Evaluation of a surveillance programme for women with a family history of breast cancer.

AIM: To establish health related costs and benefits of clinical services for women at increased familial risk of breast cancer. METHODS: Analysis of costs and outcomes for one UK regional service, supplemented with data from a multinational collaborative study. Main outcome measures were aggregate costs for regular clinical examination, mammographic screening and further investigations; breast cancer incidence; proportion of cancers detected at "early" or "late" stage, compared with corresponding data for unscreened women of comparable age; survival in relation to stage at diagnosis; itemised and aggregate costs of management for "early" and "late" stage breast cancer; hence direct health care costs per quality adjusted life-year (QALY) gained. RESULTS: The surveillance programme costs pound1500 (euro1600, US$2100) per woman (over 15 years). Breast cancer incidence is close to 6 per thousand examinations; 75% of tumours are detected through screening and 77% are "early" (path stage 1 or 2). Corresponding figures for unscreened women (including relatives of those attending the breast cancer family clinic) indicate that surveillance achieves a beneficial "stage shift", with reduction in treatment costs and improvement in survival, in about 22% of cases. CONCLUSIONS: The current clinical service for women at familial risk of breast cancer costs about pound4800 (euro5200, US$6800) per QALY gained. That figure is sensitive to the rate of detection of breast cancer and the degree of beneficial stage shift achieved. Within the realistic range of estimates for these two parameters, the cost per QALY may be as high as pound14,000 (euro15,300, US$20,000) or as low as pound1000 (euro1100, US$1400).

Prospective surveillance of women with a family history of breast cancer: auditing the risk threshold.

To evaluate current guidelines criteria for inclusion of women in special 'breast cancer family history' surveillance programmes, records were reviewed of women referred to Scottish breast cancer family clinics between January 1994 and December 2003 but discharged as at 'less than 'moderate' familial risk'. The Scottish Cancer Registry was then interrogated to determine subsequent age-specific incidence of breast cancer in this cohort and corresponding Scottish population figures. Among 2074 women, with an average follow-up of 4.0 years, 28 invasive breast cancers were recorded up to December 2003, where 14.4 were expected, a relative risk (RR) of 1.94. Eleven further breast cancers were recorded between January 2004 and February 2006 (ascertainment incomplete for this period). The overall RR for women in the study cohort exceeded the accepted 'cutoff' level (RR=1.7) for provision of special counselling and surveillance. The highest RR was found for the age group 45-59 years and this group also generated the majority of breast cancers. The National Institute for Clinical Excellence ('NICE') guidelines appear to be more accurate than those of the Scottish Intercollegiate Guidelines Network ('SIGN') in defining 'moderate' familial risk, and longer follow-up of this cohort could generate an evidence base for further modification of familial breast cancer services.

Familial breast cancer: management of 'lower risk' referrals.

Up to 40% of referrals from primary care to 'breast cancer family clinics' prove to be of women whose assessed risk falls below the guidelines' threshold for management in secondary or tertiary care, despite recommendations that they should be screened out at primary care level. A randomised trial, involving 87 such women referred to the Tayside Familial Breast Cancer Service compared two ways of communicating risk information, letter or personal interview. Both were found to be acceptable to referred women and to their family doctors, although the former expressed a slight preference for interview. Only four women returned to their family doctors with continuing concerns about breast cancer. Nevertheless, understanding of information provided by either route was unsatisfactory, with apparent confusion about both absolute and relative risks of breast cancer. Substantial minorities appear to believe that they are at no increased risk at all, or even below the population level of risk, while others remain convinced that their personal risk has been underestimated. Family history record forms, completed by the referred women, preferably with the assistance of relatives, are crucial to full assessment of familial risk but one quarter of women referred to the Tayside Familial Breast Cancer Service currently do not complete and return these forms ahead of their clinic appointment. Further collaboration between primary care and the Breast Cancer Family Service is required to improve provision for concerned women whose risks fall below the threshold for special surveillance and to maximise effective use of the family history record form.

Life expectancy in British Marfan syndrome populations.

A total of 206 patients with Marfan syndrome were ascertained throughout genetic clinics in Wales and Scotland during the period 1970-1990. There were 45 deaths representing 22% of the cohort. Mean age at death was 45.3+/-16.5 years. 50% median cumulative survival in the total cohort (n=206) was 53 years for males and 72 years for females. Multivariate analysis confirmed severity as the best independent indicator of survival. These findings and survival curves will assist in the counselling of British families and individuals with Marfan syndrome.

Group-specific component (Gc) subtypes and schizophrenia.

Two hundred and fifteen schizophrenic patients (108 males and 107 females) in north-east England have been investigated for Gc types and subtypes and compared with the frequencies in first-degree relatives and controls. Böök et al. (1978) described Gc2 allele association with schizophrenia. No such association was found with the Gc2 allele in the present study although there is a tendency towards an increased frequency in females. In subtype allele frequencies the female patients showed a significant reduction of Gc1S allele. The results suggest a difference in susceptibility associated with the Gc locus in patients of different sexes.