Skip to main content
Full AccessFull paper

Patient experiences of colonoscopy, barium enema and CT colonography: a qualitative study

Published Online:


Previous studies of patient experience with bowel screening tests, in particular CT colonography (CTC), have superimposed global rating scales and not explored individual experience in detail. To redress this, we performed qualitative interviews in order to characterize patient expectations and experiences in depth. Following ethical permission, 16 patients undergoing CTC, 18 undergoing colonoscopy and 15 undergoing barium enema agreed to a semi-structured interview by a health psychologist. Interviews were recorded, responses transcribed and themes extracted with the aim of assimilating individual experiences to facilitate subsequent development and interpretation of quantitative surveys of overall satisfaction with each diagnostic test. Transcript analysis identified three principal themes: physical sensations, social interactions and information provision. Physical sensations differed for each test but were surprisingly well tolerated overall. Social interactions with staff were perceived as very important in colouring the whole experience, particularly in controlling the feelings of embarrassment, which was critical for all procedures. Information provision was also an important determinant of experience. Verbal feedback was most common during colonoscopy and invariably reassuring. However, patients undergoing CTC received little visual or verbal feedback and were often confused regarding the test outcome. Barium enema had no specific advantage over other tests. Qualitative interviews provided important perspectives on patient experience. Our data demonstrated that models describing the quality of medical encounters are applicable to single diagnostic episodes. Staff interactions and information provision were particularly important. We found advantages specific to both CTC and colonoscopy but none for barium enema. CTC could benefit greatly from improved information provision following examination.

Symptoms of colorectal cancer, e.g. abdominal pain or change in bowel habit, are common in the general population, with the result that approximately 95% patients investigated do not actually have cancer. There is a large literature comparing the accuracy of diagnostic tests but in this study we focus on another important dimension of test performance — patients' psychological well-being. Diagnosis should be acceptable to patients in addition to being sensitive and specific. Colonoscopy and barium enema are the standard “whole-colon” tests for cancer. Colonoscopy is the reference standard but is technically demanding and associated with potentially serious adverse events. Barium enema is safer but has reduced sensitivity for cancer. It is claimed that computed tomographic colonography (CTC) is as sensitive as colonoscopy, while being safer and more acceptable to patients [14]. A case series of patients undergoing CTC and either colonoscopy (n = 696) or barium enema (n = 617) found that CTC was preferred by 72% having colonoscopy and 97% having barium enema [4]. A questionnaire study of 1233 individuals undergoing same-day CTC and colonoscopy found that CTC was thought the most convenient by 68% of responders and preferred for future screening, with more than 90% describing their experience with CTC as excellent (41%), very good (33%) or good (18%) [3].

These and similar studies [58] have assessed the acceptability of various tests using psychological questionnaires, and the results have been analysed and described by means of rating scales. However, questionnaires produced without prior explanation of the possible effects of a diagnostic method may omit dimensions that are important to patients. With specific reference to CTC, the risk of omitting a critical aspect of importance may be greater in tests that intrude on a private and sensitive part of the body. To redress this, we performed qualitative interviews with patients with the aim of characterizing their expectations and experiences in depth. We also explored whether interaction with staff during the procedure could mitigate anxiety or embarrassment, which are usually prominent features of these diagnostic tests.

Subjects and methods

Ethical permission was obtained from the relevant ethical review committee. 78 symptomatic patients who had recently undergone either CTC, barium enema or colonoscopy were identified by one of two health psychologists (C.vW, K.K). 49 ultimately agreed to be interviewed by telephone. 16 had undergone CTC, 18 colonoscopy and 15 barium enema. There were 35 (71%) women and 14 (29%) men. The overall mean age was 71 years (range, 57–92 years).


All patients underwent dietary restriction followed by full bowel purgation before each test.

Colonoscopy was performed by experienced operators using standard equipment (Olympus CF-Q240AL; Keymed, Southend, UK). An intravenous sedative (2.5 mg of midazolam), analgesic (50 mg of pethidine) and spasmolytic (10–40 mg of hycosine-N-butyl bromide) was administered in advance; nasal oxygen was administered during the procedure.

Barium enema was performed to a standard protocol by experienced radiographic technicians. An intravenous spasmolytic was administered routinely (20 mg of hycosine-N-butyl bromide) unless contraindicated. 600 ml of barium suspension (Polibar; E-Z-EM, Bichester, UK) was instilled via a rectal catheter and the colon distended via manual compression of the carbon dioxide-filled enema bag. Multiple digital fluoroscopic spot images of the colon were supplemented by lateral decubitus overcouch radiographs.

No tagging agents were used for CTC. An intravenous spasmolytic was administered routinely (20–40 mg of hycosine-N-butyl bromide) unless contraindicated. Colonic insufflation was performed by an experienced radiologist via manual compression of an enema bag containing 2500 ml of carbon dioxide, with the patient left-lateral and then supine on the CT scanner table. Both prone and supine acquisitions were obtained using a multidetector machine (LightSpeed Plus; GE Medical Systems, Abdingdon, UK) and the following parameters: 1.25–2.50 mm collimation; pitch 6; 120 kVp; 50–100 mA; 50% overlap.


Patients were interviewed by telephone by one of two health psychologists (K.K, C.vW). K.K. conducted 32 interviews with patients recruited via routine diagnostic services who had been examined using one of the three tests within 3 months. C.vW conducted 17 interviews with patients recruited to a randomized trial who had undergone either colonoscopy or CTC within the previous 48 h. Interviews were semi-structured: key aspects of each procedure were identified in advance and patients then asked about these, e.g. their experiences/perceptions before, during and after the procedure, and their general evaluation. Both psychologists had attended episodes of all three tests in advance to help develop semi-structured questions based on each. The semi-structured nature of the interview allowed patients to elaborate on those areas that interested or concerned them the most. Interviews were recorded and responses transcribed subsequently into a study spreadsheet, along with patient details, by K.K. and C.vW. Transcripts were read repeatedly by the researchers who tried to code the data by identifying recurrent themes that related to patients' perceptions of their experiences. As codes were accumulated, they were grouped into more general themes by the researchers. This procedure aimed to assimilate several individual experiences into broad descriptive groups of patient concerns and interests that could then be interpreted by the researchers, and is known as “thematic analysis” [9, 10]. Such a procedure facilitates subsequent development and interpretation of quantitative structured questionnaire surveys of overall satisfaction with each diagnostic test.


The thematic analysis of patients' responses to the semi-structured interviews identified three main themes: (i) physical sensation during the procedure; (ii) social interaction with staff; and (iii) information conveyed both during and shortly after the procedure.

Colonoscopy (Table 1)

Physical experience

Three principal components influenced experience: (i) insertion and manipulation of the colonoscope; (ii) gaseous insufflation to improve visualization; and (iii) occasional requests for the patient to alter position. Patients described colonoscopic intubation as uncomfortable and occasionally painful. One woman described colonoscopy as “worse than having a baby”. In many cases, discomfort was associated with intubation of specific segments (e.g. the sigmoid) (“That is when I really could not stand the pain. It was when they examined the top part near the chest”; S13, female). Discomfort was not constant but limited to relatively brief and intense bursts. Sedation was administered routinely and made “a big difference” for most patients. Patients referred to sedation as the main strategy to manage discomfort. One patient who could not be accompanied home was therefore ineligible for a sedative, and complained of considerable pain (“I did feel pain because I hadn't had the injection”). Some reported using additional techniques (e.g. relaxing, breath-holding) to control pain, often initiated by guidance from medical staff.

Social interaction

Colonoscopy is a practical procedure performed directly by the medical team, who thus remain in close proximity to the patient throughout. This differs from barium enema and especially colonography, which are both carried out at some distance from the patient. Although proximity could be perceived as intimidating and embarrassing, the interviews suggested the opposite. Patients benefited greatly from reassurance and comfort, and often attributed their lack of embarrassment to the way staff handled the situation. Patients described numerous examples where staff used humour (“We were all having a bit of a laugh actually”; S40, female), touch (“One of the nurses was very good, she held my arm because I was in so much pain”; S34, female) or verbal praise (“They encouraged me by saying that I was doing the right things”; S41, female).

Information provision

During colonoscopy, patients may receive visual feedback via the monitor used by the colonoscopist. Patients frequently commented on this, although reactions varied. Some found the images “fascinating” (“Really nice to look at your internal organs [laughs]”; S2, male) and a welcome distraction (“I was looking at that screen to try and take my mind off things”; S35, male). Others deliberately avoided looking (“I saw a screen but shut my eyes”; S46, female). Some felt the images too explicit and presented unwelcome information (“It would depend if you were squeamish”; S2, male). Several patients noted that they were too “sleepy” to look. Patients continually emphasised information communicated to them during or shortly after colonoscopy. This was perceived very positively (“The doctor told me that my examination had been perfectly straightforward and that they had not found anything wrong and I could go home when I was ready”; S48, male). Patients mentioned that staff often used the recovery period as an opportunity to convey information about test outcome. This was particularly helpful for those who had been markedly sedated.

Barium enema (Table 2)

Physical experience

Two principal components influenced experience: (i) instillation of barium and gas and the necessity to retain it; and (ii) frequent requests to change position on the examination table. Many patients described the procedure as uncomfortable but not particularly painful (“A little discomfort really, it was not at all painful”; S8, female). Others were more troubled — one patient felt they would “explode” (S10), whereas another described “sharp wooden pains” during instillation (S21, male). Patients were particularly detailed when describing their expectations and worries about barium enema, in many cases based on “hearsay” (“People had sort of said what it was going to be and that your dignity will go completely”; S4, female). Those with negative expectations often described their experience as “much better than expected”. One compared the discrepancy between his worry and the actual experience of the test with fear of going to the dentist (“The fear, the anticipated pain, or discomfort is greater than the actual happening”; S5, male). Patients were generally positive about their experience, focusing on the need to tolerate discomfort for health. Many used phrases such as “you have to get on with it” to emphasise their understanding that investigation was necessary and beneficial. However, the true extent of feelings was often expressed indirectly, e.g. when asked if they would undergo enema again (“I would not want to have it again. You hope there might be another way of finding out”; S4, female).

Social interaction

Barium enema was most likely to evoke embarrassment and provided a considerable challenge for both staff and patients. One patient used childbirth as an analogy. Generally, patients felt medical staff managed well. Barium/gas instillation and the postures adopted were the most embarrassing aspects. One patient stressed the importance of maintaining privacy by avoiding unnecessary personal exposure and minimizing the number of staff involved. One female patient was uncomfortable with a male performing the procedure.

Information provision

Enemas were performed by radiographic technicians and, like colonoscopy, a monitor displayed visual images. Responses suggested the monitor was less informative than during colonoscopy. Several patients stated that they could not see the screen. Others had difficulty interpreting the information, undoubtedly because it was an X-ray image (“I saw something there and I thought, 'oh god', I wonder what that is?…but because I don't understand the body it's got nothing to do with me really. I am just a patient”; S10, female). Technicians frequently detailed procedural progress but the ability to offer instant feedback regarding the result was limited by the need for subsequent radiologist interpretation. When provided, feedback was cautious, stressing the need for subsequent analysis, which provoked anxiety (“Thank God, it is all over; then it is 'I have to go back for the result'. Then of course you start to worry”; S7, female).

CT colonography (Table 3)

Physical experience

Three principal components influenced experience: (i) colonic insufflation of gas; (ii) breath-holding; and (iii) intravenous contrast. Warmth during intravenous injection was mentioned by three patients and this was frightening for those not prepared in advance. Patients generally coped with breath-holding but this was challenging for a few (“It seemed to go on and on, and once I thought I might burst in a minute. I felt a bit frightened, I thought I am going to have to breathe in a minute, I am going to ruin it all”; S26, female). Colonic distension was the primary cause of discomfort. Discomfort did not last throughout but was most intense during gas introduction, diminishing subsequently. In some, it was associated with tenesmus (“Uncomfortable in an unbearable sort of way”; S39, female). Like barium enema, patients stressed that their experience was uncomfortable rather than painful. One suggested that anxiety may have precipitated her discomfort (“It was uncomfortable but I think it was made more so by the fact that I am anxious”; S41, female). Patients who had undergone or heard about other colonic investigations previously were enthusiastic that CTC might eliminate negative experiences (“I have a friend whose mother and grandmother had bowel cancer. She can't face up to screening. I told her about this. It would be perfect, if it was not for preparation. It might save her life. I would recommend it anytime”; S1, female).

Social interactions

CTC was the most impersonal test and involved less direct interaction with clinical staff. This may reduce embarrassment — one patient described staff leaving the room during scanning as an important advantage. However, most were still embarrassed: one described the gown provided as inadequate. There was less opportunity for staff to mitigate anxiety during CTC.

Information provision

Unlike enema or colonoscopy, there was no visual feedback during CTC. No-one complained about this — some reported having their eyes closed during scanning. Verbal feedback was inconsistent. Many reported none (“Nothing at all and I was too nervous to ask. I guess they are not allowed to anyway?”; S41, female). Some were told results were conditional upon subsequent radiologist interpretation. Patients not receiving verbal feedback were most likely to be confused about follow-up (“They said they would notify the doctor, but that is the only problem, which doctor?”; S43, male). Some described anxiety regarding outcome (“No test results yet. I am just hoping they didn't find anything and there is another avenue that I might go down. I am dreading it might be cancer. Everybody dreads it”; S14, female). We also noted incidental but important comments related to patients' beliefs that CTC could detect abnormalities in organs outside the large bowel. Patients interpreted this feature as a strength (“A member of the medical team did say one might see more than just the bowel and as I am concerned about my waterworks as well. Possibly, I thought that would help me”; S41, female). Interestingly, one patient believed assurance about additional organs clearly outweighed the inconvenience of subsequent investigations (“I mean the fact that they were able to say other organs were alright, presumably the liver and you know the lower stomach, and, well, they said that the bowel looked fairly healthy too, apart from this polyp, so I suppose I had to go through it, sort of virtually a second time. I think it was the right decision”; S26, female).


We aimed to support existing quantitative findings [28] with more penetrating qualitative data based on semi-structured interviews. Transcripts were analysed and themes relating to physical experience, verbal and non-verbal feedback, and social interaction identified and extracted [9, 10]. Our primary focus was on how CTC differed from established tests.

Generally, patients found all three tests demanding but also ultimately necessary for diagnosis of their symptoms. Patients were uniformly confident that diagnostic tests had been administered professionally by medical staff, and also felt supported by them. Physical sensations associated with CTC were comparable to barium enema but accounts indicated that CTC was less physically demanding. Neither barium enema nor CTC required sedation, which formed an essential management strategy for those subjects having colonoscopy. It is well recognized that sedation administered during colonoscopy may modify patients' recollection of events, and this will influence how the test is perceived subsequently.

In contrast, feedback during or shortly after the test procedures was an important source of divergence. Patients undergoing colonoscopy received most information at the time of the test, both verbal and non-verbal, which created a diverse set of responses. At one extreme, patients shunned visual feedback to avoid potentially bad news, whereas others found the video image of their colon fascinating and a welcome distraction. During barium enema, patients either did not see the monitor or found its display confusing because they could not interpret it. There was no visual feedback during CTC. This, combined with a lack of instant reassurance about the results, meant that patients having enema or CTC could feel unduly anxious afterwards; this was never true for those having colonoscopy. This finding supports previous research highlighting the importance of communicating diagnostic test results quickly [11, 12]. Patients indicated substantial benefit from instantaneous feedback by the endoscopic team or by having the results of their test communicated shortly afterwards for those patients whose sedation was more pronounced. However, it is important to recognize that this preference may be strongly influenced by the clinical outcome. A qualitative study of mammography found a preference for hearing about abnormal findings from a patient's physician rather than the radiologist [11]. Breaking bad news, e.g. diagnosis of a colon cancer, may be inappropriate in the radiology department if the resources and specialist staff necessary to deal with patients' concerns and anxieties are not available immediately. It is also important to note that, having recognized these factors, procedures could be implemented to address them, especially because the vast majority of patients will not have cancer. For example, radiographers report barium enema examinations in some units and therefore could offer immediate reassurance if the result were normal. Also, improvements in CTC software may accelerate interpretation, which may also be performed initially by radiographers. Computer-assisted detection may also facilitate this, all of which could combine to increase the feasibility of offering instant feedback following CTC, akin to that experienced after colonoscopy.

Patients were very positive regarding social interaction and communication with staff. Our analyses suggest that this contributed greatly to a positive experience. Touch, verbal praise and even humour were frequently cited to diminish embarrassment. Encouragement also moderated pain and discomfort. Interestingly, despite a prior expectation that patients might benefit from less contact with, and a greater distance from, medical staff, patients rarely reported embarrassment during colonoscopy, arguably the most intimate test. In reality, close proximity between staff and patients allowed a more personal relationship to develop. Non-verbal behaviours such as touch and humour were described as instrumental in mitigating the physical and social challenges associated with intimate examination. This parallels models of doctor–patient interaction that stress the role of physician behaviour in facilitating patient understanding, trust and overall satisfaction with medical encounters [1316]. Qualitative studies such as ours can be especially valuable when the experiences described can be recognized by physicians [17]. In our study, patients clearly described numerous positive interactions with physicians and associated them with increased satisfaction regarding the medical encounter, which reduced physical discomfort and post-test anxiety. Encounters with medically trained staff were fewest during barium enema, many of which were performed by radiographic technicians (medically qualified staff sited the insufflation tube and intravenous access for all patients undergoing CTC).

Our study detected several important implications for CTC — the most novel test. Patients believed that CTC may tackle barriers that deter bowel screening and were also aware of and intrigued by the fact that CTC could image organs outside the large bowel, describing this as a positive advantage. Unfortunately, there is a risk of false reassurance, e.g. we found that patients believed the bladder was well imaged. Patients also highlighted some weaknesses, most of which involved the lack of information provision both during and after the test, especially delayed feedback about the outcome. CTC practitioners should consider whether a brief review could be completed before the patient leaves; our data suggest that this would have considerable benefit. Faster scanners will tackle breath-holding problems and staff should prepare patients for the sensations accompanying intravenous contrast injection. Patients were generally unaware that CTC may precipitate follow-up investigations contingent on incidental findings, and that many of these may ultimately prove unnecessary [18, 19]. We did not specifically investigate this but anticipate that some patients will find additional tests welcome for reassurance or diagnosis, whereas others will not, in line with personal attitudes. Preference for a particular test will be determined by individual concerns and expectations. Patients could be appraised of the specific advantages and disadvantages of each. For example, colonoscopy would attract individuals who want immediate results and do not mind explicit medical detail. In contrast, those for whom pain avoidance is paramount may favour CTC. Importantly, our study found no specific advantage of barium enema: it was not competitive with colonoscopy or CTC in any sphere.

Our study does have limitations. Patients were predominantly recruited from one centre, so its practices dominated, although procedures reflected standard practice. We did not investigate pre-test expectations but we anticipate these could be important. We did not interview patients with cancer, who represent approximately 5% of those investigated; their perceptions may differ. Also, all patients were symptomatic, and asymptomatic screenees may feel differently (although our patients noted that CTC may be particularly suited to screening).

In summary, we explored patient experiences during CTC, enema and colonoscopy in great depth. Staff interactions and information provision proved particularly important. We found advantages specific to both CTC and colonoscopy but none for barium enema.

Table 1. Example quotes from patients undergoing colonoscopy(CC)

Table 1.

Table 2. Example quotes from patients undergoing barium enema(BE)

Table 2.

Table 3. Example quotes from patients undergoing CT colonography(CTC)

Table 3.

The research team would particularly like to thank Reshma Kanani and David Burling for their help with recruitment, as well as the Department of Health and the HTA for funding the project (Project no. HTA/02/02).

This work was undertaken at UCLH/UCL and Imperial College who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. The views expressed in this publication are those of the authors and not necessarily those of the UK Department of Health.


  • 1 Van Dam J, Cotton P, Johnson CD, McFarland BG, Pineau BC, Provenzale D, et al. AGA future trends report: CT colonography. Gastroenterology 2004;127:970–84. Crossref Medline ISIGoogle Scholar

  • 2 Taylor SA, Halligan S, Saunders BP, Bassett P, Vance M, Bartram CI. Acceptance by patients of multidetector CT colonography compared with barium enema examinations, flexible sigmoidoscopy, and colonoscopy. AJR Am J Roentgenol 2003;181:913–21. Crossref Medline ISIGoogle Scholar

  • 3 Pickhardt PJ, Choi JR, Hwang I, Butler JA, Puckett ML, Hildebrandt HA, et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 2003;349:2191–200. Crossref Medline ISIGoogle Scholar

  • 4 Gluecker TM, Johnson CD, Harmsen WS, Offord KP, Harris AM, Wilson LA, et al. Colorectal cancer screening with CT colonography, colonoscopy, and double-contrast barium enema examination: prospective assessment of patient perceptions and preferences. Radiology 2003;227:378–84. Crossref Medline ISIGoogle Scholar

  • 5 Svensson MH, Svensson E, Lasson A, Hellstrom M. Patient acceptance of CT colonography and conventional colonoscopy: prospective comparative study in patients with or suspected of having colorectal disease. Radiology 2002;222:337–45. Crossref Medline ISIGoogle Scholar

  • 6 Ristvedt SL, McFarland EG, Weinstock LB, Thyssen EP. Patient preferences for CT colonography, conventional colonoscopy, and bowel preparation. Am J Gastroenterol 2003;98:578–85. Crossref Medline ISIGoogle Scholar

  • 7 Akerkar GA, Yee J, Hung R, McQuaid K. Patient experience and preferences toward colon cancer screening: a comparison of virtual colonoscopy and conventional colonoscopy. Gastrointest Endosc 2001;54:310–5. ISIGoogle Scholar

  • 8 Van Gelder RE, Birnie E, Florie J, Schutter MP, Bartelsman JF, Snel P, et al. CT colonography and colonoscopy: assessment of patient preference in a 5-week follow-up study. Radiology 2004;233:328–37. Crossref Medline ISIGoogle Scholar

  • 9 Richie J, Spencer L, O'Connor W. Carrying out qualitative analysis. In: Ritchie J, Lewis J, editors. Qualitative research practice. London, UK: Sage Publications, 2003:219–62. Google Scholar

  • 10 Corcoran S, Narayan S, Moreland H. "Thinking aloud" as a strategy to improve clinical decision making. Heart Lung 1988;17:463–8. Medline ISIGoogle Scholar

  • 11 Lind SE, Kopans D, Good MJ. Patients' preferences for learning the results of mammographic examinations. Breast Cancer Res Treat 1992;23:223–32. Crossref Medline ISIGoogle Scholar

  • 12 Davey HM, Lim J, Butow PN, Barratt AL, Houssami N, Higginson R. Consumer information materials for diagnostic breast tests: women's views on information and their understanding of test results. Health Expect 2003;6:298–311. Crossref MedlineGoogle Scholar

  • 13 Beck RS, Daughtridge R, Sloane PD. Physician-patient communication in the primary care office: a systematic review. J Am Board Fam Pract 2002;15:25–38. MedlineGoogle Scholar

  • 14 Berrios-Rivera JP, Street RL Jr, Garcia Popa-Lisseanu MG, Kallen MA, Richardson MN, Janssen NM, et al. Trust in physicians and elements of the medical interaction in patients with rheumatoid arthritis and systemic lupus erythematosus. Arthritis Rheum 2006;55:385–93. Crossref MedlineGoogle Scholar

  • 15 Lupton D. Consumerism, reflexivity and the medical encounter. Soc Sci Med 1997;45:373–81. Crossref Medline ISIGoogle Scholar

  • 16 Wright EB, Holcombe C, Salmon P. Doctors' communication of trust, care, and respect in breast cancer: qualitative study. BMJ 2004;328:864 Crossref MedlineGoogle Scholar

  • 17 Giacomini MK, Cook DJ. Users' guides to the medical literature: XXIII. Qualitative research in health care A. Are the results of the study valid? Evidence-Based Medicine Working Group. JAMA 2000;284:357–62. Crossref Medline ISIGoogle Scholar

  • 18 Xiong T, McEvoy K, Morton DG, Halligan S, Lilford RJ. Resources and costs associated with incidental extracolonic findings from CT colonogaphy: a study in a symptomatic population. Br J Radiol 2006;79:948–61. Link ISIGoogle Scholar

  • 19 Hara AK, Johnson CD, MacCarty RL, Welch TJ. Incidental extracolonic findings at CT colonography. Radiology 2000;215:353–7. Crossref Medline ISIGoogle Scholar

Volume 82, Issue 973January 2009
Pages: 1-e22

© The British Institute of Radiology


  • RevisedNovember 27,2007
  • ReceivedJuly 19,2007
  • AcceptedDecember 13,2007
  • Published onlineJanuary 28,2014


The research team would particularly like to thank Reshma Kanani and David Burling for their help with recruitment, as well as the Department of Health and the HTA for funding the project (Project no. HTA/02/02).

This work was undertaken at UCLH/UCL and Imperial College who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. The views expressed in this publication are those of the authors and not necessarily those of the UK Department of Health.