The scariest part of the long-term fiscal projections the Treasury released last week is the discussion of health-care costs.

The problem of babyboomers' superannuation is a doddle by comparison.

This year we will spend about $13.5 billion on the public health system - some 7 per cent of gross domestic product.

It is slightly above the 6 to 7 per cent average in the OECD, but then our GDP per capita is below par for a developed country.

On current trends health spending will rise to nearly 11 per cent of GDP by the middle of the century, the Treasury projects.

Superannuation costs are projected to rise from 4 to 8 per cent of GDP over the same period.

Health spending has been growing fast, some 7.6 per cent per annum on average over the past 15 years in nominal terms or more than 5 per cent a year in real terms.

It is also growing about half as fast again as the tax base which supports it.

If it continues to grow at that pace it will swallow up more than three-quarters of the $1.1 billion allocated for new spending over the next few Budgets, the Treasury says.

It reckons growth in public health spending needs to be more like 4 per cent a year (nominal) to be consistent with a sustainable level of public debt.

The alternative is that mounting interest costs pre-empt more and more of the taxpayer's dollar, up until the point where the rest of the world via the financial markets takes the scissors to our national credit card.

Since 1950 health spending has grown more than three times faster than per capita GDP and the gap between the two tracks has widened markedly since the mid-1990s.

It illustrates one of the factors which makes constraining health care costs a particularly intractable problem.

Demand for health services grows at least as fast as incomes. It resembles in that way a luxury good.

"Generally across the OECD the long-term trend is that for every 1 per cent increase in income there is a rise of 1 per cent or more in health spending," says New Zealand Institute of Economic Research director Jean-Pierre de Raad.

Economic growth has another effect on health costs as well.

In general real wages across the economy rise in line with productivity or the increase in per capita GDP.

Those increases are expected to flow through to health workers if only because of the international mobility of skilled people.

Indeed one of the issues the ministerial health review cites is that "our health workforce has a high dependence on overseas born and trained staff in a world of growing health workforce shortages and with our ability to pay internationally competitive salaries falling behind faster-growing economies".

But it is hard for a labour-intensive sector such as health to match the productivity gains of the economy as a whole. So if remuneration in the sector has to at least match what is happening elsewhere in the economy, relative prices (costs) for health services will rise, from that effect alone.

Economic growth, in short, does not get us off the hook.

Then there is the demographic problem. Average lifespans have increased but remain finite and we tend to cost the health system most in our final year.

Statistics NZ projects the annual number of deaths to roughly double by 2050 and there is not much that can be done about that as most of the people concerned are already around.

A focus on prevention may make all kinds of sense in terms of health outcomes, but from the narrow perspective of custodians of the public purse it does not help much.

"People living longer generally develop other ailments which increase lifetime health care costs," the Treasury says, which sounds a bit like "we go to all the trouble and expense of saving lives but people just die of something else anyway. What's the point?"

To be fair, it is a little less callous than that. The plea is rather for more rigorous cost-benefit analysis of preventive measures, as against other possible interventions.

Meanwhile, technological progress is a two-edged sword. Medical advances provide benefits for patients, the Treasury concedes, but often at a high cost. Each breakthrough creates a new source of demand.

"Treatments for heart disease have evolved from bed rest and aspirin in the 1950s to a range of treatments which includes coronary bypass surgery and angioplasty now."

But de Raad points out that technology can deliver productivity gains, citing the move from open heart to keyhole surgery. Overall, the average length of stay in hospitals halved over the 1990s.

The ministerial review pointed to productivity gains to be had from improving the way hospitals, and the system more generally, do things.

"There is much to be gained by reducing the substantial gap between the best and worst performers within and between hospitals," it said.

Measures to improve patient safety can save bed-days as well as reducing preventable harm.

"More efficient national procurement of hospital supplies and management of back-office and support services will also help," the review said.

It also seemed to favour increasing concentration of specialised services in regional hubs.

And it called for "a Pharmac-like process for assessing the cost-effectiveness of medical devices and assessing them for public funding."

No doubt there are worthwhile savings to be made in these areas.

Intuitively they seem likely to provide only a limited or one-off offset to the big, intractable drivers of health costs: demographics, innovation, the dynamics of the labour market and the link between rising incomes and rising demand.

But the Treasury contends that between 2002 and 2008 at least half of the increase in health spending arose from discretionary policy initiatives by the Government.

They may have in mind things like moves to subsidise doctors' visits and raising the income- and asset-testing thresholds for long-term care.

"Dealing with future demand pressures will ... require the Government to manage public expectations as to what the publicly funded health system can do for people."

The clear implication is an expectation that people will have to fund more of their own health care.

Doesn't that mean rationing access to health care by income?

"Well," says de Raad, "at the moment we ration by waiting lists."