Website of the Department of Health

The sample of patients

The sample of patients was drawn from hospital administrative records of hospital episodes. These were defined as periods in which a patient stays in hospital from admission to discharge, lasting over a number of nights or, in the case of day patients, as beginning and ending on the same day. For this survey the sample of patients was limited to those who had been discharged between July 1999 and June 2000, having been diagnosed with one of six different types of cancer. The diagnosis codes (ICD Version 10) were as follows:

Colorectal C18, C19, C20, C21
Lung C34
Breast C50
Prostate C56
Ovarian C61
Non-Hodgkin's lymphoma C82, C83, C84, C85

The sample of Trusts

All NHS Trusts in England estimated to have 150 or more qualifying patients within the stated period, were asked to participate in the survey. There were 172 such Trusts; all of them participated. Staff at the Trusts identified the eligible patients. Multi-Centre Research Ethics Committee (MREC) approval was granted for the survey prior to any contact being made with individual Trusts. Approval from the relevant Local Research Ethics Committee (LREC) was subsequently obtained for each Trust. Prior to mailing all lists of selected patients were submitted to the National Strategic Tracing Service (NSTS) so that the names of patients who had recently died could be removed.

Mailing procedures

Eligible patients were sent a questionnaire along with a covering note (on Trust letterhead, signed by the Medical Director at the Trust) and an explanatory leaflet about the survey. The letter and leaflet referred to patients with particular conditions. No reference was made to the patient's diagnosis and care was taken that the letter was not signed by a cancer specialist. This was considered by all involved with the planning of the survey (including the MREC) to be an essential element of the way the survey was presented, since there could be no certainty that patients or others in their household were aware of, or had accepted, their diagnosis. Patients who did not reply within three weeks of being mailed were sent reminders.

Fieldwork and response rates

Overall, at the 172 Trusts included in the survey, a total of 123,984 patients were selected across the six cancer types as eligible for the survey. There was a considerable variation in the numbers selected at each Trust. For the majority of Trusts however the selected sample was in the range 500-800 patients. Prior to mailing, the names of patients who were known to have died were removed from the sample. Relatively small numbers of selected patients were removed for other reasons (eg difficulties in matching NHS number to the NSTS files). In all, approximately one in four of selected names was removed, resulting in questionnaires being sent to 92,683 patients. Completed questionnaires were received back from 65,337 patients. This represented an overall response of 74%, after discounting returned mail, patients whose deaths had been reported after mailing and others who had been sent questionnaires but proved, in the event, to be ineligible. The response rates for patients with different types of cancer varied, as set out below:

The response rate varied between Trusts, from below 60% to over 80%. For this series of Network reports, results from patients at eight Trusts - where the response rate was below 60% or because in absolute numbers there were fewer than 100 returns - are not shown separately on grounds that these data may be statistically unreliable.