1 SPATIAL CASEWORKER ALLOCATION NSW Department of Community Services Ben Smith Senior Economist 4-5 October 2005.

1

SPATIAL CASEWORKER ALLOCATION

NSW Department of Community Services

Ben Smith

Senior Economist

4-5 October 2005

2

Spatial Caseworker Allocation

Between 2003 and 2008 the number of caseworkers will almost double to around 1,800.

All will be allocated using a Resource Allocation Model (RAM).

To produce the best outcomes for DoCS clients, the model must balance four sub-objectives:

3

Allocation Objectives

Equalising workload across NSW

1. Ensuring resources go to where they are used most effectively

2. Targeting client groups that receive the greatest benefit from intervention

3. Shifting resources as client needs change over time

4

Allocation Indicators

Different workload indicators are used to allocate each caseworker type to areas:

Early Intervention – percentage of reports of abuse or neglect for children aged 0 to 8.

Child Protection – percentage of all reports of abuse or neglect.

Out-of-home Care – percentage of children in OOHC.

5

Effectiveness

Equalising workload across areas eliminates over-servicing of low-need children in one area at the expense of higher-need children in other areas.

The allocation has been found to closely match measures of population size and socio-economic status (SEIFA rankings).

Allocations of rural CSCs are adjusted to reflect their higher travel times.

6

Changing Needs Over Time

Current indicator data must ‘predict’ future allocation needs.

But too much volatility can affect staff morale.

Average annual change in indicators is 18%, half of which is not sustained (random).

Only sustained changes should trigger a re-allocation of caseworkers.

7

Random Change

CSC A Percentage of Annual Child Protection Reports

0

0.1

0.2

0.3

1997-98

1998-99

1999-2000

2000-01

2001-02

2002-03

% o

f al

l CSC

s R

epor

ts

8

Sustained Change

CSC B Percentage of Annual Child Protection Reports

0

24

6

8

1997-98

1998-99

1999-2000

2000-01

2001-02

2002-03

% o

f al

l CSC

s R

epor

ts

9

‘Smoothing’ Re-allocations

Method Volatility p.a. Accuracy

Annual 17.2% 17.2%

1. Biennial 12.4% 18.5%

2. Reallocate every two years

8.7% 20.1%

3. Act on changes above 20% only

7.2% 19.7%

10

‘Smoothing’ Re-allocations

Option (3) has 42% less volatility (12.4% to 7.2%) than option (1), at a cost of 7% worse accuracy (18.5% to 19.7%).

This means a 42% reduction in staff movements every year, a substantial benefit.

The cost is that, on average, under-staffed CSCs will have 1.3% fewer staff than under option (1).

11

Future Improvements - Forecasting

Multiple regression could be used to find historical drivers of workload. Forecasting these could improve accuracy. The following factors may be linked to changes in workload:

Population changes in local government areas (LGAs)

Demographic changes such as fertility and ageing Demographic changes such as poverty, teen

pregnancy and drug abuse Economic changes such as unemployment The success of early intervention strategies.