Wednesday 8 May 2013

The relationship between health, life expectancy, socioeconomic deprivation and ESA WCA outcome.. Yay, stats.


Take-home points

There is significant variation in the proportion of the various WCA assessment states across English Local Authorities after accounting for population size. The percentage of Fit for Work judgements, for example, ranges from 15.00% to 43.82%.

As the life expectancy of an area increases, the proportion of claimants found fit for work decreases.
In areas with higher levels of health deprivation, there is a trend to find more people fit for work.

As an area experiences more socioeconomic deprivation, there is a trend to find more people fit for work.

Both health deprivation and socioeconomic deprivation act to lower life expectancy, which then acts to lower the proportion of people found fit for work. This model accounts for 21% of the variance in the proportion of fit for work judgements.

When comparing the 25 areas with the highest proportion of “fit for work” judgements (average 36.03%) with the 25 lowest (18.49%):

  • The average life expectancy for people living in the areas with the highest proportion of “fit for work” judgements was 3 years less than those living in the lowest-proportion areas
  • People living in the high-proportion areas suffer experience more health deprivation and disability
  • Significantly more people in the high-proportion areas (30%) are living in England’s poorest 20% of areas, compared to only 3% of those in the low-proportion areas.
From this data, I conclude that people living in the poorest and unhealthiest parts of England are being put at a significant disadvantage by the Work Capability Assessment for Incapacity Benefit reassessments.
  

Image 1. The relationship between life expectancy and %FFW with line of best fit



Image 2. Proportion of LA population living in England's 20% most deprived LSOAs

Data

I compiled data from the DWP (Quarterly official statistics bulletin, Issue: 30 April 2013, Table 11, http://statistics.dwp.gov.uk/asd/workingage/index.php?page=esa_wca) and the 2010 English Indices of Deprivation (Local Authority level) and collated this data into one database, available here.
All Ns = 324, all *s = p < .05, all **s = p < .001 and all tests are two-tailed unless otherwise specified.

Data were analysed in SPSS and Amos.

Description

Employment and Support Allowance data

A summary of the ESA caseload is below.
Work-Related Activity Group
Support Group
Fit For Work
Closed Before Assessment
Awaiting Assessment
Total Caseload
233,110
167,590
168,270
20,660
14,720
604,350
Table 1. ESA Caseload as of April 2013.

The proportion of the various ESA states across the 324 Local Authorities (LAs) is shown in Table 2. As you can see, there is a lot of variation between LAs across the various ESA state, with the percentage found Fit for Work ranging from 16% to 46%.


Assessment Complete
Awaiting Assessment
Raw data
WRAG
Support
Fit for Work
Closed Before Assessment
Awaiting Assessment
Maximum
5480
3980
5180
570
360
Minimum
30
40
20
10
0
Mean
719.48
517.25
519.35
63.77
45.43
Standard Deviation
701.53
409.55
569.35
63.40
55.72

Percentages





Maximum
57.78
40.00
43.82
10.00
10.24
Minimum
27.09
18.83
15.00
1.41
00.00
Mean
38.75
28.49
26.45
3.58
2.72
Standard Deviation
4.63
4.57
4.76
1.03
2.10
Table 2. Descriptive statistics of ESA states across English Local Authorities

Health and deprivation data

England’s mean male life expectancy at birth (‘life expectancy’) was 78.87 years with a standard deviation (SD) of 1.68. The minimum life expectancy was 73.60 years and the maximum was 85.10. The mean health deprivation score was 1.86 (SD = .63), ranging from 0.26 to 3.45. Higher scores indicate more health deprivation and disability, and includes factors such levels of work-limiting disability.

The mean LSOA deprivation rank was 14734.05 (SD = 5532.88), ranging from 2399.12 to 28228.06. Higher scores indicate higher levels of socioenomic deprivation. The second deprivation measure was the percentage of local people living England’s quintile of most-deprived areas. The minimum was 0%, with a maximum of 83% (mean = 14.75, SD = 16.97).

As expected, both health measures strongly correlate (r = -.876**), indicating that as health deprivation decreases, life expectancy increases.
Both deprivation measures were highly correlated (r = .871**).
For simplicity, the deprivation variables were factor-analysed and reduced to a single variable (“Deprivation factor”) accounting for 93.56% of the variance, with both variables having a weighting of .967. Increases in score of this variable indicate increasing levels of deprivation.
Analysis
Fit for Work Proportion
The percentage of ‘fit for work’ WCA judgements (%FFW) was significantly related to life expectancy (r = -.459**), health deprivation (r = .379**) and the deprivation factor (r = .389**).
When controlling for health deprivation and the deprivation factor, the relationship between life expectancy and %FFW remained significant (r = -.262**).
When controlling for life expectancy and deprivation, the relationship between %FFW and health deprivation was no longer significant (r = -.070, p = .209), as with the deprivation factor when controlling for both health variables (r = .046, p = .410).
Thus, it appears that health deprivation and socioeconomic deprivation are exerting an effect on %FOW by affecting life expectancy, and an analysis of this will follow. This relationship was modelled, as in Image 1. This model was a very good fit for the observed data (GFI = .998, x2=1.611, p = .447, df = 2, RMSEA = 0.0, 90%CI 0.00 to .103, Tucker-Lewis Index = 1.001). Every parameter was significant at the p <.001 level.


Image 3. Structural equation model of relationship between health, deprivation and %FFW variables (PFOW = Proportion of Fit for Work judgements)
In an independent-samples t-test the 25 areas with the highest %FFW had significantly lower life expectancy (3.06 years, t(48)=9.989**), higher levels of health deprivation and disability (t(48)=7.279**) and a higher proportion population living in England’s most deprived quintile of areas (2.63% vs 30.36%, t(26.442)=7.004**). Image 2 shows the relationship between life expectancy and %FFW