Society: AGA

Background
In the Dutch colorectal cancer (CRC) screening program, individuals aged 55 to 75 years are invited biennially for fecal immunochemical test (FIT)-based screening using a positivity cut-off of 47 microgram Hemoglobin per gram feces (µg Hb/g). The start age is higher than advised within several screening guidelines: the United States guideline advises to start screening at age 45; the European guideline advises to start screening at age 50. Being one option to extend the screening program, lowering the start age should be weighed against other possibilities to further optimize the program. We aimed to investigate the cost-effectiveness of several adjustments to the current Dutch CRC screening program.

Methods
The Microsimulation Screening Analysis model for CRC (MISCAN-Colon) was used to simulate CRC screening while varying four parameters: 1) screening interval between 1, 2 or 3 years, 2) FIT cut-off between 15, 20, 30, 40, 47, 50 or 60 µg Hb/g, 3) start age between 50, 52, 54, 55, 56, 58 or 60 years, and 4) stop age between 70, 72, 74, 75, 76, 78 or 80 years. This resulted in 1029 different screening strategies. For each of these strategies, quality-adjusted life-years gained (QALYs), costs and colonoscopy demand were calculated. Efficient screening strategies, i.e. strategies that yield the highest number of QALYs for a given cost level, were identified. In sensitivity analyses, colonoscopy capacity was restricted to the capacity required for the current strategy, as well as an additional 20% extra capacity. A willingness-to-pay threshold of €20,000 was assumed for all analyses.

Results
Without restrictions on colonoscopy capacity, only strategies with the lowest cut-off considered (i.e. 15 µg Hb/g) were efficient, and almost all efficient strategies had a stop age of 80 (Figure). The cheapest efficient strategies had a screening interval of three years and start age 60, whereas the most costly strategies had an interval of one year and start age 50. Costs and effects of efficient strategies ranged from 164 QALYs and cost-saving per 1,000 individuals to 281 QALYs at €240,403 per 1,000 individuals. The optimal screening strategy comprised annual screening at a cut-off of 15 µg Hb/g from age 50 through 80. When colonoscopy capacity was restricted to the capacity required for the current screening strategy, the current strategy (biennial screening from 55-75 years at cut-off 47 µg Hb/g) was optimal. With 20% extra colonoscopy capacity, triennial screening from 50-74 years at a cut-off 20 µg Hb/g was optimal.

Conclusion
From a cost-effectiveness perspective, increasing the stop age of CRC screening should precede over decreasing the start age. However, with limited colonoscopy capacity, it is better to lower the FIT cut-off and start age of screening, while increasing the screening interval.

Background: Diabetic gastroparesis (DGP) is a chronic disorder of the stomach characterized by delayed gastric emptying and foregut symptoms (e.g. nausea and vomiting). In 2020, the FDA approved the first metoclopramide (MCP) nasal outpatient treatment for patients with acute and recurrent DGP. We sought to compare the real-world healthcare resource utilization (HRU) among DGP patients treated with nasal MCP (NMCP) versus oral MCP (OMCP).
Methods: NMCP patients were retrospectively identified by linking specialty pharmacy data to the Symphony Integrated Dataverse®(SID) via Datavant. OMCP patients who had a diagnosis of DGP, initiated OMCP from 06/22/2020 – 12/31/2021, and were ≥18 years of age at treatment initiation were included. Patients were required to have ≥6 months pre-index (MCP dispense date) and ≥6 months post-index continuous capture. OMCP patients were matched to NMCP patients via propensity scores (“PS”, nearest neighbor) based on patient’s age, sex, region, payor, Charlson Comorbidity Score (CCI), and any 6-month hospitalization or emergency department (ED) visit pre-index. Mean, all-cause and separately DGP and symptom-related (nausea, vomiting) visits for physician office (PO), hospital outpatient (HO), inpatient hospitalization (IH), and ED were compared between cohorts during the 6 months following treatment initiation (i.e., index date). Additionally, total procedures performed in PO and HO settings were reported. Statistical significance was assessed using the Mann-Whitney test. A generalized linear model (GLM) with Poisson distribution and log-link was used to report incidence rate ratios (IRR) and 95% confidence intervals (CI) for all-cause HRU. RStudio was used for all statistical analyses.
Results: A total of 257 NMCP patients were PS-matched to 257 NMCP patients (out of 7,797 NMCP patients). Overall, patients mean age was 53.5 years, 77% were female, and over 60% were commercially insured. In the 6-month follow-up period, NMCP patients filled an average (standard deviation) of 2.1 (1.4) NMCP prescriptions and OMCP 2.1 (1.5) prescriptions. All-cause HRU are displayed in Table 1 and DGP and symptom-related are shown in Figure 1. NMCP patients were 36% less likely to experience hospital inpatient visits (IRR = 0.64 [95% CI: 0.47, 0.87]) and 61% less likely to experience ED (IRR=0.39 [95% CI: 0.31, 0.49]) compared to matched OMCP patients during the 6-month follow-up. In addition, DGP-symptom related HRU for PO, IP, and ED were significantly lower in the NMCP vs OMCP cohorts.
Conclusion: Patients treated with NMCP had a significantly lower all-cause PO, ED, and HO visits. Similarly, DGP symptom-related HRU was significantly lower for NMCP patients across all settings (except for HO) including IH.
Keywords: gastroparesis, metoclopramide, nasal, oral, healthcare resource utilization

Background: Nationwide organized gastric cancer (GC) screening programs have been running for decades in South Korea and Japan. Several individual-level and simulation studies indicated the effectiveness of such programs. However, few studies have analyzed the population impact based on observed data. We aimed to perform a quasi-experimental study and assess the population impact of organized GC screening on gastric cancer mortality in these two countries. Methods: South Korea and Japan started nationwide organized GC screening in 2002 and 1983, respectively. We estimated its effect on age-standardized mortality for GC, and other upper gastrointestinal (UGI) diseases (esophageal cancer and peptic ulcer) among people aged 40 years or above using a flexible synthetic control method (SCM). The concept of SCM is to create a synthetic control by deriving a weighted average of control countries without intervention. Weights are based on the outcome and covariates during the pre-treatment period. We compared post-treatment trends in outcome of South Korea and Japan with the counterfactual trend of synthetic controls and estimated average post-treatment rate ratios (RRs). The donor pools for South Korea and Japan comprised 38 and 23 control countries. We used World Health Organization mortality data from 1985 to 2017 and 1950 to 2017 for the analyses of South Korea and Japan. Covariates from the World Bank and the Global Burden of Diseases study were used in the analyses of South Korea. We conducted sensitivity analyses to test the robustness of our results, including: placebo test, negative control age band analysis, negative control outcome analysis, leave-one-out examination, and synthetic difference-in-difference analysis. Results: The pre-treatment fits of observed trends with synthetic control were acceptable for the analyses of South Korea and Japan’s GC mortality but poor for Japan’s other UGI disease mortality (Figure 1). The average post-treatment RRs and 95% confidence intervals (CIs) are 0.84 (CI 0.72-0.97) for GC mortality and 0.74 (CI 0.63-0.87) for other UGI disease mortality in South Korea. The effect on GC mortality increased over the post-treatment period, and the RR reached 0.71 in the 15th year after the program’s start. Results were robust to sensitivity analyses (Table 1). For Japan, the average RRs were 0.91 (CI 0.83-1.00) for GC mortality and 0.63 (CI 0.38-1.05) for other UGI disease mortality between 1983 and 2017. Sensitivity analyses indicated that the estimates for Japan might be biased. Conclusion: The benefits of nationwide GC screening programs were apparent in South Korea and promising in Japan. Our population-level analyses based on a quasi-experimental design complement the existing evidence and indicate the value of organized GC screening. Experiences from South Korea and Japan could inform other high GC-burden nations.

<b>Figure 1 Trends in age-standardized gastric cancer mortality in South Korea (A) and Japan (B) versus their synthetic control countries. The placebo tests for South Korea (C) and Japan (D).</b> The solid vertical line indicates the introduction of nationwide organized gastric cancer screening. The gray lines in C and D indicate the effects in control countries without nationwide screening. RR means post-treatment average rate ratios. CI means confidence intervals.

Figure 1 Trends in age-standardized gastric cancer mortality in South Korea (A) and Japan (B) versus their synthetic control countries. The placebo tests for South Korea (C) and Japan (D). The solid vertical line indicates the introduction of nationwide organized gastric cancer screening. The gray lines in C and D indicate the effects in control countries without nationwide screening. RR means post-treatment average rate ratios. CI means confidence intervals.

<b>Table 1 Sensitivity analyses for effects of nationwide gastric cancer screening on gastric cancer mortality in South Korea and Japan.</b>

Table 1 Sensitivity analyses for effects of nationwide gastric cancer screening on gastric cancer mortality in South Korea and Japan.

Introduction
Despite increased demand for gastroenterology services, in 2017 only 17.6% of gastroenterologists were women. On average, physicians receive approximately 19% of their income from treating Medicare patients. The aim of this study was to evaluate the total amount of CMS reimbursements for gastroenterologists and identify any differences between male and female providers after adjusting for multiple factors previously identified to affect differences in salary.

Materials and Methods
The CMS Physician and Other Supplier Public Use File (POSPUF) Database displays Part B claims organized by provider. For each unique NPI registered under “Gastroenterology” the total standardized amount reimbursed by Medicare was extracted for 2019. One univariate and one multivariate-adjusted linear regression model were used to analyze gender and CMS reimbursement. Reimbursement amounts greater than 97.5^th %ile and less than 2.5^th%ile were excluded. The multivariate model evaluated gender after adjusting for region, practice setting, number of services performed, average complexity and age of Medicare beneficiaries, and physician experience.

Results
For 2019, there were 2577 female and 11215 male gastroenterologists. Men had higher median total payments ($94416.28 vs $56014.14), higher median total services (1047 vs 633) and higher median unique HCPCS codes billed (39 vs 32). Men also had higher median years of experience since graduation (28 vs 17). Of the 13,792 gastroenterologists billing to Medicare in 2019, 13,500 (97.9%) had data regarding number of years of experience (2503 females, 10997 males). This subset was used for the linear regression models.

The univariate unadjusted model demonstrated that female gastroenterologists received less total CMS reimbursement than their male counterparts (log b = -0.42 [-0.46 to -0.39]). After adjusting for region, practice setting, number of services performed, average complexity and age of Medicare beneficiaries, and number of years of physician experience, female gastroenterologists still received less CMS payments (log b = -0.15 [-0.18 to -0.12]).

Conclusion
Much of the discussion regarding gender pay gaps can be subjective. However, even after adjustment for multiple factors, female gastroenterologists are receiving less CMS payments, which can comprise a significant portion of their annual income. Further objective data is warranted to provide a more accurate understanding of reimbursement inequity and help drive change from a national level to address gender-pay gaps.

Background: Outcome switching refers to differences found between outcomes reported in a randomized controlled trial (RCT) protocol or registration and in its final publication. Such changes may allow for the preferential reporting of positive or favorable results as primary outcomes. Previously, researchers have found that outcome switching is highly prevalent among RCTs. However, this phenomenon has not been investigated in RCTs for inflammatory bowel disease (IBD) drug therapies.

Aim: To determine the prevalence of outcome switching in RCTs for IBD drug therapies.

Methods: We performed a search on clinicaltrials.gov for all phase 3 RCTs pertaining to IBD drug therapies that had results published. Trials were excluded if results for multiple trials were reported as a single pooled analysis, or if the full-text trial publication with results could not be found. Full-text publications were identified through clinicaltrials.gov or by hand-search with confirmation using the trial registration number.

Three authors independently extracted data. Pre-specified primary and secondary outcomes on clinicaltrials.gov were collected from the version dated immediately prior to recruitment, and reported outcomes were extracted from the manuscript publications. Differences in registration and publication outcomes were recorded, and additional outcomes found in publications were extracted as well. Data was collected on whether trials were industry sponsored. Our primary outcome was the number of studies that reported primary outcomes that differed from pre-specified primary outcomes.

Results: We identified 104 trials, of which 72 (69%) were included for data extraction. Twenty-four (33%) studies reported primary outcomes that differed from those that were pre-specified in the protocol. Nineteen (26%) of these had introduced a new primary outcome, and 5 (7%) had primary outcomes that were previously pre-specified secondary outcomes. Seventy (97%) studies added additional primary or secondary outcomes in their corresponding publications. Two (3%) of the studies had no outcome switching of any kind.

Ninety-six (96%) trials had industry sponsors. Both trials which were free of any outcome switching were not industry sponsored.

Conclusion: Our results show that outcome switching among IBD RCTs is highly prevalent. Furthermore, all trials that had no outcome switching or new added outcomes had non-industry sponsors. Improvements in outcome reporting are therefore required to prevent bias from being introduced into trial results, especially if the results are used to inform clinical decisions.

Background: Telemedicine is not consistently superior to standard care in the management of inflammatory bowel disease (IBD). Non-inferiority is an acceptable outcome if telemedicine improves the efficacy and efficiency of care. Owing to the positive results of the TECCU app (Telemonitoring of Crohn’s Disease and Ulcerative Colitis) in a pilot trial¹, we aimed to evaluate the time in remission and quality of life (QoL) of IBD patients controlled by telemonitoring (G_TECCU), compared to standard care (G_Control) after 12 weeks.
Methods: A 2-arm randomized multicenter study with a non-inferiority design was performed in 29 IBD centers in Spain. Adult IBD patients who initiated therapy with immunosuppressant or biological agents for disease activity were included. Exclusion criteria were: patients with ileorectal/ileo-pouch anal anastomosis, stoma, active perianal disease, no Internet access. Time in remission was evaluated with Harvey-Bradshaw/Walmsley indexes (according to Crohn's disease [CD] / ulcerative colitis [UC], respectively), fecal calprotectin (FC) and reactive C protein (RCP). QoL was assessed with IBDQ-9 and medication adherence with Morisky-Green index and patient satisfaction.
Results: We included 157 patients, and 126 were analyzed after 12 weeks of follow-up. The demographic and clinical variables are listed in Table 1. The time in remission was not inferior in patients who used the TECCU app (mean 4.2 weeks [SD 3.8]) compared with patients who received standard care (mean 4.2 weeks [SD 3.2]; difference 0.03 [95%CI -1.21 to 1.27]; p=0.017) (Figure 1). In patients with UC, disease activity improved significantly in both TECCU (mean SCCAI improvement -3.60 [SD 3.25]; p= 0.001) and standard care groups (-3.89 [SD 4.56]; p=0.001;). In patients with CD, clinical activity also improved significantly in both groups (mean Harvey-Bradshaw improvement -1.65 [SD 2.51]; p=0.001 in TECCU; -2.24 [SD 4.15]; p=0.001 in standard care group). Similarly, FC and CRP values improved significantly in both groups. Considering QoL, the IBDQ-9 score improved significantly in TECCU (mean 11.29 [SD15.8]; p<0.001) and standard care groups (16.59 [SD 22.1]; p<0.001). Medication adherence improved significantly in TECCU group and it was superior compared with standard care. Patient satisfaction was superior to 90% in both groups at week 12.
Conclusions: In IBD patients who initiate biological agents or immunomodulators, TECCU app is not inferior to standard care to maintain remission in the short-term. Telemonitoring with TECCU app associated a higher improvement in medication adherence and an increase in QoL.
¹Del Hoyo J et al. A Web-Based Telemanagement System for Improving Disease Activity and Quality of Life in Patients with Complex Inflammatory Bowel Disease: Pilot Randomized Controlled Trial. J Med Internet Res. 2018 Nov 27;20(11):e11602.

<b>Table 1. Demographic and clinical variables</b>

Table 1. Demographic and clinical variables

<b>Figure 1. Time in remission until 12 weeks</b>

Figure 1. Time in remission until 12 weeks

Francisco Mesonero

Appearances