More than seven people a day died of mouth and throat cancer in England in 2020 to 2022, at a death rate a third higher than in 2008 to 2010. The figures for every council

What is counted. Every figure on this page is for one group of cancers: cancers of the lip, tongue, mouth, salivary glands, tonsils and pharynx (codes C00 to C14 in the international classification of diseases). The pharynx is the throat, from behind the nose down to where the food pipe begins. Cancer of the voice box (the larynx) is not in the group. The Office for Health Improvement and Disparities (OHID), which publishes the figures, calls the group oral cancer. This page calls it mouth and throat cancer, because it is wider than mouth cancer in the everyday sense. The NHS describes mouth cancer as affecting “any part of the mouth, including the gums, tongue, inside the cheeks, or lips”. Of the 8,462 deaths registered in England in 2020 to 2022, 4,016 (47%) were from cancers of the lip, tongue and mouth, 671 (8%) from cancers of the salivary glands, and 3,775 (45%) from cancers of the tonsils and pharynx or of a part of the group that was not specified, on the Office for National Statistics' count by cause of death.

November is Mouth Cancer Action Month, run by the Oral Health Foundation. OHID publishes two measures of these cancers for every council in England on its Fingertips service: new cases and deaths. This page puts both in one searchable chart and one free file, with the confidence interval the source gives for every figure, so that a council is described as above or below England only where the source's own test finds the difference statistically significant. The newest figures are deaths for 2020 to 2022 and new cases for 2017 to 2019. Nothing here describes this year.

New cases are cancers newly diagnosed and recorded by the NHS's cancer registry. The latest period published by council is 2017 to 2019. Deaths are deaths from those cancers, from death registrations. The latest period is 2020 to 2022. Both are pooled over three years, which steadies the small numbers in a single council. A rate is cases or deaths for every 100,000 people, adjusted for age, so a council with many older residents can be compared fairly with a young one. Each rate comes with a 95% confidence interval, the range the true rate is likely to sit in given how few cases there are. A council is significantly above England when its whole interval is above England's rate, and significantly below when its whole interval is below. Otherwise the figures cannot tell it apart from England.

1. More than seven deaths a day in 2020 to 2022, and a death rate a third higher than in 2008 to 2010

8,462 deaths from mouth and throat cancer were registered in England in the three years 2020 to 2022. That is about 2,800 a year, or 54 a week. For 2008 to 2010 the figure is 5,303, on OHID's count, which adjusts the earlier years for changes in how causes of death are coded. Part of that 60% rise is a larger and older population. The age-adjusted rate removes that part, and it still rose by 34%, from 3.9 to 5.2 deaths for every 100,000 people. The two periods' confidence intervals do not overlap, so the rise is larger than chance.

New cases of mouth and throat cancer and deaths from it in England, age-adjusted, by three-year period, 2007 to 2009 up to 2020 to 2022

New cases rose in the same way up to 2017 to 2019. 24,115 were diagnosed in those three years, about 8,000 a year, against 16,283 in 2007 to 2009. The age-adjusted rate rose 28%, from 12.0 to 15.4 for every 100,000. The series of new cases by council stops at 2017 to 2019, so it does not show what has happened since.

Men accounted for 68% of the deaths in 2020 to 2022. Their death rate was 7.6 for every 100,000 against 3.0 for women, and it had risen faster: by 39% since 2008 to 2010, against 23% for women.

2. Deaths by council, 2020 to 2022: 22 significantly above England, 16 significantly below

Search for a council to read its figures. District councils also show their county's figure. Every rate is given with its period, its interval and the number of deaths or cases it rests on. A rate for a place in past years is not a risk for a person who lives there now.

Deaths from mouth and throat cancer for every 100,000 people, age-adjusted, by council, 2020 to 2022

Council death rates in 2020 to 2022 ran from 2.3 for every 100,000 in Stroud (95% confidence interval 1.1 to 4.2, from 10 deaths) to 9.7 in Islington (95% confidence interval 6.8 to 13.3, from 40 deaths). That range is wider than the real differences are likely to be. The middle council's rate is built on 24 deaths over three years, and 146 councils had fewer than 25, so one council's rate can move a long way on a handful of deaths. That is what the intervals are for. On the source's own test, 22 councils were significantly above England, 16 were significantly below, and 247 were not significantly different.

Even that needs care. With 285 councils and a 95% test, chance alone would put about 7 on each side. 22 is about three times that, so the excess is unlikely to be chance alone, but some of the 22 will be. Three further checks help. Five councils were above England on the source's stricter 99.8% limits: Islington, Leicester, Liverpool, Manchester and Sandwell. Nine of the 22 were also significantly above England in 2017 to 2019, a period that shares no year with this one: Coventry, Hammersmith and Fulham, Knowsley, Leicester, Manchester, Middlesbrough, Nottingham, Sunderland and Wirral. And the tables below list only councils with at least 25 deaths in the period.

139 councils had 25 or more deaths in 2020 to 2022 and are eligible for these tables. The other 146 with a rate, and the 9 with too few deaths for one, are in the chart and the file but not in the tables.

3. New cases by council, 2017 to 2019: 24 significantly above England, 23 significantly below

New cases are more numerous than deaths, so the rates are steadier: the middle council's rests on 63 cases. They are also older. 2017 to 2019 is the latest period OHID publishes by council.

New cases of mouth and throat cancer for every 100,000 people, age-adjusted, by council, 2017 to 2019

The rate of new cases in 2017 to 2019 ran from 8.9 for every 100,000 in Eastleigh (95% confidence interval 6.2 to 12.4, from 35 cases) to 24.5 in Manchester (95% confidence interval 21.3 to 28.1, from 225 cases). 24 councils were significantly above England's 15.4 and 23 significantly below, against about 7 on each side by chance. Seventeen of the 24 were also significantly above England in 2014 to 2016, the previous period with no year in common. Ten of the 24 were in the North West and six in the North East.

4. Ten councils were significantly above England on both measures

The two measures come from different records, the cancer registry and death registrations, and cover different years. Being significantly above England on both is the firmest finding these figures can give about a single council. Ten councils were. All 10 are unitary or metropolitan councils, each with more cases and more deaths than the middle council. A small district with the same rates could have too few cases to be told apart from England, so this is not a list of the 10 highest rates. The figures record that the rates were higher in those periods. They do not say why, and they are not a measure of any service or of any person's risk.

Five were in the North West and three in the North East, the two regions that were themselves significantly above England on both measures. Two councils, Eastleigh and Tunbridge Wells, were significantly below England on both. The two measures do not always agree. Cornwall, counted with the Isles of Scilly, was significantly above England for new cases in 2017 to 2019 and significantly below it for deaths in 2020 to 2022, and across the 283 councils with both figures the two rates are only loosely related (rank correlation 0.38, where 0 is no relationship and 1 a perfect one).

5. Rates were higher in the more deprived councils, and differed by region

OHID groups councils into tenths by deprivation. In the most deprived tenth the death rate in 2020 to 2022 was 6.5 for every 100,000, and in the least deprived it was 4.5, so the most deprived tenth was 46% higher. For new cases in 2017 to 2019 the figures were 18.2 and 13.3, 37% higher. In both cases the two intervals do not overlap. Council by council, the rate of new cases has a rank correlation of 0.52 with the 2025 deprivation index and the death rate 0.43. Deprivation is associated with part of the difference between councils, and leaves much of it unaccounted for. This is a pattern across groups of councils. It does not explain any one council's figure.

New cases (2017 to 2019) and deaths (2020 to 2022) were highest in the most deprived tenth of councils

The death rate in 2020 to 2022 was higher than in 2008 to 2010 in every region. In eight of the nine the two periods' intervals do not overlap.

6. Higher than in the earliest period in most councils

The rate of new cases in 2017 to 2019 was higher than in 2007 to 2009 in 253 of the 290 councils with a figure for both periods. In 19 of them the rise is clear on a strict test, where the two periods' intervals do not overlap at all. In none is there a fall that passes the same test.

For deaths the comparison can be made in 204 councils, because 81 of the 285 had fewer than 10 deaths in 2008 to 2010 and so have no rate for that period. The death rate was higher in 2020 to 2022 than in 2008 to 2010 in 153 of the 204, and the rise passes the same strict test in nine: Bradford, Doncaster, Leicester, Luton, Portsmouth, Sandwell, Shropshire, Swindon and West Northamptonshire. Again none has a fall that passes it. For a single council the national rise is the better guide: most councils' own numbers are too small to show a change of a third as more than chance.

How to use these figures

If you report on a council or represent one, find it in either chart or in the council file. Give the period with the figure, since new cases are for 2017 to 2019 and deaths for 2020 to 2022, and give the interval and the number of cases or deaths beside any rate. Say a council was above or below England only where the file says significantly. Where it does not, the accurate line is that the council's rate was not significantly different from England's, and that England's rose by a third between 2008 to 2010 and 2020 to 2022. The rates are adjusted for age, so an older population is not the explanation for a high one. Say what is counted: cancers of the mouth and throat, which the source calls oral cancer, not mouth cancer alone.

If you work in public health for a council or an integrated care board, the 10 councils in section 4, and the North East and North West as regions, are where both measures pointed the same way. These figures cannot say why, and nothing on this page is a finding about causes in any named place.

On causes, the official sources speak about the disease in general. The NHS page on the causes of mouth cancer says “It's not always clear what causes mouth cancer”, and lists among the things that may make a person more likely to get it that “you smoke or chew tobacco, betel nut or paan” and that “you drink a lot of alcohol”. OHID's own note on the new-cases indicator calls it “a direct measure of smoking-related harm”. Those are the sources' statements about the population as a whole. They are not statements about any council or any person.

If you run a dental practice, link the NHS mouth cancer pages in preference to writing your own list of symptoms. If you quote a local figure, quote your council's with its period, its interval and the number of cases or deaths, and say whether it was significantly different from England's.

How we did this

Sources. OHID Fingertips, two indicators, taken from the Fingertips API on 12 November 2026. Indicator 1206, Oral cancer registrations: new cases of cancer of the lip, oral cavity and pharynx (ICD-10 C00 to C14) registered by NHS England's National Cancer Registration and Analysis Service, three-year periods from 2007 to 2009 up to 2017 to 2019. Indicator 92953, Mortality rate from oral cancer, all ages: deaths from the same cancers, from Office for National Statistics death registrations, three-year periods from 2008 to 2010 up to 2020 to 2022. Both are for persons of all ages. These were the latest periods on Fingertips on the day. Cited as OHID asks: Office for Health Improvement and Disparities. Public health profiles. 2026 https://fingertips.phe.org.uk/ © Crown copyright 2026.

What the group contains. The Fingertips indicators give one figure for the whole of C00 to C14. The split between the mouth, the salivary glands and the pharynx, and the yearly counts since 2022, are from the Office for National Statistics' Mortality statistics by underlying cause, taken from Nomis on 12 November 2026: deaths registered in England with an underlying cause in C00 to C06, in C07 and C08, and in C09 to C14. The three years 2020 to 2022 add to 8,462, exactly the Fingertips count.

Licence. Each Fingertips indicator carries its own terms. Indicator 1206 states: “The data may be reused referencing OHID”. Indicator 92953 has no reuse statement of its own and is marked Crown copyright, and Fingertips states that all its content is available under the Open Government Licence except where otherwise stated. Neither is marked for non-commercial use only. The Open Government Licence allows anyone to “exploit the Information commercially and non-commercially”, provided the source is acknowledged. Office for National Statistics figures are published under the same licence.

NHS wording. Every quotation from the NHS is word for word from the page linked beside it, as at 12 November 2026. Information from the NHS website is licensed under the Open Government Licence v3.0. The NHS pages quoted are about mouth cancer, which is narrower than the group counted here.

Rates and intervals. The rates are directly age-standardised to the 2013 European Standard Population, which means they show what each council's rate would be if it had the same age mix as a standard population. Every rate, count, 95% confidence interval and comparison with England in this study is the source's own. We checked that the source's comparison is the rule stated above, the whole interval against England's rate, on every council, and it is.

Councils. 296 unitary, metropolitan, London borough and district councils on April 2023 boundaries, and the 21 county councils, which have rows of their own in the file. The source combines City of London with Hackney and the Isles of Scilly with Cornwall. It gives no death rate where there were fewer than 10 deaths in the period, which applies to 9 councils in 2020 to 2022: Broxbourne, Hart, Maldon, Mansfield, Melton, Rochford, Rutland, Tamworth and Torridge. Cumberland and Westmorland and Furness have no new-cases figure, because the series ends before they were created. The file notes the figure for the former county of Cumbria in 2017 to 2019, 14.9 for every 100,000 (95% confidence interval 13.1 to 16.8, from 256 cases). For new cases in North Yorkshire and Somerset the source applies the former county's figure to the new unitary council, which covers the same area. Nothing withheld by the source has been estimated.

Change over time. The earliest period on Fingertips is 2007 to 2009 for new cases and 2008 to 2010 for deaths. OHID has adjusted the earlier years' deaths for three changes in how the cause of death is coded, so that the series can be compared over time. A change is called a rise or a fall only where the two periods' 95% intervals do not overlap. That rule is ours, not the source's. It is stricter than a formal test of the difference and will miss some real changes.

Controls we ran. Chance: with a 95% test about one council in forty lands on each side of England by luck, which is the 7 quoted above, so we also report the 99.8% limits for deaths (the source gives none for new cases) and whether a council was also significantly above England in the previous period with no year in common. Small numbers: the ranked tables leave out any rate built on fewer than 25 cases or deaths, which was the source's own floor for a death rate until it lowered it to 10. Level of council: among the 153 counties and unitary councils, where numbers are larger, 17 were significantly above England for deaths and 22 for new cases, and the same 10 councils were above on both. Deprivation and region: section 5. Age: both rates are standardised for age, so differences in how old a council's population is are already removed.

What this is not. It is not a picture of this year: the newest figures are deaths registered in 2020 to 2022 and cases diagnosed in 2017 to 2019. It is not a measure of any dental practice, hospital or cancer service, and it does not show how early cancers were found or how people fared after diagnosis. Deaths in 2020 to 2022 are not the same people as the new cases of 2017 to 2019. People are counted where they lived, not where they were treated. And a rate for a place says nothing about any one person who lives there.

Questions we get asked about this data

How many people died of mouth and throat cancer in England in 2020 to 2022?

8,462 deaths were registered in the three years, about 2,800 a year, counting cancers of the lip, mouth, salivary glands and pharynx. The age-adjusted death rate was 5.2 for every 100,000 people, 34% higher than in 2008 to 2010. Cancers of the lip, tongue and mouth alone accounted for 4,016 of the 8,462.

Was the rate higher where I live?

Search for your council in the charts in sections 2 and 3. Each readout gives the rate, its confidence interval, the number of cases or deaths and whether the source classes it as significantly above or below England, for deaths in 2020 to 2022 and new cases in 2017 to 2019. For most councils the answer is that the rate was not significantly different from England's. No figure here says what the rate is now, and a rate for a place is not a risk for a person.

Why are the figures several years old?

They are the latest that OHID publishes for every council on Fingertips, checked on 12 November 2026. Small areas need three years of data pooled together. New cases by council run to 2017 to 2019, in an indicator last updated in April 2022, and deaths to 2020 to 2022, last updated in September 2024. Newer national counts exist. The Office for National Statistics counts 3,133 deaths from the same group of cancers registered in England in 2023, 3,117 in 2024 and 3,176 in 2025. Those are counts for single years. They are not adjusted for age, so they cannot be set beside the council rates here. NHS England publishes national counts of new cases for later years in its Cancer Registration Statistics.

Does a high rate mean local dentists or doctors were missing cases?

No. These figures count how many people were diagnosed or died. They say nothing about how or when a cancer was found, and they cannot be used to judge any service.

For editors

The data is free to reuse with attribution to Practice Radar, under Creative Commons BY 4.0. Source: Office for Health Improvement and Disparities, Fingertips indicators 1206 (Oral cancer registrations, from NHS England) and 92953 (Mortality rate from oral cancer, all ages, based on Office for National Statistics data). The council file, mouth-cancer-rates-by-council-2026.csv, has one row per council with its ONS code, and rows for each region and England.

Four things to carry into any report. The group counted is cancers of the lip, mouth, salivary glands and pharynx (ICD-10 C00 to C14), which the source calls oral cancer. It is wider than mouth cancer alone, which accounted for 47% of the deaths. The newest figures are deaths for 2020 to 2022 and new cases for 2017 to 2019, so please give the period with any figure. Please give the interval and the number of cases or deaths with any council's rate, and describe a council as above or below England only where the file says significantly. And with about 285 councils tested, chance alone would mark about 7 as significantly above England and about 7 as significantly below. Patient information: nhs.uk/conditions/mouth-cancer.

Jack Henty, who compiled this analysis, co-owns Bollington Dental Practice in Cheshire and founded Practice Radar. He can answer questions about how the figures were put together.

Contact: hello@practiceradar.co.uk