UK Chapter of the International System Dynamics Society 12 th - PowerPoint PPT Presentation

Applying a System Dynamics Approach to Inform the Direction of Offender Pathways and Services Savas Hadjipavlou & Samantha Hinks: Ministry of Justice Douglas McKelvie: Symmetric SD UK Chapter of the International System Dynamics Society 12 th February 2009

Overview of the presentation  What is the context we’re working in?  What was the Problem & why did we need a model?  Why use System Dynamics?  What does the model look like?  What does it include?  What are the outputs?  How has it helped in problem solving?  What other applications has it had?  New model and applications

What is the context we are working in?  Traditionally units within government have worked on their own specific objectives rather than looking at the bigger picture.  Modelling is still a very new approach, particularly within the Ministry of Justice.

What was the problem & why did we need a model?  Need to explore impacts of an intervention for long stay serious offenders, on the wider prison/hospital estate.  Need to plan complex (and expensive) services where impacts take a long time.  Need to take account of ‘recycling’ (e.g. breach and reconviction) into the system.  Need to be able to assess costs and benefits.

Why a System Dynamics Approach? Allows us to : Model a whole system visually. • Tell a ‘story’. • Look at stocks (offenders) and flows through the • system. Use data and assumptions about services and • policies to help predict outcomes. Carry out ‘what if’ analysis; identify bottlenecks. • ITHINK software allows a highly interactive • approach.

Overview of Model Mental Health Act route 3 Programme 1 Screening 2 Assessment 4 Pre Tariff 5 Post Tariff 6 Community Supervision

What does the Model include?  Sentence rates  % ‘treatable’/ long stay  Sentences lengths  Treatment effectiveness  Proportion needing  Post treatment capacity assessment  Release rates  Assessment and treatment  Length of community capacity supervision  Assessment and treatment  Breach rates length  Reconviction rates

What are the Outputs? Numbers (until 2024):  In custody  Under community supervision  Assessed  Waiting for assessment  In treatment  Waiting for treatment  Treated  Reconvicted

Using the model to problem solve Issue: Waiting list for DSPD  Currently: capacity of 130 Numbers Waiting for the DSPD Prison DSPD prison treatment Programme places; 5 year programme. 700  Results in a waiting list for Numbers waiting 600 services over time. 500 400 300  The model can help us 200 inform what might be done to 100 0 reduce the waiting list and 3 5 7 9 1 3 5 0 0 0 0 1 1 1 0 0 0 0 0 0 0 quantify the impact of 2 2 2 2 2 2 2 Time (July) different options. Numbers do not necessarily represent reality

Option 1: Increase capacity – but how far? Numbers Waiting for the DSPD Prison Programme by Capacity current capacity 700 Numbers waiting 600 500 double capacity 400 300 200 triple capacity 100 0 quadruple capacity 2003 2005 2007 2009 2011 2013 2015 multiply initial Time (July) capacity by 5 • Multiplying the initial treatment capacity by 5 would eliminate a waiting list, but is this financially and practically viable? Numbers do not necessarily represent reality

Option 2: Increase treatment capacity and reduce treatment length – but how far? Numbers Waiting for the DSPD Prison Programme by Scenario 700 Numbers waiting 600 current capacity & 500 5yr prog 400 triple capacity & 5yr 300 prog 200 100 triple capacity & 3yr 0 prog 3 5 7 9 1 3 5 0 0 0 0 1 1 1 0 0 0 0 0 0 0 triple capacity & 2yr 2 2 2 2 2 2 2 prog Time (July) • Tripling the initial capacity and reducing the treatment length to 2 years, could eliminate a waiting list. • This then enables us to think through the practicalities of this. Numbers do not necessarily represent reality

Other applications  Contributed to the annual Prison Population Projections  Parole Board Projections  Informing changes to legislation for serious, high risk offenders.

Using the model to problem solve The Problem:  IPPs (Indeterminate Public Protection Sentences) were putting pressure on the Prison system.  Large number of receptions each month;  Short sentences (tariffs);  Wide range of needs;  Unable to assess and move IPPs into treatment within tariff so many likely to stay (unnecessarily) beyond tariff.  The model could be used to help explore different options for altering the application of the IPP sentence.

Using the model to problem solve: Scenario to be tested  Setting a ‘2 year’ seriousness threshold for IPPs  Where threshold met, sentencers have more discretion to give other sentences as well as IPPs which would result in:  Fewer IPP receptions p/m  Higher average tariff

Results: Impact on Prison Population  Implementing the IPP Prison Population by Scenario proposed changes could reduce the IPP IPP Prison Population 12,000 Prison Population by about 6,000 by 2014. 10,000 current  But… those who 8,000 situation would previously 6,000 proposed scenario have got IPPs would 4,000 now get different 2,000 sentences. So we 2008 2009 2010 2011 2012 2013 2014 used the model to consider the impact Year (April) on the wider Prison population. Numbers do not necessarily represent reality

Results: Impact on Prison Population  Implementing the Combined IPP; EPP; Det 4yrs+ Prison Population proposed changes by Scenario could reduce the combined Prison Combined Prison 37000 Population population by more 35000 current situation 33000 than 2,000 by 2014. proposed scenario 31000  But… what about 29000 2008 2009 2010 2011 2012 2013 2014 impacts elsewhere in the system? Year (April)  The model enables us to look at this. Numbers do not necessarily represent reality

Results: Impact on Probation Supervision  Implementing the IPPs under Probation Supervision by Scenario proposed changes could reduce the No. of IPPs in Community 2500 number of IPPs under 2000 Probation supervision 1500 current situation by about 800 by 2014. proposed scenario 1000  But… those who would 500 previously have got 0 IPPs would now get 8 9 0 1 2 3 4 0 0 1 1 1 1 1 0 0 0 0 0 0 0 different sentences. So 2 2 2 2 2 2 2 we needed to consider Year (April) the impact on the wider Probation caseload. Numbers do not necessarily represent reality

Results: Impact on Probation Supervision  Implementing the proposed changes was Combined IPP; EPP; det 4yrs+; Lifer Combined Probation Caseload by Scenario unlikely to result in a large No. under Probation saving to Probation. By 25000 Supervision 2014, the combined 20000 current situation number needing Probation proposed scenario 15000 Supervision could be reduced by about 300 . 10000 2008 2009 2010 2011 2012 2013 2014  So… it is important to Year (April) consider the system as a whole, which the model enabled us to do. Numbers do not necessarily represent reality

Need for a New Model  Need to consider level of need (mental health; alcohol/drug misuse) for potential diversion to interventions for all offenders across the whole of the Criminal Justice System.  Want to be able to segment the population by sentence type and offence type to better target need with appropriate interventions.

HEALTH PARTNERSHIP: POSSIBLE POINTS OF INTERVENTION The Offender/Health Pathway Preventing Offending or reoffending Statutory agencies Voluntary groups Community SOCIALLY Prevention Appropriate Placements – getting people in the resettlement/ EXCLUDED – access to right place at the right time supervision health NOMS services for DRUGS + + identical Charge groups Arrest + + + + + MENTAL Ensure + Court Conviction Sentence Parole Expiry HEALTH continuity OFFENCE CJS of of care Sentence PERSONALITY + + DISORDERS Prison Care in Custody Transfer Transfer Transfer PREVIOUS police, prison, hospital Life continues High/ OTHERS + Medium + Security Goes straight Hospital May back into Start offend being receiving managed in healthcare community – Mental Health in the receiving Tribunal community healthcare Community Existing health services Receiving healthcare in community Care in the community Hospital Possible points of transfer for offenders (where appropriate) into hospital Prison Possible points of transfer back into prison Community + Possible Health Partnerships intervention

Next Steps  Have a national picture of ‘need’, but can use the infrastructure to develop regional models and a model for women.  Can introduce capacity constraints to consider the reality of rolling out new interventions.  ITHINK provides a user-friendly medium for doing this.

Summary  Although things are changing, Government departments still have a tendency to focus on their individual unit objectives rather than see the bigger picture.  The use of modelling techniques is still quite new.  We have hopefully shown the need for modelling techniques and how a System Dynamics approach has been used to problem solve and guide policy decision-making.