Predicting inpatient mortality in patients with inflammatory bowel disease: A machine learning approach.

BACKGROUND AND AIM: Data are lacking on predicting inpatient mortality (IM) in patients admitted for inflammatory bowel disease (IBD). IM is a critical outcome; however, difficulty in its prediction exists due to infrequent occurrence. We assessed IM predictors and developed a predictive model for IM using machine-learning (ML). METHODS: Using the National Inpatient Sample (NIS) database (2005-2017), we extracted adults admitted for IBD. After ML-guided predictor selection, we trained and internally validated multiple algorithms, targeting minimum sensitivity and positive likelihood ratio (+LR) ≥ 80% and ≥ 3, respectively. Diagnostic odds ratio (DOR) compared algorithm performance. The best performing algorithm was additionally trained and validated for an IBD-related surgery sub-cohort. External validation was done using NIS 2018. RESULTS: In 398 426 adult IBD admissions, IM was 0.32% overall, and 0.87% among the surgical cohort (n = 40 784). Increasing age, ulcerative colitis, IBD-related surgery, pneumonia, chronic lung disease, acute kidney injury, malnutrition, frailty, heart failure, blood transfusion, sepsis/septic shock and thromboembolism were associated with increased IM. The QLattice algorithm, provided the highest performance model (+LR: 3.2, 95% CI 3.0-3.3; area-under-curve [AUC]:0.87, 85% sensitivity, 73% specificity), distinguishing IM patients by 15.6-fold when comparing high to low-risk patients. The surgical cohort model (+LR: 8.5, AUC: 0.94, 85% sensitivity, 90% specificity), distinguished IM patients by 49-fold. Both models performed excellently in external validation. An online calculator (https://clinicalc.ai/im-ibd/) was developed allowing bedside model predictions. CONCLUSIONS: An online prediction-model calculator captured > 80% IM cases during IBD-related admissions, with high discriminatory effectiveness. This allows for risk stratification and provides a basis for assessing interventions to reduce mortality in high-risk patients.

Colonoscopy videos on YouTube: Are they a good source of patient education?

Background and study aims Colonoscopy is an effective tool to prevent colorectal cancer. Social media has emerged as a source of medical information for patients.YouTube (a video sharing website) is the most popular video informative source. Therefore, we aimed to assess the educational quality of colonoscopy videos available on YouTube. Methods We performed a YouTube search using the keyword "colonoscopy" yielded 429 videos, of which 255 met the inclusion criteria. Colonoscopy Data Quality Score (C-DQS) was created to rate the quality of the videos (-10 to +40 points) based on a colonoscopy education video available on the Ameican Society of Gastrointestinal Endoscopy (ASGE) website. Each video was scored by six blinded reviewers independently using C-DQS. The Global Quality Score (GQS) was used for score validation. The intraclass correlation coefficient (ICC) was used to assess the similarity of the scores among reviewers. Results Professional societies had the highest number of videos (44.3 %). Videos from professional societies (6.94) and media (6.87) had significantly higher mean C-DQS compared to those from alternative medicine providers (1.19), companies (1.16), and patients (2.60) ( P  < 0.05). Mean C-DQS score of videos from healthcare providers (4.40) was not statistically different than other sources. There was a high degree of agreement among reviewers for the videos from all sources (ICC = 0.934; P  < 0.001). Discussion YouTube videos are a poor source of information on colonoscopy. Professional societies and media are better sources of quality information for patient education on colonoscopy. The medical community may need to engage actively in enriching the quality of educational material available on YouTube.

Clinical efficacy and safety of mucosal incision-assisted biopsy for the diagnosis of upper gastrointestinal subepithelial tumors: A systematic review and meta-analysis.

BACKGROUND: Endoscopic ultrasound-guided fine-needle aspiration and biopsy (EUS-FNA/FNB) has been traditionally used for making a tissue diagnosis. Several newer techniques are emerging as a viable alternative to EUS-FNA/FNB, including mucosal incision-assisted biopsy (MIAB), with a view to increasing the diagnostic yield for upper gastrointestinal (GI) subepithelial tumors (SETs). We conducted a systematic review and meta-analysis to describe the overall diagnostic yield of MIAB for upper GI SETs. METHODS: Multiple electronic databases (MEDLINE, EMBASE and Google Scholar) and conference abstracts were comprehensively searched. The primary outcome of our meta-analysis was the overall diagnostic yield of the MIAB. The secondary outcome was to study complications in terms of perforation and clinically significant bleeding. The meta-analysis was performed using a DerSimonian and Laird random-effect model. RESULTS: Seven studies were included in the final meta-analysis, reporting a total of 159 patients (male 86, female 73) with a mean age of 58 years. The overall pooled diagnostic yield of MIAB was 89% (95% confidence interval [CI] 82.65-93.51, I 2=0.00). Histologically, GI stromal tumor was the reported diagnosis in 38.62% (95%CI 22.29-56.24, I 2=77.51%) of tumors, followed by leiomyoma 25% (95%CI 18.02-32.62, I 2=4.42%). The overall rate of clinically significant bleeding following the procedure was 5.03% (95%CI 0.36-12.86, I 2=57.43%) and no perforations were reported. CONCLUSIONS: MIAB is a safe and effective technique for the diagnosis of upper GI SETs and can be considered as a viable alternative to EUS-FNA/FNB. MIAB can be performed during routine endoscopy and no advanced equipment is required.

Opioid Use Disorder Increases 30-Day Readmission Risk in Inflammatory Bowel Disease Hospitalizations: a Nationwide Matched Analysis.

BACKGROUND AND AIMS: The opioid epidemic has become increasingly concerning, with the ever-increasing prescribing of opioid medications in recent years, especially in inflammatory bowel disease [IBD] patients with chronic pain. We aimed to isolate the effect of opioid use disorder [OUD] on 30-day readmission risk after an IBD-related hospitalization. METHODS: We retrospectively extracted IBD-related adult hospitalizations and 30-day, any-cause, readmissions from the National Readmissions Database [period 2010-2014]. OUD and 30-day readmission trends were calculated. Conventional and exact-matched [EM] logistic regression and time-to-event analyses were conducted among patients who did not undergo surgery during the index hospitalization, to estimate the effect of OUD on 30-day readmission risk. RESULTS: In total, 487 728 cases were identified: 6633 [1.4%] had documented OUD And 308 845 patients [63.3%] had Crohn's disease. Mean age was 44.8 ± 0.1 years, and 54.3% were women. Overall, 30-day readmission rate was 19.4% [n = 94,546], being higher in OUD patients [32.6% vs 19.2%; p < 0.001]. OUD cases have been increasing [1.1% to 1.7%; p-trend < 0.001], while 30-day readmission rates were stable [p-trend = 0.191]. In time-to-event EM analysis, OUD patients were 47% more likely (hazard ratio 1.47; 95% confidence interval [CI]:1.28-1.69; p < 0.001) to be readmitted, on average being readmitted 32% earlier [time ratio 0.68; 95% CI: 0.59-0.78; p < 0.001]. CONCLUSION: OUD prevalence has been increasing in hospitalized IBD patients from 2010 to 2014. On average, one in five patients will be readmitted within 30 days, with up to one in three among the OUD subgroup. OUD is significantly associated with increased 30-day readmission risk in IBD patients and further measures relating to closer post-discharge outpatient follow-up and pain management should be considered to minimize 30-day readmission risk.

Opioid use disorder in admissions for acute exacerbations of chronic pancreatitis and 30-day readmission risk: A nationwide matched analysis.

BACKGROUND: The opioid epidemic in the United States has been on the rise. Acute exacerbations of chronic pancreatitis (AECP) patients are at higher risk for Opioid Use Disorder (OUD). Evidence on OUD's impact on healthcare utilization, especially hospital re-admissions is scarce. We measured the impact of OUD on 30-day readmissions, in patients admitted with AECP from 2010 to 2014. METHODS: This is a retrospective cohort study which included patients with concurrently documented CP and acute pancreatitis as first two diagnoses, from the National Readmissions Database (NRD). Pancreatic cancer patients and those who left against medical advice were excluded. We compared the 30-day readmission risk between OUD-vs.-non-OUD, while adjusting for other confounders, using multivariable exact-matched [(EM); 18 confounders; n = 28,389] and non-EM regression/time-to-event analyses. RESULTS: 189,585 patients were identified. 6589 (3.5%) had OUD. Mean age was 48.7 years and 57.5% were men. Length-of-stay (4.4 vs 3.9 days) and mean index hospitalization costs ($10,251 vs. $9174) were significantly higher in OUD-compared to non-OUD-patients (p < 0.001). The overall mean 30-day readmission rate was 27.3% (n = 51,806; 35.3% in OUD vs. 27.0% in non-OUD; p < 0.001). OUD patients were 25% more likely to be re-admitted during a 30-day period (EM-HR: 1.25; 95%CI: 1.16-1.36; p < 0.001), Majority of readmissions were pancreas-related (60%), especially AP. OUD cases' aggregate readmissions costs were $23.3 ± 1.5 million USD (n = 2289). CONCLUSION: OUD contributes significantly to increased readmission risk in patients with AECP, with significant downstream healthcare costs. Measures against OUD in these patients, such as alternative pain-control therapies, may potentially alleviate such increase in health-care resource utilization.

Trends in Hospitalization and Mortality of Venous Thromboembolism in Hospitalized Patients With Colon Cancer and Their Outcomes: US Perspective.

INTRODUCTION: Venous thromboembolism (VTE) is a major cause of morbidity and mortality in hospitalized patients with colon cancer. We assessed nationwide population-based trends in rates of hospitalization and mortality from VTE among patients with colon cancer to determine its impact. METHODS: We queried the Nationwide Inpatient Sample (NIS) database entries from 2003 to 2011 to identify patients with colon cancer. Bivariate group comparisons between hospitalized patients with colon cancer with VTE to those without VTE were made. Multivariate logistic regression analysis was used to obtain adjusted odds ratios. The Cochrane-Armitage test for linear trend was used to assess occurrences of VTE and mortality rates among patients with colon cancer. RESULTS: The total number patients with colon cancer was 1,502,743, of which 41,394 (2.75%) had VTE. The median age of the study population was 69 years; 51.5% were women. After adjusting for potential confounders, compared with those without VTE, patients with colon cancer with VTE had significantly higher inpatient mortality (6.26% vs. 5.52%, OR 1.15, P < .001) and greater disability at discharge (OR 1.38, P < .001), but were not associated with longer length of stay (LOS) or cost of hospitalization. From 2003 to 2011, despite an increase in hospitalization rate with VTE in patients with colon cancer, their mortality steadily declined. CONCLUSION: VTE in hospitalized patients with colon cancer is associated with a significantly higher inpatient mortality and greater disability, but not with longer LOS or cost of hospitalization. Furthermore, even though there has been a trend toward more frequent hospitalizations in this patient population, their mortality continues to decline.