The 'inverse care law', population ageing and the hospital system: a distributional analysis

Economic Papers (Economic Society of Australia), March, 2005 by Linc Thurecht, Agnes Walker, Ann Harding, Jim Pearse

This study examines the socioeconomic status of NSW hospital patients in 1999-2000 and projects likely hospital costs to 2009-2010. It draws upon unique patient based datasets from NSW public and private hospitals that include hospital separations, as well as the associated treatment costs in each of the four years to 1999-2000. First, we examine whether patients of similar age had similar per patient hospital costs by socioeconomic status. Second, we examine whether patients requiring similar treatment had similar per patient hospital costs by considering patients treated for coronary heart disease. Third, we consider the impact that population ageing and changes in treatment propensities are likely to have on hospital usage and costs by 2009-2010, assuming that no changes occur in per unit treatment costs. Finally, we have estimated the additional impact of rising medical costs on these projections.

Keywords: Health, Hospitals, Inequality, Ageing

JEL Codes: H51, J11

1 Introduction

Among the many important debates taking place in health policy is the question of whether patients of higher socioeconomic status (SES) receive more costly and extensive treatment once they present for treatment (the 'inverse care law') and the likely impact of population ageing on future health costs. On the first issue, a recent paper investigated the possibility that the 'inverse care law' may apply to general practitioner (GP) services in Australia (Furler et al., 2002). That 'law' was described in the 1970s by a United Kingdom GP as the principle that "the availability of good medical care tends to vary inversely with the need of the population served" (Hart, 1971, p. 412). Furler et al. found that those living in the least disadvantaged postcode areas were significantly more likely to receive long or prolonged consultations with their doctor. They concluded that this represented an example of care provision being in inverse relationship to 'need'. Studies with similar conclusions include Hall and Holman (2003)--women in higher SES groups were significantly more likely to receive breast reconstructive surgery after surgery for breast cancer than lower SES women; and Robertson et al. (1998)--the likelihood of receiving angiography and revascularisation was significantly greater for residents from high SES locations.

A second debate which has received extensive attention since the Federal Treasurer published the Intergenerational Report (2002) and followed it with Australia's Demographic Challenges (2004), is the likely impact of population ageing on future health care costs. The Treasury has identified rising health-care costs as the major area of risk to the Commonwealth budget over the next forty years.

In this article we use a unique dataset to examine both of these issues, using NSW hospitals as a case study. First, we use a 1999-2000 dataset to examine whether patients had treatments amounting to similar per patient hospital costs, regardless of the patient's SES. As a case study involving patients with a similar principal diagnosis, differences in per-patient treatment costs by SES are studied for coronary heart disease (CHD). CHD is a sub-set of cardiovascular disease, which is one of the National Health Priority Areas. AIHW (2001) identify cardiovascular disease as the most costly disease for the health system in Australia while NSW Health (2002) finds that CHD (along with stroke) is the leading form of cardiovascular disease in NSW. While average costs per admission have previously been reported (e.g., see Table 6.2 in AIHW, 2002a), the results presented in this paper are prepared on the basis of the average cost per-patient for different socioeconomic groups. Second, we examine the impact that population ageing, changes in treatment propensities and likely increases in per unit treatment costs within hospitals are likely to have on NSW hospital usage and costs by 2009-2010.

2 Data Description

The findings presented in this paper are based on in-patient separations from all non-psychiatric public and private hospitals in New South Wales over the 1996-1997 to 1999-2000 period. (1) The data was sourced from the NSW Health Inpatient Statistics Collection and the NSW Hospital Cost Data Collection. Included are demographic and administrative information, diagnoses details, procedures performed during the episode of care, and the cost of each separation. The cost of treatment was calculated at the episode level (clinical costing) or as an average cost for a diagnosis-related group (DRG) within the medical facility (cost modelling).

A key feature of this data is that separations have been linked at the patient level. This was achieved by a probabilistic linking of separation records by variables such as address and date of birth. It is therefore possible to trace for individual patients the total cost of treatment even if there have been multiple separations. (2) This enables average patient costs to be calculated by treatment type, socioeconomic status, or other variables of interest. Table 1 shows the number of separations and patients in each year. A full description of the data, the patient linkage methodology, and steps taken to maximise data integrity is provided in Thurecht et al. (2003a).