We conducted a retrospective cohort study on adult patients hospitalized for acute severe ulcerative colitis to assess the impact of extended monitoring on corticosteroid 30-day readmission rates and transition failures. Several studies have also looked at factors associated with readmission in ulcerative colitis patients [9,10,11], our study found that extended monitoring on corticosteroids was not associated with a decreased likelihood in 30-day readmission. In addition, our study found a low rate (8%) of oral steroid transition failure overall, suggesting that most patients will do well after transitioning to oral steroids. Patients who were deemed clinically appropriate to forego extended inpatient monitoring had particularly low rates of oral corticosteroid failure (3%). Extended inpatient monitoring for ≥ 24 h on oral steroids was not associated with a decreased likelihood of transition failure or 30-day readmission. In fact, we found that extended inpatient monitoring was paradoxically associated with an increased likelihood of oral steroid transition failure.

There are many potential explanations for this. For one, a prior study showed that patients with clinical markers of more severe inflammatory bowel disease (i.e., tachycardia, elevated CRP, or severe Mayo scores at the time of admission) were associated with longer hospitalizations [12]. Therefore patients who were clinically selected for extended inpatient monitoring by their provider team were more likely to experience oral steroid transition failure and trialed on oral corticosteroids 1.5 days later than the accelerated inpatient monitoring group. Moreover, there was a trend toward more need for rescue therapy in the extended monitoring group, 73 vs 60% (p = 0.12) and less bio-naïve patients, 33 vs 46% (p = 0.10) which may reflect that biologic-experienced patients are more likely to start inpatient rescue therapy due to more refractory or difficult to control disease, however, this was a statistically insignificant finding.

Due to the small overall number of patients who experienced oral steroid transition failure, we lacked statistical power to identify the predictors driving these phenomena. The association between extended monitoring and transition failure may be the result of astute clinician decision-making, wherein patients deemed at higher risk for clinical decompensation were preferentially observed for a longer period of time after transitioning to oral steroids. Indeed, the observation that patients selected for extended monitoring had a tendency to spend more time on IV steroids supports this hypothesis.

The association between extended monitoring and transition failure may also be confounded by the fact that prolonged length of stay increases a hospitalized patient’s risk of adverse events such as hospital-acquired infections, deconditioning, poor nutrition, adverse drug effects, and venous thromboembolism [13, 14]. An alternative explanation is that individual clinician practice styles, rather than objective patient disease severity, contribute to the timing of steroid transition and subsequent duration of observation. Clinicians with a more “conservative” practice style may keep patients on IV steroids for longer and also watch patients on oral steroids for longer prior to discharge. It has also been demonstrated that patients admitted to tertiary medical centers are more likely to undergo more extensive work-up and treatment when compared to community sites, thus contributing to prolonged hospitalizations [15]. Our study did not survey clinicians regarding their practice style, and this is a topic for future investigation.

Healthcare expenditures continue to rise at unsustainable rates; over the past two decades, the cost of caring for IBD has nearly doubled, in part from the increased use of expensive but effective biologic agents in the maintenance of disease [16, 17]. IBD patients incur annual costs that are three-fold higher than non-IBD patients ($22,987 compared to $6956) [7]. This financial burden impacts both the health system and the individual, as IBD patients experience paying twice the out-of-pocket costs of a non-IBD patient, in addition to likely lost wages and elevated insurance premiums [7, 18].

In light of this, our study aimed to contribute to high-value care of an increasingly expensive disease and improve hospital-wide quality outcomes by reducing unnecessarily prolonged lengths of stay. Understandably, some patients admitted for ASUC would benefit from extended inpatient monitoring post-transition to oral steroids. However, given our study’s finding of overall low rates of transition failures, we argue that such a practice should be selective rather than routine, as it appears that most patients do not benefit from extended monitoring. Improving how we select patients for prolonged inpatient monitoring on oral steroids would improve cost-effective care [19] by reducing wasteful overtreatment in patients who would otherwise fare well with a shorter monitoring duration.

This study had several limitations. Although we discovered that subjects in the extended monitoring group were more likely to experience transition failure, our study was not able to meaningfully identify specific clinical predictors of steroid failure, likely reflecting insufficient power rather than a true lack of such predictors. Therefore, a multi-center study would be able to better identify clinical predictors of steroid transition failures. Another limitation of this study is its retrospective approach. In a prospective cohort study, patients would ideally be assigned a priori or randomized to either extended or accelerated monitoring groups. As a result, transition failures in the extended monitoring group may potentially be exaggerated as patients with more severe illness during their hospitalization are more likely to have an extended hospitalization as a result. Moreover, patients in this cohort study were cared for by a heterogeneous group of both hospital medicine and gastroenterology providers, whose own intrinsic practices around steroid transitions and discharge may have varied. Future studies should assess clinician practice styles as a potential novel target for quality improvement interventions.

In summary, patients with ASUC who were transitioned to oral steroids had low rates of inpatient failures and 30-day readmissions. Accelerated (< 24 h) inpatient monitoring after transitioning to oral steroids appears safe in carefully selected patients, and can potentially be cost-saving without worsening clinical outcomes. This precedent can be further explored and iteratively investigated through quality improvement projects aimed at reducing prolonged hospitalization and overtreatment. Further studies are needed to identify clinical predictors of steroid failure that would warrant extended monitoring.