Predicting the numbers of older people “helped to live at home”

Predicting the numbers of older people “helped to live at home”Predicting the numbers of older people “helped to live at home”

A needs-based model and its implications

Project undertaken for ADSS Eastern Region – Project Chair Hugh Gault (Cambridge CC Social Services)

Eastern Region Performance Network

BSPS Conference, University of Kent, 13 Sept 2005

© 2005 Tribal Plc Page 2

Who is “Tribal”?

A leading provider of professional support services, offering a broad range of consultancy and managed services.

Employs over 2,000 people and operates from a network of 60 offices across the UK and Eire.

Operates across the public sector, within education; health and social care; local government, housing & regeneration; and central government.

Works with 80% of secondary schools; 75% of local authorities; 90% of further education colleges; and the majority of strategic health authorities and NHS trusts.


What does Tribal do?

Provides a wide range of services:

• Change management

• Performance improvement

• Benchmarking

• Executive resourcing & interim management

• Organisational development

• Direct service

• Strategy development & implementation


HTLAH – starting point


Our hypothesis

Are authorities actually providing services at a level appropriate for the needs of their population?

If so, why is this not reflected in the performance judgements?

Can we identify predictors of need?


Social care needs model

Older People

No need for social care

Private care Need for public social care

Family networks

Housing quality

Health – mortality/ morbidity

Income (IS)

Voluntary organisations

“Social Capital”

Transport/access

Preventive services

Availability of private sector

Self-funders/income


People who come in to the Social Services ‘domain’

Referrals Assessments

Care provision

Res/NH care

Intensive

Other HTLAH

No service

Voluntary organisations/direct access?

Eligibility Criteria

NO

YES


Are there any predictors?

Correlations with data from 2001 census and DWP - links to deprivation:

Older people living alone

Older people in rented accommodation

Index of Multiple Deprivation (IMD 2004)

Pensioners receiving Income Support

Pensioners in overcrowded accommodation


What about CSSR response?

FSS is allocated according to deprivation factors

£££ - The more you get, the more you

can give

…and so you should!


The technical analysis

Started with full set of demographic data (IMD, Census and DWP) at LA level.

SPSS (stepwise regression) to identify the strongest predictors

Statistical tests to check robustness and statistical validity

Strongest combination of predictors: “lone pensioners” and “LLTI”


Predictive model demonstrates that you can use the demographic data to calculate indicative data for HTLaH

Compared this to the actuals for 2003 & 2004

Predicted vs. actual PAF C32 PI


Predicted vs. actual PAF C32 PI


C32 and banding of performance


So What?

Not all authorities have the same levels of community need

LA funding is based on assumptions of different need levels

The model isn’t perfect, but need levels can be predicted

Provision is generally consistent with predicted need levels


Messages for LAs and CSCI

“More does not necessarily equal better”

It’s wrong to assume that there is a universal linear target

Performance must be judged in context of community needs.

Trying too hard to “improve” may lead to wasteful allocation of scarce resources.

Research, benchmarking and needs mapping are essential to understanding performance

Predicting the numbers of older people “helped to live at home”

Documents

tribal plcpage

predictors of need

predicted need levels

social care local government

managed services

aloneolder people

local authorities

social careprivate careneed