Life expectancy by sexual orientation, England and Wales: March 2021 to March 2024

• The England and Wales Census 2021 included a voluntary question on sexual orientation for the first time; we have linked census data to death registrations to estimate life expectancy according to sexual orientation.

At age 20 years:

• The life expectancy for men identifying with an LGB+ orientation ("gay or lesbian", "bisexual" or "other sexual orientation") was 1.2 years lower compared with men identifying as straight or heterosexual; life expectancy was 59.4 years for LGB+ men and 60.7 years for straight or heterosexual men.

• The life expectancy for women identifying with an LGB+ orientation was 0.9 years lower compared with women who identified as straight or heterosexual; life expectancy was 63.0 years for LGB+ women and 64.0 years for straight or heterosexual women.

• The life expectancy for individuals identifying with a gay or lesbian sexual orientation was 59.3 years for men and 62.8 years for women; for those identifying with a bisexual orientation, life expectancy was 59.6 years for men and 62.9 years for women.

• These differences in remaining life expectancy between people who identify as LGB+ and people identifying as straight or heterosexual are based on large samples with consistent patterns across age groups and previous findings; however, we cannot say from this analysis whether sexual orientation is a contributing risk factor for decreased life expectancy.

Confidence intervals

A confidence interval is a measure of the uncertainty around a specific estimate. As confidence intervals around estimates widen, the level of uncertainty about the true value increases. Larger population groups, such as our sampled straight or heterosexual population, will have a lower level of error compared with smaller population groups, such as our sampled LGB+ population. Therefore, the widths of confidence intervals reported in our data have sizeable differences.

The differences in life expectancy between groups that are reported in this bulletin are not necessarily statistically significant. No formal statistical significance testing was applied.

Period life expectancy

The life expectancy estimates reported here are period based. Period based life expectancy at a given age is the average number of additional years a person would live if he or she experienced the age-specific mortality rates for that time period, throughout the rest of their life. It is not the number of years a person will be expected to actually live. The measure of life expectancy reported in this release reflects mortality rates for this linked population for a particular period in time.

For this publication we report life expectancy at age 20 years, not at birth, because sexual orientation data are available in Census 2021 from age 16 years only. Life expectancy can therefore be interpreted as the average number of additional years of life at age 20 years.

More information can be found in our Period and cohort life expectancy explained methodology.

Sexual orientation

Sexual orientation is an umbrella term covering sexual identity, attraction, and behaviour. For an individual Census 2021 respondent, these may not be the same. For example, someone in an opposite-sex relationship may also experience same-sex attraction, and vice versa. Therefore, these statistics should be interpreted only as showing how people responded to the census question, rather than necessarily being about who they are attracted to or their actual relationships.

This analysis does not make any assumptions about sexual relationships or behaviour, nor does it represent whether someone's sexual orientation is publicly disclosed or is private.

The sexual orientation groups included in this release were "gay or lesbian", "bisexual" and "other sexual orientation"; these groups were aggregated as an LGB+ group. For "other sexual orientation", respondents were asked to write in the sexual orientation with which they identified. Common answers included pansexual, asexual, queer, or the field was left blank. The LGB+ groups were then compared with the straight or heterosexual orientation group.

No information on gender identity was used in this release. Given the level of uncertainty associated with the responses to the Census 2021 gender identity question, the data cannot be used to give precise estimates of the size of the relationship between gender identity and health outcome data.

The estimates reported in this bulletin are period-based life expectancies, estimated using an abridged life table. Our Life expectancy estimates template shows how the abridged life table is used to derive life expectancy estimates. 

Abridged life tables, which group ages into five-year intervals, are used in preference to complete life tables for smaller populations. This is because death counts can be too low for examining mortality for single years of age.

More quality and methodology information on life expectancy is available in our Health state life expectancies, UK quality and methodology information (QMI). 

Data sources

This analysis used anonymised data from Census 2021 as our population base. Census 2021 data were linked to NHS numbers in the NHS Personal Demographics Service (PDS) 2019, with a linkage rate of 94.4%. More information about Census 2021 linkage to the PDS is available in our Census 2021 to PDS linkage report. To adjust for non-linkage to the NHS PDS, we derived post-stratification weights by sexual orientation, age group, sex, region, and ethnic group, which were applied throughout the analysis.

Census 2021 data were then linked to our death registrations by NHS number, covering deaths occurring between 21 March 2021 and 20 March 2024, and registered by 1 December 2025. We were able to link 90.0% of deaths in our death registrations to the NHS PDS. For this release, we used date of death occurrence, not date of registration. Because of death registration delays in England and Wales, not all deaths during the analysis period have been included.

Our population estimates for March 2022 to March 2023, and March 2023 to March 2024 were calculated by using our Census 2021 population, subtracting deaths occurring in the preceding year and increasing the age of the population by one year. This analysis does not account for emigration of the population sample.

The census is the most representative data source available to produce statistics about the population by sexual orientation in England and Wales. It covers an estimated 97% of the population. However, not all people living in England and Wales in March 2021 were included in Census 2021 (for example, because of non-response). Quality considerations, along with the strengths and limitations of Census 2021 more generally, can be found in our Quality and methodology information (QMI) for Census 2021.

Inclusion criteria

Individuals were included in our population if they had a record in the Census 2021 dataset and had a valid NHS number linked to their Census ID. The population was then restricted to individuals who:

• were aged 18 years and over on Census Day (21 March 2021)
• were either alive at the end of our analysis period (21 March 2024) or died between 21 March 2021 and the end of this period
• had a response to the sexual orientation question in Census 2021 (including proxy responses)

Of Census 2021 respondents aged 18 years and over, 92.6% had a response to the sexual orientation question and the remaining 7.4% did not. Our 2021 population included 40,079,960 individuals (84.9% of all people aged 18 years and over in Census 2021).

In the linked population, 38,758,400 people (96.7%) identified as straight or heterosexual and 1,321,555 people (3.3%) identified with an LGB+ orientation ("gay or lesbian", "bisexual" or "other sexual orientation"). Within the LGB+ population, 655,870 people described themselves as gay or lesbian, 526,110 people described themselves as bisexual, and 139,570 people selected "other sexual orientation".

To check whether there was any possible bias for reporting of sexual orientation in proxy responses (29.5% of our sample were proxy responses), we ran further analysis. This only included individuals who had completed the census questionnaire themselves (not by proxy) (see Table 3 in our accompanying dataset).

Of the non-proxy respondents, 96.2% identified as straight or heterosexual, and 3.8% of people identified as LGB+. This analysis revealed the same pattern in our life expectancy estimates by sexual orientation but overestimated remaining life expectancy (by around 2.6 and 2.0 years for male and female life expectancy at age 20 years, respectively; and by around 1.8 and 1.6 years for male and female life expectancy at age 65 years, respectively).

This overestimation is larger than that for the overall sample (including proxy responses, the overestimation is around 1 and 0.5 years for male and female life expectancy at age 20 years, respectively; and by around 0.4 years for both male and female life expectancy at age 65 years).

These estimates of remaining life expectancy are likely to be overestimated partly because of limitations with our linked census sample which underestimates mortality. People who cannot be linked are, on average, in poorer health, and deaths of people who have emigrated are not accounted for.

The difference in life expectancy is larger when we exclude proxy respondents because they tend to be those in poorer health (mortality rates were 2.4% and 5.7% for non-proxy and proxy-only samples, respectively). We cannot say by how much life expectancy for each of the sexual orientation groups is overestimated.

More information on our sample flow can be found in our accompanying dataset.