Royal Borough of Windsor and Maidenhead LSCB Annual Conference20th January 2010, Moor Hall Cookham,

Learning together to safeguard children:

a ‘systems’ model for case reviews

Dr Sheila Fish Senior Research Analyst, SCIE

Overview

1. Background to the SCIE-led project

2. How it fits with revised ‘Working Together’ & Ofsted’s evaluation criteria

3. Underpinning theories of causality & implications for learning

4. Putting it into practice – key aspects of the process

1. Why do we need new methods of learning in child welfare?

A long history of investigations into child abuse tragedies – Serious Case Reviews & public enquiries

yet findings are familiar and repetitive this raises questions about their value for

improving practice similar circumstances in aviation and other high-

risk industries led to development of the ‘systems approach’

Implausible to put the error down to laziness or stupidity

gets to the bottom of ‘why’ accidents occur allowing for more effective solutions

Developments in the health sector

“An organisation with a memory” (DH, 2000)

No equivalent in child welfare Relevance to child welfare in theory Munro (2005) ‘A systems approach to investigating child

abuse deaths’ British Journal of Social Work, 35, 531- 546

Nothing on how it would work in practice Victoria Climbié tragedy gave extra impetus SCIE project was born

Blaming not explaining: problem exemplified in recent high profile case

“as a social worker people keep asking me about Baby P – ‘how could this possibly have happened?’ And all I can say is – there must have been reasons … it’s complicated”

Without adequate explanations…

incredulity quickly turns to anger and condemnation of those involved - hard to believe that a motivated, well-meaning, competent worker could act this way

so conclude must be the result of stupidity, malice, laziness or incompetence

YET reasonable to assume that most people come to work each day wanting to help children, not to allow them to be harmed;

practitioners rarely intend to make mistakes so better explanations are required the systems approach is explicitly designed to

address these ‘why’ questions

How has the model been developed?

not ‘off the shelf’; detailed developmental work to adapt it

builds on Managing risk and minimising mistakes (Bostock et al 2005)

a review of the safety management literature (Munro 2008

tried out in practice; two pilot case reviews valuable feedback provided by staff at all stages fine-tuned the model based on this experience

SCIE Guide 24Authors: Sheila Fish, Eileen Munro & Sue Bairstow

Learning together to safeguard children: developing a multi-agency systems approach for case reviews

All available free on SCIE’s website

www.scie.org.ukTitle: Learning together to

safeguard children Guide 24 ‘at a glance’ summary Social Care TV film Research report 19

2. What can the model be used for?

‘The SCIE model is intended to be used in any circumstance where practice needs to be reviewed, not just in the cases of serious harm or death’

Community Care “blueprint for serious case reviews” 16 February 2009

can be used to examine any case, not just those with tragic outcomes

also good reason to focus our curiosity on a) routine practice, b) practice that practitioners and/or families are happy with and c) innovations that seem to be working welld) others?

SCRs & Working Together guidance

The systems approach ... .. IS a conceptual framework / way of thinking

and a structured and systematic process for conducting case reviews

... is NOT an alternative to SCRs but a methodology to support the SCR process

Revised Ch8 consultation; ADCS response

“The current approach to serious case reviews promotes that rather reductive, linear, analytic process ... Our view is that in the revised guidance the process should be opened up and, in particular, that the approach set out by SCIE should be promoted as one of the core approaches to undertaking a serious case review”.

Govt response to consultation

“The Government recognises that there is a range of methodologies that can assist in learning from SCRs and from cases which do not meet the SCR criteria, and does not propose to prescribe in Working Together any particular methodology for undertaking SCRs”.

SCIE position: transparency of methods used Currently little transparency about the

methods used in the conduct of SCRs This hampers:

the process of quality assurance and the possibility of reflection and learning in

order to foster continual improvement Compare with field of ‘evidence based

practice’ Journal articles, conferences, text books

Different methods for different kinds of cases?

Nb. A conceptual framework for understanding front-line practice

In health: RCA used in by NPSA to analyse a range of

‘patient safety incidents’ Endorsed by the WHO

In engineering, originated as a method for accident investigation developed to incorporate ‘near misses’

Strengths of the systems approach

Builds on considerable history in other fields Clear, transparent method that requires reviewers to

show their working out In built quality control

creating a shared, multi-agency process from the start minimises change of strong agency dominating interpretation; provides opportunity for challenge from an early stage

drawing on basic social science research methods for avoiding confirmation bias

Actively involves frontline staff, making the process itself a learning exercise

Aids cumulative learning from a

series of SCRs

Developing the evidence base

SCIE currently collaborating with regions to conduct pilot case reviews using the systems model (NW, West Mids, South West)

Ofsted has offered to evaluate them as if the were SCRs

SCIE keen to support a community of practice & share learning from other case review examples that people are conducting.

3. Underpinning theories of causality & implications for learning

Why do things go wrong? Lessons from aviation

Traditional person-centred investigation

vs.

System-centred investigation

Why do things go wrong? The person-centred approach

We analyze the causal sequence until we get to a satisfactory explanation.

Human error provides a satisfactory explanation.

If only the social worker had done …..

then the tragedy would not have happened. Conclusion: Erratic people degrade a safe

system so that work on safety requires protecting the system from unreliable people.

To reduce human error, we

1. Put psychological pressure on workers to perform better.

2. Reduce human factor as much as possible.formalize/mechanize/proceduralize.

3. Increase surveillance to ensure compliance with instructions etc.

Appealing but a false charm

Hindsight bias leads us to grossly overestimate how reasonable this action would have looked at the time and how easy it would have been for the worker to do it.

It is only with hindsight

that the world looks linear

because we know which causal chain actually operated. The domino theory of

causation

Why do things go wrong? The alternative system-centred approach

Individuals are not totally free to choose between good and problematic practice

We are all part of the multi-agency systems and our behaviour is shaped by systemic influences

The standard of performance is connected to features of tasks, tools and operating environment.

Implications for case reviews

A case review needs to provide a ‘window on the system’ identifying which factors are supporting good practice and which factors are, inadvertently making poor

practice more likely.

How? For both good and poor practice, need to understand

the ‘local rationality’ Why did this action/inaction/decision seem the

sensible thing to do at the time? Target recommendations at:

making it harder to safeguarding poorly and easier to do so well

Involves tackling the ‘latent conditions’ that make poor practice more likely

Active failures are like mosquitoes. They can be swatted one by one, but they still keep coming.

The best remedies are to create more effective defences and to drain the swamps in which they breed.

The swamps, in this case, are the ever present latent conditions.James Reason

Reason’s Swiss cheese model

Complex emergent model

Starkly contrasting views of how to understand the human role

Replace and substitute human beings Emphasis on fallibility and irrationality Requirement for procedural interventions and

standardisation Increase use of technical solutions

People create safety Emphasis on flexibility and adaptability Recovery from error High reliability organisations -- mindfulness, anticipation,

teamwork, respect expertise, intolerance of failure

4. Key aspects of the process

Who needs to learn and from whom?

learning should be everyone's business a system wide approach: vertical & horizontal front-line workers from different agencies and

professions need opportunities to learn about and from each other

senior managers and policy makers need to be open to learning from those at the ‘sharp end’ in a multi-agency context, it is increasingly difficult to predict

with any certainty how new policies and guidance, strategic and operational decisions impact on direct work with children and young, their carers and families.

What can’t it help with?

Legality issues access to files, especially of adults health

records Fine judgements about when poor practice

unearthed becomes a disciplinary matter

Broader blame climate Specifications about time & resources

required

Is this ‘root cause analysis’ ?

Closely related As a name RCA is misleading

implies a single ‘root’ cause when often causes are complex (non-linear dynamics)

highlights the search for causes of particular incidents rather than emphasising the need to understand strengths & weaknesses of whole systems

Chosen to put the word ‘system’ in the name because this draws attention to that key feature of the model – the opportunity to study the whole system

LT model does differ

3 key adaptations

Limitations of the knowledge base: we do not know how to mend damaged children

or families like we know how to fix broken cars

Working with families Professionals interact with families and these

relationships can influence their thinking positively as well as negatively

Multiple-agencies, professions and locations Adds to the complexity of the analysis of how

they work together and contribution each makes to the final outcome

Is it a ‘no blame’ approach?

Trying to avoid hindsight bias does not make this a “soft” approach

Not all practice is equal; quality of performance does matter and task of reviewers includes to make their judgement explicit

If ‘no blame’, equally ‘no exoneration’ Purpose is to explain & learn, not to ascribe

culpability or accountability

How do we go about it?

Preparation 1. Identifying a case for review

2. Selecting the review team

3. Identifying who should be involved

4. Preparing participants •Introductory letter (Appendix 1)

Data collection

1. Selecting documentation

2. One-to-one conversations

•Example of explanatory communication to participants (Appendix 2)

•Conversation structure (Appendix 3)

Organising & analysing the data

1. Producing a narrative of multi-agency perspectives

2. Identifying and recording key practice episodes and their contributory factors

3. Reviewing the data and analysis

4. Identifying and prioritising generic patterns

5. Making recommendations

•Template for table of key practice episodes (Appendix 4)

•Framework for contributory factors (Appendix 5)

•Typology of underlying patterns (Appendix 6)

Key features of the process

1. Involves high degree of collaboration with staff directly involved with the family Get their input to define the terms of reference. E.g. ask

practitioners who key people to speak to are

2. No IMRs at the start; instead analysis by multi-agency group and of M-A working, Allows for a focus on the interactions between agencies

from the start.

3. Draws on 2 data sources: includes in-depth 1-1 conversations, as well as documentation Nb. Instead of each manager interviewing staff in their

own agency, the same review team members speak to all the staff, from across all agencies

Key aspects of organising and analysing the data

1) Expanding the ‘chronology’. To understand ‘local rationality’ we need to go beyond the facts & highlight people’s differing perspectives

Assemble narrative of multi-agency perspectives

2) Identifying “key practice episodes” (significant to the way the case developed or was handled),

Judging the adequacy of practice in these episodes And highlighting contributory factors

3) Continual checking back & exploring further participants provide a vital check on basic accuracy also need to validate the analysis & prioritisation of

issues by the reviewers

4th aspect: moving beyond case specific details …

Aim is to make one case act as a “window on the system” (Charles Vincent 2004)

Good or problematic practice may look the different in different cases but the sets of underlying influences may be the same

6-part typology of such patterns for child welfare (more detail later) not all patterns can be covered so selection is

necessary different patterns will stand out to differing extents

for different people so debate is necessary there is no magic formula

Typology of patterns

1. human-tool operatione.g. the influence of assessment forms2. family-professional interactionse.g. dominance of the mother in social care involvement & losing focus

on the child3. human judgement/reasoninge.g. failure to review judgements and plans 4. human-management system operation e.g. resource-demand mismatch5. communication and collaboration in multi-agency working

in response to incidents/crisese.g. referral procedures and cultures of feedback6. communication and collaboration in multi-agency working

in assessment and longer-term worke.g. understanding the nature of the task; assessment and planning as

one off event or on-going process?

Benefits of such a typology

provides a conceptual framework for organising all the layers of interaction influencing the work done with a family

so that comparisons across cases can be easily conducted

and greater opportunity for cumulative learning from the series of SCRs

Developing recommendations

Systems models suggest three different kinds :

1. clear cut • E.G. creating a consistent rule re. cc-ing people into letters

• Ensuring that all voices are heard, especially workers who are actually going into families’ homes

2. require judgement and compromise • E.G. more attention in supervision to detecting errors of human

reasoning requires more time – can that be obtained by cutting back on other tasks?

3. need further research • E.G. difficulties in capturing risk well in a Core Assessment indicates

a need to research how widespread the problem is and if necessary experiment with alternative frameworks and forms. Would take time but be of national benefit

Feedback from participants

“not always a comfortable process” “gained a lot of insights about the work of my

colleagues from other agencies” “got a better understanding of how influential

and pervasive organisational culture is on face to face practice”

Feedback from pilot participants

“This way of carrying out reviews does feel much more empathetic both to professionals and family, also more wide ranging and about normal human behaviour rather than endless policies and procedures – were they present, and who didn’t follow them?”

“The recommendations feel much more constructive and practical – the aim to address real difficulties of shopfloor workers – not to make a whole lot more work developing new processes almost for the sake of being seen to do something”.

High levels of interest from sector

“Provides an evaluative framework that is derived from the nature of safeguarding work (as distinct from say education)”

“The core focus on interaction; on people and things; not only on a particular individual’s actions”;

“It will engage frontline practitioner and managers in contrast to now - whatever people say SCRs still feel done to you”

“the strength of starting off with the multi-agency team” “The importance of moving past a traditional chronology

and capturing the different perspectives” “The typology would mean that we could aggregate

learning very easily across cases and make judgements about where strategic action can take place from a common ground”

Next steps

Need to build up pool of people with experience and expertise in

using the systems approach the repertoire of case review examples

SCIE is offering: Regional programme of pilots (training &

supporting SCR Panel members) 1-day training event in this approach (regional) Training for Vol Orgs and Indep Consultants hoping also to facilitate a community of practice

network to share the learning

Contact: sheila.fish@scie.org.uk

Thank you.