to the reduction in ‘non‐physical’ mental health related presentations. There was an increase in overdose presentations; however, where the drug involved was one that is considered readily accessible, particularly paracetamol and/or ibuprofen. Both of these medications continued to be available via supermarkets, which remained open during lockdown, and COVID‐19 media reports regularly referred to the value of having paracetamol as part of a pandemic response kit, leading to reports of stockpiling and fear of shortages.15, 16

Limitations

The study focussed on a single hospital and a single set of comparative groups. As a result of changes to the hospital’s data collection system, it was not possible to look at a time series of the cohort prior to 2019. New Zealand coronial suicide data are not available for the study period, and so it is not possible to comment on any changes in suicide rates in Christchurch.

Conclusions

During the COVID‐19 lockdown, both overall ED presentations as well as mental health‐related presentations decreased, but with a proportional increase in overdoses and self‐harm, particularly involving easily accessible medications, such as paracetamol and ibuprofen.

Acknowledgements

The authors would like to acknowledge Melanie Browne, Information Analyst, and Soledad Labbe‐Hubbard, Project Specialist (Canterbury District Health Board) who assisted with data access and presentation.

Author contributions

The paper was conceived by LRJ and MWA. Data collection and analysis was performed by LRJ, AM and GJH. Interpretation and write‐up by LRJ, SKR, AM, GJH and MWA. All authors have approved the final version submitted herein.

Competing interests

None declared.

    Data availability statement

    The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

    Data availability statement

    The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

    References