Victim Not Suspect

CPS

CPS Area Inspection Programme

The Crown Prosecution Service

HM Crown Prosecution Service Inspectorate assessed casework quality in all 14 CPS Areas, setting a baseline across 2021-22 and following up in 2024. These pages carry the published scores for each theme, alongside separate inspections into rape and stalking casework.

Look up a CPS Area

Every Area's full record, across all eight themes, with baseline and follow-up scores.

Browse the 14 Areas

Across all 14 Areas the pattern is consistent. Areas scored well on applying the Code for Crown Prosecutors and on selecting the right charges, above 85% on every measure. They scored poorly on the analysis and case strategy behind those decisions, below 51% before charge and below 59% after it. The legal test was applied correctly. The reasoning behind it was thin.

What happened to the programme

  1. Baseline

    2021-22

    Published October 2021 to November 2022

    All 14 Areascoverage

    Casework quality assessed across magistrates' courts, Crown Court and RASSO casework. 90 cases per Area. Every Area scored on added value and grip in all three strands.

  2. Follow-up

    2024

    Published January 2025

    All 14 Areascoverage

    The same assessment repeated for magistrates' courts and Crown Court casework only. RASSO was not re-assessed.

  3. Phase 3

    2025-26

    Published September 2025 and January 2026

    2 Areascoverage

    Yorkshire and Humberside and Cymru-Wales, chosen because they showed significant improvement or decline between the baseline and the follow-up. HMCPSI describes this as a targeted risk-based approach. Phase 3 asked what drives casework quality rather than scoring it, so it produced no new added value or grip figures for either Area.

  4. 2026-27 programme

    2026-27

    Published March 2026

    No Areascoverage

    HMCPSI's business plan for 2026-27 lists 15 CPS inspections. None is an Area inspection.

For 12 of the 14 Areas there are two measurements and, as at the 2026-27 business plan published in March 2026, no third is scheduled. Phase 3 covered two Areas selected for the size of their movement, not as a representative sample, so their findings should not be read as typical of the other 12.

Pre-charge

Post-charge

Victims and witnesses

Other aspects

Inspections by offence type

HMCPSI has also inspected CPS casework in two specific offence areas. These sit outside the Area Inspection Programme and use different methods and samples.

Checking the published figures

HMCPSI publishes national average scores alongside the highest and lowest Area score for each measure. We checked those ranges against all fourteen individual Area reports. Four of the six figures match. Two do not.

Errors in the composite report's published ranges

VNS analysis

The lowest magistrates' courts and Crown Court added value figures are printed against the wrong strand. The lowest Crown Court grip figure does not match any Area's score.

The RASSO strand

Rape and serious sexual offences casework was assessed once, in 2021-22, and was not re-assessed at follow-up.

Why there is no follow-up figure

RASSO casework was assessed in the 2021-22 baseline and was not re-assessed in the 2024 follow-up. HMCPSI's stated reason is that too few RASSO cases from the relevant period had reached a conclusion for a meaningful sample to be drawn, a consequence of the Crown Court backlog. The 2025 thematic inspection of early advice and pre-charge decision making in adult rape cases does not fill the gap: HMCPSI had intended to compare its findings against the baseline and abandoned that comparison, stating the approach had changed too much for a direct comparison to be fair.

Sources

  • HM Crown Prosecution Service Inspectorate, baseline assessments of all 14 CPS Areas, published October 2021 to November 2022.
  • HM Crown Prosecution Service Inspectorate, Area Inspection Programme, follow-up of the 14 CPS Areas in England and Wales, January 2025, publication CP001:1320.

We checked the composite report's published ranges against all fourteen Area reports and found two figures that do not match. See errors in the composite report's published ranges.