Evaluation of the Better Access to Psychiatrists, Psychologists and GPs through the Medicare Benefits Schedule initiative: component B: an analysis of Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Scheme (PBS) administrative data

5.6 What is the relationship between mental health need and Better Access uptake, service use and benefits paid at the Division level?

Page last updated: August 2010

Having determined the 'performance' of each Division on the measure of mental health need and the two Better Access indicators (total Better Access and allied health Better Access services used), it was possible to develop regression models in which Better Access services used were predicted by various Division level factors (GP workforce supply, potential to access services, and other Division characteristics). The two sets of models are considered below.

The best fitting models for the data were obtained using a hierarchical model–building process comprising 4 steps. Each step comprised one or more candidate variables. Step 1 included the GP workforce supply factor variable: the rate of full–time weighted equivalent GPs in the Division (GP FWE) per 1,000 population. Step 2 included factors relating to potential to access services: state or territory; and remoteness (% of the Division population residing in remote localities, as defined by RRMA categories 6 and 7). Step 3 included the measure of mental health need: percentage of population meeting criteria for need. Step 4 included other Division characteristics: the percentage of Division population (aged 15 years and over) participating in the labour force; the percentage of Division population unemployed; the percentage of Division population living in localities of greatest relative socioeconomic disadvantage, as defined by IRSED deciles 1 to 2.

The successive contribution of the variables in each step to the explanatory power of the model was examined using the R2 statistic. Importantly, this analysis strategy enabled estimation of the independent contribution of each predictor once other factors had been accounted for. Variables that were associated with the outcome variables in univariate analyses at or below the 0.15 probability level were considered for inclusion in the models. In addition, each predictor was retained only if it contributed at least an additional 1% to the variance explained by the model. Variables were excluded if there was evidence of multicollinearity.

5.6.1 Total Better Access services used
5.6.2 Total Allied Health Better Access services used

5.6.1 Total Better Access services used

The final model predicting total Better Access services used is shown in table 5.6. The final model explained 54.70% of the variation in Better Access services used.

Each successive step contributed at least 1% additional variation explained in total persons using Better Access services in 2009. In step 1, the rate of full–time workload equivalence (FWE) of GPs per 1,000 population in each Division was found to be positively associated with Better Access services used, explaining 11.07% of the variation (adjusted R2 0.1107). Of the step 2 candidate variables, being a Division in Victoria was positively associated with services used. Being a Division with relatively high percentage of the population residing in remote locations was negatively associated with services used. Being a Division in South Australia or the Northern Territoryi was not significantly associated with Better Access services used after the subsequent addition of other variables in the model. The step 2 variables together contributed an additional 32.19% to the variation explained (giving an adjusted R2 of 0.4326). The step 3 measure of mental health need was positively associated with services used, adding an additional 3.27% to the variation explained (giving an adjusted R2 of 0.4653). Of the step 4 variables, having arelatively higher percentage of the population living in areas of greater relative socioeconomic disadvantage was negatively associated with services used, adding an additional 8.17% to the variation explained (giving an adjusted R2 of 0.5470).
Top of page
In summary, higher rates of Better Access services used were found in Divisions that had relatively higher levels of mental health need, after adjusting for all other variables in the model. However there were other factors that play a part in explaining rates of Better Access services used at a Divisional level. Higher rates of Better Access services used were also found in Divisions that had higher rates of GP supply, and Divisions located in Victoria. Lower rates of Better Access services used were found in Divisions with relatively more people living in socioeconomically disadvantaged areas and Divisions with relatively more people living in remote locations.

Table 5.6 Final model showing adjusted associations between predictors and total Better Access services used in 111 Divisions of General Practicea in 2009

CoefficientSEtP95% CI
Step 1: GP workforce supply
12.10
2.66
4.55
<0.001
6.83 to 17.38
Step 2: Division in Victoria
64.25
11.03
5.82
<0.001
42.37 to 86.14
Step 2: Divisions in SA or NT
-20.80
17.30
-1.20
0.232
-55.10 to 13.51
Step 2: Remoteness
-1.72
0.46
-3.77
<0.001
-2.63 to -0.82
Step 3: Mental health need
10.11
4.88
2.07
0.041
0.43 to 19.78
Step 4: Relative socioeconomic disadvantage
-1.39
0.32
-4.46
<0.001
-2.01 to -0.77

a Two influential outliers removed

5.6.2 Total Allied Health Better Access services used

The final model predicting total allied health Better Access services used is shown in table 5.7. The final model explained 50.99% of the variance in allied health Better Access services used.

In step 1, the rate per 1,000 population of full–time workload equivalence (FWE) of GPs in each Division was found to be positively associated with allied health Better Access services used, explaining 6.68% of the variation (adjusted R2 0.0668). Of the step 2 variables, being a Division in Victoria was positively associated with allied health services used, whereas being a Division with a relatively high proportion of the population living in remote locations was negatively associated with allied health services used. Being a Division in South Australia or the Northern Territory was not significantly associated with allied health Better Access services used after the addition of subsequent variables into the model. The step 2 variables together contributed an additional 30.12% to the variation explained (adjusted R2 0.3670). The step 3 measure of mental health need was positively associated with services used, adding an additional 6.10% to the variation explained (adjusted R2 0.4280). Of the step 4 variables, percentage of population percentage of the population living in areas of greater relative socioeconomic disadvantage was negatively associated with service use, adding an additional 8.19% to the variation explained (adjusted R2 0.5099).

In summary, higher rates of allied health Better Access services used were found in Divisions that have relatively higher levels of mental health need, after adjusting for all other variables in the model. However there were other factors that played a part in explaining rates of allied health Better Access services used at a Divisional level. Higher rates of allied health Better Access services used were also found in Divisions that have higher rates of GP supply, and Divisions located in Victoria. Lower rates of Better Access services used were found in Divisions with relatively more people living in socioeconomically disadvantaged areas and Divisions with relatively more people living in remote locations.

Table 5.7 Final model showing adjusted associations between predictors and total allied health Better Access services used in 113 Divisions of General Practice in 2009

CoefficientSEtP95% CI
Step 1: GP workforce supply
7.09
2.08
3.40
0.001
2.95 to 11.22
Step 2: Division in Victoria
45.86
8.64
5.30
<0.001
28.7 1 to 63.01
Step 2: Divisions in SA or NT
-22.51
13.57
-1.66
0.100
-49.41 to 4.38
Step 2: Remoteness
-1.19
0.36
-3.31
0.001
-1.90 to -0.48
Step 3: Mental health need
10.69
3.81
2.81
0.006
3.14 to 18.2
Step 4: Relative socioeconomic disadvantage
-1.05
0.24
-4.35
<0.001
-1.53 to -0.57

Footnotes

i The Northern Territory and South Australia were combined for analysis based on results of univariate analyses and the fact that the Northern territory comprises only one Division of General Practice.