Long-term prognosis in Crohn's disease: factors that affect quality of life

Authors


Dr C. Canavan, Digestive Diseases Centre, University Hospitals of Leicester, Leicester General Hospital, Gwendolen Road, Leicester LE5 4PW, UK.
E-mail: cc71@le.ac.uk

Abstract

Summary

Background

There are many studies investigating quality of life in recently diagnosed patients and following surgery for Crohn's disease, but there are none investigating quality of life changes with disease duration. The response shift model suggests quality of life improves with time following diagnosis.

Aim

To assess how well the model applies to patients with Crohn's disease.

Methods

The Cardiff Crohn's disease database contains data on all patients diagnosed there since 1934. Three hundred and ninety four patients diagnosed before 1 January 1985 were traced and the mortality status on 31 December 2004 established. Two hundred and eleven still living were sent quality of life questionnaires. Two hundred and eighty five questionnaires were sent to patients with varying disease duration attending out-patient clinics in Leicester.

Results

Eighty-nine valid replies were received from Cardiff, 63 from Leicester patients diagnosed over 20 years, 69 from Leicester patients diagnosed <10 years. There was no difference in quality of life between newly diagnosed and established patients. Of greatest concern was possible need for ostomy, uncertain nature of disease, and lack of energy. Stepwise regression showed that increased disease severity, older age and smoking adversely affect quality of life.

Discussion

Quality of life is equally poor in patients with established disease as in those newly diagnosed, and directly correlates with disease severity. The response shift model may not be applicable in Crohn's disease.

Introduction

The objective of treatment in a chronic condition such as Crohn's disease is to allow patients to live normally and assessing quality of life (QoL) is one way of assessing whether this objective is being met. There have been a number of studies of QoL in inflammatory bowel disease. Patients with Crohn's disease generally experience worse QoL than those with ulcerative colitis.1 As would be expected, QoL is worse during relapse than remission, but even when in remission studies have found patients’ QoL is poorer than that of healthy controls.1, 2 QoL in Crohn's disease is adversely affected by smoking and taking steroids.3 It is also worse in women.3

The most important concerns for patients are the uncertain nature of the disease, adverse effects of medication, having to use an ostomy bag, low-energy levels and the possible need for surgery.2, 4 These major concerns are the same across different cultures.4 It is important for clinicians to be aware of them so that they can be addressed within consultations. Patients with Crohn's disease also suffer from depression more frequently and this can adversely affect QoL. Patients with Crohn's disease often take more time off work than other employees and one study has shown that over a period of 6 years 59% of patients had to change their job or modify their hours as a direct result of their disease.5 This underlines the huge social and economic impact of Crohn's disease.

Research to date concentrates predominantly on patients diagnosed in the previous 5–10-year period. This provides a good picture of QoL and patient concerns at an early stage and how this might be affected by disease activity and medical or surgical intervention as well as the influence of various demographic factors and coping mechanisms. There have been no studies into how the patient's QoL and their concerns are affected by the chronic course of the disease. This study of patients with Crohn's disease for more than 20 years will establish how such concerns are affected by disease duration. The response shift model of QoL suggests that over a period of time experience of a chronic disease will alter patients’ expectations of health.6 As a consequence the discrepancy between expectations and experience is less and their perceived QoL is better than when first diagnosed, even though the clinical picture of their condition may be worse.6, 7 This study aims to determine the validity of this model within Crohn's patients who have been diagnosed for 20 or more years.

Methods

Cases were defined from a database of all patients diagnosed in Cardiff from 1932 until 1985. The database was set up in 1976 for an incidence and mortality study8, 9 and kept up to date to provide data for quinquennial incidence and prevalence studies. All patients had to meet at least one of the following criteria.

  • 1Histological confirmation of Crohn's disease in either resected or biopsy samples.
  • 2Radiological evidence of Crohn's disease with a suitable clinical history.

Cases of ‘acute terminal ileitis’ or ‘unclassified inflammatory bowel disease’ were not included. Patients were traced in the first place using the current patient database from University of Wales Hospital, Cardiff. This flagged up those people known to have died and those known still to be alive. The next method used was the people-searching programme, http://www.192.com, based on the electoral role of 2003. This indicated those patients still alive and living at the same address. The relevant Primary Care Trusts (Health Groups in Wales) were contacted in order to follow-up patients thought to be alive but who were not traced using http://www.192.com. The patients who remained untraced were then searched for on the national database of births, deaths and marriages provided on the national register and accessed using http://www.bmd.com. These tracing methods found 97% of the cohort with only 13 patients whose mortality status was unknown.

The QoL data were gathered by sending anonymous questionnaires to those patients still alive regarding their concerns about Crohn's disease. After 1 month a second mailing of the questionnaire was sent on pink paper and followed after a further month by a third copy on blue paper. This technique is used in marketing and allows assumptions to be drawn about non-responders.10 Patients in Leicester were also sent the questionnaire. These patients formed two groups, patients attending clinics at the Leicester General Hospital and those belonging to the Leicestershire branch of the National Association of Colitis and Crohn's (NACC). These responses allowed the Cardiff responses to be compared with other patients diagnosed more than 20 years ago to assess if it is appropriate to generalize results. It also allowed direct comparison between newly diagnosed patients and those with established disease and analysis of any difference in the concerns of patients who join NACC. Patients were split into those diagnosed <10 years ago and more than 20 years ago because there is a lot of literature regarding the QoL and disease prognosis in patients diagnosed <10 years ago with which data from this study can be compared. All the Cardiff patients have been diagnosed longer than 20 years and there have been no studies of QoL in patients who have lived with the condition for that long. Patients from Cardiff were matched to those from Leicester according to age, gender, age at diagnosis, disease severity and ethnicity. This meant that questionnaires from south Asian patients from Leicester were excluded, as all the Cardiff patients were Caucasian. These factors were matched as they have all been shown to affect a subject's experience of QoL.2

An instrument frequently used to measure QoL in inflammatory bowel disease is the Inflammatory Bowel Disease Questionnaire (IBDQ).11 This questionnaire was the first to be published and is still the most widely used, but it is most suitable for use as an outcome measure in clinical trials.12 It is best administered by a clinician, rather than self-administered and focuses on disease-related outcomes.12 The Rating Form for Inflammatory Bowel Disease Patients’ Concerns (RFIPC)2 is more focused on patients’ concerns with regard to their condition and can be easily and accurately self-administered,12 making it a more suitable instrument for this study.

The RFIPC is a 25-item measure scored from 1 to 100 using a 10 cm individual item visual analogue scale.2 It measures four factors: impact of disease, sexual intimacy, complications of the disease and body stigma. This questionnaire responds better to psychological aspects of QoL than the more widely used IBDQ.13 It has been well validated, demonstrates discriminative validity, is internally consistent and correlates well with the Sickness Impact Profile and Symptom Check List.2 The RFIPC is sensitive to disease activity, so the questionnaire also included the Crohn's Disease Index for Survey Research.14 This instrument was developed from the Crohn's Disease Activity Index (CDAI), designed specifically for epidemiological surveys, sent out by post and completed by the patient. It measures three variables, number of liquid stools, abdominal pain and general well being. These were the elements found to be most statistically significant from the CDAI in assessment of disease activity. The results from this measure correlate well with the CDAI and the opinion of expert clinicians.14 Patients were also asked additional personal information about age, sex, date of diagnosis, smoking habits and two questions about hospital admissions and previous surgery. These later questions were used in a surgical audit at Leicester General Hospital to assess patients’ understanding of their surgery. Answers were correct in 80% of patients when the responses were compared with patients’ notes, confirming the accuracy and utility of patients’ responses (C. Canavan, M. Kingston, M. Kelly, M. Thomas; unpublished data).

Responses to each question were quantified by measuring the length in millimetres along the baseline at which the patient had marked their concern. This gave a mark out of 100. Drossman et al. found that people experiencing ‘normal’ QoL would rate their concerns with a mean value of 38.8.2 A score higher than this represents poor QoL and a lower score indicates the patient is less concerned and can be considered to be experiencing a better QoL. As well as calculating the level of concern reported for individual questions, overall QoL can be assessed by adding the scores from the individual questions and dividing by 25 (the number of questions). This gives a score out of 100, where 38.8 would again represent a ‘normal’ QoL, with higher scores indicating worse QoL.

The questions can be grouped into categories of concerns. The four categories are:

  • 1disease-related concerns – questions addressed: financial difficulties; pain or suffering; developing cancer; dying early; having surgery; energy level;
  • 2concerns about body stigma – questions addressed: loss of bowel control; being a burden or depending on others; attractiveness; feeling dirty or smelly; having an ostomy bag; feelings about body; intimacy;
  • 3concerns about the impact of Crohn's disease on relationships – questions addressed: feeling alone; feeling out of control; passing disease onto others; being treated as different;
  • 4concerns regarding sexual intimacy – questions addressed: sexual performance; ability to have children; loss of sexual drive.

Disease severity was calculated using the scoring system devised by Pallis and Mouzas13 for survey research where a score is generated from the following equation:

image

The disease severity rating is a continuous variable and can be converted into mild (rating = 0–24), mild/moderate (rating = 25–49), moderate/severe (rating = 50–74) and severe categories (rating = 75+) and this was investigated within subgroups.

The mean scores for each of these categories were calculated in the same way as the overall mean. The mean scores were then calculated for different subgroups within the responders. These mean values were compared using independent two sample t-tests for discrete data and Pearson's method of correlation for continuous data. Stepwise multiple linear regression was used to identify the variables that had a statistically significant impact on QoL scores. Data were plotted to assess departure from normality and showed that it was reasonable to use parametric methods for analyses.

Ethical approval for this project was gained from the Leicestershire and Rutland Multiple Region Ethics Committee (reference number 04/Q2501/119).

Results

Two hundred and eleven questionnaires were sent to patients with Crohn's disease in Cardiff. One hundred and twenty one replies were received (57% response rate), of which 89 were valid responses, i.e. the questionnaire had been returned and was completed correctly, 49 replies from the first posting, 26 from the second posting and 14 from the third posting. Invalid questionnaires included: five dead; 10 moved; nine not completed correctly; six declined to complete questionnaires; two responses were too late. Two hundred and eighty five questionnaires were sent to patients with Crohn's disease in Leicester (78 to patients through the NACC). One hundred and ninety nine replies were received (70% response rate), 178 were valid, 63 from patients diagnosed at least 20 years ago and 69 diagnosed <10 years ago. Forty-six had been diagnosed between 10 and 19 years; these patients are only included in the regression analyses into patients undergoing surgery and analysis of duration of disease. Invalid questionnaires included: six dead; three moved; nine not completed correctly; two declined to complete questionnaires; one response was too late. Table 1 shows the demographics of patients who replied. The Cardiff cohort is closely matched to the Leicester patient group diagnosed for more than 20 years so direct comparison between these groups is possible. The patients diagnosed <10 years are well matched in most areas, although fewer have had any surgery for Crohn's disease and the cohort is generally younger. The groups are matched suitably well to allow comparison as gender and disease severity are closely matched as is culture, with south Asian patients being excluded.

Table 1.  Demography of patients who returned valid questionnaires
 CardiffLeicester patients diagnosed 20+ yearsLeicester patients diagnosed <10 years
  1. The percentage values are given in parenthesis.

  2. * 46 of the Leicester patients who replied have been diagnosed for between 10 and 19 years.

Total valid replies included896369
Gender
 Male35 (39)30 (48)25 (36)
 Female54 (61)33 (52)44 (64)
Have had surgery68 (76)52 (83)33 (48)
Age at diagnosis (years)
 Under 3032 (36)19 (30)31 (45)
 30 or older57 (64)44 (70)38 (55)
Years since diagnosis (years)*
 <100069 (100)
 20–2956 (63)45 (72)0
 30–3929 (33)14 (22)0
 40 or longer4 (4)4 (6)0
Current age (years)
 Under 300018 (26)
 30–5038 (43)26 (41)34 (49)
 51–7039 (44)29 (46)12 (17)
 71 or older12 (13)8 (13)5 (8)
Disease severity
 Mild27 (30)21 (33)23 (33)
 Moderate/mild34 (38)24 (38)22 (32)
 Moderate/severe22 (25)11 (17)13 (21)
 Severe6 (7)7 (12)10 (14)
Current smokers26 (29)16 (25)20 (29)

The demography of the patients who responded to each mailing in Cardiff is broadly similar. There is no statistically significant difference in the overall mean QoL reported by patients, first posting 44.7 (95% CI: 38.6–52.1), second posting 46.1 (95% CI: 36.4–55.3) and third posting 51.9 (95% CI: 46.0–59.8). The concern groupings also show no statistically significant differences in mean response scores for patients replying at different mailings. The similarity and lack of a trend in demographics and responses from each mailing suggests that the data collected are representative of the cohort, even though over 40% did not respond, as it suggests that if they had, their responses would have been similar to those of the patients who had. This technique is well recognized and widely used within marketing to make assumptions of a target population.10, 15

Duration of disease

The overall mean QoL score for patients who have had Crohn's disease for more than 20 years is 46.2 (95% CI: 40.1–51.3), which is not significantly different from newly diagnosed patients, 47.8 (95% CI: 42.0–53.5).

There is no statistically significant difference in the concerns of patients who have been diagnosed for over 20 years compared with newly diagnosed patients, except for question 13. This addresses the patient's concerns about having children. Newly diagnosed patients are more concerned about this (mean score: 32.4; 95% CI: 28.1–37.8) than those who have had the disease for more than 20 years (mean score: 15.2; 95% CI: 12.3–19.1). It is likely that this difference reflects the age of individuals in the two groups as 60% of the patients diagnosed more than 20 years are over 51, whilst only 25% are in the group diagnosed <10 years. The highest level of concern in both groups was about disease complications (47.7; 95% CI: 40.9–54.5) and body stigma (51.8; 95% CI: 44.5–59.1), with fewer concerns reported about sexual intimacy (35.2; 95% CI: 27.2–40.3) and the impact of disease on daily life and relationships (40.2; 95% CI: 32.8–44.0). Patients in both groups were least concerned about passing on Crohn's disease to others (32.4; 95% CI: 28.6–34.0), being treated as different by others (31.9; 95% CI: 28.2–33.3) and possible financial difficulties because of the disease (35.7; 95% CI: 32.4–37.1). The greatest concerns were about the need for an ostomy bag in the future (66.4; 95% CI: 63.8–70.1), producing unpleasant odours (58.7; 95% CI: 42.3–68.0), lack of energy (62.2; 95% CI: 46.4–70.6), side-effects of medication (57.9; 95% CI: 43.8–67.2) and the uncertain nature of the condition (65.5; 95% CI: 60.2–68.3).

Gender

Men and women diagnosed recently and with established disease were most concerned about body stigma-associated issues, and report least concern about sexual intimacy. There is a trend for women to be more concerned about all factors of their disease, but this in not statistically significant (mean overall QoL for women 50.84; 95% CI: 45.0–56.5 compared with 43.0; 95% CI: 36.1–49.9).

Surgery

Patients who have undergone surgery for Crohn's disease report worse overall QoL than those patients who have not. This is statistically significantly worse for patients from Cardiff, 50.1 (95% CI: 45.9–54.3) for those who have had surgery, 33.7 (95% CI: 24.1–43.4) for those who have not. Responses from the Leicester patient cohorts show the same trend but do not achieve statistical significance. A two sample t-test of the combined data from the returned and valid questionnaires shows that surgery is more likely with increased duration of disease. Patients who required surgery had been diagnosed for 6.3 years longer on average than those who have not had surgery (95% CI: 3.5–9.1 years longer; P < 0.001), showing that need for surgery increases with disease duration, as one would expect. Need for surgery is more likely in patients diagnosed at younger ages, patients who required surgery were diagnosed 4.0 years earlier than those who have not had surgery (95% CI: 0.6–7.4 years younger at diagnosis; P = 0.02).

Age at diagnosis

There is a trend for patients diagnosed >30 to experience worse QoL (48.0; 95% CI: 40.8–55.4) than those diagnosed over the age of 30 years (44.3; 95% CI: 41.7–46.4), but this is not statistically significant.

Current age

Current age did not significantly affect QoL for patients aged 70 years or younger. Patients over 71 years and diagnosed <10 years report fewer concerns and a QoL that is statistically significantly better than all younger patients (P = 0.005). However, it is not statistically significantly different to QoL reported by those over 70 years diagnosed at least 20 years ago. The QoL reported by these older patients with long established disease is not statistically significantly better than for younger patients (Table 2).

Table 2.  Mean quality of life (QoL) scores on 10 cm analogue scale according to current age with 95% confidence intervals in brackets
Current age (years)Mean QoL for patients diagnosed <10 yearsMean QoL for patients diagnosed more than 20 years
Under 3048.5 (40.0–56.9)No patients
30–5052.9 (47.0–58.7)50.1 (43.3–56.9)
51–7050.7 (42.9–58.6)46.6 (39.2–53.4)
71 or older24.0 (13.0–35.0)36.0 (25.1–46.9)

Smoking

There is a trend suggesting better QoL in all domains for patients who have been diagnosed for 20 years or more and smoke. The opposite trend is seen in patients diagnosed in the last 10 years; however, neither trends are statistically significant. Pearson's correlation method showed no statistically significant correlation between number of cigarettes smoked and disease severity in patients who had been diagnosed for more than 20 years (−0.08, P = 0.31), but showed a highly statistically significant positive correlation between amount smoked and increased disease severity in patients diagnosed <10 years ago (0.36, P < 0.0001).

Disease severity

Patients reported higher levels of concern and worse QoL in all dimensions with increasing disease severity (Table 3). QoL reported by patients with severe disease was statistically significantly worse than patients with mild disease in both recently diagnosed and established cases. There was no statistically significant difference in the QoL reported by recently diagnosed and established patients with Crohn's disease of comparable severity.

Table 3.  The difference in questionnaire responses in Cardiff and Leicester according to disease severity; mean scores on a 10 cm visual analogue scale with 95% confidence intervals in bracket
 Overall quality of life scoreConcerns about complications of diseaseConcerns about body stigmaConcerns about impact of diseaseConcerns about sexual intimacy
Cardiff
 Mild35.5 (26.6–44.5)38.7 (29.8–47.7)39.8 (29.8–47.7)31.1 (21.1–41.0)22.4 (13.5–31.4)
 Moderate/mild45.3 (39.8–50.8)48.3 (41.3–55.2)50.9 (43.7–58.0)38.4 (31.7–45.1)32.7 (25.0–40.3)
 Moderate/severe57.9 (51.7–64.2)60.3 (52.6–68.0)64.3 (56.3–72.3)51.2 (42.1–60.2)45.1 (33.4–56.8)
 Severe57.1 (42.1–72.1)66.5 (46.8–86.1)64.1 (45.2–83.1)48.8 (25.1–72.5)33.0 (13.5–52.5)
Leicester: diagnosed at least 20 years
 Mild36.1 (26.3–45.9)36.8 (26.4–47.1)38.3 (28.8–47.9)31.5 (21.0–42.0)22.5 (13.5–31.4)
 Moderate/mild46.0 (39.7–52.2)44.0 (36.5–51.5)54.3 (44.5–63.0)36.8 (28.8–44.7)34.1 (24.3–45.9)
 Moderate/severe58.3 (49.5–67.1)54.6 (48.8–61.1)69.0 (56.9–81.2)48.5 (38.3–58.7)47.7 (29.3–65.9)
 Severe65.3 (47.8–87.9)66.1 (45.8–86.3)67.6 (45.9–89.3)61.5 (39.1–83.9)57.0 (36.5–77.3)
Leicester: diagnosed <10 years
 Mild29.1 (23.0–35.3)29.1 (22.2–36.0)32.7 (25.2–40.2)21.2 (14.8–27.5)20.2 (13.3–27.1)
 Moderate/mild53.5 (47.2–59.8)52.6 (46.1–59.1)57.7 (50.0–58.4)47.1 (39.5–54.7)44.3 (34.7–53.8)
 Moderate/severe58.8 (51.4–66.2)61.1 (56.8–69.0)61.6 (52.0–71.2)49.7 (40.0–59.2)48.4 (35.5–61.3)
 Severe64.4 (58.5–70.3)63.2 (56.8–69.5)68.7 (60.6–76.9)61.1 (51.3–70.8)50.8 (38.2–63.5)

Belonging to a support group

A two sample t-test shows NACC patients do not have significantly worse disease severity than non-members (P = 0.72). However, NACC members reported worse QoL in all areas than non-members, but these results are not statistically significant.

Stepwise regression analysis

Stepwise regression analysis supports the univariate analyses described above and shows that the factors associated with worse QoL are increasing disease severity, older age and having surgery. Smoking showed a different trend in QoL in newly diagnosed patients to those who had been diagnosed more than 20 years, but was not statistically significant. There is also no statistically significant difference in QoL with increasing disease duration.

Additional concerns

In the section of the questionnaire in which patients were asked to make comments, 38% expressed additional concerns (Figure 1). The most frequent was about maintaining a diet that reduced symptoms associated with Crohn's disease (4% of all patients). Flare-up of symptoms and osteoporosis because of long-term steroid use were also common concerns across the three patient groups. Of the people who were concerned about extracolonic Crohn's disease, half were specifically concerned about occurrence of Crohn's in their mouth, even though none of them had a history of this. Three patients who have been on long-term treatment with infliximab (10 or more treatments) wrote that they were concerned that the efficacy of the drug reduced with each treatment.

Figure 1.

Graph of additional concerns expressed by patients in all three cohorts.

Discussion

This study investigates the QoL experienced by people who have been diagnosed with Crohn's disease for more than 20 years. Previous studies have concentrated on patients who have been diagnosed <10 years. The results from Cardiff and Leicester patients are statistically similar, so were combined and then compared with the newly diagnosed patients from Leicester. This has shown that on the whole, the concerns reported by newly diagnosed patients do not change with disease duration. This undermines the response shift model that suggests that patients with chronic illness experience better QoL as the duration of their disease increases because their expectations of health decrease,6, 7 although to better assess this model within Crohn's disease longitudinal study data would be necessary. It is possible that this model does not fit Crohn's disease well because the condition often follows a remitting and relapsing pattern and the response shift model may be more suitable for diseases that remain constant over many years or gradually deteriorate.

The ideal way to address the question of how QoL changes with duration of illness would be to identify a group of patients at time of diagnosis and follow them prospectively for many years. This, of course, would prove to be very difficult, especially to follow patients for up to 50 years, as some were in this study, not least because many patients would be lost to follow-up. In this study we instead took two groups of patients, those diagnosed <10 years and those diagnosed for more than 20 years and assessed the various factors that affected their QoL. The groups were then merged and the effect of disease duration on QoL was assessed as the dependant variable.

Disease-related concerns and anxieties regarding body image were rated highest in all subgroups, whilst sexual concerns were lowest. It is possible that patients were reluctant to answer questions about sexual intimacy even though questionnaires were anonymous. This problem was encountered by Moody16 in a study of sexual function in Crohn's disease, she found that concerns about sexual function are often much higher than reported and individual interviews are needed to elicit patients’ full level of concern.

The individual concerns that were greatest across all patient groups included the possible need for an ostomy bag, the uncertain nature of the disease and lack of energy. These concerns have also been rated amongst the most important in other studies of patients with Crohn's disease.2–4 It is possible that concerns about ostomy bags could be reduced by better patient education about the likely need for such surgery and what it would entail.

Concerns about cancer registered a mean score of over 50.0 in both groups. This indicates a moderate level of anxiety and a recent meta-analysis of cancer risk in Crohn's disease shows that this is a valid concern (C. Canavan, K. Abrams, J. Mayberry; unpublished data). Such worries are much higher in patients with ulcerative colitis.2 This may be because more research has been carried out into the link between ulcerative colitis and colorectal cancer, thus heightening awareness of this risk in patients. This effect may be re-enforced by offering patients screening. As screening has not yet proven effective in improving prognosis17 and may increase patient concerns, thus reducing QoL, this reduces the cost–benefit ratio.

A previous study has shown that women experience poorer QoL than men3 and a similar trend was seen in this study, both in established disease and in the newly diagnosed. Patients diagnosed under 30 years of age reported more concerns than those diagnosed older in both patient groups. Neither trend was statistically significant; possibly due to small sample size.

The only factors shown to be significantly important in affecting QoL are surgery, being aged <70 years and disease severity. About 80% of people diagnosed over 20 years ago had undergone surgery compared with only 48% of recently diagnosed patients. Patients who had undergone surgery had Crohn's disease for a mean of 6.3 years longer than those who had not. They were also 4.0 years younger when diagnosed. Patients who had surgery reported a worse QoL, with their highest level of concern regarding body image whereas the greatest concerns of patients who had not undergone surgery were about disease complications. This would suggests that rather than reducing concerns about disease complications, having surgery increases concerns about body stigma. It may be possible to reduce level of concern regarding body stigma by counselling patients about their surgery preoperatively. Only one study by Blondel-Kucharski et al. reports reduced QoL following surgery and this was seen to improve with time.3 The study by Blondel-Kucharski et al. was prospective, so likely to elicit more accurate responses than this retrospective study. However, this study represents a longer follow-up period than that by Blondel-Kucharski et al. and would suggest that whilst QoL improves postsurgically, it fails to reach the level experienced by patients who have never had surgery. This may be because patients who require surgery have more severe disease.

In patients diagnosed <10 years, smoking is associated with more severe disease and consequently poorer QoL, a finding reported on several occasions.2–4 In patients diagnosed more than 20 years ago there was no significant increase in reported disease severity amongst smokers and their QoL is similar to newly diagnosed patients. This might be because in patients diagnosed for a longer time the negative effect of smoking is countered by management regimes that take time to establish so this effect would not seen in more recently diagnosed smokers. It may also be that the effect smoking has on the underlying pathology in early Crohn's disease does not occur in established disease because over time the pathological process changes.

The demography of patients responding to each mailing shows no significant differences between responders, suggesting no inequality of representation between subgroups. There is no significant difference or trend in the concerns and QoL reported by patients replying to each posting, which suggests that patients who did not respond are likely to experience a similar QoL to those who did. This technique reduces the limitations imposed by the 43% non-response rate.10 Coupled with the fact that there was no significant difference between patients responding in Cardiff or Leicester the results of this study can be generalized to patients with Crohn's disease across the UK. It cannot be applied to South Asian patients; however, as they were excluded from this study because culture can affect QoL.2, 4 Further work could include the use of a generic QoL instrument such as the EuroQoL-5-dimension (EQ-5D),18 or SF-36, which would allow a comparison with the general population.

The QoL may be seen as a ‘soft’ outcome measure with ill-defined end points but one aim of treatment in conditions such as Crohn's disease is to improve QoL. To establish whether this is being achieved, it must be measured. The RFIPC is a well-validated and consistent measure of QoL in Crohn's disease2 and provides rich data about the concerns of patients. Its visual analogue scale means that qualitative answers can be manipulated in a quantitative manner. This study shows that patient concern does not lessen with duration of disease and QoL does not improve. This means that current management regimes are of limited success. Better and realistic counselling of patients with regard to their disease concerns during consultations and easy access to other health professionals, such as specialist nurses, may help address this issue.

Acknowledgement

This study received funding from UHL Trust. C. Canavan was supported by a grant from the Jean Shanks Foundation.

Ancillary