Variation in the provision of chemotherapy for colorectal cancer

Abstract

Methods-Multilevel logistic regression to separate eVects of patients, areas and hospitals.Results-During the study period, 8% (n=626) of the study population received chemotherapy within six months of their first admission. Adjusting for comorbidities and emergency admissions, both age and deprivation were significantly associated with the treatment. The odds ratios (OR) of chemotherapy relative to patients aged 65-74 were 2.13 and 4.50 for patients aged 55-64 and under 55 respectively. Relative to patients resident in the most aZuent areas, the OR of chemotherapy for patients resident in the most deprived areas was 0.73. Area level availability of the treatment was not significantly associated with a patient's odds of receiving the treatment while on site provision of chemotherapy at the hospital of first admission was (OR=4.32). There was significant unexplained variation between hospitals of first admission but not between areas of residence; between hospital variation decreased by 22% during the study period. Conclusion-DiVerences according to age may reflect both clinical and patient decisions regarding the benefits of the treatment relative to its toxicity. Lower treatment rates in deprived areas may indicate inequitable access to services. Hospital diVerences may reflect consultant eVects and it would be expected that these should decrease now that the eYcacy of the treatment has been recognised and guidelines have been issued. (J Epidemiol Community Health 1999;53:775-781)

Access to specialist cancer care: is it equitable?

“…To permit comparison with the previous study by McLeod (1999), several of the same definitions were used. Patients with no recorded hospital admissions were excluded.…”

“…Patients with no recorded hospital admissions were excluded. Consistent with McLeod (1999), 6 months from date of first hospital admission was taken as the period in which uptake of chemotherapy was recorded, as almost 90% of patients who received chemotherapy did so within this period. Patients admitted after the 30th June 1996 were therefore excluded (n=1517).…”

“…Depcat 1 was categorised as 'most affluent', Depcat 2 and 3 as 'mid-affluent', Depcat 4 and 5 'mid-deprived' and Depcat 6 and 7 as 'most-deprived' quartile. Rurality was based on population density per hectare (McLeod, 1999). Postcode sectors with density 56 were termed 'rural' whilst those 56 were 'urban'.…”

“…Dukes' staging, which would have been the severity indicator of choice, was not available in cancer registration data at the time. Instead, 'death within the first 6 months' was used as an indication of severity of illness (McLeod, 1999). Since the date of death was not included in the linked data, days from diagnosis to death was used.…”

“…During the early 1990s, chemotherapy was still relatively new, and was likely to be more sensitive to variation in uptake than it may be now. McLeod (1999) showed that from 1990 to 1994 in Scotland the onsite provision of chemotherapy at hospital of first admission was significantly positively associated with patients' odds of receiving chemotherapy. A further aim was to update the analysis to the period immediately before the introduction of clinical guidelines for colorectal cancer management (SIGN, 1997) and the Calman -Hine recommendations so that it could serve as a baseline study against which change could be monitored.…”