Criminal Designs

1
Criminal Designs

• Sheila M. Bird
• (sheila.bird_at_mrc-bsu.cam.ac.uk)
• MRC Biostatistics Unit, CAMBRIDGE CB2 2SR

2
What works in UK criminal justice?

• Large RCTs
• essentially untried . . .

3
Judges prescribe sentence on lesser evidence than
doctors prescribe medicines

• Is
• public
• aware?

4
Drug Treatment Testing Orders (DTTOs)
evaluations charade (1)

• England and Wales 210 clients
• Scotland 96 clients
• Targets for DTTO clients in EW 6000 per annum
• DTTO clients gt gt 21,000 by end 2003

5
RSS Court DTTO-eligible offenders do DTTOs work ?

• Off 1 DTTO
• Off 2 DTTO
• Off 3 alternative
• Off 4 DTTO
• Off 5 alternative
• Off 6 alternative
• Database linkage to find out about major harms
  offenders deaths, re-incarcerations .
  . . Serious further offences

6
Judicial trials 3-way randomisation D 400 to
DTTO, A 400 to alternative, J 800 to judge
decides between DTTO versus alternative

• Is J. better than
• AD?

7
Performance Monitoring in the Public Services
8
Electronic tag as worn by offenders
9
Electronic surveillance tagging

• ? diversion from prison, RLO
• (Restriction of Liberty Order)
• early release from prison, HDC
• (Home Detention Curfew)
• diversion from remanded to jail, EM-bail
• (Electronically-Monitored bail versus remand may
  influence trial date and/or judges sentence . .
  . )
• ? addition to community sentence, ISSP
• (Intensive Supervision and Surveillance Program)

10
Linked fatalities per 10,000 offender-years
11
Cost per linked fatality?

• 1.5 millions pounds
• mostly, life-years lost and impact on families ?
  costed, but not re-imbursed,
• by Criminal Justice!

12
Restriction of Liberty Orders

• Scotlands CJ study . . .
• Public consultation . . .
• Private tender for electronic tagging . . .
• Post-marketing surveillance . . . (diverted from
  jail?)

13
Application of scientific method

• Randomisation to compare like with like
• Adequate study size for precise estimation
• Reporting standards as in medical journals
• Efficacy and costs rational, prior estimates
• Peer scientific review of
• Study/trial protocol

14
Scottish courts RLO-eligible offenders ? guess

• Off 7 RLO ?
• Off 8 RLO ?
• Off 9 RLO ?
• Off10 RLO ?
• Off11 RLO ?
• Off12 RLO ?
• Off13 RLO ?
• Off14 RLO ?
• Breach rates costs versus . . .
  ?

15
Evaluations-charade (2)funded guesswork
electronically monitored RLOs

• Scotland 143 clients (152 RLOs 422 assessed)
• 75 RLOs for 3-6m
• Failure in 40/143 RLOs prison for 23/40
  11/103.
• Assessed, not given RLO 40 sent to prison !!
• Assume equal s of 3m 6m RLOs 3,680, annual
  cost in Scotland of 3.68m ?1000 RLOs!
• Saving of ? 300,000 if RLOs displace equal
  length prison terms, or . . .
• 1.7m if all displaced prison terms were 6m
  . . .

16
Public Consultation v. private contractpublic
opinion re electronically monitoring for . . .

• Cost of national RLO roll out could be in the
  region of 4 millions annually, compared to a
  long term prison saving of 1.7 millions!
• 1000 RLOs - used as in pilot - could mean
  reduction of 400 in custodial terms, but add
  back 100 for breach actions . . .

17
Linked fatality Scotland

• Callum Evans (born 1987, Restriction of Liberty
  Order, June 2005) with Peter Clark (also 18) in
  October 2005, murdered 23-year old John Hatfield
  (132 injuries knife/machete axe) outside CEs
  home.
• CEs being outside his residence did not trigger
  alarm because tag had been wrongly set by Reliance

18
Was CEs RLO a diversion from jail ? or Tagging
added to community order ? Not known !
19
DESIGN Electronically-Monitored bail v. remanded
to prison

• When to randomise?
• What to compare?
• Costs?

20
Post-RCT Outcomes Re index offence Time to
trial/sentence Actual sentence Within y
years Days in custody Days on community
order Serious further offence Survival _at_ 3
years Costs
21
Randomised controlled trials

• to
• police
• Policy
• by
• Home Office
• Prisons
• Criminal Justice

22
Evaluations-charade

• Failure to randomise
• Failure to find out about major harms
• Failure even to elicit alternative sentence ?
  funded guesswork on relative cost-effectiveness
• Volunteer-bias in follow-up interviews
• Inadequate study size re major outcomes . . .

23
Five PQs for every CJ initiative

• PQ1 Minister, why no randomised controls?
• PQ2 Minister, why have judges not even been
  asked to document offenders alternative sentence
  that this CJ initiative supplants re CE?
• PQ3 What statistical power does Ministerial
  pilot have re well-reasoned targets?
• or just kite flying . .
  .
• PQ4 Minister, cost-effectiveness is driven by
  longer-term health CJ harms, how are these
  ascertained ? database linkage?
• PQ5 Minister, any ethical/consent issues?