The policy and health service background to this discussion are the radical changes in cancer services currently underway in the wake of the Calman-Hine Report and the wider changes ushered in by the NHS and Community Care Act 1990 (UK). Using the changing face of hospice care as the focus, the authors explore some of the potential issues and dilemmas involved in providing supportive care for cancer patients and their families. Three 'themes', or areas of concern, are highlighted: links between services, changing organizational factors, and increasing 'medical imperialism'. Potential benefits and drawbacks of the changing ethos and organizational structures are discussed. Interview data are used as 'triggers' for the presentation of the authors' own reflections on developments in the hospice and cancer services' arenas. The paper draws on interview data collected in the pilot phase of a 3-year study on the psycho-social needs of cancer patients and their informal carers in north-west England. Twenty-nine interviews were conducted with a range of professionals involved in the provision of cancer services in Lancaster and Kendal. In the spirit of 'gathering thoughts' and facilitating debate, a commentary on developments in the hospice sector is offered rather than any firm conclusion.

A SURVEY of 861 children aged 11 to 15 plus was carried out in five schools in one educational auth ority using an anonymous questionnaire. It was found that children who had viewed the Master's Snooker Tournament sponsored by Benson and Hedges were more likely to recall the brand than non-viewers. This replication of previous findings was extended to include the effects of sponsored sport on smoking, after allowing for personality and social factors known to influence the frequency of smoking. Associations between recall, other adver tising variables and smoking were positive in all but one instance after allowing for the effects of controls.

BASED on previous work from the USA, a five lesson smoking education plan was developed and used with Secondary 1 pupils (aged 12-13 years) in three schools. A survey by anonymous questionnaire six months after the intervention showed that, among pre-test non-smokers, 27 claimed to be smokers (7.3%) in the intervention schools compared with 11 (16.7%) in a control school (p<.001), while there was no statistically significant effect of the intervention on pre-test smokers. Multivariate analyses showed similarity of concurrent and true prospective predic tions of risk factors for smoking, but failed to show the prospective value for peer group example in those who had never smoked at pre-test. The conclu sion drawn on the efficacy of the intervention must be qualified since no objective validation of smoking testimony was attempted, while the control school was not chosen strictly at random and had a somewhat higher initial smoking rate. Nevertheless, there are good grounds for arguing that the educa tional intervention had a beneficial effect.

our data is that the three groups were alike from the start. (3) We believe that Dr Taylor overrates the importance of a few data points, say the eight points around -6 50o in the medium group. Were four out of these points instead scattered from -7 to -9O, then no one would think of bimodality. In fact we cannot determine from the present study whether this population is bimodal. A large number of values are necessary to disprove a Gaussian distribution. Rather arbitrarily we divided "fast" from "slow" losers in early menopause at -3°% per year. With the three groups thus divided a X2 test showed no difference between the groups. (4) We compared the 14 women with an increased bone mineral content with other women, and our point was that the "positive difference" was not related to calcium intake. Furthermore, when we compared these 14 women with the 14 with the greatest losses we found that the latter weighed more and had higher concentrations of oestrogens. This is described elsewhere and may show how "fast" losers can be picked up by biochemical means. Again, however, this does not seem to be connected to calcium intake. We know of one way, and one only, of preventing postmenopausal bone loss, and that is hormone replacement therapy. LISBETH NILAS C CHRISTIANSEN

Surveys of adolescents' smoking habits in three regions, the north of England, Sheffieldand the east of Scotland were conducted concurrently with a national survey (O.P.C.S.) in late autumn 1982. Considerable variations were found in smoking prevalence between regions and also between individual schools. Since the timing of school education about smoking has been found to be critical, and the conclusion drawn suggests that the national data on smoking prevalence is not a sufficiently precise guide, each school may need to conduct its own survey to estimate numbers of children in each age group who have tried smoking or have already become regular smokers.

A REPRESENTATIVE survey of 880 children in first, third, and fifth years was carried out in five secon dary schools in one education authority using an anonymous questionnaire. It was found that children were most aware of the cigarette brands which are most frequently associated with sponsored sporting events on TV. Children's TV viewing of a recent snooker championship sponsored by one cigarette manufacturer was positively correlated with the proportion of children associating that brand, and other brands used in TV sponsorship, with sport. Following a snooker championship sponsored by another cigarette manufacturer, a second survey was carried out on a new sample showing that awareness of this brand, and the proportion of children associating it with sport, had increased from the first survey. This demonstrates that the TV sports spon sorship by tobacco manufacturers acts as cigarette advertising to children and therefore circumvents the law banning cigarette advertisments on TV.

An investigation was made of the effects on choice reaction time and movement time of hypoxia (oxygen deprivation) induced by a reduction in total air pressure inside a decompression chamber. The subjects were adult humans who were given no pre training on the task used, were tested once at a pressure level equivalent to some altitude between ground level and 15,000 ft, and were given no information on the altitude attained. A variety of tasks were used with simple and complex spatial and code relationships between stimulus and response. In three experiments on 92 subjects in all the most consistent finding was that from gTOund level to 7,000 ft or 10,000 ft there was a significant increase in reaction time and a significant decrease in movement time. Above 10,000 ft the results were less consistent, reaction time showing further increase or a decrease to that found at ground level and movement times varying in the opposite direction to these changes. It was suggested that in young adult subjects and in tasks of low ‘ mental load ’ the inverse variation of reaction time and movement time was quite consistent and was possibly a result of an unimpaired total response time plus a selective impairment of the time taken to initiate a response found only at intermediate altitudes. In tasks with mora ’ mental work ’ involved, hypoxic impairment of total response time was found at altitudes as low as 5,000 ft. Above 10,000 ft there was little further impairment, possibly as a result of some compensatory adjustment in drive or arousal level which might also have been the reason for the lack of impairment of reaction time at the highest altitudes in the simple tasks. The great discrepancy between these results and previous literature, and the additional finding of quicker response times with prolonged exposure to the experimental conditions, were attributed to the separate analysis of reaction time and movement time and the use of naive subjects in a novel experimental design,

Three groups of 10 university students each were tested on a cognitive task of “information-reduction” at ground level, 7,000 ft or 14,000 ft equivalent altitude in a decompression chamber. The task involved three levels of difficulty of transformation of digits read out The Ss were tested once after 20 or 50 min exposure to the condition set and given no prior training on the task. The two more difficult tasks showed significant linear decline with attitude but the easiest task showed no change.