Main findings

Total IBD hospital admission rates have more than doubled from 2003–2004 to 2012–2013 driven by substantial rises in elective admissions particularly for lower GI endoscopy and cytokine inhibitor infusions. While medical admissions have risen steeply, surgical admissions have plateaued with a significant reduction in the rate of emergency and elective surgery for CD and a reduction in the rate of elective surgery for UC over the decade despite an increasing incidence and severity of disease phenotype in European populations.4 5 29 Mean length of stay has decreased for both elective and emergency admissions. These trends suggest a shift from reactive emergency management of IBD towards more planned day case based activity. There has also been a reduction in elective lower GI endoscopy for CD and UC as a proportion of all IBD admissions over the last decade. Emergency endoscopy rates fell significantly for UC but not CD over the study period. The rise in cytokine inhibitor infusions (infliximab) reflects the emergence and increasing use of these agents over the period initially for CD and more recently for UC. These findings reveal important changes in the approach to the management of IBD over the past decade with a shift to shorter, elective, day case medical admissions.

Findings in relation to other studies

Hospital admission rates for IBD have largely been reported to be rising in Europe and North America for some years but are not consistent.30–34 All these studies predate our findings which analysed a period during which there has been significant changes in the availability of newer medications. We found overall IBD admissions more than doubled in the decade between 2003–2004 and 2012–2013. This is striking compared with previous studies that found much smaller increases in admission in the nineties.9 Overall, with the advent of national screening measures, lower GI endoscopy rates have previously been reported to be rising with an indication that this is associated with reduction in mortality at the population level.35 36 Regarding lower GI endoscopy rates for patients with IBD, there is a lack of published literature. We found a significant fall in the proportion of lower GI endoscopy rates for both emergency and elective UC admissions and for elective CD admissions over the 10-year period. This may reflect improved disease control as a result of medical advancements with a reduction in disease flares requiring endoscopic evaluation. These findings suggest that the overall increase in demand for endoscopic services over recent years is a result of the increasing incidence and prevalence of IBD.

In the USA, there have been three analyses using the National inpatient sample (NIS), in many ways comparable to HES, with all these studies indicating a rise in admissions.32–34 Most recently, non-surgical admission rates increased by 68% and 52% for CD and UC, respectively, between 2004 and 2007.32 However, a study from Northern California, from the health maintenance organisation Kaiser Permanente, which has arguably similar constraints on healthcare provision to the NHS, found a 33% and 29% decrease in hospital admission rates for CD and UC, respectively, between 1998 and 2005 but importantly the study did not include day case activity.31. An Irish study, of both day cases and inpatients, also reported an overall rise in national IBD admission rates between 1996 and 2001.31 The rises are consistent with a general trend of rising hospital admissions in developed countries,37 among an ageing population with increasingly complex multimorbidity. Superimposed on this trend is rising incidence of IBD including a proportionally greater rise in children and young people with IBD among whom the condition has a more severe phenotype.2 4 5 38 An additional explanation for rises in admissions may be improved ascertainment over the study period. Finally, the rises may reflect changing clinical practice, in particular, the advent of biologic drugs that may have driven the shift to shorter more frequent elective day case activity.

We examined surgery rates between 2003-2004 and 2012-2013, as a proportion of all IBD admissions and found a significant reduction in CD surgery rates, both perianal and major surgery, for emergency and elective admissions. UC elective surgery rates also fell significantly but there was no significant reduction in emergency surgery. An analysis of rates of surgery in the USA between 1998 and 2012 showed a similar decrease in admissions for CD-related and UC-related bowel surgery of 4.9% and 3.3% with the greater proportion related to a reduction in emergency surgery.32 39

The proportion of day case activity rose and there was an overall reduction in inpatient length of stay suggesting patients now have more frequent but shorter hospital spells. In keeping with our findings, day case admissions have also risen substantially in Ireland.31 This may, in part, reflect the greater use of day case admissions for anticytokine infusions in recent years. The greater rise in elective admission rates compared with emergency admissions may also reflect a more general shift in chronic disease management in the UK national health service and elsewhere directed towards better disease control through outpatient care and planned elective procedures replacing emergency activity.40–42 We found length of stay for IBD emergency admissions fell from 9.2 to 6.8 days and 10.8 to 7.6 days for CD and UC during the study period depending on the admission type. In the USA, over the same time period, average length of stay for patients with either UC or CD has also fallen by a similar magnitude.32 34

The observed trends in our study and others are reflected in a meta-analysis of studies from Europe, North America and Asia that found a significant drop in IBD-related surgery over the past five decades.6 A national Danish population-based cohort also report a decreasing requirement for IBD surgery over the past 30 years.43 A more recent analysis from 2002 to 2010 in Canada also showed an overall reduction in surgical resection rates with elective operations being more commonly performed than emergency surgery.44 Studies show the increasing and earlier use of thiopurines and maintenance cytokine inhibitor therapy is associated with a reduced need for surgery in CD.16 45 We postulate that as biological therapies become increasingly established, we will see further evidence of a fall in rates of surgery in IBD in the future.14–16 46–49

We found an increase in the use of infusions largely related to anticytokine infusions (infliximab) in England during the study period. Our study is the first nationwide study to characterise temporal trends in infliximab use. The rising trend reflects the increasing use of infliximab in the latter half of the study period although maintenance therapy for CD was not finally approved for use in the UK until towards the end of the study period.11 The use of cytokine inhibitor agents as a maintenance treatment in the management of UC was not approved until after the end of the study period in 2015.13 However, the use of infliximab, as a second-line option, for the induction of remission in acute severe colitis was approved in the UK in 2008.12 This difference in the availability of anti-TNFs for the treatment of UC compared with CD is reflected in the much smaller number of cytokine inhibitor infusions for UC. Our findings mirror the interim UK biologics audit, a study reporting data derived from the majority of the nation’s IBD units, indicated a relatively modest uptake of cytokine inhibitor use in 2010 but temporal trends were not reported.50 51

Accumulating evidence suggests that long-term cytokine inhibitor treatment may significantly reduce the need for surgery in CD,45 52 53 although the evidence in UC is possibly less compelling.54 55 Data from Canada, Hungary and the UK show that surgical rates were already falling before biological therapy came into widespread use.14 15 46 47 We and others have reported data supporting an important role for thiopurines in reducing the long-term need for surgery.16

Study strengths and limitations

HES maintains a comprehensive data set of all admissions to NHS hospitals in England and has previously been used in the field of IBD research.56 57 Our national study includes all hospitals in England, and its size and representativeness mean that it overcomes potential biases of previous regional referral centre cohort studies.56 However, in common with other observational studies using large administrative databases, the quality of the coding of diagnoses and surgery relies on data entry performed locally at hospital level with the potential for misclassification. The precision and depth of coding may have changed over the decade of study and might, at least in part, account for the observed rise in admissions.58 Furthermore, some procedures may not have been coded with the relevant accompanying IBD code. We have attempted to improve the ascertainment of IBD-related admissions by including IBD admissions coded as either primary or secondary diagnoses and, in the case of the latter, admissions were only included where the primary diagnosis was considered relevant to IBD by a clinical steering group. Unfortunately, HES does not provide data for adalimumab which is a self-administered subcutaneous injection that does not require admission. Finally, we did not detail trends in specific surgical procedures but we have distinguished major abdominal (intestinal) surgery from perianal surgery in CD which is arguably the most clinically relevant categorisation. HES is an administrative dataset so causal relations between trends in admissions and changes in IBD management cannot be drawn merely observed associations.

Implications and future research

Our findings, indicating a shift towards more planned and shorter elective inpatient care is a considerable improvement for patients who previously required long hospital admissions. The general fall in endoscopy and surgery rates coupled with a rise in cytokine inhibitor use may suggest improved disease control. The shift from emergency to elective activity with shorter length of stay will have reduced costs substantially. That being said, cost savings are likely offset by the increasing expense of biologics as underscored by the Dutch IBD healthcare cost analysis COIN study as well as the rising incidence and prevalence of IBD.59 Further health economic studies, particularly with regard to avoidable emergency care, would be enlightening.

Our findings address a gap in information identified by the quality agenda set by the published UK IBD standards60 61 and provide valuable information for future health service planning. The fact that surgery rates are falling despite the rising incidence and prevalence of this chronic disorder suggests that there have been improvements in the quality of care and efficacy of treatment provided to patients with IBD but further work is needed to elucidate this further. It is probable that recent advances in the management of IBD including the substantial increase in the use of thiopurines and cytokine inhibitor therapy may have impacted on long-term outcomes in this condition and with it a reduction in long-stay admissions. Further, IBD time trends analysis in the forthcoming decade will shed further light on the impact of these agents.