Research and analysis

Premature mortality during COVID-19 in adults with severe mental illness

Published 19 October 2023

Applies to England

Introduction

During the COVID-19 pandemic, people with severe mental illness (SMI) were more likely to have contracted COVID-19 and more likely to have died as a result than people without SMI (reference 1). However, less is known about premature mortality from all causes in people with SMI during the pandemic. In particular, consideration of how this differs across subgroups, across regions in England, or how this compares to premature mortality in people without SMI.

Previous reports by the Office for Health Improvement and Disparities (OHID, 2022) and Public Health England (2018) have found that people with SMI are more likely to die before 75 years of age than people without SMI and have worse physical health than the general population.

People with SMI may also be more excluded or isolated than the general population, and may be more affected in times of adversity or change (reference 2). For example, the pandemic may have impacted adversely on both physical and mental wellbeing in this population, and led to more disrupted access to healthcare than for other people (references 3 and 4).

The aim of this report is to investigate premature mortality (deaths in those aged 18 to 74) during the early part of the COVID-19 pandemic (January 2020 to March 2021) in people with SMI (for example, who have schizophrenia or bipolar disorder).

Important to know

To enhance understanding of the report some important terms and aspects of the work are explained below. Further details on how these terms were defined and used, and on how the results of this report were generated, are provided in the methodology section.

Severe mental illness

There is no easy way to define the size of the population with SMI at any one point in time. SMI tends to affect individuals over a long period. For the purposes of this study, contact with secondary care mental health services, any time during the preceding 5 years, is used as a proxy indicator for a diagnosis of SMI. Having contact with secondary care mental health services is defined as evidence of a referral to or contact with secondary care mental health services in the mental health services datasets.

The term severe mental illness is used to describe people with a group of conditions that are often chronic and so debilitating that their ability to engage in functional and occupational activities is severely impaired. SMI generally includes diagnoses such as schizophrenia, bipolar disorder or other psychotic illnesses that cause severe functional impairment. However, people in contact with secondary care mental health services may also have severe depression, post-traumatic stress disorder, personality disorder, eating disorder or other mental health problems. People in contact with learning disability or autism services only were excluded from this analysis.

Mortality with SMI

A deceased person is considered to have had SMI if they had a referral to or contact with secondary care mental health services in the 5 years before their death.

Premature mortality

In this report premature mortality refers to a death between the age of 18 and 74 years.

Timeframe of report

The majority of the analyses in this report covers the period from January 2020 to March 2021. This is to capture the earliest possible effects of the COVID-19 pandemic, before testing was available, and to also cover the second wave.

Increase in premature mortality during the pandemic

This is a measure of the increase in premature mortality in people with SMI during the pandemic (April 2020 to March 2021) compared to an average of annual counts and rates for 3 years before the pandemic (2017 to 2019).

Excess premature mortality during the pandemic

This measures how much more likely premature mortality is in people with SMI than people without SMI during the pandemic.

Main findings

This report highlights the continued inequality experienced by people with SMI and how they may be especially vulnerable to premature death during a pandemic. It is part of a series of reports considering aspects of premature mortality in people with SMI.

Increase in premature mortality

The premature mortality rate for people with SMI increased by 16% during the first year of the COVID-19 pandemic (April 2020 to March 2021) compared to the annual average before the pandemic (2017 to 2019). This was a similar rate of increase to those without SMI (15%). However, the ongoing underlying higher rate of premature mortality in people with SMI meant the impact in those with SMI was greater. There were an extra:

• 244 deaths for every 100,000 people with SMI (from 1,505 deaths per 100,000 before the pandemic to 1,749 per 100,000 during the pandemic)

• 47 deaths for every 100,000 people without SMI (from 315 deaths per 100,000 before the pandemic to 362 deaths per 100,000 during the pandemic)

When considering absolute numbers of deaths the impact is more apparent for adults with SMI. 42,815 people aged 18 to 74 with SMI died during the first year of the pandemic compared to an annual average of 35,025 people with SMI in the 3 years before the pandemic (2017 to 2019). This is an extra 7,790 deaths, an increase of 22.2%.

For people without SMI, 132,140 people aged 18 to 74 died during the first year of the pandemic compared to an annual average of 112,850 people in the 3 years before the pandemic (2017 to 2019). This is an extra 19,290 deaths, an increase of 17.1%.

Variation by deprivation

People with SMI living in the most deprived areas of England had the highest premature mortality rates before the pandemic, the highest premature mortality rates during the pandemic, and the greatest increase in premature mortality during the pandemic.

Variation by ethnic group

People of white ethnicity with SMI had the highest premature mortality rate before the pandemic, the highest premature mortality rate during the pandemic, but the lowest increase during the pandemic of those with known ethnicity. In contrast, while people of black or Asian ethnicity with SMI had lower premature mortality rates before the pandemic and lower premature mortality rates during the pandemic, the increase in premature mortality during the pandemic was much larger for both these groups (people of black ethnicity 43%, people of Asian ethnicity 50%).

Variation by region

In people with SMI, the premature mortality rate was highest in the North East both before and during the pandemic. However, the greatest increase in premature mortality during the pandemic in people with SMI was in London.

Excess premature mortality

During the pandemic, the premature mortality rate was almost 5 times higher in people with SMI (an excess premature mortality of 387%) than people without SMI.

Findings and interpretation

In this study the population aged 18 to 74 in contact with secondary mental health services is used as a proxy to identify people with SMI. In the period of interest (January 2020 to March 2021) there were approximately 39.5 million adults aged between 18 and 74 in England. Of these, around 3.5 million had a referral to, or were in contact with secondary mental health services in the previous 5 years. These people were considered as ‘the population at risk’, and there were 53,660 deaths in this group during the period. There were around 36 million people who did not have a referral to or contact with secondary mental health services in the previous 5 years. Within this population there were 164,090 deaths between January 2020 and March 2021.

Increase in premature mortality during the COVID-19 pandemic compared to before the pandemic in people with SMI

People with SMI were at a higher risk of premature mortality during the first year of the pandemic (April 2020 to March 2021) than people without SMI. However, people with SMI were already at increased risk of premature mortality before the pandemic. It is therefore important to know if the already high mortality rate in this population increased even further during the pandemic.

There was a 16% increase in premature mortality in people with SMI during the pandemic (April 2020 to March 2021), compared to the average annual rate before the pandemic (2017 to 2019). People with SMI had a premature mortality rate of:

• 1,505 deaths per 100,000 people with SMI per year before the pandemic

• 1,749 deaths per 100,000 people with SMI during the pandemic

For every 100,000 people with SMI, an extra 244 people died prematurely during the first year of the pandemic, compared to before the pandemic. 

Table 1 shows that men with SMI had a higher premature mortality rate than women with SMI before and during the pandemic. The increase in premature mortality rates in people with SMI during the pandemic was similar in men (17%) and women (16%), though this varied with age. It also shows that in people with SMI aged 18 to 49 the premature mortality rate increased by 8% in women during the pandemic, but there was little change in the mortality rate for men. In contrast, in people with SMI aged 50 to 74 the increase in premature mortality during the pandemic period was similar in men (19%) and women (17%).

Table 1: premature mortality rate before and during the COVID-19 pandemic and increase in premature mortality rate during the pandemic in people with SMI, by age and sex

	Mortality rate (per 100,000) before the pandemic (95% confidence intervals)	Mortality rate (per 100,000) during the first year of the pandemic (95% confidence intervals)	Increase in mortality
Men aged 18 to 74	1,828 (1,803 to 1,854)	2,130 (2,103 to 2,157)	17%
Women aged 18 to 74	1,217 (1,197 to 1237)	1,408 (1,387 to 1,429)	16%
Men aged 18 to 49	466 (457 to 474)	472 (458 to 486)	1%
Women aged 18 to 49	256 (251 to 263)	277 (267 to 288)	8%
Men aged 50 to 74	3,723 (3,690 to 3,757)	4,435 (4,374 to 4,497)	19%
Women aged 50 to 74	2,552 (2,526 to 2,578)	2,980 (2,933 to 3,029)	17%

Comparing the increase in premature mortality during the COVID-19 pandemic between people with SMI and people without SMI

Figure 1 shows that the rate of premature mortality before the pandemic was much higher in people with SMI than people without SMI. The increase in premature mortality during the first year of the pandemic (April 2020 to March 2021) was similar in those with SMI (16%) and without SMI (15%). There were an extra:

• 244 deaths for every 100,000 people with SMI (from 1,505 deaths per 100,000 before the pandemic to 1,749 per 100,000 during the pandemic)

• 47 deaths for every 100,000 people without SMI (from 315 deaths per 100,000 before the pandemic to 362 per 100,000 during the pandemic)

Figure 1: premature mortality rate in people with and without SMI before and during the COVID-19 pandemic

When considering absolute numbers, 42,815 people aged 18 to 74 with SMI died during the first year of the pandemic compared to an annual average of 35,025 people with SMI in the 3 years before the pandemic (2017 to 2019). This is an extra 7,790 deaths, an increase of 22.2%.

For people without SMI, 132,140 people aged 18 to 74 died during the first year of the pandemic compared to an annual average of 112,850 people in the 3 years before the pandemic (2017 to 2019). This is an extra 19,290 deaths, an increase of 17.1%.

Variation in the increase in premature mortality during the COVID-19 pandemic in people with SMI

Differences in the increase in premature mortality rates during the COVID-19 pandemic in people with SMI, by deprivation

More people with SMI live in areas in the most deprived quintile of England (1,042,555) compared to the least deprived quintile (419,565). From April 2020 to March 2021 there were 14,225 premature deaths in people with SMI in the most deprived quintile compared to 4,775 deaths in the least deprived quintile.

People with SMI living in the most deprived areas of England had:

• the highest premature mortality rates before the pandemic (1,815 deaths per 100,000 people with SMI)

• the highest premature mortality rates during the pandemic (2,183 deaths per 100,000 with SMI)

• the greatest increase in premature mortality during the pandemic (20%)

In comparison, people with SMI living in the least deprived areas of England had:

• the lowest premature mortality rates before the pandemic (1,225 deaths per 100,000 with SMI)

• the lowest premature mortality rates during the pandemic (1,382 deaths per 1000,000 with SMI)

• the lowest increase in premature mortality during the pandemic (13%)

The highest premature mortality rates during the pandemic in people with SMI living in the most deprived areas were in:

• the North East with 2,660 deaths per 100,000 people with SMI

• the North West with 2,244 deaths per 100,000 people with SMI

• the East Midlands with 2,321 deaths per 100,000 people with SMI

For people with SMI living in the most deprived areas of the country, the premature mortality rate during the pandemic was lowest in those living in London (1,920 deaths per 100,000 people with SMI). However, the increase in premature mortality during the pandemic in people with SMI living in the most deprived areas was much higher in London (53%), compared to the North West (22%), North East (14%), Yorkshire and Humber (10%), East Midlands (17%), West Midlands (19%), East of England (15%), South East (11%) and South West (11%).

Differences in the increase in premature mortality rates during the COVID-19 pandemic in people with SMI, by ethnicity

Figure 2 shows that premature mortality rates in people with SMI before the pandemic (2017 to 2019) and during the pandemic (April 2020 to March 2021) were highest in people of white ethnicity. However, Figure 3 shows that those of white ethnicity had the lowest increase in premature mortality during the pandemic (15%). In contrast, a larger increase in premature mortality during the pandemic was observed in people of Asian ethnicity (50%) and those of black ethnicity (43%). The premature mortality rate for people of black ethnicity during the pandemic reached the level of that in people of white ethnicity before the pandemic.

People with SMI of white ethnicity had:

• a premature mortality rate before the pandemic of 1,599 deaths per 100,000 people with SMI

• a premature mortality rate during the pandemic of 1,840 deaths per 100,000 people with SMI

In contrast, people with SMI of black or Asian ethnicity had:

• premature mortality rates before the pandemic of 1,108 deaths per 100,000 people with SMI of black ethnicity and 957 deaths per 100,000 people with SMI of Asian ethnicity

• premature mortality rates during the pandemic of 1,583 deaths per 100,000 people with SMI of black ethnicity and 1,438 deaths per 100,000 people with SMI of Asian ethnicity

Figure 2: premature mortality rate before and during the COVID-19 pandemic in people with SMI, by ethnicity

Figure 3: increase in premature mortality during the COVID-19 pandemic in people with SMI, by ethnicity

Regional differences in the increase in premature mortality during the COVID-19 pandemic in people with SMI

This section describes regional variation in premature mortality during the COVID-19 pandemic. Some of this variation will relate to regional differences in levels of deprivation and ethnic make-up of the population and the impact these had on increased rates of premature mortality in people with SMI.

The premature mortality rate in people with SMI increased during the first year of the pandemic (April 2020 to March 2021) in all regions of England. Figure 4 shows that the highest premature mortality rate before (1,937 deaths per 100,000 people with SMI per year) and during (2,140 deaths per 100,000 people with SMI) the pandemic in people with SMI was in the North East, an increase of 10%. The premature mortality rate in people with SMI in London before (1,259 deaths per 100,000 people with SMI per year) and during (1,579 deaths per 100,000 people with SMI) the pandemic was lower than in the North East. However, despite the comparatively low mortality rates, Figure 5 shows that London had the highest increase in premature mortality during the pandemic (25%).

Figure 4: premature mortality rate before and during the COVID-19 pandemic in people with SMI, by region

Note: the width of lines for England represents the range of the estimate (95% confidence interval), for all bars this range is shown by error bars.

Figure 5: increase in premature mortality during the COVID-19 pandemic compared to before the pandemic in people with SMI, by region

Note: the width of lines for England represents the range of the estimate (95% confidence interval), for all bars this range is shown by error bars.

Regional differences in increased premature mortality during the COVID-19 pandemic in people with SMI compared to people without SMI

People with SMI had a higher premature mortality rate than people without SMI both before and during the pandemic in every region of England. However, the increase in premature mortality during the pandemic was similar for both those with and without SMI both in England and regionally.

The increase in premature mortality was highest in London in both those with and without SMI. During the pandemic in London:

• an extra 320 deaths per 100,000 people were reported in people with SMI

• an extra 90 deaths per 100,000 people were reported in people without SMI

The increase in premature mortality during the pandemic in people aged 18 to 49 with SMI was under 10% for all regions of England except for London, where the increase was 21%. In people aged 50 to 74 with SMI, the increase was also greatest in London.

Excess premature mortality during the COVID-19 pandemic in people with SMI compared to people without SMI

The premature death rate was far higher in people with SMI. In the 15-month study period, people aged between 18 and 74 with SMI had a premature mortality rate of 1,756 per 100,000, compared to a rate of 361 per 100,000 for people without SMI. This means people with SMI were 4.9 times more likely to die before the age of 75 in this period (an excess premature mortality rate of 387%).

Men with SMI had a higher premature mortality rate than women with SMI during the pandemic. However, women with SMI had a higher excess premature mortality compared to their peers without SMI.

Women with SMI had a premature mortality rate of 1,414 deaths per 100,000 during the pandemic. This equated to being 5.1 times more likely to die (an excess of 408%) compared to women without SMI.

Men with SMI had a premature mortality rate of 2,137 deaths per 100,000 during the pandemic. This equated to being 4.8 times more likely to die (an excess of 378%) compared to men without SMI.

Regional differences in premature and excess premature mortality during the COVID-19 pandemic in people with SMI compared to people without SMI

Figure 6 shows regional differences in premature mortality rates during the early part of the pandemic in people with SMI compared to people without SMI.

The premature mortality rate in people with SMI during the pandemic (January 2020 to March 2021) was higher than in people without SMI in every region of England.

The data largely shows a north to south divide with premature mortality rates being significantly higher than the England average in the North East, North West, East Midlands and West Midlands regions. The exception was premature mortality in people with SMI in Yorkshire and Humber where the rate was significantly lower than the England average.

Data on excess premature mortality shows a mixed geographical picture with 4 regions (North East, East Midlands, South East and South West) having significantly higher excess premature mortality than the England average. These are the regions where the greatest additional mortality was seen in people with SMI compared to any increase in mortality in people without SMI, albeit from a different baseline.

The North East region is of particular interest. During the pandemic it was the region with the highest rate of premature mortality in people with SMI and one of the regions with a high rate of premature mortality in those without SMI. It also had one of the highest excess premature mortality figures, suggesting that although premature mortality during the pandemic impacted on the whole population, its effect on those with SMI was even greater.

In the North East:

• the premature mortality rate during the pandemic in people with SMI was 2,155 deaths per 100,000 people with SMI compared to 402 deaths per 100,000 in people without SMI

• the excess premature mortality in people with SMI was 438%

The lowest excess premature mortality in people with SMI was in London (335%) and Yorkshire and Humber (329%), and the premature mortality rate for people in these regions for people with SMI were below the England rate. However, people with SMI in these regions were still over 4 times more likely to die prematurely than people without SMI.

Figure 6: premature mortality rate during the COVID-19 pandemic in people with SMI compared to people without SMI, by region

Note: the width of lines for England represents the range of the estimate (95% confidence interval), for all bars this range is shown by error bars.

Figure 7: excess premature mortality during the COVID-19 pandemic in people with SMI compared to people without SMI, by region

Note: the width of lines for England represents the range of the estimate (95% confidence interval), for all bars this range is shown by error bars.

Conclusions

This report draws attention to the heightened mortality risk experienced by people with SMI during the COVID-19 pandemic compared to the years before the pandemic. People with SMI were also almost 5 times more likely to die during the pandemic than people without SMI.

During the pandemic, for every 100,000 people aged 18 to 74 with SMI, an extra 244 premature deaths per 100,000 were reported, compared to before the pandemic. In contrast, in people without SMI it was an extra 47 premature deaths per 100,000.

In terms of absolute numbers, this translates to 42,815 people aged 18 to 74 with SMI died during the first year of the pandemic, compared to an annual average of 35,025 people with SMI in the 3 years before the pandemic (2017 to 2019). This is an extra 7,790 people, an increase of 22.2%.

For people without SMI, 132,140 people aged 18 to 74 died during the first year of the pandemic compared to an annual average of 112,850 people in the 3 years before the pandemic (2017 to 2019). This is an extra 19,290 people, an increase of 17.1%.

The stark difference in premature mortality between people with and without SMI pre-pandemic, and the sharp rise in premature mortality in people with SMI during the COVID-19 pandemic, highlights the vulnerability of people with SMI to premature death.

When considering how risk factors inter-relate it should be noted that the increase in premature mortality during the pandemic in people with SMI was highest in people living in deprived areas, in older age groups and in the black and Asian ethnic groups. This highlights that people with SMI may have multiple factors which increase their risk of premature mortality.

The increase in premature mortality for people with SMI during the pandemic was particularly high in London. This may reflect the regional impact of COVID-19 early in the pandemic and high levels of deprivation and proportions of people from black and Asian ethnic groups in London. The North East region also deserves attention as it had the highest premature mortality rates both before and during the pandemic.

This report does not describe the causes of death contributing to the premature mortality, for example the split between COVID-19 and other major causes of death. From previous work it is known that in pre-pandemic circumstances 2 out of 3 deaths in people with SMI are from physical illnesses and many of these deaths could be prevented.

Condition-specific premature mortality in people with SMI will be the subject of future reports. Liver disease is used here to illustrate that there are likely to be changes in mortality by cause during the pandemic. It is known that people with SMI are at a substantially higher risk of dying from liver disease than people without SMI - this is often alcohol related. Other work describes the sharp increase in mortality from liver disease during the first year of the COVID-19 pandemic, which was mostly due to a rise in deaths from alcohol related liver disease. Evidence also suggests that alcohol consumption increased during the pandemic especially among people who pre-pandemic drank most heavily. Further investigation of liver disease and other major causes of premature mortality during the pandemic is required.

Other important causes of premature death in people with SMI include cancer, respiratory disease and cardiovascular disease. It is known that for the general population there were delays in cancer diagnosis, or for lung cancer some were not even diagnosed prior to death.

The role of dual or multiple diagnoses in people with SMI and its impact on premature mortality is known but the impact of the pandemic should be evaluated further.

However, interpretation of trends in specific causes of death can be challenging because of displacement of causes of death such as chronic obstructive pulmonary disease (COPD) by COVID-19.  

People with SMI are likely to have been more vulnerable to premature mortality during the COVID-19 pandemic because of their high prevalence of risk factors such as smoking and obesity.

People with SMI under non-pandemic circumstances face challenges in accessing health services and evidence suggests this is likely to have been exacerbated in accessing both GP and in-patient services (reference 4).

While this analysis examines the impact of the COVID-19 pandemic on premature mortality, it is likely that people with SMI would also be vulnerable to other societal changes, such as changes in economic circumstances and climate change. It is therefore important that people with SMI should be given special consideration in pandemic and major incident planning as a high risk group. It will also be important to consider how best to provide health services and support to this hard to reach group during times of adversity or societal change.

Severe mental illness has been recognised as a key clinical area in the NHS England Core20PLUS5 approach. Mental illness will be one of the key major conditions in the Department of Health and Social Care’s major conditions strategy. Both highlight the need to improve clinical outcomes and reduce premature mortality in people with SMI.

Next steps

The findings of this report highlight the vulnerability of people with SMI to infectious disease pandemics. Similar vulnerability to premature mortality may occur in response to other extreme events and times of societal change. This analysis highlights the need for recognition of the vulnerability of this group by health and social care professionals and others in contact with people with SMI. The findings should also inform actions of emergency planners, public health specialists, service commissioners and providers, and policy makers, to mitigate the impact of future pandemics and other potential extreme events.

For policy makers, it is important that the needs of people with SMI are considered in line with the Equality Act 2010 and this includes provisions of appropriate and timely guidance.

For service commissioners and providers, it is important that people with SMI are considered in strategic and emergency planning. For example:

• securing access to health care for this vulnerable group during lockdowns

• consideration of people with SMI as a priority group for vaccination during future pandemics and other significant communicable diseases outbreaks

• mitigation of the risks of extreme hot or cold weather to health and wellbeing in this population

The analytical and research community also has an important role in providing new data and evidence to support the actions of service commissioners and providers, and policy makers. Further work is required to understand:

• how much of the excess premature mortality in people with SMI during the pandemic was due to COVID-19 and how much was due to other causes

• impact of other societal or extreme environmental events on mortality rates in this population

• the role of dual or multiple diagnoses in people with SMI and impact on premature mortality during the pandemic

Methodology

Strengths and limitations

Strengths of this analysis include:

1. This is a novel analysis, highlighting the negative impact, in terms of premature mortality, of the COVID-19 pandemic on people with SMI.

2. The impact of the COVID-19 pandemic on people with SMI is analysed by age, sex, ethnicity and region. This demonstrates differential effects of the pandemic on specific subgroups and also geographically.

3. The methodology accounted for and attempted to compensate for differences in delays in reporting deaths between people with and without SMI.

Limitations include:

1. There are differences in how SMI can be defined. In primary care SMI is defined as people with a diagnosis of schizophrenia, bipolar disorder and other psychotic illnesses, whereas definitions in other services might use a wider range of conditions. This analysis uses “people referred to or in contact with secondary care mental health services in the preceding 5 years” to identify those with SMI. This definition may include people who do not have SMI but are accessing care for other mental health conditions.

2. This analysis defines people who died with SMI as those with evidence of a referral to or contact with secondary care mental health services in the 5 years before death. Not all people with SMI will have regular treatment in secondary care. People with these illnesses with stable disease being managed by their GP were not included.

3. Directly standardised mortality rates were used throughout, age-standardised to the European Standard Population, 2013. The overall analyses therefore do not take account of other causes of variation in mortality rates during the pandemic such as sex, ethnicity or deprivation.

4. Analysis by region does not account for variation in mortality rates at a smaller scale. There are likely areas of both high and low mortality in all regions of England.

5. The period from January 2020 to April 2021 was a period of changing social restrictions, healthcare access and COVID-19 rates. The pandemic period used in this report covers January 2020 to April 2021. Variation in mortality over this period is not assessed. A previous paper found that in London the main increase in mortality was in the second quarter of 2020. By the third quarter and fourth quarter mortality rates returned to near normal for that time of year. The first and second wave of the COVID-19 pandemic hit regions differently in terms of spread and mortality (reference 1).

6. While reporting periods were adjusted to account for delays in death registration in people with SMI, these delays may still have resulted in fewer deaths being counted in this group.

Cohort definition

This work concentrates on the impact of COVID-19 on the mortality rate of people with SMI. However, there is no national data set with both mental illness diagnoses and linkage to individual death records. The data source used for this report (Mental Health Services Dataset (MHSDS)) contains limited information about patient diagnoses. This means that people with a referral to or contact with secondary care mental health services was used as a substitute for people with SMI.

Defining contact with mental health services at time of death

Death records were used to identify a cohort of people that died from January 2017 to June 2021. These were linked to records of secondary care mental health services. Those referred to or in contact with these services in the 5 years before death were defined as having SMI. Some groups of people were likely in contact with mental health services for reasons other than complex or severe mental illnesses and were therefore included. People who were only in contact with learning difficulties or autism services were excluded.

People with no record of contact with or referral to secondary care mental health services in the 5 years before death were defined as not having SMI.

Ethnic groups

Ethnicity was obtained from the MHSDS and its predecessor datasets, as it is not routinely recorded as part of death registration. Therefore, ethnicity was only available for the people with a referral to, or contact with, mental health services. This was available for 75% of this cohort. Latest known ethnicity record was used. The ethnicity record was self-defined by the individual and was collected using the national code definition (see MHDS v4.1 Technical Output Specification).

Because of small populations for some ethnic groups (especially in some regions), aggregated ethnic groups were used in line with Race Disparity Unit recommended standards for use by the government and its departments.

Deprivation quintiles

Area level deprivation was used and based on place of residence of the individual as recorded at the death registration or as the latest record within the MHSDS and its predecessor datasets. Index of Multiple Deprivation 2019 (IMD2019) national quintiles were used and were assigned based on the rank of IMD2019 score as quintile 1 for most deprived and quintile 5 for least deprived.

Death registrations

People with a date of death between January 2020 and March 2021 were defined as dying during the pandemic. For comparison, people with deaths between January 2017 and December 2019 were defined as dying before the pandemic.

The registration of some deaths may be delayed due to referral of the death for investigation by a coroner. The deaths of people with SMI are more likely to be referred to a coroner than the deaths of people without SMI, because of the causes and circumstances of their deaths. To account for this, different cut-offs between date of death and registration of death were used.

The 15-month period of January 2020 to March 2021 was chosen for the pandemic period to enable as many months as possible of pandemic cases to be included in the analysis.

For people in contact with mental health services, death records during the pandemic period were included if the death was registered up to 30 June 2021. Deaths in the period before the pandemic were included if they were registered by 31 March of the following year. For example, deaths in 2018 were included if they were registered by the 31 March 2019.

For people not in contact with mental health services, deaths during the pandemic period were included if they were registered up to 30 April 2021. Deaths in the period before the pandemic were included if they were registered by 31 January of the following year. For example, deaths in 2018 were included if they were registered by the 31 January 2019. Even after doing this, the deaths of people in contact with mental health services were estimated to be less complete than those not in contact.

Another source of delay in the registrations was the COVID-19 pandemic itself. At the beginning of the pandemic (April to September 2020), the high number of deaths led to longer delays than in previous years. However, deaths early in the pandemic will have been captured later during the study period. A report by the Office for National Statistics on the impact of registration delays on mortality statistics in England and Wales in 2020 found a similar proportion of deaths with delays over 2 weeks than in previous years.

Indicators used in this report

The following indicators are used to measure mortality in this report.

The mortality rate

This is the number of people who die in a specified population and time period, per 100,000 people. This indicator is calculated as a directly standardised rate, meaning it shows the number of people which would die per 100,000 people if the age structure of the population was the same as a reference population (European Standard Population, 2013). This allows comparison over time and across geographical areas.

The mortality rate for those with SMI was calculated as the number of people in a given time period who died and had been referred to or had contact with secondary care mental health services in the 5 years before death, divided by the total number of people who had been referred to or had contact with secondary mental health services in the 5 years before that time period. This number was then multiplied by 100,000.

The mortality rate for those not in contact with mental health services was calculated as the number of people in the time period who died and did not have contact with mental health services in the 5 years before death. This was divided by the total number of people in the general population minus those who had contact with secondary mental health services in that time period. This number was then multiplied by 100,000.

Increase in mortality rate

This measures the increase in mortality rate during the pandemic period in people with SMI compared to the average annual mortality rate in people with SMI in the 3 years before the pandemic. This was calculated by taking the death rate during the pandemic in people with SMI and subtracting the average annual death rate in people with SMI for the period before the pandemic. This gives the difference between the 2 groups. This is then divided by the average annual rate in the period before the pandemic and multiplied by 100 to give the percentage increase in mortality rate.

Excess mortality

This is the percentage difference between people in contact with mental health services and those not in contact. A difference of 0% would indicate similar rates of mortality. A difference of 100% would indicate twice the mortality rate in the population of interest and may also be expressed that people in that population are twice as likely to die.

Excess mortality during the COVID-19 pandemic period is people in contact with mental health services compared to people not in contact with mental health services. This was calculated by taking the death rate during the pandemic in people in contact with services and subtracting the death rate in people who were not in contact with these services. This gives the difference between the 2 groups. This is then divided by the rate in those not in contact with services and multiplied by 100 to give the percentage excess.

Estimating mortality

Throughout this report, a range is provided alongside estimates of mortality. In graphs these are represented as error bars, or by the width of horizontal bars. This is based on the 95% confidence intervals, a way of showing uncertainty, and the likely range that contains the true estimate of mortality.

Data sources

This report used the following data sources:

• Office for National Statistics Civil Registration of Deaths

• Mental Health Services Data Set (MHSDS) and its predecessor datasets

• Office for National Statistics mid-year population estimates

References

1. Jayati Das-Munshi, Chin Kuo Chang, Ioannis Bakolis, Matthew Broadbent, Alex Dregan, Matthew Hotopf and others. All-cause and cause-specific mortality in people with mental disorders and intellectual disabilities, before and during the COVID-19 pandemic: cohort study. The Lancet Regional Health - Europe, 2021.

2. Michael W Best and Christopher R Bowie. Social exclusion in psychotic disorders: an interactional processing model. Schizophrenia Research, 2022.

3. Barrett EA, Simonsen C, Aminoff SR, Hegelstad WTV, Lagerberg TV, Melle I, Mork E, Romm KL. The COVID-19 pandemic impact on wellbeing and mental health in people with psychotic and bipolar disorders. Brain and Behavior, 2022.

4. Boyer L, Fond G, Pauly V, Orléans V, Auquier P, Solmi M, Correll CU, Yon DK, Llorca PM, Baumstarck-Barrau K, Duclos A. Impact of the COVID-19 pandemic on non-COVID-19 hospital mortality in patients with schizophrenia: a nationwide population-based cohort study. Molecular Psychiatry, 2022.

Acknowledgements

This report was produced by OHID’s National Mental Health Intelligence Network.

We are particularly thankful for the contributions from Alison Brabban, Wendy Burn, Alan Cohen, Carl Child, David Fisher, Joseph Hayes, Sarah Holloway, Alison Hooper, Misha Imtiaz, Tim Kendall, David Kingdon, J Kirkbride, Alexandra Lazaro, Naomi Launders, Beth McGeever, Athira Manoharan, David Osborn, Giovanna Polato, Leila Reyburn, Lucy Schonegevel, Robert Stewart, Alex Stirzaker, Emma Tiffin, Mark Wagner and Matthew Wickenden.

This report has been co-funded by NHS England.