SummaryThere is evidence that patients living in outlying areas have poorer survival from cancer. This study set out to investigate whether they have more advanced disease at diagnosis. Case notes of 1323 patients in north and northeast Scotland who were diagnosed with lung or colorectal cancer in 1995 or 1996 were reviewed. Of patients with lung cancer, 42% (69/164) living 58 km or more from a cancer centre had disseminated disease at diagnosis compared to 33% (71/215) living within 5 km. For colorectal cancer the respective figures were 24% (38/161) and 16% (31/193). For both cancers combined, the adjusted odds ratio for disseminated disease at diagnosis in furthest group compared to the closest group was 1.59 (P = 0.037). Of 198 patients with non-small-cell lung cancer in the closest group, 56 (28%) had limited disease (stage I or II) at diagnosis compared to 23 of 165 (14%) of the furthest group (P = 0.002). The respective figures for Dukes A and B colorectal cancer were 101 of 196 (52%) and 67 of 172 (39%) (P = 0.025). These findings suggest that patients who live remote from cities and the associated cancer centres have poorer chances of survival from lung or colorectal cancer because of more advanced disease at diagnosis. This needs to be taken into account when planning investigation and treatment services.

More than 20% of the UK population live in rural areas (Cox, 1995) but there is little information on rural-urban patterns of cancer survival (Watt et al, 1993). Studies in other countries suggest that rural residence is associated with poorer survival, which could reflect more advanced stage at diagnosis and less adjuvant treatment (Bonett et al, 1990;Liff et al, 1991;Launoy et al, 1992). In the UK, the few studies of rural health in general have produced conflicting results but, overall, challenge the widespread belief that rural people have a health advantage over their urban counterparts (Phillimore and Reading, 1992;Watt et al, 1993; Cox, 1998).This study set out to investigate whether survival from cancer differed for patients resident in rural and urban areas. Two main rural indicators are associated with health: size of the local population and distance from health services (Weinert and Boik, 1995). In this paper, the hypotheses to be tested were that: (1) settlement size and (2) distance to the nearest cancer centre were associated with poorer survival. Excludes 3579 cases who died on the first day and 18 other cases who were followed up for less than 1 day.

For common cancers, survival is poorer for deprived and outlying, rural patients. This study investigated whether there were differences in treatment of colorectal and lung cancer in these groups. Case notes of 1314 patients in north and northeast Scotland who were diagnosed with lung or colorectal cancer in 1995 or 1996 were reviewed. On univariate analysis, the proportions of patients receiving surgery, chemotherapy and radiotherapy appeared similar in all socio-economic and rural categories. Adjusting for disease stage, age and other factors, there was less chemotherapy among deprived patients with lung cancer (odds ratio 0.39; 95% confidence intervals 0.16 to 0.96) and less radiotherapy among outlying patients with colorectal cancer (0.39; 0.19 to 0.82). The time between first referral and treatment also appeared similar in all socio-economic and rural groups. Adjusting for disease stage and other variables, times to lung cancer treatment remained similar, but colorectal cancer treatment was quicker for outlying patients (adjusted hazard ratio 1.30; 95% confidence intervals 1.03 to 1.64). These findings suggest that socio-economic status and rurality may have a minor impact on modalities of treatment for colorectal and lung cancer, but do not lead to delays between referral and treatment.

Stage at diagnosis and survival from cancer vary according to where people live, suggesting some may have delays in diagnosis. The aim of this study was to determine if time from presentation to treatment was longer for colorectal and breast cancer patients living further from cancer centres, and identify other important factors in delay. Data were collected on 1097 patients with breast and 1223 with colorectal cancer in north and northeast Scotland. Women with breast cancer who lived further from cancer centres were treated more quickly than those living closer to cancer centres (P ¼ 0.011). Multilevel modelling found that this was largely due to them receiving earlier treatment at hospitals other than cancer centres. Breast lump, change in skin contour, lymphadenopathy, more symptoms and signs, and increasing age predicted faster treatment. Screen detected cancers and private referrals were treated more quickly. For colorectal cancer, time to treatment was similar for people in rural and urban areas. Quicker treatment was associated with palpable rectal or abdominal masses, tenesmus, abdominal pain, frequent GP consultations, age between 50 and 74 years, tumours of the transverse colon, and iron medication at presentation. Delay was associated with past anxiety or depression. There was variation between general practices and treatment appeared quicker at practices with more female general practitioners.

We have investigated c-erbB-2 protein expression in a large cohort of well-characterized colorectal tumours, and in a subset of lymph node metastases. We have also evaluated a Val 655 Ile single nucleotide polymorphism, which is associated with an increased risk of breast cancer, in a subset of the colorectal cancer patients and in healthy control subjects. Immunohistochemical studies revealed that while 81.8% of tumours expressed c-erbB-2, in the majority of cases equivalent levels of c-erb-B2 were seen in adjacent normal mucosa. Colon tumours were significantly more likely to express c-erbB-2 than rectal tumours (P=0.015). Only 52.4% of the metastases displayed staining patterns concordant with their primary tumour, indicating that determination of c-erbB-2 protein in colorectal tumours cannot predict the status of lymph node metastases. PCR -RFLP analysis of the Val 655Ile single nucleotide polymorphism demonstrated that allele frequencies were identical between colorectal cancer patients and a control group of Caucasian subjects (Ile=0.80 and Val=0.20 in each case), indicating that it is not related to the risk of developing colorectal cancer in this population. Furthermore, there was no relationship between c-erbB-2 protein expression and gene polymorphism (P=0.58). In terms of prognosis, no association was seen between either c-erbB-2 protein expression or the presence of the Val allele and patient survival (P40.05 in each case), suggesting that c-erbB-2 is not a prognostic marker in colorectal cancer.