Gastrointestinal Unit, Royal Hampshire County Hospital, Winchester, UK
Received date: 9 October 2003; Accepted date: 18 November 2003
Iron-deficiency anaemia (IDA) is a very common disorder. In 1998, an analysis of historical data revealed that 20 patients harbouring gastrointestinal cancer who were initially referred with IDA had an average wait of 11 weeks from referral to diagnosis.Recent government recommendations have brought this subject to the fore and are driving change in an already pressured service by insisting that IDA should be included in the indications for urgent referral as part of the two-week wait rule. With the paucity of diagnostic symptoms and clinical signs seen in most patients with IDA, a wait for consultant outpatient review prior to endoscopic evaluation seemed an unnecessary delay. In December 1998 a fast track protocoldirected nurse-led IDA clinic was established and here we evaluate the results of the first consecutive 100 patients. All 100 patients referred over 15 months were seen within two weeks. A nurse specialist followed an established protocol to elicit a structured medical history and undertake a limited physical examination. The presence of alarm symptoms or signs led to rapid investigation within two weeks. Nineteen patients were found to have gastrointestinal cancer (17 colonic, two oesophageal). In these 19 patients the mean time from general practitioner referral letter to diagnosis was four weeks and 79% had a diagnosis within three weeks. In conclusion, this nurse-led open access IDA clinic attracted appropriate referrals from primary care and greatly improved the referral-to-diagnosis time for patients harbouring gastrointestinal malignancy. In all cases it was apparent that a clinical nurse specialist was entirely capable of safely assessing these patients and planning appropriate investigations.
anaemia, colonic neoplasm, iron deficiency, nurse practitioner
Iron deficiency is the commonest cause of anaemia in the United Kingdom. Estimates suggest that 3–5% of men and post-menopausal women have irondeficiency anaemia (IDA) with the level rising to 7% in elderly hospital inpatients.[2–4] Unless there is a clinically obvious cause, occult gastrointestinal bleeding should always be considered, with studies demonstrating gastrointestinal abnormalities in 50– 90% of IDA patients.[2,5–9] The incidence of carcinoma (oesophageal, gastric or colonic) in these studies is consistently reported to be greater than 10% and even as high as 50%. This has been influential, with government recommendations aimed at improving the diagnosis of colorectal cancer suggesting that IDA should be included in the indications for urgent referral as part of the two-week wait rule.
Traditionally IDA patients have been referred to a number of different specialities including surgeons, gastroenterologists, haematologists and geriatricians. Waiting times for clinic appointments and the urgency of subsequent investigation are likely to have varied considerably. In 1998, an analysis of historical data from this hospital revealed that 20 consecutive patients specifically harbouring gastrointestinal cancer, who were initially referred with IDA, had an average wait of 11 weeks (median eight weeks) from referral to diagnosis. With the relative paucity of diagnostic symptoms and clinical signs found in patients presenting with IDA, a wait for consultant outpatient review prior to endoscopic evaluation seemed an unnecessary delay. In December 1998, in collaboration with consultant colleagues who frequently saw IDA patients, we established a fast track protocolbased nurse-led IDA clinic. This clinic had open access to the local primary physicians and was widely advertised. This paper reports the findings of the first 100 referred patients.
In October 1998 the local primary physicians received notification of the new IDA clinic and were encouraged to use the facility. Between December 1998 and March 2000, 100 new cases of IDA were referred for evaluation. All patients were assessed by a clinical nurse specialist using a specially developed protocol focusing on the relevant medical history and a basic clinical examination (see Appendix 1). The nurse had attended the consultant clinic and observed and been taught basic examination for a three-month period. Structured questioning was used to identify alarm symptoms. The clinic was run in parallel with the gastroenterology outpatient clinic and all patients were discussed with the consultant or specialist registrar (SpR) and a management plan decided upon.
Conformation of IDA was obtained (haemoglobin [Hb] < 13.0 g/dL and ferritin < 30 g for males, Hb < 11.5 g/dL and ferritin < 14g for females with a mean corpuscular volume [MCV]< 80 F1) and subsequent management was based on the recently published British Society of Gastroenterology guidelines.10 If any alarm symptoms or signs were elicited (e.g. dysphagia, weight loss, abdominal mass) then the aim was to perform appropriate investigations within two weeks. For other patients with no such features, investigations were booked within the following six weeks. Provision was made in the endoscopy department for two extra cases per week. If a likely source of significant bleeding other than the gastrointestinal tract was apparent (e.g. menorrhagia), then the patient was referred to the appropriate speciality. Most asymptomatic patients were booked for outpatient combined oesophagogastroduodenoscopy (OGD) and colonoscopy at the same session. A minority of patients with significant co-morbidity who were felt to be at risk from colonoscopy had a barium enema instead. Patients with definite localising symptoms to the upper or lower gastrointestinal tract (e.g. dysphagia, altered bowel habit) had that area investigated first. Endoscopic findings taken to be significant causes of anaemia were grade 3 or 4 oesophagitis, portal hypertensive gastropathy, haemorrhagic gastritis, chronic peptic ulceration, coeliac disease, colonic polyps >1 cm, colonic angiodysplasia and carcinoma of the oesophagus, stomach and colon.
Between one and four patientswere referred each week and all 100 consecutive patients were seen within one to two weeks of receiving the referral letter. There was generally spare capacity in the clinic and it was envisaged that up to six patients could have been seen and investigated without a major impact on endoscopy provision.
The median age was 66 years (range 29–88) with a female predominance (63% female vs 37% male) and a mean haemoglobin of 9.4 mmol/l (range 5.7–12.0). The outcome for these 100 patients and the investigations performed are shown in Figure 1. Two patients were admitted as emergencies prior to their outpatient investigations. The first had a haematemesis and at OGD a large duodenal ulcer was found. The second was admitted with abdominal pain and had an inpatient OGD and colonoscopy. Both procedures were normal and subsequent investigations revealed an acute promyelocytic leukaemia. Two elderly patients had such significant co-morbidity that they were too unfit for further investigation. Following discussion and against our recommendation five patients declined further investigation. One patient was referred to the ear, nose and throat surgeons for treatment of recurrent epistaxis, and eight young females with no gastrointestinal symptoms were referred for gynaecological assessment of possible menorrhagia. Six of these patients were felt to have menorrhagia sufficient to account for their anaemia and the other two returned for OGD and colonoscopy.
At the initial assessment, 15 patients were thought to have alarm symptoms and were investigated ‘urgently’ within two weeks. The other patients had initial investigations within six weeks with an overall mean time from referral to diagnosis of five weeks. In no patients were relevant clinical signs elicited at the initial nurse examination or subsequently by the attendant physician in those cases with malignancy.
Results of the investigations are shown in Table 1. Of the 86 patients who had gastrointestinal investigations, 45 (52%) had significant abnormalities detected. Nineteen cancers were found, and all patients thought initially to have alarm symptoms were found to have neoplastic disease. For these patients the mean time from referral letter to diagnosis was four weeks (range 2–8) with 79% diagnosed within three weeks. It is noteworthy that of the 17 patients with colonic cancer, seven patients had Dukes stage D disease. Only one patient had both an upper and lower gastrointestinal cause for their IDA (gastritis and colonic cancer).
It has been known for many years that IDA is a marker for gastrointestinal malignancy and should prompt urgent assessment of iron-deficient patients. The establishment of a rapid access nurse-led clinic enabled the assessment of IDA patients within two weeks of referral and more importantly resulted in the majority of cases (80%) of malignant disease being diagnosed within three weeks. This clearly compares very favourably with the historical data from this hospital when the mean time from referral with IDA to diagnosis of cancer was 11 weeks. Retrospective analysis of data is fraught with the problems of case selection and bias. However, comparison with historical cancer cases is probably more reliable then a comparison with retrospective endoscopy patients, who undoubtably would represent a selected population.We were unable to find data in the literature looking specifically at waiting times for IDA, although these may become available as the two-week wait recommendations are audited.
There is the potential for delay in the diagnosis of neoplasia in IDA patients, as the importance placed on investigation of such patients seems to vary enormously. This hospital serves a population of 235 000 and examination of the haematology computer database revealed that over a year 1651 individuals (993 females, 658 males) had results suggestive of IDA (low Hb and lowMCV) and yet only 100 patients were referred to this clinic over 15 months. The clinic was new at that time and referral numbers have increased since then, but it still seems likely that many patients with IDA are managed solely in the community. Recent published data suggest that 50% of IDA patients in the community have either no investigations at all or just testing for faecal occult blood.11 Furthermore, faecal occult blood testing in IDA patients may have a sensitivity of only 50% for detecting significant pathology compared to endoscopic evaluation.8,12 It is possible that general practitioners only refer patients to hospital who they perceive to be at high risk of malignancy due to their symptoms or clinical signs. However, 21% of the patients withmalignancy detected in this study were entirely asymptomatic, and several studies have documented a poor correlation between symptoms and endoscopic findings.5,9 Despite this documented poor correlation, alarm symptoms did appear to be a very reliable predictor of malignancy in this patient cohort.
Delay in diagnosis may also result from lack of centralisation of referrals, with patients referred to different specialities each with variable clinic waiting times and variable access to investigation. This study suggests that if referral patterns are centralised then prompt investigation can be ensured. The costs of this clinic are difficult to ascertain exactly, as it was run as an adjunct to the high-throughput parallel consultant clinic. Minimal receptionist and clerical support were required and the only major cost was the nurse specialist’s time (£11/h, H grade of payscale).
It could be argued that all IDA patients should be sent directly for investigation without an initial assessment. However, the advantage of an outpatient review are several-fold: 1) investigations can be targeted (e.g. small bowel studies if the history suggests Crohn’s disease); 2) investigations can be instigated urgently if alarm symptoms are present. Conversely, patients with long-standing but not previously investigated anaemia could have more leisurely investigations; 3) patients with other causes of IDA can be directed to the appropriate specialist; 4) investigations can be discussed and the consent process begun and 5) patients’ co-morbidity can be assessed to establish the most appropriate investigation. We found that a clinical nurse specialist using a specially developed protocol was entirely capable of assessing these patients and planning their investigations. Certainly patient satisfaction appeared to be high, probably in part because of the time available for the nurse to explain the investigations and to begin the consent process (average consultation 30 min). Because the overall number of patients requiring investigation was small (<2/week), and only a minority needed investigation within two weeks, the impact on endoscopy provision was minimal.
The findings at investigation of this group of IDA patients were similar to those of previous series.[2,5–9] In the 86 patients investigated, significant abnormalities were found in 52%, with malignancy in 22% and coeliac disease in 6%. A degree of variation in the frequency of abnormalities reported in studies may represent differences in the patient cohort, criteria taken for a significant positive finding and the method of investigation employed. Barium enema, for example, can miss small polyps or angiodysplasia, but it appears to be comparable to colonoscopy at demonstrating colonic carcinoma in IDA patients. All patients having an OGD had duodenal biopsies taken to look for coeliac disease and five cases had typical histological features. It is interesting to note that the endomysial antibody test was positive in only two of these patients. This apparent low sensitivity of the endomysial antibody test is at variance with most published studies, although in partial villous atrophy the sensitivity may be only 30%.[14,15] It is possible that the low endomysial antibody sensitivity may have related to a problemwith the local assay, and screening for coeliac disease is now done using the tissue transglutaminase antibody.
In this study a nurse-led open access protocol-based IDA clinic greatly improved the referral-to-diagnosis time for patients harbouring gastrointestinal malignancy. It is hoped that initiatives towards earlier cancer diagnosis and treatment will result in enhanced long-term survival. Certainly, a reduced waiting time for investigation will help to allay many patients’ anxiety regarding their possible occult malignancy and allownon-malignant conditions (e.g. peptic ulceration, severe oesophagitis) to be treated before complications ensue. This clinic was a safe efficient clinical pathway that centralised the referral of patients from primary care and freed consultant clinic space. As recent recommendations include IDA as an indication for urgent referral (i.e. within two weeks), the development of this type of clinic may be one useful approach to service provision.