Inflammatory bowel disease: epidemiology and management in an English general practice population


Prof. G. P.Rubin Benedict Building, St George’s Way, Sunderland, SR2 7BW, UK. E-mail:



Inflammatory bowel diseases have significant long-term morbidity and healthcare resource consequences. Studies based on secondary care records may have underestimated the contribution of general practitioners (GPs) to its management.


To describe the epidemiology and management of inflammatory bowel disease using GP records as the primary data source.


A systematic search of GP clinical records in northern England, identifying cases of inflammatory bowel disease, patient consultation behaviour, prescribing patterns, and extent of specialist care.


In a population of 135 723, the incidence of ulcerative colitis was 13.9/100 000 per year (CI: 7.5–20.3) and for Crohn’s disease 8.3/100 000 per year (CI: 3.4–13.2). The age–sex adjusted point prevalence for ulcerative colitis on 1st January 1995 was 243.4/100 000 (CI: 217.4–269.4) and for Crohn’s disease 144.8/100 000 (CI: 124.8–168.8). The mean number of consultations (s.d.) with specialists and GPs were similar, both in the first 12 months after referral (specialists 3.94 ± 3.15, GPs 3.34 ± 3.55) and in the most recent 12 months (1.02 ± 2.02, 1.04 ± 2.04). Only 29.9% of all patients were definitely under specialist care.


Prevalence rates, but not incidence rates, for inflammatory bowel disease are substantially higher than previously described in UK populations. General practitioners make a significant contribution to meeting the healthcare needs of these patients.


The non-specific inflammatory bowel diseases, ulcerative colitis and Crohn’s disease, are chronic relapsing disorders. They are important because they result in long-term morbidity and place significant demands upon healthcare resources. Although generally perceived as uncommon conditions largely managed by specialists, there is some evidence that GPs may play a significant role in the long-term care of patients with inflammatory bowel disease.1

Many epidemiological studies of inflammatory bowel disease have used hospital records and databases to identify cases.2[3][4][5][6]–7 These may miss patients who have been lost to follow-up, who have not been hospitalised, or who have received hospital treatment for inflammatory bowel disease in another area.8, 9 Some studies have used regional or national morbidity registers based on in-patient diagnoses in an attempt to overcome these problems.10[11]–12 Others have used a supplementary approach to general practitioners (GPs) in the study area, asking them to search their disease registers for cases, to report new cases or to refer possible new cases to the study centre.9, 13[14][15][16][17]–18

The clinical records of UK GPs are recognized as a potentially rich source of epidemiological and morbidity data.19 Over 95% of the population in the UK is registered with a named GP who provides primary healthcare and acts as a gatekeeper to specialist care. As a result GP records contain comprehensive and cumulative information on patient healthcare for precisely defined populations. Their completeness and accuracy has improved greatly with computerization.20, 21 No epidemiological study of inflammatory bowel disease has used these records as the primary source for case identification. Our aim was to describe the epidemiology and management of inflammatory bowel disease in one health district using GP records as the primary data source.


Population and study area

The study was conducted in North Tees health district, comprising Stockton-on-Tees and its surrounding localities. This is a geographically defined, predominantly urban area with a population ethnicity which is 98.39% white, 1.01% Asian and 0.6% other.22 The 24 general practices located in this area had a total population of 179 496 permanently resident patients. Practices that did not routinely record morbidity data on a computerized clinical database were excluded from the study. Fifteen practices fulfilled the study requirements, representing a population of 135 723 or 75.6% of the total. The average list size for these practices was 9077. The study area is served by a single District General Hospital, which has a specialist gastroenterology unit.

Case finding

All practices based in the study area were invited to participate in the study. Non-responders were followed up by a second written invitation and personal contact by the author. Case finding used predefined read codes to systematically search computer diagnostic and prescribing records. These codes covered all the likely diagnostic terms applicable to inflammatory bowel disease, and those drugs that might be used in its treatment. Possible cases were then verified by scrutiny of the paper records. Random case reviews were made by GPR and APSH, looking for observer error, with negative results. Data were double entered into the EPI-Info statistical programme.23 Incidence and prevalence rates were calculated using standard methodologies. Confidence intervals were calculated using standard techniques and reported where appropriate.24

Case definition

The strategy for case definition was based on that described by Garland.25 A specialist diagnosis of ulcerative colitis, Crohn’s disease or indeterminate inflammatory bowel disease, based on clinical, investigative or histological findings, was a prerequisite to inclusion in the study. Patients with secondary (ischaemic, radiation) colitis, infective colitis and apparent acute self-limiting colitis were excluded from the study.

Confirmatory evidence for the diagnosis was sought in the form of histological, endoscopic, radiological and operative findings, at least one of which had to support the clinical diagnosis and be present in the GP records. This could be in the form of an original report or contained within the specialist’s correspondence with the GP. Diagnoses made on purely clinical grounds were excluded from the study.

Extent of disease

Ulcerative colitis was defined as idiopathic colitis confined either to the rectum (proctitis), not extending beyond the mid-transverse colon (left-sided) or involving the entire colon (extensive). The extent of disease was defined as the greatest extent recorded at any stage in the course of the illness by colonoscopy or radiography. Crohn’s Disease was categorized as ileal, small bowel, large bowel and rectal. Involvement at one or more of these sites at any time was recorded. A diagnosis of indeterminate inflammatory bowel disease was assigned to those patients for whom no subsequent information was available to allow specific categorization.


Only those GP consultations with a record relating to inflammatory bowel disease were included. Patients were considered to be under specialist care if they had attended out-patient appointments in the previous 12 months and were considered to be discharged if the most recent letter received from the responsible specialist stated this. Current drug therapy was defined as the issue of a GP prescription within the previous 6 months.


We identified 568 patients with inflammatory bowel disease in a population of 135 723, giving a crude prevalence for inflammatory bowel disease of 418.5/100 000 (CI: 384.2–452.8). The ratio of ulcerative colitis:Crohn’s disease was 1.67:1. There were 334 patients with ulcerative colitis; 170 males, (mean age 53.2 years, median 53 years), and 164 females (mean age 50.7 years, median 48 years); ratio male : female, 1:1.04). There were 200 patients with Crohn’s disease; 83 males (mean age 45.3 years, median 44 years), and 117 females (mean age 51.5 years, median age 49 years); ratio male : female 1:1.41. There were 34 patients with non-specific inflammatory bowel disease; 12 males and 22 females (mean age 51.9 years, median 48 years).

A confirmatory histology report was available in the GP records for 287 (85.9%) of those patients with ulcerative colitis, and for 159 (79.5%) of patients with Crohn’s disease. For the remainder, confirmatory evidence came from radiological, colonoscopic or surgical findings present in the records.

A further 45 patients, 30 with ulcerative colitis and 15 with Crohn’s disease, were excluded from the study because confirmatory evidence of investigations was absent from the GP records.

Incidence of inflammatory bowel disease

Between 1985 and 1994 there were 179 new cases of ulcerative colitis, 104 new cases of Crohn’s disease and 24 cases of indeterminate inflammatory bowel disease. The incidence of ulcerative colitis in the years 1990–1994 was 13.9/100 000 per year (CI: 7.5–20.3), and for Crohn’s disease 8.3/100 000 per year (CI: 3.4–13.2) (Table 1).

Table 1.  Incidence and age–sex adjusted point prevalence of inflammatory bowel disease on 1st January 1995Thumbnail image of

Prevalence of inflammatory bowel disease

The age–sex adjusted point prevalence on 1st January 1995 for ulcerative colitis was 243.4/100 000 (CI: 217.4–269.4), for Crohn’s disease 144.8/100 000 (CI: 124.8–168.8) and for indeterminate inflammatory bowel disease 24.0/100 000 (CI: 18.0–30.0) (Table 1). Prevalence was independent of year or place of diagnosis. Ulcerative colitis was most commonly confined to the rectum (150, 44.9%), although 70 patients (21.0%) had extensive colitis (Table 2). The peak age of onset was 20–29 years (interquartile range 28–53 years). There were no significant differences in the extent of ulcerative colitis by sex (χ2 1.55, 2 d.f., P=0.46) or age of onset (one way analysis of variance P=0.81), although the interquartile range for patients with total colitis was slightly lower at 25–49 years.

Table 2.  Extents of disease and mean age at diagnosis for inflammatory bowel diseaseThumbnail image of

Crohn’s disease was most commonly present in the colon (115, 57.5%) or ileum (98, 49.0%). The peak age of onset was 20–29 years (interquartile range 28–49 years). There were no significant differences in the extent of disease between the sexes, except that large bowel Crohn’s disease was significantly less common in men than women (P < 0.001, odds ratio 0.31; CI: 0.16–0.58). The interquartile range for age of onset was lower in those patients with ileal and small bowel disease.

Patient management

The clinical care of patients with inflammatory bowel disease was almost equally shared between gastroenterologists and GPs. In the first 12 months after diagnosis there was a mean (s.d.) of 3.94 (± 3.15) specialist consultations, together with 3.34 (± 3.55) GP consultations specifically related to inflammatory bowel disease. Those patients whose diagnosis had been made 2 years or more prior to the study date had a mean of 1.02 (± 2.02) specialist and 1.04 (± 2.04) GP consultations in the previous 12 months (Table 3).

Table 3.  Specialist and GP consultations in the first year after referral and in the most recent yearThumbnail image of

Only a minority of patients were clearly under continuing specialist care. A total of 185 out of 329 patients with ulcerative colitis, and 86 out of 97 with Crohn’s disease, had definitely been discharged from hospital follow-up. (Table 4). Patients definitely discharged from follow-up were more likely not to be taking drug therapy than those under continuing care (190 out of 293 vs. 54 out of 170, P=0.0001). Nevertheless, in the former group, seven patients were receiving prescriptions for oral steroids and six for rectal steroids. The most common drug therapy was an oral salicylate, although 161 (48.2%) patients with ulcerative colitis and 121 (60.5%) with Crohn’s disease were not receiving therapy, and 32 (9.6%) and 17 (8.5%), respectively, were on dual therapy (Table 5).

Table 4.  Extent of provision of continuing specialist careThumbnail image of
Table 5.  Inflammatory bowel disease patients in general practice: current drug therapyThumbnail image of

Seventy patients had extensive colitis, of whom 34 had a history of total colectomy. Of the remainder, 24 were on drug therapy, although only 13 were definitely under specialist care. Overall, 81 (14.3%) of the patients in this study had undergone total colectomy. Among patients with Crohn’s disease, 125 (62.5%) had undergone at least one surgical resection; in 36 patients, (18.0%) this was a total colectomy.


This is the first study into the epidemiology and management of inflammatory bowel disease to use GP records as the primary data source. We found that inflammatory bowel disease is more prevalent in this population than previously described and that its management is shared equally between specialists and GPs.

In the UK, GP records are comprehensive, cumulative and, because they follow the patient throughout their life, usually contain the entire medical history. This contrasts with hospital based records which remain in a single centre and relate primarily to the care provided there. The UK system of patient registration also means that general practice populations can be precisely defined.

Although the clinical information held on computerized GP records is known to correlate closely with external sources of patient information, we identified several potential sources of error in the course of this study.26

First, identification of subjects relied on a prescribing or diagnostic entry in the computerized clinical record. Because many diagnostic entries are made in response to the need for drug prescription, patients who have been in remission for several years were less likely to be represented. GPs also vary in the extent to which they enter clinical diagnoses in the computerized medical record. Furthermore, some clinical record systems cannot be searched for patients who have previously been treated for inflammatory bowel disease but are no longer being treated.

Second, the quality of clinical information contained in letters to GPs from hospital specialists has improved over time. In some instances it was only possible to infer from the correspondence that an investigation had been carried out, the results not being provided to the GP.

Lastly, the true size of the population served by a general practice is subject to inaccuracies that can arise during the process of addition and removal of patients from the list. Using population figures for Cleveland based on 1991 census data, this resulted in ‘list size inflation’ of 3.6%.22

These factors would all result in failure to identify affected patients, particularly those with longer-established disease, or those in remission and not requiring treatment. The outcome would be an underestimate of disease prevalence.

Ulcerative colitis

Because of the retrospective nature of this study, and the quality of computerized record keeping prior to 1988, the derivation of annual incidence rates from this data becomes unreliable with increasing time since diagnosis. The age adjusted incidence for ulcerative colitis for the period 1990–1994 was 13.9/100 000 per year. This is comparable to previously reported rates for other northern European populations and for that reported in 1980 for the same geographical area as this study.7 Most recently, an incidence rate of 24.5/100 000 has been reported in Iceland.27 Ekbom found an increased incidence of ulcerative colitis from 7 to 12/100 000 between 1965 and 1983, which was mainly accounted for by an increase in the incidence of proctitis.12 Tysk described similar findings but considered this to be an artefact due to improved recording of diagnoses.10 We found the incidence of proctitis for the period 1990–1994 to be 7.2/100 000 per year (50.5% of all incident cases of ulcerative colitis). This is an increase from 4.7/100 000 per year (31.1% of all incident cases) between 1970 and 1977 for the same area.7

The age–sex adjusted prevalence for ulcerative colitis was 243.4/100 000. This is considerably higher than the figure of 99/100 000 reported for the same area in 1977, and for other European studies, but comparable with that recently reported in Minnesota.7, 10 28, 29 This high prevalence may be due better case ascertainment, although Langholz considered increasing prevalence and stable incidence to be an expected finding for a disease with low mortality and a good prognosis.28

The case mix by extent of disease is comparable with that found in other recent studies, and we conclude that our high prevalence does not result from the discovery of a cohort of patients with mild disease who have been lost to specialist care. Specifically, the prevalence of proctitis is comparable to other studies.

Crohn’s disease

The age adjusted incidence of Crohn’s disease for the period 1990–1994 was 8.3/100 000 per year. This is comparable to rates for similar populations in northern Europe and to a figure of 5.3 for the same geographical area reported for 1971–1977.5, 6, 12, 14, 15 The trend for Crohn’s disease incidence to increase between the 1950s and the 1970s and stabilize thereafter has been explained in the UK by the increased availability of specialist gastroenterology units, increased awareness of Crohn’s disease, recognition of Crohn’s colitis and transfer from ulcerative colitis.30 All of these factors apply to the area under study.

The prevalence of Crohn’s disease was 144.8/100 000. This is comparable to figures reported for Scotland but considerably higher than those for other areas of England and Wales or for Scandinavia.5, 6 The interpretation of these differences in the extent of Crohn’s disease is limited by differences between individual studies in their classification. However, we found a higher percentage (57.5%) of patients to have large bowel involvement.

Patient management

In recent years a shared care approach to the management of patients with inflammatory bowel disease has been advocated.31 GPs have expressed a wish to share care with the support of guidelines and consensus guidelines which have been developed in collaboration with gastroenterologists in order to support this process.32, 33 In a study of routine vs. open access follow-up hospital care, patients with inactive or mild stable inflammatory bowel disease had a mean of 7.73 (routine) to 9.23 (open access) all cause GP surgery visits over 24 months, roughly twice the number of out-patient visits for the same period.27 Our findings are broadly in keeping with this and indicate that the reality of shared care is likely to precede the development of a jointly agreed management strategy. That 16 out of 70 patients with extensive colitis were lost to specialist follow-up is cause for concern, since these would be candidates for surveillance colonoscopy. Also of concern are the 13 patients on oral or rectal steroids who were not under specialist supervision.

Our results demonstrate that inflammatory bowel disease is substantially more prevalent than previously stated, and that general practice makes a significant contribution to meeting its demands on healthcare resources. The study highlights the need for GP training in the management of inflammatory bowel disease and the direction of resources to primary care. Programmes of care which make use of open access to specialists in order to avoid duplication and unnecessary routine consultation, but which ensure continued follow-up for those who need it, should be developed and evaluated.


The authors wish to acknowledge the support given in data collection and analysis by the Cleveland Medical Audit Advisory Group and the Northern Primary Care Research Network (NoReN). The study received limited funding from SmithKline Beecham Pharmaceuticals.