Rectal bleeding: prevalence and consultation behaviour

Subjects and methods QUESTIONNAIRE SURVEY

A postal questionnaire survey was carried out in a random sample of adult patients registered with two group general practices on Tyneside. The practices were of roughly equal size. One practice was situated in a predominantly middle class suburb of Newcastle upon Tyne, and the other served a large inner city population. The combined list size of adult patients for the two practices was 16600. Each practice list was divided by sex and stratified into 10 year age bands from 20 to 80. Each patient in each band was sequentially allocated a number from 1 to 100, and a sample of 12 in each 100 was drawn by means of tables of random numbers.

Patients in the sample were sent a letter from their general practitioner and a brief postal questionnaire, validated in a previous study,1 which asked whether they had consulted a doctor for any of a variety of lower bowel symptoms. A single reminder was sent to nonrespondents after four weeks. Results from the questionnaire were coded and analysed. A list of all respondents who indicated that they had experienced rectal bleeding and whether they had consulted was given to their general practitioners for further action if necessary.

INTERVIEW

Of the respondents who had indicated that they were prepared to be contacted for interview, those who had experienced rectal bleeding in the previous 12 months and who had not undergone hospital investigations for rectal bleeding were divided into consulters and non-consulters. Patients who had undergone hospital investigations for rectal bleeding were excluded as the experience might have affected their knowledge and perceptions of rectal bleeding. The non-consulters were subdivided according to whether they had seen blood mixed with stools (which has a higher predictive value for colorectal cancer than superficial bleeding11) or blood on toilet paper only. Random samples of non-consulters (15 with blood mixed with stools and 15 who had seen blood on toilet paper only) were then selected for interview. A control group of 30 respondents who had consulted a general practitioner and who were matched for age, sex, and characteristics of bleeding (blood mixed with stools or on the paper only) was also selected and invited for interview. The selected patients were contacted by telephone or letter according to their expressed preference on the questionnaire, and an interview was arranged in their home at a convenient time.

The interviews lasted between 40 and 60 minutes and consisted of a combination of structured and semistructured questions that had been developed in a previous study.12 The interviews focused on five main areas: basic demographic data; a detailed account of the duration, frequency, and characteristics of rectal bleeding and associated symptoms and family history of bowel disease; a semistructured examination of subjects' knowledge of causes of rectal bleeding and their beliefs about their own bleeding; details of the actions taken by the respondents in relation to the bleeding, including any seeking of advice; and a general examination of use of health services in the previous 12 months and more specific details about consultation for rectal bleeding, including an exploration of cues to consult.

The interviews were audiotaped and transcribed for analysis. Various quantitative non-parametric methods were used to analyse the demographic and background clinical data. The remaining data on perceptions, beliefs, and health care seeking behaviour were analysed with qualitative methods of content analysis, in which conceptual themes were identified and then revised and confirmed.13 14

The study was approved by Gateshead Ethics Committee and Newcastle and North Tyneside Joint Ethical Committee.

QUESTIONNAIRE

Of the 2000 questionnaires sent out, 119 were returned undelivered and 1200 (60%) of the remaining 1881 were completed and returned for analysis. Of the respondents, 287 (150 men, 137 women) had noticed rectal bleeding at some time in their lives and 231 had noticed it within the previous 12 months. Bleeding was most commonly reported by those aged under 50 (table I). Of the respondents who reported rectal bleeding, 90 (31%) indicated having seen blood mixed with stools and 197 (69%) reported blood only on toilet paper. Only 118 (41%) of these respondents had sought medical advice for their bleeding. Patients aged over 60 were most likely to have consulted a doctor, and those aged 40-60 were least likely to have done so (χ²=7.67, P<0.022) (table II). While blood on toilet paper was more commonly reported than blood mixed with stools, those patients with blood in their stools were more likely to have consulted a doctor than were those who had seen blood on the paper only (53 v 64, χ²=17.109, P<0.0001).

INTERVIEW

Altogether 1749 of the respondents to the questionnaire agreed to be interviewed. When those who had not experienced rectal bleeding within the past year and those who had previously undergone hospital investigations had been excluded, 61 consulters and 119 non-consulters were left for sampling. For each group, 12 men and 18 women were interviewed. The mean age of the non-consulters was 45, and that of the consulters was 47. While all the people interviewed had experienced rectal bleeding within the previous year, the range of duration of symptoms at the time of interview was four months to 12 years (median 36 months) for consulters and six months to 13 years (median 51 months) for non-consulters.

Knowledge of causes of bleeding—The subjects were asked to list all the possible causes of rectal bleeding. Knowledge of causes was similar for both groups, with haemorrhoids being the most commonly mentioned cause. The second most common known cause was cancer, mentioned by 23 consulters and 19 non-consulters (table III). Only one of the non-consulters had no knowledge of causes of rectal bleeding.

Action taken before consultation—Before consulting a doctor, 25 consulters had discussed their rectal bleeding with a relative or friend, most commonly their spouse, and nine had attempted self treatment by purchasing preparations for haemorrhoids. Of the non-consulters, 22 had so far discussed their bleeding and 11 had tried self treatment. Of these non-consulters, only four had not seen their general practitioner for some other health problem in the previous 12 months.

Reasons for consulting a doctor—Delays between the start of symptoms and seeking medical advice ranged from none to a delay of 12 years (median 2 months; interquartile range 2 weeks to 12 months). Table IV shows that the most common reason given for consultation was worry that rectal bleeding might be a sign of serious disease (“I was concerned that it could be something that could get more serious if it wasn't dealt with”). Of the 16 subjects who gave this as the main reason for seeking advice, six had delayed for at least a year. The next most common reason given was that the bleeding and associated symptoms were causing pain, discomfort, or embarrassment (“I couldn't stand the pain it was like sitting on broken glass”). For others, the opportunity arose while consulting for another reason.

Reasons for not consulting a doctor—The main reason given for not seeking advice was the belief that the bleeding was not serious (“I just don't class it as serious…I think they've got enough to do without that”) (table IV). Of the 21 informants who gave this as their main reason, 12 had previously mentioned knowledge of a link between rectal bleeding and cancer. For others, the symptoms cleared up spontaneously, and two informants were too embarrassed to mention it (“The last twice I've seen the doctor I didn't pluck up courage … embarrassment more than anything”). Non-consulters were also asked what they thought the cause of their bleeding was: 18 thought that it was caused by haemorrhoids, eight thought that constipation and associated straining was the cause, one thought that the “skin was broken,” and three had no idea of what the cause could be.