Part of COVID-19 Population Risk Assessment Summary Statistics: High Risk Population
Observations
Risk thresholds
England’s CMO, in consultation with senior clinicians, used the results of the QCovid® research to define the risk threshold that should be used to determine if someone may be high risk.
The research indicated that the 2% of the study population with the highest risk results were most likely to die from coronavirus. This was translated from the combined risk of catching and dying from coronavirus with results higher than or equal to:
- an Absolute Risk of 0.5% (or 5 in 1,000)
- a Relative Risk of 10 (or 10 times the baseline risk)
Absolute Risk (AR) is the overall risk based on the health outcome of people with the same characteristics and risk factors, who contracted COVID-19 and were admitted to hospital or died as a result.
Relative Risk (RR) is the level of risk compared to a person of the same age and sex at birth, without any other risk factors.
Figure 1 to Figure 3 show the three risk profiles of the high risk cohort by age group and sex at birth.
The Absolute Risk profile appears to be predominantly males and in the older age groups.
The Relative Risk profile appears to be predominantly female and in the younger to middle-age groups.
The majority of people identified as high risk on both (Absolute Risk and Relative Risk) risk profiles appear to be males for lower age groups and females for the higher age groups.
Summary
Most people on the Absolute Risk profile are male and aged 70 years and over. This compares to the Relative Risk profile where most are female and aged below 70 years.
People on the Absolute and Relative Risk profiles were mostly males for the younger age groups and females for the older age groups.
Use of default values on QCovid®
The QCovid® calculation engine requires specific pieces of information about each person to generate their risk assessment results. If key information is missing from their record or has been recorded as unknown or outside the minimum/maximum thresholds accepted by the model, default values are used as a substitute in the calculations.
A precautionary approach was taken in the COVID-19 PRA when processing patient records. This was to ensure the calculated risk of anyone catching coronavirus and dying was not underestimated, which would have resulted in them being excluded and not added to the SPL.
To mitigate this risk, default values used in the COVID-19 PRA had higher-than-average risk associated with them, apart from postcode where the default value had an average risk, based on the QCovid® model. This means that the COVID-19 PRA is likely to be overestimating the level of risk for people with information missing from their records, and therefore identifying more people as potentially high risk.
| Type and quantity of default data used | Number with valid postcode | Number with invalid postcode | % in high risk cohort |
|---|---|---|---|
| One default used | 362,670 | 212 | 23.6% |
| BMI missing only | 140,083 | 92 | 9.1% |
| Ethnic Category missing only | 152,736 | 77 | 9.9% |
| BMI (outside min/max range) only | 69,287 | 43 | 4.5% |
| Sex at birth missing only | 564 | 0 | 0.0% |
| Two defaults used | 64,426 | 87 | 4.2% |
| BMI missing and Ethnic Category missing | 55,097 | 81 | 3.6% |
| BMI missing and Sex at birth missing | 715 | 1 | 0.0% |
| BMI (outside min/max range) and Ethnic Category missing | 8,589 | 5 | 0.6% |
| BMI (outside range) and Sex at birth missing | 9 | 0 | 0.0% |
| Sex at birth missing and Ethnic Category missing | 16 | 0 | 0.0% |
| Three defaults used | 705 | 0 | 0.0% |
| No defaults used | 1,107,875 | 860 | 72.1% |
| TOTAL | 1,535,676 | 1,159 |
More than 28% of people in the high risk cohort have missing data for at least one attribute across ethnic category, sex at birth or BMI, or have data outside the acceptable minimum-maximum range (a BMI below 15 or above 47), which resulted in the use of default data in the Covid-19 PRA calculation.
Overall, less than 0.1% have a missing or invalid postcode.
Distributions by age and sex at birth
Figure 5 and Figure 6 show the number of people in each 5-year age group for female and male adults.
People whose sex at birth is unknown are excluded from the figures below since numbers are small, with less than a half of one percentage point of each cohort have missing sex at birth.
There are approximately 1.5 million people in the high risk cohort, with a 51:49 split between female and male.
The largest difference between the number of males and females is between ages 75 to 84, with more males compared to females and, conversely, between ages 30 to 39 with more females compared to males.
There are approximately 17 million people in the Total PRA cohort, with a 55:45 split between female and male.
The highest difference between the number of males and females is between ages 70 to 79, with more males compared to females.
Summary
Both cohorts appear to be slightly skewed towards females. The Total PRA cohort appears balanced for most age groups apart from those aged between 70 to 79 years.
This pattern is not reflected for the same age group in the high risk cohort indicating that, although many people in this age group were identified as potentially high risk, many of them were not assessed as high risk based on the QCovid® model.
Ethnic category
There are three figures for each of the cohorts. Figures with suffix 'a' show the overall distribution of people by ethnic category, figures with suffix 'b' show the number of people by age group and ethnic category, and figures with suffix 'c' show the proportion of people within each ethnic category by age group.
Figures with suffix 'b' and 'c' are displayed as a 3-colour heat map to show variation between the highest (red), middle (amber) and lowest (green) numbers or proportions.
People whose ethnic category has not been recorded are included in this analysis and are categorised as ‘Not Stated or Unknown'.
High-Risk cohort
Just under a half of all people in the high risk cohort identify as White British (49%).
Ethnic category Black African has the next highest representation at 6%, followed by those who identify as Indian or Pakistani, each accounting for 5% of the high risk cohort.
The two ethnic categories with the lowest representation in the high risk cohort are White and Asian Mixed or White and Black African Mixed, each representing less than 0.5% of this cohort.
People in the high risk cohort whose ethnic category is not stated or unknown affects 14% overall.
Figure 7b: High risk cohort – Population distribution by ethnic category and age group
Figure 7c: High risk cohort – Percentage age distribution within each ethnic category
Figures 7b and 7c show that around 50% who are White British and 60% who are White Irish in the high risk cohort were aged 80 years or older.
For other ethnic categories, more than 50% were aged between 30 and 59 years except for those who are Caribbean where they accounted for 40% within this age group. For those identifying as Caribbean, more than a quarter (28%) were aged between 80 and 89 years.
Total PRA cohort
More than a half of all individuals in the total PRA cohort identify as White British (56%).
Ethnic category Indian has the next highest representation at 5%, followed by those who identify as Other White Background, Other Ethnic Group including Arab, or Pakistani, each accounting for 4% of the total PRA cohort.
The two ethnic categories with the lowest representation in the total PRA cohort are White and Asian Mixed or White and Black African Mixed, each representing less than 0.5% of the cohort.
People in the total PRA cohort whose ethnic category is not stated or unknown affects 15% overall.
Figure 8b: Total PRA cohort – Population distribution by ethnic category and age group
Figure 8c: Total PRA cohort – Percentage age distribution within each ethnic category
In the total PRA cohort, around 18% who identify as White British and 25% who are White Irish are aged 80 years or older.
For other ethnic categories, at least 60% of them are aged between 19 and 49 years, younger than the age profile seen in the high risk cohort. There are two exceptions, for those who identify as Caribbean or Other White Background, where around 40% are aged between 19 and 49 years.
Summary
Both cohorts are predominantly White British, which constitutes 49% of the high risk cohort and 56% of the total PRA cohort. The proportion of people who identify as White Irish or Any Other Black African Caribbean are similar across both cohorts, accounting for approximately 1% of the total cohort.
White and Asian Mixed or White and Black African Mixed ethnic categories have the lowest representation on both cohorts, accounting for less than 0.5% of the total cohort.
Missing ethnic category data accounts for 14% and 15% of the high risk and total PRA cohorts respectively.
The age group distribution of ethnic categories in the total PRA cohort is skewed towards young adults except for those who are White British, White Irish or those in Other White Background. Comparatively, the distribution appears to be more towards middle-age groups in the high risk cohort, except for those who are White British or White Irish.
Body Mass Index (BMI)
BMI is a measure used to calculate whether a person’s weight is healthy. The general BMI categories are:
- Underweight – BMI is less than 19 kg/m2
- Healthy weight – BMI is 20 kg/m2 and below 25 kg/m2
- Overweight – BMI is 25 kg/m2 and below 30 kg/m2
- Obese – BMI is 30 kg/m2 and below 40 kg/m2
- Severely (or morbidly) obese - BMI is at least 40 kg/m2
The below analysis reports on two broad categories - whether a person is obese or not. This is because people who are obese or ‘excessively’ overweight are at an increased risk of serious complications and death from COVID-19, compared to those with a healthy weight, a report from Public Health England (PHE) has concluded.
A person is categorised as obese if their reported BMI is at least 30 kg/m2 Similarly, a person is categorised as not obese if their reported BMI is less than 30 kg/m2.
People with an unknown BMI are included in this analysis (categorised as Missing), as are those whose BMI value falls outside the required range, as this affects a large proportion of people within each cohort. People with a BMI <15 kg/m2 are categorised as not obese and people with a BMI >47 kg/m2 are categorised as obese.
There are six figures for each of the high risk and total PRA cohorts. Figures with suffix 'a' show the overall distribution of people by all BMI categories. Figures with suffix 'b' show the number of people by age group and reported BMI categories. Figures with suffix 'c' show the proportion of people in the reported BMI categories by age group. Figures with suffix 'd' and 'e' are a subset of figures with suffix 'b' and 'c' and include people whose sex at birth is female only. Similarly, figures with suffix 'f' and 'g' include people whose sex at birth is male only.
Figures with suffix 'b' to 'g' are displayed as a 3-colour heat map to show variation between the highest (red), middle (amber) and lowest (green) numbers or proportions.
High Risk cohort
Approximately 40% of people in the high risk cohort were either obese or severely obese, with more than 30% of them classified as severely obese. Whereas approximately 6% and 18% were underweight or within the healthy range respectively. A further 22% were overweight.
Figure 9b: High risk cohort – Population distribution by BMI range and age group
Figure 9c: High risk cohort – Percentage age distribution within each BMI range
Obesity is highly represented in middle-aged adults (35 to 64 years) in the high risk cohort with a third of obese people in age groups 45 to 59 years.
Approximately 13% of people in the high risk cohort had missing BMI data, most of them in the older age groups, 80 years and above.
Figure 9d: High risk cohort (Female only) – Population distribution by BMI range and age group
Figure 9e: High risk cohort (Female only) – Percentage age distribution within each BMI range
Obesity among females in the high risk cohort is highly represented between ages 35 to 59. Missing or unknown BMI values are concentrated towards the oldest age group, at least 90 years, as well as those aged between 30 and 44 years.
Figure 9f: High risk cohort (Male only) – Population distribution by BMI range and age group
Figure 9g: High risk cohort (Male only) – Percentage age distribution within each BMI range
Obesity among males in the high risk cohort is highly represented between ages 45 to 59. Missing or unknown BMI values are concentrated toward higher age groups, 80 years and above, which accounts for more than 50% of those in the missing category.
Total PRA cohort
Approximately 26% of the total PRA cohort were categorised as either obese or severely obese, with 16% of them categorised as severely obese. A further 5% of them were categorised as underweight, 22% within the healthy range, and 27% were overweight.
Figure 10b: Total PRA cohort – Population distribution by BMI range and age group
Figure 10c: Total PRA cohort – Percentage age distribution within each BMI range
Obesity is highly represented in middle to higher age groups (50 to 74 years) in the total PRA cohort.
Approximately 20% of people in the total PRA cohort had missing BMI data, most of them in the younger to middle- age groups (59 years and below).
Figure 10d: Total PRA cohort (Female only) – Population distribution by BMI range and age group
Figure 10e: Total PRA cohort (Female only) – Percentage age distribution within each BMI range
Obesity among females in the total PRA cohort is highly represented between ages 50 to 59. Missing or unknown BMI values are concentrated for ages between 35 and 59.
Figure 10f: Total PRA cohort (Male only) – Population distribution by BMI range and age group
Figure 10g: Total PRA cohort (Male only) – Percentage age distribution within each BMI range
Obesity among males in the total PRA cohort is highly represented between ages 50 to 79 with more than 13% of them aged between 70 to 74 years. Missing or unknown BMI values are concentrated toward the younger age groups (34 years and below).
Summary
There is a higher proportion of obese or severely obese people in the high risk cohort compared to the total PRA cohort with 27% and 13% of the total compared to 22% and 4% respectively.
The age profile of obese adults in the high risk cohort are generally younger, with more than half of them in age groups between 35 and 59 years compared to the total PRA cohort where almost half of them were aged between 50 and 74 years.
There is a higher proportion of adult females classified as obese in the high risk cohort compared to those in the total PRA cohort, 45% compared to 28%.
There is little difference overall in the age profile of females who are obese in the high risk cohort and the total PRA cohort. In both cohorts, obesity is over-represented in all age groups between 40 and 74 years.
The proportion of missing or unknown BMI data is lower in the high risk cohort, 13% compared to 20% in the total PRA cohort. Missing or unknown BMI data is more evident in older people aged 80 and above in the high risk cohort, whereas it is the younger age groups below 50 years in the total PRA cohort.
Last edited: 26 November 2021 5:34 pm