Chapter 3: Data Sources
3.5 Aggregate Measures which Change over Time
There were several important factors which changed over time but were not appropriate to apply to small areas.
3.5.1 Total Number of GPs
The total number of active GPs was a result of government policy decisions on numbers to be trained, and regulations influencing migration and registration of migrant doctors.
39 See
DoHA data (Department of Health and Ageing 2006)40 provided state level data which included all GPs who claimed on the MBS. As this included all GPs accessing
Medicare, including those with very low activity levels, it was unlikely that the market significantly affected the measure, and it was assumed the total numbers of GPs reflected policy decisions and was exogenous to market changes. GP density by state was used rather than the national figure, as the GP densities varied markedly by state.
3.5.2 State Levels of Unemployment
Both the use of health services, and the health of the community, have been linked to the overall state of the economy in many studies (Ruhm 2003; Granados 2004; Jin, Shah
et al. 1995; Aakvik & Holmas 2006). The role of the state unemployment measure was
to reflect the impact of the economy overall. As there can be differences between states (for example the current Western Australian mining boom), annual state level
unemployment rates were used in the panel modelling in preference to national unemployment rates.41
3.5.3 The Level of the MBS Rebate
To reflect the policy decisions regarding rebates, the rebate for standard GP
consultations (Item 23 for VR GPs, Item 53 for OMP GPs, see Section 2.2.2) were used. Item 23 services accounted for 82 per cent of consultation items claimed by VR GPs in 2001, with similar proportions of Item 53 consultations for OMP GPs.42
The average level of the rebate was confounded to a degree by the different rebates available to VR GPs and OMP GPs, and the declining proportion of OMP activity. During the period 1996 to 2003, the MBS rebates for OMP services were fixed in
40Data available directly from DoHA at
http://www.health.gov.au/internet/wcms/publishing.nsf/Content/health-pcd-statistics-gpnos.htm last accessed 16 October 2007.
41
This data was directly available from the ABS at
http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/6202.0.55.001Sep%202006?OpenDocument. Data for June each year was used. For the States, the seasonally adjusted unemployment rate was
extracted. For the Territories this was not available and the trend unemployment rate was used. Last accessed 16 October 2007.
42 Extracted from Medicare Australia on-line Medicare Statistics
http://www.medicareaustralia.gov.au/providers/health_statistics/statistical_reporting/medicare.shtml last accessed 16 October 2007.
nominal terms. However, during this period there were also policy changes which permitted certain OMP GPs (e.g. some of those practising in remote areas, and some practising in undersupplied outer metropolitan areas) to claim MBS rebates at the higher Group A1 level which was set for VR GPs. The derived level of the rebate used was weighted according to the relative quantum of services by VR and OMP GPs
(Department of Health and Ageing 2007b). Table 3.2 shows MBS rebates in nominal and real terms for the period 1996 to 2007.43
Table 3.2: MBS rebates for standard GP consultations, 1996-2007
Calendar Year Item 23 (VR GP Item 53 (OMP) AWE
adjustment Overall rebate
Date Set Nominal Rebate* Nominal Rebate* VR/OMP combined, AWE adjusted to 1996 prices 1996 Nov-95 $20.85 $17.85 1.000 $20.48 1997 Nov-96 $20.85 $17.85 0.966 $19.78 1998 Nov-97 $21.00 $17.85 0.925 $19.07 1999 Nov-98 $21.30 $17.85 0.895 $18.72 2000 Nov-99 $22.00 $17.85 0.859 $18.52 2001 Nov-00 $22.95 $17.85 0.819 $18.44 2002 Nov-01 $24.45 $17.85 0.780 $18.68 2003 Nov-02 $25.05 $17.85 0.732 $18.00 2004 Nov-03 $25.70 $17.85 0.712 $17.98 2005 Nov-04 $26.25 $17.85 0.672 $17.35 2006 Nov-05 $31.45 $17.85 0.649 $19.99 2007 Nov-06 $32.10 $21.00 0.624 $19.72
* MBS rebate for VR GP services changed from 85% to 100% of Schedule Fee in November 2005, and for OMP GPs changed in November 2006.
Sources:
AWE based on male full time ordinary time seasonally adjusted weekly earnings for May of the relevant calendar year (Australian Bureau of Statistics 2007).
MBS rebates: Biggs (2004) and Medicare Benefits Schedule for relevant years (e.g. Australian Government Department of Health and Ageing 2006).
VR/OMP relative weights Department of Health and Ageing (2007b).
3.5.4 Time Trends
McRae (2006) showed that over the period 1996 to 2003 the declining trend in services per GP was greater than can be explained by GP demography. This trend was likely to be due to both economic factors and attitudinal change (e.g. in relation to work/life balance). To accommodate the possibility of unmeasured attitudinal change, a time
43 The overall real value of the rebate declines steadily with AWE price indexation until 2005. With CPI
trend was included in the equation for the number of services per GP in which this attitudinal shift was relevant. Time trends were also potentially relevant for mortality, where there was strong long term downward trend which may have been brought about by societal and technological changes which could not be captured by variables in this dataset (Deaton & Paxson 2001).
Appendix 3.3 shows different patterns in trends in GP numbers and GP activity by region, and shows that this was likely to have been due to government policy effects, in particular the endeavours to encourage GPs in training to spend periods in rural practice. Interactions between trend and region were, therefore, also introduced into the
modelling where relevant.
3.5.5 Indemnity Indicators
Another possible cause of changing GP patterns of behaviour was that they were
initiated by GP concerns about the medical indemnity system. These concerns began to arise in 2000 when indemnity insurance rates began to rise, and the major national medical insurer appeared to be likely to collapse, and did indeed collapse in 2002 (Zinn 2003). Anecdotal evidence suggests that at this time GPs began to increase fees to cover potential increases in medical indemnity insurance premiums, and this became the tipping point for a wider fee increase. A dummy variable was defined which reflected the time up to and including 2000, and the time beyond 2000, to test this hypothesis.