Applying system dynamics to health and social care commissioning in - - PowerPoint PPT Presentation

Applying system dynamics to health and social care commissioning in the UK

Professor Eric Wolstenholme, David Monk, Gill Smith & Douglas McKelvie, OLM Consulting Coping but not coping – masking the reality of running organisations beyond their design capacity

Modelling Patient Pathways for Older People Across Health and Social Care – National and Local Modelling Disease Progression and Mental Health Services

What we have been doing

Aims To demonstrate the benefits of applying systemic polices across long patient pathways involving multiple health and social care agencies

pr imar y c ar e hos pital as s es m enttr … medic al beds tr ans fer s from medic … fas t s tr eam medic al … s ur gic al emer genc ies s ur gic al beds interm ediate c are nur s ing - r es idential domic illiar y c ar e nhs c ontinuing c are elec tiv e w ait lis t hos pital av oidenc e medic al em ergenc ies

medical beds (fast) hospital assessment medical beds slow medical emergencies transfers medical to surgical elective wait list primary care hospital avoidance surgical emergencies surgical beds intermediate care nursing residential NHS continuing care domicilliary care Routes to home exist from each sector

Primary Care Intermediate Care Acute Care Post Acute Care

Whole systems pathway modelling in health and social care

Modelling patient pathways for older people

• Delayed hospital discharges
• Investment in new capacity
• Reducing elective wait times
• Increasing elective episodes
• Allocating community beds
• Improving assessment efficiency

• Systemic answers?
• Insights?
• Learning?

What is it we provide?

• Validation – addressing inconsistencies
• Surfacing of hidden policies?

What might be: What is: Implications for:

• Health and social care practice
• The meaning of data
• System dynamics modelling

OLM

SD is unique in its ability to draw out inconsistencies

Organisation

Actual Processes Actual Policies Actual Data Actual behaviour

• ver time

Actual structure Management World Perceived

Data Perceived Processes Perceived Policies Perceived behaviour

• ver time

Perceived structure

Mismatches in behaviour Data/Structure inconsistencies and revisions Data used in models

Model structure

Policies used in models Processes used in models

Modelling World

Model behaviour

• ver time

Findings

• Gaps in data
• Inconsistencies in data
• No bottlenecks where we might expect to see

bottlenecks

Hypothesis from findings? Hypothesis: Health and social care organisations:

• 1. have to meet demand and political performance

targets for their services, irrespective of their supply capabilities and so have to work beyond their design capacity

• 2. have informal, well-intended coping actions to

achieve performance targets

• 3. these underlying actions mask the ‘over

capacity’ situation and have serious unintended consequences – future time bombs

Example of inconsistency in data - capacity modelling

await service receive service service admission rate start service rate service discharge rate length of service capacity

• f

service spare capacity demand community care purchaser care supplier

20 patients/day 20 patients/day 20 patients/day 20 patients /day

200

patients

200 patients P 10 days 200 patients

For structure/data consistency: Service discharge rate = receive service / length of service BUT often find: Service discharge rate not equal to receive service / length of service AND no bottlenecks where we might expect bottlenecks Structure wrong or data wrong or both???

Sometimes waiting has to be minimal (emergencies, government performance measures) Design capacity has to be exceeded Informal structures come into play Informal structures creates their own data

In some circumstances there is no such thing as a capacity constraint

Surfacing of alternative structure

await s ervic e receiving s ervic e service admission rate start s ervic e rate service dis charge rate length of service capac ity

• f service

spare capacity demand unmet need diversion rate to

• ther services

community care purchaser care s upplier

P1 P2 P3 P4

Implications for health and social care practice

Unintended outcomes:

• demand absorbed by stocks outside the health and

social care system

• creation of cumulative unmet need
• responsibilities for care pushed back on families,

charities and communities.

• ultimately

kick-back

• n

services, with a higher proportion of people entering as emergencies Examples of coping strategies for health and social care and their unintended consequences

Changes in gate keeping thresholds in primary care: Intended outcome: to reduce demand,

Examples of coping strategies for health and social care and their unintended consequences

Reducing lengths of stays in acute hospitals: Intended outcome: creation of spare capacity Unintended outcomes:

• creates more incomplete episodes of care and

increases chance of readmissions.

• increases the ‘revolving door phenomena’, a small

population of people recycle continuously through hospital and become a significant problem.

Examples of coping strategies for health and social care and their unintended consequences

Unintended outcomes:

• disruptive bed shifts for patients
• inefficiencies for medical staff - locating

patients. Institutionalising the practice of outliers in hospitals: Intended outcome: minimise A and E wait time.

Unintended consequence:

• patient dissatisfaction
• high cost intervention
• corrupts

future investment plans.

Examples of coping strategies for health and social care and their unintended consequences

Rationing home help hours in domiciliary care: intended outcome: creation of capacity in post acute

All these actions are well known in each agency along the patient pathway for older people

So what has our work done:

• provided a non-career limiting forum for talking about

coping actions

• documented the extent of the use of these mechanisms
• surfaced the unintended consequences of such actions –

deferment of problems through time

• suggested that these actions be addressed if sustainable

change is the aim Conclusions for health and social care and their unintended consequences

Hypothesis for data here:

data is a reflection of management actions, not a characteristics of the entities measured.

Implications for the meaning of data

Current ethos:

• 1. Data is absolute
• 2. More data is better
• 3. Data is the source of evidence-based thinking used for the

purpose of organisational review and change management.

E.g. data collected on lengths of service during periods of coping :

• reflects nothing more than management overload
• bears no mathematical relationship to other parameters or to patient

characteristics.

Does data create action OR does action create data??

• 1. System dynamics is one of the few ways to uncover

inconsistencies and informal policies.

• 2. Informal actions are a major source of hidden feedback in
• rganisations
• 3. ‘What is’ analysis: When validating models against past

data we must know the process/policy structure that existed at both the formal and informal levels

implications for system dynamics modelling

• 1. ‘What might be’ analysis: the first stage here should be to

expose the real unmasked behaviour of the system when coping policies are withdrawn.

• 2. Only then is it sensible to try to demonstrate the effects of

systemic policies to redesign the system.

• 3. The ‘what might be’ phase of system dynamics studies

requires ‘best practice’ data. Not past data associated with past practice, which we wish to replace.

implications for system dynamics modelling

Does data validate a model OR does a model validate data??

Data driving actions OR actions driving data