A practice-based survey of mortality: what impact has Covid-19 had on mortality patterns?

Introduction

The gathering of mortality data in general practice is a useful exercise both educationally and in terms of highlighting local health needs; we also believe it enhances transparency about clinical practice. In 2018 we published in this journal a practice-based mortality survey, co-authored with two senior nurse colleagues with managerial responsibility for care homes, which was a retrospective analysis of 50 consecutive deaths, in which data was gathered on the age at death, cause and place of death, whether or not the patient had expressed a desired place of death, and how that correlated with actual place of death.¹

In the light of the Covid-19 pandemic, however, we wished to investigate whether, and to what degree, the latter has changed mortality patterns. This is a similar retrospective survey of 50 consecutive deaths for which complete data was available, analysing all the parameters in our original survey except for desired place of death, which was not considered this time.

Further reading: A practice-based survey of mortality

Chorleywood Health Centre is a four-doctor teaching practice with a list size of 7,000 located in a mixed suburban/semi-rural commuter zone to the North-West of London. Though the area is affluent, there are high levels of healthcare need driven by a substantial elderly population. Our original survey in 2018 took place when two nursing homes account were active locally, but this one coincided with the temporary closure of one of them for refurbishment; nevertheless, the single remaining nursing home, though accounting for only 0.7% of the practice population, generates a large proportion of end-of-life care. The practice has a culture of autonomous patient management, including palliative care, and has published two other mortality surveys.^2,3

Aims

The two broad classifications were based on pathology (deaths from malignancy vs. non-malignant disease) and place of death. As our broad ethos is to manage end-of-life care in the community whenever it is reasonable to do so, deaths occurring in patients’ own homes and nursing homes, whilst considered separately, were jointly classified as ‘home’ deaths. This data was compared to the earlier survey of 2018.¹

Methods

The practice records all notified deaths in the electronic patient record, as well as manually. Reference was made to death certificates, hospital death notices, and post-mortem examinations. The subjects were 50 patients of the practice who died consecutively in a period from August 2020 to May 2021, and for whom complete data was available, which required analysing 56 records as six were incomplete. For each patient the age, place of death and cause of death were determined.

The cause of death was recorded with a focus on any underlying chronic disease, which in practice meant sometimes using line 1b of the death certificate rather than 1a. For example, if a patient died from pulmonary oedema secondary to ischaemic heart disease (IHD), the latter was recorded as the cause of death. The causes of death were classified by organ systems. The malignant neoplasms are listed in decreasing order of frequency.

Results

Table 1 shows the numbers, ages, and place of death of this cohort. Amongst the 50 deaths, 22 were men and 28 women, with an average age at death of 82.1 years. 18 (36%) died in a nursing home, and 10 (20%) in their own homes, meaning that a total of 28 (56%) died at home. 19 (38%) died in hospital and 3 (6%) in hospices.

Total number of deaths

All patients: 50 (100%)

Men: 22 (44%)

Women: 28 (56%)

Average age at death

All patients: 82.1 years

Men: 80.3 years

Women: 83.5 years

Place of death

Own home: 10 (20%)

Nursing home: 18 (36%)

Hospital: 19 (38%)

Hospice: 3 (6%)

Cause of death

Deaths were initially classified by organ systems, and a detailed breakdown is given in Table 2. Deaths from malignant neoplasms are listed in Table 3.

Table 1. Numbers, ages and place of death of patients in survey

Table 2 shows the cause of death by organ systems. Covid-19 pneumonia emerged as the single most common cause of death in this cohort with 10 cases, and with three cases of non-Covid-19 pneumonia and four cases of lung cancer, it means that respiratory disease accounted for 17 (34%) of deaths, far ahead of the next highest organ system which was neurological with 7 (14%). Overall, 37 (74%) died from non-malignant disease and just 13 (26%) from cancer.

Cardiovascular

Ischaemic heart disease: 4

TOTAL: 4

Respiratory

Covid-19 pneumonia: 10

Other pneumonia: 3

Lung cancer: 4

TOTAL: 17

Gastrointestinal

Colorectal cancer: 2

Diverticular disease: 1

TOTAL: 3

Kidney and urinary tract

Pyelonephtitis: 1

TOTAL: 1

Male reproductive system

Prostate cancer: 1

TOTAL: 1

Breast

Breast cancer: 2

TOTAL: 2

Neurological

Dementia: 6

Parkinson’s disease: 1

TOTAL: 7

Dermatological

Malignant melanoma: 1

Infected venous ulcer: 1

TOTAL: 2

Connective tissue:

Liposarcoma: 1

TOTAL: 1

Others:

Cancer (unknown primary): 1

Frailty of old age: 10

TOTAL: 11

Table 2. Classification of death by organ system

Table 3 lists the malignant causes of death, showing lung as the most common (4 cases) followed by 2 each of colorectal, prostate and breast.

Lung:	4
Colorectal:	2
Prostate:	2
Breast:	2
Malignant melanoma:	1
Liposarcoma:	1
Unknown primary:	1
TOTAL:	13

Table 3. Deaths due to malignant neoplasms (decreasing incidence)

Comparison with our earlier data from 2018¹ shows four striking similarities:

The sex ratio is identical.
The places of death were almost identical, being 18% in own home, 40% in nursing home, 36% in hospital and 6% in hospice in the 2018 cohort.
Deaths from malignancy were comparable (24% in earlier cohort).
Lung cancer continues to be the leading cause of cancer death.

By contrast, three significant differences appear to be:

A reduction in average life expectancy for both men and women (85 in the 2018 cohort vs. 82.1 in this re-evaluation).
The emergence of Covid-19 pneumonia as the leading cause of death in this cohort.
An increase in frailty of old age as the cause of death.

Conclusion and discussion

We acknowledge, as we did in our earlier study, the limitations of a small sample size, where small numbers may significantly skew percentages.¹ Regarding the place of death, our findings show data remarkably consistent across the two studies. Given that dying at home is viewed as important by patients, with 50-60% expressing a wish to do so,^4,5 it is pleasing that our practice manages to achieve this rate.

By contrast, Gomes and Higginson’s analysis of national data has shown an inexorable decline in home deaths to around 18%,⁶ with around 10% dying in nursing homes;⁶ by contrast a significant majority of 58% died in hospital,⁶ a figure far higher than for our practice.

Nursing homes continue to provide much of the end-of-life care. Our data is also consistent across both studies for the steady, and low, death rate from cancer, and the fact that lung cancer remains the commonest fatal cancer, given both its high incidence and poor (albeit slowly improving) prognosis: this is consistent with national data⁷.

The cohort of deaths we studies coincided with the peak of the second wave of deaths from Covid-19, which emerged as the leading cause of death of patients of this practice during this period. There appears to have been a significant reduction on life expectancy relative to our earlier data. Nationally, a reduction in life expectancy has been noted during the Covid-19 pandemic and attributed directly to excess deaths from it.⁸ However, in our study the deaths from Covid-19 had an average age of 85.2 (range 70 to 93) which was higher than average for the whole cohort, so this does not appear to be a relevant factor here.

All these patients had significant comorbidities, so that apparent reduction in cases of dementia and cerebrovascular disease must be interpreted with caution. These were in most cases secondary causes of death where Covid-19 pneumonia was deemed the primary cause. It is well recognised that there has been a tailing-off of life expectancy since 2011,⁹ well before the Covid-19 pandemic, in addition to which the unintended consequences of the pandemic and its restrictions, including exacerbation of inequalities and reduced access for non-Covid health problems may independently reduce life expectancy,⁹ though our experience is that these were not relevant factors in our study. Indeed, it is likely that our cohort’s reduced life expectancy was mainly attributable to the 4 deaths in younger patients (range 47 to 56).

Our data indicates that at its peak, Covid-19 proved to be immensely consequential for primary care. Whilst it is widely thought as well as hoped that future outbreaks will be less potent,¹⁰ it is also a constantly mutating virus that is likely to remain endemic,¹⁰ a fact rarely acknowledged by the media where it is still invariably referred to as a pandemic.¹¹

Going forward, health services must reconcile coping with Covid-19’s ongoing presence, without neglecting its duties in all other areas of healthcare. With more than six million currently waiting for elective NHS care in England, and the National Audit Office predicting that this could double to 12 million by 2025, it is clearly something the NHS is lamentably failing to do.

Edin Lakasing, General Practitioner, Chorleywood Health Centre, Chorleywood, Hertfordshire. E-mail: [email protected]

Alison E. James, General Practitioner, Chorleywood Health Centre, Chorleywood, Hertfordshire

Conflict of interest: none.

References

Lakasing E, James AE, Parry SB, Mutero F. A practice-based survey of mortality. GM, 48(6), June 2018. https://www.gmjournal.co.uk/a-practice-based-survey-of-mortality (accessed 26th January 2022).
Lakasing E, Mahaffey W. A practice-based survey of mortality patterns and terminal care provision. British Journal of Community Nursing 2005; 10: 378-80.
Lakasing E, Sparkes C. A practice-based survey of mortality revisited: what trends in end-of-life care are emerging? British Journal of Community Nursing 2010; 15 (5): 236-240.
Seamark D, Blake S, Brearley SG, Milligan C, Thomas C, Turner M, Wang X, Payne S. Dying at home: a qualitative study of family carers’ views of support provided by GPs community staff. Br J Gen Pract 2014; 64 (629): e796-e803.
Higginson IJ, Sen-Gupta GJA. Place of care in advanced cancer: a qualitative systemic literature review of patient preferences. J Palliat Med 2000; 3(3): 287-300.
Gomes B, Higginson I. Where people die [1974-2030]: past trends, future projections and implications for care. Palliat Med 2008; 22(1): 33-41.
Neal RD, Sun F, Emery JD, Callister ME. Lung cancer. BMJ 2019; 365:
Public Health England. Health Profile for England 2021. https://fingertips.phe.org.uk/static-reports/health-profile-for-england/hpfe_report.html (accessed 26th January 2022).
Raleigh V. What is happening to life expectancy in England? King’s Fund, 6 December 2021. https://www.kingsfund.org.uk/publications/whats-happening-life-expectancy-england (accessed 26th January 2022).
Torjesen I. Covid-19 will become endemic but with decreased potency over time, scientists believe. BMJ 2021; 372:
Tayag Y. From pandemic to endemic: this is how we might get back to normal. The Guardian, 5^th December 2021. https://www.theguardian.com/us-news/2021/dec/05/covid-19-from-pandemic-to-endemic-this-is-how-we-might-get-back-to-normal (accessed 26th January 2022).
National Audit Office. NHS backlogs and waiting times in England. 1^st December 2021. https://www.nao.org.uk/press-release/nhs-backlogs-and-waiting-times-in-england/ (accessed 26^th January 2022).

covid-19 mortality patterns