Article Text

Download PDFPDF

Evidence or eminence: optimising patient outcomes by choosing the right doctor?
  1. Peter Laszlo Lakatos1,
  2. Gabor Veres2
  1. 11st Department of Medicine, Semmelweis University, Budapest, Hungary
  2. 21st Department of Paediatrics, Semmelweis University, Budapest, Hungary
  1. Correspondence to Dr Peter Laszlo Lakatos, 1st Department of Medicine, Semmelweis University, Koranyi st. 2/A, Budapest H-1083, Hungary; lakatos.peter_laszlo{at}

Statistics from

Ulcerative colitis (UC) is a chronic inflammatory disease of the large bowel, characterised by a relapsing-remitting course, frequently requiring hospitalisation.1 Despite significant advances in therapy over the past decades, a significant proportion of the patients still come ultimately to colectomy. Relatively few data (around 25%) are available on the hospitalisation rates in UC; however, it appears that despite advances in treatment algorithms, hospitalisation and colectomy rates have not decreased in the last decade in UC. In general, hospitalisation rates are two–threefold higher compared with the general population.2 Moreover, stable hospitalisation rates have been reported from the US by using 1990 to 2003 National Hospital Discharge Survey data (8.2–12.4 per 100 000 people)1 and in a population-based cohort from Canada between 1994 and 2001 (12.6–13.3 per 100 000 people), with major surgery being an important reason for hospitalisation in the latter study (55%).3 By contrast, increasing hospitalisation rates have been reported in another study done in the US by analysing the nationwide inpatient sample data between 1998 and 2004,4 with an annual 3% increase in hospitalisation rates, although surgical rates remained stable.

The reasons for hospitalisation in UC may be several, including infectious diseases, but these patients typically have acute severe colitis (ASC). ASC is a significant clinical condition affecting about 20% of UC patients during the disease course. Overall, the management of these patients is challenging. Murthy et al report an important study of the impact of gastroenterology specialist care compared with care by other providers on the mortality and colectomy rates of hospitalised UC patients.5 The findings of their study are fascinating—and not a little reassuring to specialists. Patients with UC admitted under non-gastroenterologists had a higher in-hospital mortality rate (1.1% vs 0.2%) compared with care by a gastroenterologist from admission. Increased mortality risk was observed in patients admitted to both, internists (OR 5.49) and general practitioners (OR 6.02), with a trend towards greater mortality risk among patients admitted under general surgeons (OR 3.49), after controlling for age and comorbidity burden. Moreover, the risk of death remained twofold greater over the year following discharge, and consultation with a gastroenterologist did not decrease this increased mortality risk. This matters. As might be expected, the presence of comorbidities was identified as an independent predictor of adverse outcome. However, due to several reasons, a significant proportion of these patients were looked after by a variety of healthcare providers, with substantially different levels of knowledge of inflammatory bowel disease patient care in everyday clinical practice. By contrast, in-hospital rate, 1-year colectomy rate and re-hospitalisation, were not affected by the specialty of inhospital physician provider. The consequence of the findings is dramatic, since it may result in recommendations that future care of hospitalised UC patients should be undertaken specifically by gastroenterologists or specialists with expertise in the management of patients with inflammatory bowel diseases.

Although the mortality in ASC has decreased from 7% after the introduction of steroids in 1955 to <1% in specialist centres today,6 hospitalisation rates and need for colectomy has remained broadly stable, with appreciable geographic variation, approaching as high as 30–60% in patients with ASC during longer term follow-up. Unfortunately, there are only a limited number of confirmed factors for the prediction of disease course in UC. Certainly, age at diagnosis (especially paediatric UC), disease extent, smoking status and presence of comorbidities have to be considered. A more recently recognised risk factor is hospitalisation. In a paper from USA, medical hospitalisation for the management of UC (OR 5.37, 95% CI 2.00 to 14.46) and the need for infliximab therapy (OR 3.12, 95% CI 1.21 to 8.07) were independent predictors of colectomy in a multivariate analysis.7 Another important predictor is mucosal healing. Apart from being part of activity indices, mucosal healing was associated with a lower subsequent risk of colectomy in the population-based study by the IBSEN group.8 In ASC, the monitoring of the patients should be far tighter. The condition of the patient has to be monitored frequently on at least a daily basis, including the appropriate assessment of the above variables, physical signs, disease activity, laboratory data, imaging and endoscopy results, where available. Based on the landmark data from Oxford, an elevated C reactive protein >45 mg/l or more than eight stools per day on day 3 following hospital admission for severe colitis, may necessitate rescue therapy and in case of failure, is highly predictive of the need for urgent colectomy. In a more recent publication by the same group, a bloody stool frequency more than six per day at index admission and just one or more of the additional Truelove and Witts' criteria resulted in a colectomy rate of 8.5%, while this was 31%, 45% and 48% if two, three or four additional criteria were present.9 Similarly, colectomy rates increased with subsequent admissions, from 19.9% on the first admission to 29–38% on second or subsequent admission. However, colectomy saves lives. Notably, in the study by Murthy et al, 5 authors were unable to adjust colectomy and mortality rates on disease-severity scores, although colectomy rates were adjusted for prior hospitalisation, hospital type and volume, and number of acute care beds. In addition, the lack of difference in in-hospital or postdischarge colectomy risks among patients admitted to different in-hospital providers suggests that colectomy risk is likely to be more influenced by patient and disease-related factors (including age and prior hospitalisation). It also suggests that colectomy is generally indicated based on the above Oxford criteria or the Swedish index and that evaluation of these data are less affected by the primary care physician if they consult with a gastroenterologist. By contrast, in Crohn's disease, where the indication for surgery may be more complex, specialist gastroenterology care within the first year following a diagnosis of Crohn's disease has been associated with decreased long-term surgery rates.10

Can we, therefore, better optimise health outcomes in UC by using specialist care? Despite the low overall mortality rate (<1%), the lack of ability to control for treatment options and possible confounding effect of hospital setting or better access to consultation with other specialists, this study emphasises that in defined clinical scenarios, such as hospitalised patients with UC, patients benefit from specialist gastroenterology care from the start. In such patients, especially those with comorbid conditions, intense vigilance and critical analysis of clinical, laboratory and imaging data is of utmost importance for the prompt recognition of disease-related complications or side effects of medical therapy, ultimately leading to decreased mortality rates. This is, after all, the outcome that matters most. In other words, the combination of advanced, up-to-date, evidence-based knowledge accompanied by long term, high-level clinical experience is needed in complex, hard-to-treat patient populations with UC, to facilitate better outcomes and is best provided by gastroenterology specialists. This conclusion is not necessarily generalisable to populations with milder UC, but the finding that physician speciality-related differences are associated with diverse health outcomes for complex medical conditions is not novel. It has been established for a variety of disorders in a range of healthcare disciplines. By contrast, the necessity of colectomy is less affected by the type of the primary care provider. It is probable that initial and rescue therapies in hospitalized patients with UC can be optimised with regard to avoiding colectomy as long as there is adequate consultation with gastroenterology specialists, but that complications leading to death can best be avoided by early specialist decision-making.


View Abstract


  • Linked article 301978.R1.

  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles