Skip to content

Crime · United Kingdom · Updated 2026-10-26

Is knife crime rising in England and Wales?

Between the years ending March 2020 and March 2026, the period over which ONS says police figures are comparable, offences involving a knife or sharp instrument recorded by the police in England and Wales fell by 11.6%, hospital admissions in England for assault by a sharp object fell by 29.7% (the latest year is provisional) and homicides by a sharp instrument fell by 33.6%.

Every figure in this summary is calculated from the sources listed on this page.

Knife crime, direction since the year ending March 2020

Change in violence involving a knife or sharp instrument, measured by three separate collections: police recorded offences, hospital admissions for assault by a sharp object, and homicides by a sharp instrument. England and Wales, Year ending March 2020 to year ending March 2026. Unit: % change.

Falling% change, 2026

90Trust scoreHigh certainty

Score breakdown

Quality34.4 / 40

How well the sources are made

Agreement30 / 30

How close the sources are to each other

Independence16 / 20

How many separate ways the number was measured

Freshness10 / 10

How recent the newest data is

How much each source counts

  • Police47%
  • Hospital24%
  • Homicide Index29%

Sources

SourceValueDefinitionQualityWeight
Police recorded offences involving a knife or sharp instrumentHome Office, published by the Office for National Statistics · Police and justice recordsRetrieved 2026-10-26Falling −11.6%2026
Selected violent and sexual offences, including homicide, in which a knife or sharp instrument was used to injure, attempt to injure or threaten, recorded by all 43 forces in England and Wales plus British Transport Police.

Year ending March 2026 (48,774 offences) compared with year ending March 2020 (55,170), the first year of the NDQIS-based series (ONS Appendix tables, year ending March 2026, Worksheet B5, Table B5a). The chart also shows the years from March 2016 (Worksheet B6, total including Greater Manchester): there is a break in April 2019, when forces began moving to NDQIS, and ONS says earlier data are not directly comparable; Greater Manchester figures before its December 2017 review are also undercounts.

ONS does not exclude Devon and Cornwall Police from the year ending March 2026 tables. Homicides in this count come from the Home Office Homicide Index.

4 of 5 checks
  • Yes: Method and definitions published. Coverage, offence list and the NDQIS identification method are published in the ONS user guide and methodology article.
  • Yes: Official statistics or peer-reviewed. Official statistics, but not accredited: OSR removed the accreditation of police recorded crime in January 2014.
  • Yes: Full count or representative sample. Full count of the selected offences recorded by all 43 forces plus British Transport Police; it covers only crimes reported to and recorded by the police.
  • Yes: Raw data downloadable. Yearly counts are downloadable as a spreadsheet (ONS Appendix tables).
  • No: No known break or quality problem. Within the period, Metropolitan Police data from April 2024 were corrected and are not directly comparable with earlier years, and not every force uses NDQIS (41 forces do), so identification still varies between forces.
47%
Hospital Episode Statistics: admissions for assault by a sharp objectNHS England, published by the Office for National Statistics · Death and health registersRetrieved 2026-10-26Falling −29.7%2026
Finished admission episodes in NHS hospitals in England (English providers, any country of residence) with an external cause of ICD-10 X99, assault by sharp object; inpatients only, counted in the year the episode ends.

Year ending March 2026 (3,285 admissions, provisional) compared with year ending March 2020 (4,674). ONS Appendix tables, year ending March 2026, Table B8 (source: Hospital Episode Statistics). Directness 0.5: it counts injuries from sharp-object assaults that led to a hospital admission, in England only, not knife offences in England and Wales.

Covers England only; Wales admissions come from a separate register (PEDW) and are not included.

4 of 5 checks
  • Yes: Method and definitions published. The ICD-10 code, counting unit and coverage are set out in the notes to Table B8.
  • Yes: Official statistics or peer-reviewed. Official statistics compiled from Hospital Episode Statistics, NHS England's administrative register of hospital care.
  • Yes: Full count or representative sample. Full count of inpatient admission episodes in NHS hospitals in England; it counts admissions, not people, and excludes injuries treated without admission.
  • Yes: Raw data downloadable. Yearly counts are downloadable as a spreadsheet (ONS Appendix tables, Table B8).
  • No: No known break or quality problem. Data for the year ending March 2026 are provisional, and ONS says provisional counts are likely to be lower than final ones; changes in data quality, coverage and NHS practice also affect the trend (Table B8 notes 4 and 5).
24%
Home Office Homicide Index: homicides by a sharp instrumentHome Office, published by the Office for National Statistics · Police and justice recordsRetrieved 2026-10-26Falling −33.6%2026
Offences currently recorded as homicide (one per victim) where the apparent method of killing was a sharp instrument, such as a knife, machete, axe or sword.

Year ending March 2026 (176 offences) compared with year ending March 2020 (265). ONS Annual supplementary tables, year ending March 2026, Table A15; checked against Homicide in England and Wales, year ending March 2025, Worksheet 8. Directness 0.5: it covers only the most serious outcome, homicide, which was 0.4% of police recorded knife offences in the year ending March 2026.

A separate Home Office collection from the main police recorded crime return; it takes account of court outcomes, so its homicide totals differ from police recorded crime.

5 of 5 checks
  • Yes: Method and definitions published. Definitions, the apparent method of killing and revision rules are published in ONS's Homicide in England and Wales articles and the user guide.
  • Yes: Official statistics or peer-reviewed. ONS states that statistics based on the Home Office Homicide Index are accredited official statistics.
  • Yes: Full count or representative sample. Record-level count of every homicide recorded by the police in England and Wales, one offence per victim.
  • Yes: Raw data downloadable. Yearly counts are downloadable as a spreadsheet (Annual supplementary tables, Table A15).
  • Yes: No known break or quality problem. No break in the series. Counts are revised as cases go through the courts (this edition is as at 12 June 2026); against the February 2026 Homicide in England and Wales tables the changes were year ending March 2015: 186 to 185, year ending March 2024: 261 to 262, year ending March 2025: 205 to 204.
29%
Knife crime, direction since the year ending March 2020, by source · % of the year ending March 2020 level
  • Police recorded offences involving a knife or sharp instrument
  • Hospital Episode Statistics: admissions for assault by a sharp object
  • Home Office Homicide Index: homicides by a sharp instrument
050100150201620182020202220242026Police recorded offences involving a knife or sharp instrument, 2016: 59.3Police recorded offences involving a knife or sharp instrument, 2017: 71.5Police recorded offences involving a knife or sharp instrument, 2018: 86Police recorded offences involving a knife or sharp instrument, 2019: 95.2Police recorded offences involving a knife or sharp instrument, 2020: 100Police recorded offences involving a knife or sharp instrument, 2021: 81.1Police recorded offences involving a knife or sharp instrument, 2022: 89.2Police recorded offences involving a knife or sharp instrument, 2023: 93.8Police recorded offences involving a knife or sharp instrument, 2024: 97.9Police recorded offences involving a knife or sharp instrument, 2025: 96.2Police recorded offences involving a knife or sharp instrument, 2026: 88.4Hospital Episode Statistics: admissions for assault by a sharp object, 2016: 86.7Hospital Episode Statistics: admissions for assault by a sharp object, 2017: 93.1Hospital Episode Statistics: admissions for assault by a sharp object, 2018: 106.7Hospital Episode Statistics: admissions for assault by a sharp object, 2019: 108.5Hospital Episode Statistics: admissions for assault by a sharp object, 2020: 100Hospital Episode Statistics: admissions for assault by a sharp object, 2021: 85.9Hospital Episode Statistics: admissions for assault by a sharp object, 2022: 88Hospital Episode Statistics: admissions for assault by a sharp object, 2023: 79.5Hospital Episode Statistics: admissions for assault by a sharp object, 2024: 81.9Hospital Episode Statistics: admissions for assault by a sharp object, 2025: 74Hospital Episode Statistics: admissions for assault by a sharp object, 2026: 70.3Home Office Homicide Index: homicides by a sharp instrument, 2016: 79.2Home Office Homicide Index: homicides by a sharp instrument, 2017: 81.1Home Office Homicide Index: homicides by a sharp instrument, 2018: 106.4Home Office Homicide Index: homicides by a sharp instrument, 2019: 97.7Home Office Homicide Index: homicides by a sharp instrument, 2020: 100Home Office Homicide Index: homicides by a sharp instrument, 2021: 89.1Home Office Homicide Index: homicides by a sharp instrument, 2022: 105.3Home Office Homicide Index: homicides by a sharp instrument, 2023: 91.7Home Office Homicide Index: homicides by a sharp instrument, 2024: 98.9Home Office Homicide Index: homicides by a sharp instrument, 2025: 77Home Office Homicide Index: homicides by a sharp instrument, 2026: 66.4
Show the data as a table
YearPolice recorded offences involving a knife or sharp instrumentHospital Episode Statistics: admissions for assault by a sharp objectHome Office Homicide Index: homicides by a sharp instrument
201659.386.779.2
201771.593.181.1
201886106.7106.4
201995.2108.597.7
2020100100100
202181.185.989.1
202289.288105.3
202393.879.591.7
202497.981.998.9
202596.27477
202688.470.366.4

Police recorded knife or sharp instrument offences, latest year

Selected violent and sexual offences in which a knife or sharp instrument was used to injure, attempt to injure or threaten, recorded by the police. England and Wales (43 police forces plus British Transport Police), Year ending March 2026 (April 2025 to March 2026). Unit: offences.

48,774offences, 2026

63Trust scoreModerate certainty

Only one source – not confirmed by others. Agreement can't be measured, so the score can't reach High.

Score breakdown

Quality32 / 40

How well the sources are made

Agreement15 / 30

Unknown with one source, so set to half

Independence6 / 20

How many separate ways the number was measured

Freshness10 / 10

How recent the newest data is

How much each source counts

  • Police100%

Sources

SourceValueDefinitionQualityWeight
Police recorded offences involving a knife or sharp instrumentHome Office, published by the Office for National Statistics · Police and justice recordsRetrieved 2026-10-2648,7742026
Selected violent and sexual offences (attempted murder, threats to kill, assault with injury or with intent to cause serious harm, robbery, rape, sexual assault, and homicide from the Homicide Index) in which a knife or sharp instrument was used to injure, attempt to injure or threaten, including where the weapon was not seen but was believed to be present.

All 43 forces in England and Wales plus British Transport Police. ONS Appendix tables, year ending March 2026, Worksheet B5, Table B5a.

The chart starts in the year ending March 2020, the first year ONS publishes on the NDQIS method for all forces. ONS does not exclude Devon and Cornwall Police from these tables.

4 of 5 checks
  • Yes: Method and definitions published. Coverage, offence list and the NDQIS identification method are published in the ONS user guide and methodology article.
  • Yes: Official statistics or peer-reviewed. Official statistics, but not accredited: OSR removed the accreditation of police recorded crime in January 2014.
  • Yes: Full count or representative sample. Full count of the selected offences recorded by all 43 forces plus British Transport Police; it covers only crimes reported to and recorded by the police.
  • Yes: Raw data downloadable. Yearly counts are downloadable as a spreadsheet (ONS Appendix tables).
  • No: No known break or quality problem. Metropolitan Police data from April 2024 were corrected and are not directly comparable with earlier years, and not every force uses NDQIS (41 forces do), so identification still varies between forces.
100%
Police recorded knife or sharp instrument offences, latest year, by source · offences
020K40K60K202020212022202320242026Police recorded offences involving a knife or sharp instrument, 2020: 55,170Police recorded offences involving a knife or sharp instrument, 2021: 44,728Police recorded offences involving a knife or sharp instrument, 2022: 49,189Police recorded offences involving a knife or sharp instrument, 2023: 51,741Police recorded offences involving a knife or sharp instrument, 2024: 53,995Police recorded offences involving a knife or sharp instrument, 2025: 53,096Police recorded offences involving a knife or sharp instrument, 2026: 48,774
Show the data as a table
YearPolice recorded offences involving a knife or sharp instrument
202055,170
202144,728
202249,189
202351,741
202453,995
202553,096
202648,774

Homicides by a sharp instrument (Homicide Index), latest year

Offences currently recorded as homicide where the apparent method of killing was a sharp instrument. England and Wales, Year ending March 2026 (April 2025 to March 2026). Unit: homicide offences.

176homicide offences, 2026

71Trust scoreModerate certainty

Only one source – not confirmed by others. Agreement can't be measured, so the score can't reach High.

Score breakdown

Quality40 / 40

How well the sources are made

Agreement15 / 30

Unknown with one source, so set to half

Independence6 / 20

How many separate ways the number was measured

Freshness10 / 10

How recent the newest data is

How much each source counts

  • Homicide Index100%

Sources

SourceValueDefinitionQualityWeight
Home Office Homicide Index: homicides by a sharp instrumentHome Office, published by the Office for National Statistics · Police and justice recordsRetrieved 2026-10-261762026
Offences currently recorded as homicide (one per victim) in the Home Office Homicide Index where the apparent method of killing was a sharp instrument, such as a knife, machete, axe or sword.

As at 12 June 2026. ONS Annual supplementary tables, year ending March 2026, Table A15.

The lowest count since the year ending March 1991 (175) in the same table, which starts in the year ending March 1979.

5 of 5 checks
  • Yes: Method and definitions published. Definitions, the apparent method of killing and revision rules are published in ONS's Homicide in England and Wales articles and the user guide.
  • Yes: Official statistics or peer-reviewed. ONS states that statistics based on the Home Office Homicide Index are accredited official statistics.
  • Yes: Full count or representative sample. Record-level count of every homicide recorded by the police in England and Wales, one offence per victim.
  • Yes: Raw data downloadable. Yearly counts are downloadable as a spreadsheet (Annual supplementary tables, Table A15).
  • Yes: No known break or quality problem. No break in the series. Counts are revised as cases go through the courts (this edition is as at 12 June 2026); against the February 2026 Homicide in England and Wales tables the changes were year ending March 2015: 186 to 185, year ending March 2024: 261 to 262, year ending March 2025: 205 to 204.
100%
Homicides by a sharp instrument (Homicide Index), latest year, by source · homicide offences
0100200300201620182020202220242026Home Office Homicide Index: homicides by a sharp instrument, 2016: 210Home Office Homicide Index: homicides by a sharp instrument, 2017: 215Home Office Homicide Index: homicides by a sharp instrument, 2018: 282Home Office Homicide Index: homicides by a sharp instrument, 2019: 259Home Office Homicide Index: homicides by a sharp instrument, 2020: 265Home Office Homicide Index: homicides by a sharp instrument, 2021: 236Home Office Homicide Index: homicides by a sharp instrument, 2022: 279Home Office Homicide Index: homicides by a sharp instrument, 2023: 243Home Office Homicide Index: homicides by a sharp instrument, 2024: 262Home Office Homicide Index: homicides by a sharp instrument, 2025: 204Home Office Homicide Index: homicides by a sharp instrument, 2026: 176
Show the data as a table
YearHome Office Homicide Index: homicides by a sharp instrument
2016210
2017215
2018282
2019259
2020265
2021236
2022279
2023243
2024262
2025204
2026176

Why the numbers differ

  • Police recording practice changed. From April 2019 forces began moving to the National Data Quality Improvement Service (NDQIS), which scans crime records for words such as 'knife' or 'stab' instead of relying on officers to flag them, and the offences in scope were redefined. ONS says data before April 2019 are not directly comparable with later years. (source, retrieved 2026-10-26)
  • ONS attributes the rise in police recorded knife offences between the years ending March 2015 and March 2020 to a mix of real change and better police recording, and says the trend is supported by hospital admissions data. (source, retrieved 2026-10-26)
  • Hospital data count only injuries from an assault by a sharp object that led to an inpatient admission (ICD-10 code X99). Threats, robberies where nobody is hurt and injuries treated without admission are not counted, and one person can be admitted more than once. (source, retrieved 2026-10-26)
  • The Homicide Index covers only the most serious cases: in the year ending March 2026, homicides were 0.4% of the knife or sharp instrument offences recorded by the police (176 of 48,774). Small numbers make its yearly changes more volatile. (source, retrieved 2026-10-26)
  • All three lines use financial years (April to March), but each dates events differently: hospital episodes by the year the episode ends, homicides by the year police recorded them, even years after the death. ONS's quarterly bulletins use rolling years ending in other months and the Metropolitan Police's monthly reports use data extracted on other dates, so their figures will not match these. (source, retrieved 2026-10-26)

Common mistakes

  • Comparing police knife crime across the April 2019 method change. Over 10 years, police recorded knife offences in ONS's long-run table (excluding Greater Manchester) rose by 48.5%, from 30,927 in the year ending March 2016 to 45,919 in the year ending March 2026, but ONS says data from before April 2019 are not directly comparable with later years, so that change is not a valid trend. (source, retrieved 2026-10-26)
  • Reading the latest hospital figure as final. Admissions for the year ending March 2026 are provisional, and ONS says final counts are likely to be higher. (source, retrieved 2026-10-26)
  • Treating every police recorded knife offence as a stabbing. The count includes threats and robberies where a knife was seen or believed to be present but nobody was injured. (source, retrieved 2026-10-26)