Overheads for a presentation to an Independent Practitioners’ Association Seminar, April 2002.
Keywords: Distributional Economics; Health;
This work comes from a report on research being prepared by Brian Easton and Suzie Ballantyne.
This research arises from a limited grant from the Health Research Council.
It uses Statistics New Zealand data. However they only provided the data, and the results presented in this study are the work of the author, not Statistics New Zealand.
The data is based on unit records from the Household Economic Survey for the three years between 1994/5 to 1996/7. Access to the data used in this study was provided by Statistics New Zealand in a secure environment designed to give effect to the confidentiality provisions of the Statistics Act 1975.
The results reported here are preliminary, and should not be quoted without permission of the researchers. The intention of presenting the preliminary data is to illustrate some general principles. The final report will look at a much wider range of issues.
Note that the samples are large:
But the explanatory power is limited, suggesting there are many elements, other than those the survey records, which contribute to the observed behaviour.
The Household Economic Survey is a random sample of New Zealand Households, in which various household characteristics are surveyed including:
– age, gender and ethnicity.
– holder of a Community Services Card (CSC), High Use Card (HUC), or Medical Insurance.
– utilisation of 13 types of health services in the previous 12 months, including visits to GPs.
– health status (excellent, good, not so good or poor). (This health status and utilisation data is used to construct a health status index..)
– household disposable income
– housing tenancy, location, size.
– house hold expenditure, including 12 types of health care, one of which is outlays on primary care. Almost 50 percent of the households outlayed on primary care. (Only 23 percent spent nothing on medical care.)
WHO USES A GENERAL PRACTITIONER?
These results come from an econometric equation. It predicts the probability of a person with a particular specific characteristic going to the doctor, relative to another with a different state of that specific characteristic. All other measured characteristics are taken to be the same.
The results are ‘odds ratios’ which tell of the likelihood of the individual utilising the doctor, relative to the base case.
asterisks show statistical difference at level of
1% (**), and
Age-Gender Determinants of Number of GP Visits: Odds Ratio females 65+ Reference Group)
statistical test is on gender difference at same age.
Ethnicity Determinants of Number of GP Visits: Odds Ratio (Pakeha Reference Group)
Eligibility Determinants of Number of GP Visits: Odds Ratio (None Held Reference Group)
|High Use Card||4.99***|
|Communisty Services Card||1.53***|
Health Status Determinants of Number of GP Visits: Odds Ratio (Excellent Health Reference Group)
|Not so Good or Poor||10.75***|
Income Determinants of Number of GP Visits:
Odds Ratio for an additional $10,000 p.a. of Disposable Income = 1.13
The overall conclusion is that access to GPs seems to be primarily on the basis of health need
1. Note CSC and Medical Insurance holders high use. Is this a price effect? (They may have above average poor health in a manner not picked up by the other measures.)
2. There are personal characteristics which we have/can not measure.
What does this tell us about the “gateway” role of GPs?
WHO PAYS THEIR GP?
These results come from a (log linear) econometric equation which predicts how much additional spending is caused by a change in a variable.
These results reflect household – not individual – spending, and they apply for all primary care services – not just GPs.
Generally these estimates are for the percentage increase/decrease if the household is not the base case.
Age-Gender Determinants: Percent Increase on Primary Care Spending (Females 65+ Reference Group)
The econometric equation specification meant it is not possible to do statistical tests on gender difference at same age. However the overall gender difference is statistically significant.
Ethnicity Determinants: Percent Increase on Primary Care Spending: All Pakeha Household Reference Group
None of the differences are significant.
Eligibility Determinants: Percent Increase on Primary Care Spending (None in Household Reference Group)
|High Use Card||-4.9|
|Community Services Card||-17.7***|
Health Status Determinants:
Percent Increase on Primary Care Spending of a household all of whom are in 0% excellent health, compared to a household all of whose members are in 100% excellent health = 64.2%***
1% increase in Household Disposable Income increases spending by .17%**
The following conclusion uses other results from the research – especially total household spending on medical care.
Underpinning the financial analysis is the earlier conclusions about utilisation.
A household with sick people spend relatively more on primary services (and all medical care) than one in which all are well.
High Use Card
Households with people with high use cards spend much the same as others on primary services, even though they have higher utilisation.
However they spend less on prescriptions, more on specialists, and more overall.
Why do high income households spend relatively more on primary services (and all medical care) than low income ones?
1. Utilisation does not seem to be a factor in regard to GP use. It may be relevant for other primary services.
2. GPs might charge them more or they might go to more expensive GPs.
3. The might use free provided services less.
Because the coefficient is less than unity, the rich spend a lower proportion than the poor of their income on primary care. (This is true for all medical care spending.)
Why do Maori and Pacific Island households spend relatively less on primary care (and most other medical care) than Pakeha and Asian ones?
1. GP utilisation is same as average (but lower utilisation for many other services).
2. GPs might charge them less or they might go to less expensive GPs.
3. The might use free provided services more.
Community Services Card
CSC spend relatively less on primary care and on each and all medical services, despite being higher utilisers of health services.
Does this reflect a success of the CSC?
High utilisation may reflect better access to free services or possibly a higher rate of sickness which the other measures are not picking up. (Also possibly they are more likely to be not-in-employment, which makes visiting easier.)
Those with medical insurance use more primary services and pay more than those without. The recorded payment is often before refunds are deducted.
They also utilise relatively more of all medical services and pay more for all their medical care if medical insurance and refunds are included.
Higher utilisation may reflect cheaper access to services.
There is also some evidence that the insured are relatively sicker. (Is ‘moral hazard’ – the effect of the sick insuring – stronger than ‘adverse selection’ – where the company discourages the sick from joining its schemes.)
The sick pay more. Primary care charges contribute to their higher payments, but other components make substantial contributions too.
The Community Services Card appears relatively effective in reducing their holder outlays, despite greater utilisation rates. The High User Card reduces outlays, but holders till pay more overall.
Those poor who do not have the CSC or HUC pay a larger share of their income in medical expenses than the rich.
Medical insurance seems to generate utilisation behaviour a bit like the CSC.