OBJECTIVES
There are many published clinical guidelines for acute pancreatitis (AP). Implementation of these recommendations is variable. We hypothesized that a clinical decision support (CDS) tool would change clinician behavior and shorten hospital length of stay (LOS).
METHODS
Design/Setting: Observational study, entitled, The AP Early Response (TAPER) Project. Tertiary center emergency department (ED) and hospital. Participants: Two consecutive samplings of patients having ICD-9 code (577.0) for AP were generated from the emergency department (ED) or hospital admissions. Diagnosis of AP was based on conventional Atlanta criteria. The Pre-TAPER-CDS-Tool group (5/30/06–6/22/07) had 110 patients presenting to the ED with AP per 976 ICD-9 (577.0) codes and the Post-TAPER-CDS-Tool group (5/30/06–6/22/07) had 113 per 907 ICD-9 codes (7/14/10–5/5/11). Intervention: The TAPER-CDS-Tool, developed 12/2008–7/14/2010, is a combined early, automated paging-alert system, which text pages ED clinicians about a patient with AP and an intuitive web-based point-of-care instrument, consisting of seven early management recommendations.
RESULTS
The pre- vs. post-TAPER-CDS-Tool groups had similar baseline characteristics. The post-TAPER-CDS-Tool group met two management goals more frequently than the pre-TAPER-CDS-Tool group: risk stratification (P < 0.0001) and intravenous fluids > 6L/1st 0–24 h (P =0.0003). Mean (s.d.) hospital LOS was significantly shorter in the post-TAPER-CDS-Tool group (4.6 (3.1) vs. 6.7 (7.0) days, P= 0.0126). Multivariate analysis identified four independent variables for hospital LOS: the TAPER-CDS-Tool associated with shorter LOS (P =0.0049) and three variables associated with longer LOS: Japanese severity score (P =0.0361), persistent organ failure (P =0.0088), and local pancreatic complications (< 0.0001).
CONCLUSIONS
The TAPER-CDS-Tool is associated with changed clinician behavior and shortened hospital LOS, which has significant financial implications.
Purpose To identify computed tomographic (CT) findings that are predictive of recurrence of colonic diverticulitis. Materials and Methods Institutional review board approval was obtained for this HIPAA-compliant, retrospective cohort study. Six abdominal fellowship-trained radiologists reviewed the CT studies of 440 consecutive subjects diagnosed with acute colonic diverticulitis between January 2004 and May 2008 to determine the involved segments, maximum wall thickness in the inflamed segment, severity of diverticulosis, presence of complications (abscess, fistula, stricture, or perforation), and severity of the inflammation. Electronic medical records were reviewed for a 5-year period after the patients' first CT study to determine clinical outcomes. Predictors of diverticulitis recurrence were assessed with univariate and multiple Cox proportional hazard regression models. Results Colonic diverticulitis most commonly involved the rectosigmoid (70%, 309 of 440) and descending (30%, 133 of 440) colon segments. Complicated diverticulitis was present in 22% (98 of 440) of patients. On the basis of the results of univariate analysis, significant predictors of diverticulitis recurrence were determined to be maximum colonic wall thickness in the inflamed segment (hazard ratio [HR], 1.07 per every millimeter of increase in wall thickness; P< .001), presence of a complication (HR, 1.75; P = .002), and subjective severity of inflammation (HR, 1.36 for every increase in severity category; P value for linear trend = .003). The difference in maximum wall thickness in the inflamed segment (HR, 1.05 per millimeter; P = .016) and subjective inflammation severity (HR, 1.29 per category; P = .018)remained statistically significant in a Cox multiple regression model. Conclusion Maximum colonic wall thickness and subjective severity of acute diverticulitis allow prediction of recurrent diverticulitis and may be useful for stratifying patients according to the need for elective partial colectomy. RSNA, 2017 Online supplemental material is available for this article.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.