Predicting demographic change is a tricky business. The one thing certain, says John Appleby, is that people will live longer and this will have a considerable impact on healthcare costs

Of all the statistics gathered by governments about their citizens, population counts are perhaps the most basic - and in many ways the most important. Censuses of the population are vital in political terms, enabling governments to check electoral registers and devise electoral boundaries, for example. They also provide the basis for allocating public resources in health and for services provided by local authorities. And of course, when combined with other figures (to produce rates, for example), they serve as a fundamental statistic for epidemiological studies, and comparison and monitoring of the performance of public organisations.

Carrying out censuses is time-consuming and expensive; as in most countries, the UK has a census every ten years (the 1991 census cost£175m). But in order to plan properly, many organisations need to have some idea of how the population is likely to change in the future. Predicting population changes is tricky, and any projection is certain to be wrong to some degree.

Nevertheless, the decennial census provides an essential basis for demographers' predictive models.

The UK's population currently stands at about 59.3 million - 'about' because the figure for 1998 is based on modelled projections which draw on the 1991 decennial census plus figures for fertility and death rates, as well as immigration and emigration. Predictions from the government actuary's department suggest that the UK population will rise to around 62.8 million by 2031, and then start to fall over the next 35 years to 58.9 million (by 2066) - slightly less than the current population.

Of importance to the health service is not just total population, but also its age structure - particularly the extremes of the age ranges, where per capita healthcare consumption is much higher than average. Some changes are startling. This year, for instance, there are around 8,000 centenarians. In seven years' time this will have doubled; and by 2066 there are likely to be over 100,000 people aged 100 or over. The number of people aged over 75 will increase by 80 per cent over the next 68 years (from 4.3 million to 7.7 million).

Although there will be some decreases in the number of younger people, the combined impact of all population changes on healthcare costs will be considerable.

On the basis of the estimated per capita hospital and community health service costs for various age groups, a-back-of-the-envelope calculation suggests that demographic change alone will increase current costs by nearly 24 per cent by 2066.

But looking so far into the future raises some questions. The average 85-year-old is likely to be fitter and healthier in 2066 than their 1998 equivalent. Will they need the same level of healthcare? On the other hand, what kind of health services will be seen as 'normal' in 2066? Judging by changes in health services over the past 70 years, and current trends in demand management (essentially, expect less), we are all likely to expect big improvements in the type and level of care the NHS delivers.

John Appleby is senior lecturer in the department of health policy and practice at East Anglia University.

Population projections and related data used in this briefing are available from:

Government Actuary's Department, Division B2, 22 Kingsway, London WC2B 6LE.

Tel: 0171 211 2661.