The diagnosis of IBS in primary care: consensus development using nominal group technique

Transcription

1 Ó The Author (2006). Published by Oxford University Press. All rights reserved. For permissions, please doi: /fampra/cml050 The diagnosis of IBS in primary care: consensus development using nominal group technique Greg Rubin a, Niek De Wit b, Villy Meineche-Schmidt c, Bohumil Seifert d, Nicola Hall a and Pali Hungin e Rubin G, De Wit N, Meineche-Schmidt V, Seifert B, Hall N and Hungin P. The diagnosis of IBS in primary care: consensus development using nominal group technique. Family Practice 2006; 23: Background. The criteria used to identify and diagnose irritable bowel syndrome (IBS) in primary care are unclear, even though most patients are managed entirely in this setting. Objective. To use a validated method of consensus development [Nominal Group Technique (Rand version)] (NGT-R) in order to construct a diagnostic framework for IBS appropriate to primary care. Methods. NGT-R is a formal method of consensus development, which uses structured interaction within a group combined with statistical derivation of group judgements. The group comprised 10 GPs with a special interest in gastroenterology and two gastroenterologists, from 10 European countries. Mailed questionnaires elicited judgements on 242 scenarios for the diagnosis of IBS, within four domains of symptoms, social and lifestyle features, psychological features and investigations. Feedback of group decisions was followed by structured face-to-face interaction and private rescoring of the questionnaire. Consensus was defined as 10/12 ratings within one of three bands, 1 3 (disagreement), 4 6 (equivocal) or 7 9 (agreement). Results. The defining features of IBS in primary care are alteration in bowel habit, bloating and abdominal pain, or discomfort or annoyance (the last reflecting important cultural differences in symptom description). These symptoms need to be present for at least 4 weeks. Supportive characteristics include female sex, family history of IBS, frequent clinic attendances, a recent major life event and a history of somatization behaviours. Abdominal examination was considered necessary in all patients and rectal examination, haemoglobin estimation and colonoscopy in those aged >55 years. The subtypes of IBS are recognized, but the diagnostic process differs only in minor ways. Final consensus was reached on 46% of statements. Conclusion. The basis of IBS diagnosis in primary care differs from, and is less exclusive than, existing criteria. Few features are deemed essential for diagnosis, while psychosocial features, patient characteristics and contextual factors are important in increasing diagnostic probability. There are important cultural differences in the description of key symptoms. These results provide information on the defining characteristics of IBS and the diagnostic process, as it occurs in primary care, and can guide clinical practice. Keywords. Irritable bowel syndrome, diagnosis, primary care, nominal group technique. Introduction Adult patients who present to the GP with disorders of the lower GI tract can pose a difficult diagnostic problem. They account for 1 in 20 of all GP consultations 1 and their symptoms are frequently ill-defined. Although functional disorders such as irritable bowel disease are the most prevalent, the possibility of Received 3 May 2006; Revised 30 July 2006; Accepted 13 September a Centre for Primary and Community Care, University of Sunderland, Sunderland, UK, b Julius Centre for Health Sciences and Primary Care, University of Utrecht, Utrecht, Netherlands, c Copenhagen, Denmark, d Department of General Practice, Charles University, Prague, Czech Republic and e Centre for Integrated Health Care Research, University of Durham, Durham, UK. Correspondence to: Greg Rubin, Professor of Primary Care, Centre for Primary and Community Care, University of Sunderland, Priestman Building, Green Terrace, Sunderland SR1 3PZ, UK. greg.rubin@sunderland.ac.uk 687

2 688 Family Practice an international journal colorectal cancer or inflammatory bowel disease must always be considered. The uncertainty that this creates may result in reluctance on the part of the doctor to attribute symptoms to a specific diagnosis. 2 In order to facilitate a positive approach to diagnosis, criteria for irritable bowel syndrome (IBS) based on specific symptoms 3 of defined duration and frequency have been developed. 4 These are derived from the characteristics of patients in secondary care and they are little used in primary care. 5 Though useful for the definition of research populations, they have been challenged as both lacking external validity for the primary care population 6 and being unnecessarily restrictive for clinical practice. 7 In some studies of primary care patients diagnosed with IBS, only a minority fulfilled the Rome II criteria. 8,9 Nevertheless, GPs have been criticized as damaging patient care by failing to conform to this disease-centred approach to IBS. 10 This may reflect a different diagnostic approach on their part, with diagnosis being based on risk estimations based on the prevalence of symptoms in primary care, balancing the relative risks of serious and non-serious disease and combining this with a limited number of investigations. Symptoms and physical findings, psychosocial background, previous disease history and consultation behaviour all play a role in the diagnostic process and may be weighted differently according to the cultural or economic setting. In brief, GPs tend to make a positive diagnosis of IBS when the risk profile for that condition is high, the characteristics of the patient fit the profile for functional disease and the risk for serious bowel disease is low. 11 This profiling approach to diagnosis is quite distinct from a criteriabased approach, though its key features and their relative importance are unknown. We used a structured consensus process to identify these key features and took a multinational approach in order to understand the diagnostic process from a European perspective. Methods We used the Rand form of Nominal Group Technique (NGT-R). This is a formal method of consensus development that uses structured interaction within a group combined with statistical derivation of group judgements. 12 It is a method of synthesizing research evidence and opinion, and is valuable when the evidence base is limited. It comprises the individual rating by members of the group of a series of scenarios that explore the subject of interest, the assembly of participants and facilitated group discussion of the scenarios and finally individual and confidential re-rating. Ten GPs with a special interest in gastroenterology and two gastroenterologists, from 10 European countries (UK, Denmark, Spain, Poland, Holland, Italy, Turkey, Slovakia, Greece and Norway) were invited to participate in the process. The study was done on behalf of the European Society for Primary Care Gastroenterology (ESPCG). An expert panel, comprising six academic GPs with a background in gastroenterology research, all members of the ESPCG, defined the cues for diagnostic decision-making, recognizing domains of symptoms, social and lifestyle features, psychological features and investigations (where appropriate, the IBS subtypes [diarrhoea-predominant (D-IBS), constipationpredominant (C-IBS) and alternating stool consistency (A-IBS)] were separately considered). From these, they constructed a series of 242 scenarios. These were mailed to all the participants, who rated each scenario for level of agreement on a 9-point Likert scale. Participants also received a pack comprising the key peer-reviewed literature relevant to the diagnosis of IBS. The group was then convened over 2 days to discuss the scenarios, facilitated by members of the expert panel. Participants were provided with the results of the first rating round (their own rating and frequency distribution per scenario). Each of the scenarios was discussed in turn, followed by a confidential re-rating. The expert panel observed but did not participate in the discussion. Literature resources were available for consultation, minutes were taken and the contribution of each participant to the discussion was monitored. In a small number of situations where it became apparent that the scenarios provided did not fully address an area of uncertainty, the group formulated additional scenarios and then confidentially rated them. The results of the second rating round were analysed by the project team in a follow-up meeting. The level of consensus was analysed for each of the scenarios and summarized. Results were reported in three categories: consensus (agree or disagree) or no consensus. Consensus was defined in advance as at least 10 out of 12 participants scoring in the same three point band (1 3, 4 6 and 7 9) of the 9-point Likert scale. Results In the first round consensus was obtained in 28% of the scenarios (63 of 225). During the group discussion the expert panel agreed on adding 23 scenarios for scoring in the second round. These explored in greater depth linguistic differences in the description of abdominal sensations, the significance of previous health problems, the history of consultation behaviour and family history of colorectal cancer. In the second round consensus was reached in 46% of the scenarios (113 of 248). Symptomatology There was consensus that abdominal sensations varied widely and could be characterized as pain or

3 The diagnosis of IBS in primary care 689 discomfort or annoyance, depending on culture and language. Annoyance was widely recognized as most typical by participants from Mediterranean countries. No consensus was reached concerning the location, type, frequency or duration of the abdominal sensation. Agreement was reached that exacerbation of symptoms after eating meals makes the diagnosis more likely and that symptom-free intervals must be present. Consensus was reached that change in bowel habit and bloating were essential for diagnosis. Associations between pain and defecation were considered supportive but not conclusive. Onset of pain with change in stool frequency or consistency, pain associated with the urge to defecate and relief of pain with defecation were all supportive of D-IBS and A-IBS, while onset of pain with change in stool frequency and pain associated with the urge to defecate supported a diagnosis of C-IBS. Participants were all fluent in English and their responses reflected the terms used by their patients to describe symptoms. There was consensus that a minimum duration of symptoms was essential and that 4 weeks was too short, but no agreement existed on a specific minimum length of time. There was consensus that the percentage of time with symptoms could be <50%, but no consensus could be obtained for a precise minimum percentage. No distinction was made between IBS subtypes. Diagnostic tests Consensus was reached that all patients should have an abdominal examination and that those aged >55 years with C-IBS or D-IBS should have a digital rectal examination. Recommendations for the use of laboratory investigations were few, particularly in those aged <55 years (Table 1). Consensus was reached that, before diagnosing D-IBS, ESR or CRP should be estimated. Faecal occult blood testing (FOBT) was not considered useful in patients aged <55 years. No agreement could be reached on the use of FOBT in patients aged >55 years. There was consensus that patients aged >55 years and those with alarm symptoms (altered blood per rectum, weight loss, abdominal or rectal mass) should have a diagnostic lower GI endoscopy before a diagnosis of IBS is made. For patients aged <55 years, endoscopy was not advocated. Consensus could not be reached on its use in this age group when additional features, such as bright red bleeding, mucus pr or a family history of colorectal cancer are present (Table 2). Comorbidity and health care seeking behaviour There was consensus for the diagnostic significance of a limited number of the associations that have been described between IBS and other comorbidities TABLE 1 Consensus on procedures required for the diagnosis of IBS Procedure IBS Type Age FOBT Sigmoidoscopy/ Colonoscopy ESR or CRP Stool culture Endo- mysial antibody test Hb/full blood count No physical exam required Digital rectal exam Abdominal examination D-IBS <55 Agree No consensus Disagree No Consensus Agree No Consensus No Consensus Disagree No consensus >55 Agree Agree Disagree Agree Agree No Consensus No Consensus No Consensus No consensus C-IBS <55 Agree No consensus Disagree No Consensus No Consensus Disagree Disagree Disagree No consensus >55 Agree Agree Disagree Agree No Consensus Disagree Disagree No consensus Agree A-IBS <55 Agree No consensus Disagree No Consensus No Consensus Disagree No Consensus Disagree No consensus >55 Agree No consensus Disagree Agree No Consensus No Consensus Disagree No consensus Agree

4 690 Family Practice an international journal TABLE 2 Consensus on need for endoscopic investigation in relation to symptoms and signs Symptoms and Signs Bright red blood per rectum Altered blood per rectum Mucus in stool Weight loss Abdominal mass Mass on rectal examination Age <55 years No consensus Agree Disagree Agree Agree Agree Age >55 years Agree Agree No consensus Agree Agree Agree TABLE 3 Patient characteristics, comorbidity and illness behaviour considered to be associated with a diagnosis of IBS BOX 1 Criteria for diagnosis of irritable bowel syndrome in primary care Consensus, Supportive Dyspareunia Family history of IBS Female sex Frequent consultations for all problems in the past Frequent consultations for functional problems in the past Recent major life event Somatization behaviour History of neurotic complaints Illness behaviours in relation to stress or characteristics (Table 3). Frequent consultations, somatization features and major life events were considered important in this respect. In the absence of a history of functional problems, reported intolerance to gluten, lactose and other food products was considered to make a diagnosis of IBS less likely, regardless of the subtype. Discussion No consensus Low back pain Tiredness Gastro-eosophageal reflux Asthma Bladder instability Dysmenorrhoea Anxiety or depression Fibromyalgia Previous hysterectomy Previous cholecystectomy Childhood sexual or physical abuse History of enteric infection (specific or non-specific) Male sex Reported food intolerance Expressed concern about health Personality type We have characterized IBS as it is recognized by practitioners in European primary care (Box 1). Alteration in bowel habit and bloating or distension, with symptom-free intervals are characteristics essential for diagnosis. Abdominal pain per se is not an essential characteristic, though participants did consider essential a feature of disordered abdominal sensation. This is expressed differently, as pain, discomfort or annoyance, according to culture and language. Symptom characteristics and inter-relationships, such as relief of abdominal pain/discomfort/annoyance with IBS NGT criteria Essential criteria Abdominal pain or discomfort or annoyance Alteration in bowel habit Symptoms free intervals Symptoms present for at least 4 weeks Bloating Supportive criteria Symptoms worse with eating Pain/discomfort/annoyance associated with defecation Onset of pain/discomfort/ annoyance associated with change in stool frequency and/or consistency Rome III criteria Essential criteria Recurrent abdominal pain/ discomfort associated with 2 or more of 1) improvement in defecation, 2) change in stool frequency at onset or 3) change in stool form at onset Symptoms on at least 3 days per month for the previous three months Symptoms present for at least 6 months Supportive criteria Bloating Passage of mucus Abnormal stool frequency (3 or more bowel movements per week or >3 per day Abnormal stool form (lumpy/ hard stool or loose/ watery stool) Urgency Straining at defecation defecation were considered supportive, as were a wide range of extra-colonic, psychosocial and behavioural features. Contextual characteristics such as the patient s personality, illness behaviour and somatization disturbances in the past assist the primary care physician in diagnosing IBS. Strengths and weaknesses This is the first time that the defining characteristics of IBS have been articulated by clinicians providing primary health care. This is significant because IBS presents first to GPs and they are the sole providers of care to the majority of sufferers. The pan-european nature of the consensus group extends the validity of the statements to cultures and languages other than English. The multi-national composition of the group might also mean that no one language or culture was sufficiently represented to allow true differences to become apparent. It is also uncertain whether there are qualitative differences between countries in the

5 The diagnosis of IBS in primary care 691 symptom experience that are hinted at by differences in the terms used to describe them. The strength of the NGT-R is its structured, detailed discussion, though this may also lead to unrepresentative and unreliable judgements. Furthermore, contextual differences, such as those between national health care systems may not be ignored by participants. 13 The reliability of recommendations reached by this method has been assessed. In formulating criteria for colonoscopy there was good agreement between two separate panels, in the US and Switzerland, 14 while a methodologically flawed comparison of recommendations for upper gastrointestinal endoscopy between those derived by NGT-R and the ASGE criteria using 2300 patients found the two to be of comparable efficiency in identifying patients who would prove to have significant lesions. 15 Relation to other research The group recognized disordered abdominal sensation as a feature of IBS but found that, first, this was characterized differently between countries and languages and, second, that it varied in severity from mere annoyance to pain. This contrasts with the Rome criteria 4 for which pain or discomfort is a cardinal symptom. Specific measures of frequency and persistence of symptoms also characterize these criteria, in contrast to our own findings. The very recent Rome III consensus statement diverges even further. 16 (Box 2) The importance of consultation behaviours, notably frequent consultation, somatization and abnormal illness behaviours in response to stress, in supporting the diagnosis of IBS may appear to be a new dimension to the diagnostic process. They are key elements in the decision-making process in general practice, however, where they are integrated with medical symptoms and the selective use of investigations to both limit somatization and achieve a positive diagnosis. An association with inappropriate consultations for minor illness and multiple somatic complaints has been previously recognized. 17 Extra-colonic symptoms, however, have less prominence in making the diagnosis and in most instances, there was no consensus on their significance. Apart from being associated with IBS, symptoms such as tiredness, urinary frequency and backache are commonly encountered in general practice and may be perceived as lacking sensitivity. Only a limited number of investigations were deemed essential in the diagnosis of IBS, in order to exclude the more serious alternative diagnoses. Notably, lower GI endoscopy was restricted to patients >55 years and serological tests for coeliac disease were not advocated, even in patients with D-IBS. Indeed the extensive investigation advised by the BSG guidelines on management of D-IBS was not supported. 18 Nominal group technique (NGT-R) is now a wellestablished technique for consensus development. BOX 2 Illustrative scenarios with participant responses Q3. In patients with abdominal pain and diarrhoea, the following symptoms confirm a diagnosis of IBS: 3.1 The onset of pain is associated with a change in frequency of stool Q16. In patients with abdominal pain and change in frequency or consistency of stool, a diagnosis of IBS is more likely if: 16.3 There is a family history of IBS Q20. In patients with abdominal pain and change in frequency or consistency of stool, a diagnosis of IBS is more likely in the presence of: 20.2 Expressed concern about health Q. 23 A patient with abdominal pain and alternating bowel habit can only be diagnosed as having IBS after the following have been done and are normal: 23.6 Stool culture

6 692 Family Practice an international journal Our methodological process was comparable with other published examples, where the number of scenarios has ranged from 264 to 1536 and the proportion achieving consensus from 39 to 82%. 19,20 That consensus was reached on only 46% of scenarios may reflect the limited evidence base and lack of gold standard for diagnosis in IBS and the unselected nature of patients seen in primary care. Implications for practice If the validity of the characteristic features of IBS, as identified though this consensus development process, can be confirmed in long-term, prospective studies, they may help GPs to make a more confident, positive diagnosis of IBS. We found that symptoms, illness behaviour, comorbidity and psychosocial factors all contribute to the diagnostic process. Our conclusions help to rationalize the use of additional investigations and, thereby, lessen the risk of somatization. This is likely to enhance the doctor/patient relationship and enable a more effective therapeutic relationship. There are significant differences between these findings, derived largely from GPs with gastroenterological expertise, and the Rome III criteria for diagnosis of IBS. Further research is needed to more clearly define the characteristics of the syndrome over time in patients in the community. The criteria we have identified need to be evaluated in the general practice setting in order to determine their test characteristics and utility. Finally, our findings have implications for the future development of drugs for IBS, since they more accurately characterize those patients with IBS as it is recognized in primary health care. Acknowledgements The authors thank the participants in the consensus group: Alberto Bozzani, Louis Ferrant, Tibor Hlavaty, Paal Kristensen, Christos Lionis, Juan Mendive, Jean Muris, Fabio Pace, Malgorzata Palka, Richard Stevens, Mehmet Ungan and Peter Whorwell. Roger Jones provided advice to the expert group. This study was sponsored by the ESPCG and funded by Novartis Pharmaceuticals. Conflicts of interest: GR and PH have received research funding, consultancy fees and speaker s honoraria from several companies with products for the treatment of IBS. References 1 Thompson WG, Heaton KW, Smyth GT, Smyth C. Irritable bowel syndrome in general practice: prevalence, characteristics and referral. Gut 2000; 46: May C, Allison G, Chapple A, Chew-Graham C, Dixon C, Gask L et al. Framing the doctor-patient relationship in chronic illness: a comparative study of general practitioners accounts. Soc Health and Illness 2004; 26(2): Manning AP, Thompson WG, Heaton KW, Morris AF. Towards positive diagnosis of the irritable bowel. BMJ 1978; 2: Thompson WG, Longstreth GF, Drossman DA, Heaton KW, Irvine EJ, Muller-Lissner SA. Functional bowel disorders and functional abdominal pain. Gut 1999; 45 (Suppl II): Lea R, Hopkins V, Hastleton J, Houghton LA, Whorwell PJ. Diagnostic criteria for irritable bowel syndrome: utility and applicability in clinical practice. Digestion 2004; 70: Starmans R, Muris JW, Fijten GH, Schouten HJA, Pop P, Knotterus JA. The diagnostic value of scoring models for organic and non-organic gastrointestinal disease, including the irritable bowel syndrome. Med Decis Making 1994; 14: Boyce PM, Koloski NA, Talley NJ. Irritable bowel syndrome according to varying diagnostic criteria: are the new Rome II criteria unnecessarily restrictive for research and practice. Am J Gastro 2000; 95: Mearin F, Badia X, Balboa A, Baro E, Caldwell E, Cucala M et al. Irritable bowel syndrome prevalence varies enormously depending on the employed diagnostic criteria: comparison of Rome II versus previous criteria in a general population. Scand J Gastroenterol 2001; 36(11): Bijkerk CJ, de Wit NJ, Stalman WA, Knotternus JA, Hoes AW, Muris JW. Irritable bowel syndrome in primary care: the patients and doctors views on symptoms, etiology and management. Can J Gastroenterol 2003; 17: Longstreth GF, Burchette RJ. Family practitioners attitudes and knowledge about irritable bowel syndrome. Fam Pract 2003; 20: Knottnerus JA, van Weel C, Muris JWM. Evidence base of clinical diagnosis: evaluation of diagnostic procedures. BMJ 2002; 324: Murphy MK, Black NA, Lamping DL, McKee CM, Sanderson CFB, Askham J et al. Consensus development methods, and their use in guideline development. Health Technol Assess 1998; 2(3): i iv, Raine R, Sanderson C, Black N. Developing clinical guidelines: a challenge to current methods. BMJ 2005; 331: Burnand B, Vader JP, Froehlich F, Dupriez K, Larequi-Lauber T, Pache I et al. Reliability of panel-based guidelines for colonoscopy: an international comparison. Gastrointest endosc 1998; 47: Bersani G, Rossi A, Suzzi A, Ricci G, De Fabritiis G, Alvisi V. Comparison between the two systems to evaluate the appropriateness of endoscopy of the upper digestive tract. Am J Gastro 2004; 99: Longstreth GF, Thompson WC, Chey WD, Houghton LA, Mearin F, Spiller RC. Functional bowel disorders. Gastroenterol 2006; 130: Whitehead WE, Winget C, Fedoravicius AS, Wooley S, Blackwell B. Learned illness behavior in patients with irritable bowel syndrome and peptic ulcer. Dig Dis Sci 1982; 27(3): Jones J, Boorman J, Cann P, Forbes A, Gomborone J, Heaton K et al. British Society of Gastroenterology guidelines for the management of the irritable bowel syndrome. Gut 2000; 47: Dubois RW, Lim D, Hebert P, Sherwood M, Growe GH, Hardy JF et al. The development of indications for the preoperative use of recombinant erythropoietin. Can J Surg 1998; 41: Stovelaar HJ, McDonnell J, Stals H, Smets L. Belgian expert panel on Gastro-protective treatment in NSAID users. Gastro-protective treatment in patients using NSAIDs. Development of appropriateness criteria by a multidisciplinary expert panel. Scand J Rheumatol 2003; 32: