A Digital Health Approach to Improve Compliance With Surveillance Colonoscopy Guidelines: The SCOPES Program

Surveillance Guidelines Should Be Updated to Recognize the Importance of Serrated Polyps

Differentiating surveillance from non-surveillance colonoscopy for colorectal cancer in patients with inflammatory bowel disease (IBD) using electronic medical records (EMR) is important for practice improvement and research purposes, but diagnosis code algorithms are lacking. The automated retrieval console (ARC) is natural language processing (NLP)-based software that allows text-based document-level classification. The purpose of this study was to test the feasibility and accuracy of ARC in identifying surveillance and non-surveillance colonoscopy in IBD using EMR. We performed a split validation study of electronic reports of colonoscopy pathology for patients with IBD from the Michael E. DeBakey VA Medical Center. A gastroenterologist manually classified pathology reports as either derived from surveillance or non-surveillance colonoscopy. Pathology reports were randomly split into two sets: 70 % for algorithm derivation and 30 % for validation. An ARC generated classification model was applied to the validation set of pathology reports. The performance of the model was compared with manual classification for surveillance and non-surveillance colonoscopy. A total of 575 colonoscopy pathology reports were available on 195 IBD patients, of which 400 reports were designated as training and 175 as testing sets. Within the testing set, a total of 69 pathology reports were classified as surveillance by manual review, whereas the ARC model classified 66 reports as surveillance for a recall of 0.77, precision of 0.80, and specificity of 0.88. ARC was able to identify surveillance colonoscopy for IBD without customized software programming. NLP-based document-level classification may be used to differentiate surveillance from non-surveillance colonoscopy in IBD.

Colorectal cancer is the second leading cause of cancer-related deaths in the United States. Colorectal cancer risk can be effectively managed through early detection and removal of precancerous lesions, known as colorectal polyps, with routine colonoscopy screening. The current guidelines for colonoscopy screening and surveillance do not consider detailed clinical information and polyp characteristics from prior colonoscopies. Developing a colonoscopy surveillance plan based upon a patient’s personalized polyp recurrence risk is important for preventing the progression of colorectal cancer. To address this clinical need, in this paper, we proposed and developed a natural language processing and machine learning model to predict colorectal polyp recurrence risk using features derived from patient colonoscopy and pathology reports in electronic medical record systems. Colonoscopy records and the associated pathology reports from 952 patients in a tertiary academic care center in New Hampshire were obtained from 2011 to 2017. Polyp characteristics were extracted from these records using a natural language processing pipeline. The extracted features from these records along with other demographic and anthropometric information were used to develop and compare six machine learning models for their ability to predict polyp recurrence. Our evaluation of these models revealed a range of performance advantages, such as an area under the curve as high as 65%, and it further highlighted important features in predicting polyp recurrence from demographic and medical health record sources. Our predictive analysis highlights the potential of personalized risk modeling for colorectal cancer screening, which can reduce unnecessary screenings, healthcare costs, and psychological stress, while improving patient health outcomes.

The US Multi-Society Task Force on Colorectal Cancer published guidelines for colonoscopy screening and surveillance in 2008 and affirmed them in 2012. Characteristics associated with guideline adherence among US gastroenterologists have not been assessed. Assess awareness and adherence of US gastroenterologists with national guidelines for colonoscopy screening and surveillance and predictors of adherence to guidelines. A Web-based survey was administered to gastroenterologists in various practice settings across the USA. A total of 306 gastroenterologists completed the survey; 86 % reported awareness of the guidelines. Low-volume colonoscopists (<20/month) were less likely to be aware of the guidelines (OR 0.26, p = 0.03) compared to high-volume colonoscopists (>100/month). Those completing training before 1990 were less likely to report following guidelines (OR 0.37, p = 0.01). Adherence with guidelines was then assessed via clinical scenarios. Compared to physicians finishing training in 1991-2010, less adherence was seen in those finishing before 1990 (OR 0.75, p < 0.001) or currently in training (OR 0.72, p = 0.004). Compared to the Western USA, less adherence was seen in the Midwest (OR 0.69, p = 0.001), Northeast (OR 0.63, p < 0.001), and South (OR 0.59, p < 0.001). Lower adherence was seen among non-academic physicians (OR 0.72, p = 0.001) and low-volume colonoscopists (OR 0.52, p < 0.001). There is poor adherence with colonoscopy screening and surveillance guidelines among US gastroenterologists. Poor adherence was associated with being in training or finishing training before 1990, practicing in the South, non-academic settings, and low colonoscopy volume. These findings can target interventions for quality improvement in colorectal cancer screening and surveillance.

Rapid growth in adoption of electronic health records (EHRs) has led to an unprecedented expansion in the availability of large longitudinal datasets. Large initiatives such as the Electronic Medical Records and Genomics (eMERGE) Network, the Patient-Centered Outcomes Research Network (PCORNet), and the Observational Health Data Science and Informatics (OHDSI) consortium, have been established and have reported successful applications of secondary use of EHRs in clinical research and practice. In these applications, natural language processing (NLP) technologies have played a crucial role as much of detailed patient information in EHRs is embedded in narrative clinical documents. Meanwhile, a number of clinical NLP systems, such as MedLEE, MetaMap/MetaMap Lite, cTAKES, and MedTagger have been developed and utilized to extract useful information from diverse types of clinical text, such as clinical notes, radiology reports, and pathology reports. Success stories in applying these tools have been reported widely. Despite the demonstrated success of NLP in the clinical domain, methodologies and tools developed for the clinical NLP are still underknown and underutilized by students and experts in the general NLP domain, mainly due to the limited exposure to EHR data. Through this tutorial, we would like to introduce NLP methodologies and tools developed in the clinical domain, and showcase the real-world NLP applications in clinical research and practice at Mayo Clinic (the No. 1 national hospital ranked by the U.S. News & World Report) and the University of Minnesota (the No. 41 best global universities ranked by the U.S. News & World Report). We will review NLP techniques in solving clinical problems and facilitating clinical research, the state-of-the art clinical NLP tools, and share collaboration experience with clinicians, as well as publicly available EHR data and medical resources, and finally conclude the tutorial with vast opportunities and challenges of clinical NLP. The tutorial will provide an overview of clinical backgrounds, and does not presume knowledge in medicine or health care. The goal of this tutorial is to encourage NLP researchers in the general domain (as opposed to the specialized clinical domain) to contribute to this burgeoning area. In this tutorial, we will first present an overview of clinical NLP. We will then dive into two subareas of clinical NLP in clinical research, including big data infrastructure for large-scale clinical NLP and advances of NLP in clinical research, and two subareas in clinical practice, including clinical information extraction and patient cohort retrieval using EHRs. Around 70% of the tutorial will review clinical problems, cutting-edge methodologies, and real-world clinical NLP tools while another 30% introduce use cases at Mayo Clinic and the University of Minnesota. Finally, we will conclude the tutorial with challenges and opportunities in this rapidly developing domain.

<h3>Introduction</h3> On recommencement of surveillance after the first wave of the Covid-19 pandemic, the BSG guidelines for colonoscopy surveillance had been updated in October 2020. This created an opportunity to reassess all 1243 colonoscopy surveillance patients by developing a new surveillance nurse led pathway at Wye Valley NHS Trust Hereford (WVT). <h3>Methods</h3> A new Band 7 Non-medical Endoscopist was appointed on transition from Bowel Scope to BCSP and set up a new assessment process to review surveillance patients against the new BSG criteria. Patients who no longer met the new guidelines were written to explaining the change in the BSG criteria. Patients still meeting the surveillance criteria were invited to a nurse led clinic where suitability for colonoscopy was assessed, specific instructions regarding bowel preparation given, renal function assessed and arrangements made for anticoagulation ± diabetic control. Clinics were either face to face or via telephone according to the patient’s wishes. <h3>Results</h3> 1243 patients on the initial colonoscopy surveillance waiting list underwent initial validation and 39% of patients were discharged as they no longer met the surveillance criteria. A further 8% were discharged for other reasons (moved away, new health concerns). After clinic assessment, only 34% of the original surveillance cohort were suitable for surveillance colonoscopy, 3% were not fit enough for any surveillance, 2% declined any colonic investigations, 2% agreed to CTC, 2% DNAd, 1% had moved away and 4% were Covid anxious and requested a delay in surveillance. 5% were upgraded to C2WW due to new symptoms. With a combination of reducing demand and increasing capacity back to pre-Covid levels, surveillance waiting times have reduced from 12 months to 6 weeks. The discharges have released 151 colonoscopy lists (758 colonoscopies) to the C2WW and screening programmes. Using NHS tariff for colonoscopy of £372, these 758 discharges achieved a financial benefit of £281,976. The surveillance clinic cost has been 2 Band 7 nursing clinics per week over an 8 month period annualised to £8960 pa. <h3>Conclusions</h3> Nurse led clinics can thoroughly assess colonoscopy surveillance patients against the new BSG guidelines. This substantially reduces demand, reduces waiting times and generates a cost benefit. Quality is also enhanced by specific information of the procedure and anticoagulation issues. This has reduced surveillance waiting times down to JAG recommendations within 8 months. Our new surveillance process has released a total of 151 colonoscopy lists by demonstrating only 34% of the initial cohort of patients remained suitable for colonoscopy surveillance.

Several studies have highlighted poor compliance with surveillance colonoscopy guidelines. The National Health and Medical Research Council (NHMRC) guidelines were revised in 2018 and were more complex than the previous iteration (2011). The aim of this study was to determine the impact of 2018 NHMRC polyp surveillance guidelines on compliance with colonoscopy surveillance intervals. A multicentre retrospective clinical audit was conducted between January 2020 and February 2021. Patients awaiting a colonoscopy for polyp surveillance at two public tertiary care hospitals in South Australia were included. Compliance rates of recommended polyp surveillance colonoscopy intervals after implementation of 2018 NHMRC guidelines were compared with 2011 NHMRC guidelines. The projected impact on colonoscopy bookings of the change in guideline intervals was modelled to 5 and 10 years, factoring in differences in compliance. Of 3996 patients awaiting colonoscopy services at two public hospitals in South Australia, 1984 patients (60% male, median age 61 years) were waitlisted for polyp surveillance. Overall compliance with surveillance guidelines was >60%. Implementation of the 2018 NHMRC guidelines significantly reduced compliance from 65.8% (2011 guidelines) to 50.8% (2018) (χ2 <0.001, OR 0.5). Modelling projections to 5 and 10 years demonstrated that application of the 2018 guidelines significantly increases the projected number of colonoscopy bookings per year. The revised 2018 NHMRC guidelines have resulted in significantly poorer compliance post-implementation, possibly due to their increased complexity. This has potential to increase the surveillance colonoscopy waiting list burden.

The Colonoscopy Screening and Surveillance Guidelines were developed by the Japan Gastroenterological Endoscopy Society as basic guidelines based on the scientific methods. The importance of endoscopic screening and surveillance for both detection and post-treatment follow-up of colorectal cancer has been recognized as essential to reduce disease mortality. There is limited high-level evidence in this field; therefore, we had to focus on the consensus of experts. These clinical practice guidelines consist of 20 clinical questions and eight background knowledge topics that have been determined as the current guiding principles.

<h3>Introduction</h3> On recommencement of surveillance after the first wave of the Covid-19 pandemic, the BSG guidelines for colonoscopy surveillance had been updated in October 2020. This created an opportunity to reassess all 1243 colonoscopy surveillance patients by developing a new surveillance nurse led pathway at Wye Valley NHS Trust Hereford (WVT). <h3>Methods</h3> A new Band 7 Non-medical Endoscopist was appointed on transition from Bowel Scope to BCSP and set up a new assessment process to review surveillance patients against the new BSG criteria. Patients who no longer met the new guidelines were written to explaining the change in the BSG criteria. Patients still meeting the surveillance criteria were invited to a nurse led clinic where suitability for colonoscopy was assessed, specific instructions regarding bowel preparation given, renal function assessed and arrangements made for anticoagulation ± diabetic control. Clinics were either face to face or via telephone according to the patient’s wishes. <h3>Results</h3> 1243 patients on the initial colonoscopy surveillance waiting list underwent initial validation and 39% of patients were discharged as they no longer met the surveillance criteria. A further 8% were discharged for other reasons (moved away, new health concerns). After clinic assessment, only 34% of the original surveillance cohort were suitable for surveillance colonoscopy, 3% were not fit enough for any surveillance, 2% declined any colonic investigations, 2% agreed to CTC, 2% DNAd, 1% had moved away and 4% were Covid anxious and requested a delay in surveillance. 5% were upgraded to C2WW due to new symptoms. With a combination of reducing demand and increasing capacity back to pre-Covid levels, surveillance waiting times have reduced from 12 months to 6 weeks. The discharges have released 151 colonoscopy lists (758 colonoscopies) to the C2WW and screening programmes. Using NHS tariff for colonoscopy of £372, these 758 discharges achieved a financial benefit of £281,976. The surveillance clinic cost has been 2 Band 7 nursing clinics per week over an 8 month period annualised to £8960 pa. <h3>Conclusions</h3> Nurse led clinics can thoroughly assess colonoscopy surveillance patients against the new BSG guidelines. This substantially reduces demand, reduces waiting times and generates a cost benefit. Quality is also enhanced by specific information of the procedure and anticoagulation issues. This has reduced surveillance waiting times down to JAG recommendations within 8 months. Our new surveillance process has released a total of 151 colonoscopy lists by demonstrating only 34% of the initial cohort of patients remained suitable for colonoscopy surveillance.

Surveillance colonoscopy guidelines following adenomas or sessile serrated adenomas/polyps (SSPs) are based on pathology features known to be associated with risk of future colorectal cancer. A synchronous conventional adenoma may increase the malignant potential of SSP, but current guidelines do not address this combination of pathologies. The aim was to assess the risk of advanced neoplasia after SSP with or without synchronous adenoma compared to that following a conventional adenoma. An audit was conducted on colonoscopies performed between 2000 and 2014 as part of a surveillance program. Index colonoscopy findings were classified as: low-risk SSP and high-risk SSP (size ≥ 10mm or with cytological dysplasia) with and without synchronous adenoma; high-risk adenoma and low-risk adenoma. Risk of advanced neoplasia was determined at subsequent surveillance colonoscopies. In total, 2157 patients had adenoma or SSP found at index colonoscopy-low-risk adenoma (40%), high-risk adenoma (54%) and SSP (4%). Synchronous adenomas were seen with 47% of SSP. The median follow-up was 50.3months (interquartile range 28.1-79.3). Compared to an index finding of low-risk adenoma, index findings of high-risk adenoma, as well as SSP with synchronous adenoma, were independent predictors of future advanced neoplasia (high-risk adenoma: hazard ratio (HR) = 2.04 (95% CI 1.70-2.45); high-risk SSP + adenoma HR = 3.20 (95% CI 1.31-7.82); low-risk SSP + adenoma: HR = 2.20 (95% CI 1.03-4.68)). Synchronous adenoma increases the risk of advanced neoplasia for SSP equivalent to that seen following high-risk adenoma. Guidelines for surveillance should take into account concurrent pathologies with SSP.

An appropriate post-polypectomy surveillance program requires the effectiveness of reducing colorectal cancer and safety. In addition, the post-polypectomy surveillance program should consider the burden of limited medical resource capacity, cost-effectiveness, and patient adherence. In this sense, a risk-stratified surveillance program based on baseline colonoscopy results is ideal. Major international guidelines for post-polypectomy surveillance, such as those from the European Union and the United States, have recommended risk-stratified surveillance programs. Both guidelines have recently been updated to better differentiate between high- and low-risk individuals. In both updated guidelines, more individuals have been downgraded to lower-risk groups that require less frequent or no surveillance. Furthermore, increased attention has been paid to the surveillance of patients who undergo serrated polyp removal. Previous guidelines in Japan did not clearly outline the risk stratification in post-polypectomy surveillance. However, the new colonoscopy screening and surveillance guidelines presented by the Japan Gastroenterological Endoscopy Society include a risk-stratified post-polypectomy surveillance program. Further discussion and analysis of unresolved issues in this field, such as the optimal follow-up after the first surveillance, the upper age limit for surveillance, and the ideal method for improving adherence to surveillance guidelines, are warranted.

To assess the compliance of the surveillance colonoscopy waiting list with ACPGBI/BSG guidelines for colonoscopy follow-up and to measure the impact of adjusting referrals to be inline with the guidelines. This is a quantitative five-stage clinical audit cycle involving a large patient cohort from the Kent and Medway Cancer Network, which includes seven hospitals across four NHS Hospital Trusts and an estimated population of 1.8 million. 3020 patients were waiting for a surveillance colonoscopy. Their notes were reviewed and the indications for colonoscopy were compared with the ACPGBI/BSG 2002 guidelines. Those patients whose referral to the surveillance colonoscopy waiting list was not found to be compliant were adjusted to be inline with the guidelines. The impact of adjusting the surveillance colonoscopy waiting list on the diagnostic colonoscopy service was assessed by measuring the average waiting times for a colonoscopy before and after the intervention. Around 22% (n = 664) of surveillance colonoscopy referrals were inline with the guidelines, 51% (n = 1540) could be cancelled from the list and 27% (n = 816) could be given a new date. Implementing these recommendations reduced the average wait for a diagnostic colonoscopy from 76.8 to 56.0 days (P = 0.0022). Following guidelines for surveillance colonoscopy can reduce waiting times for diagnostic colonoscopy. This allows a faster patient journey for diagnostic colonoscopy and a uniform plan for duration and frequency of surveillance colonoscopy. However, this action promoted serious debate on the social, moral and ethical issues.

Purpose: To prevent colorectal cancer in ulcerative colitis (UC), major gastroenterology societies recommend an intensive program of colonoscopic screening/surveillance every 1–2 years beginning 8 years after disease onset. In patients without UC, rates of participation in colonoscopic screening programs have been well described. However in patients with UC, rates of participation in colonoscopic surveillance programs have not. Methods: We studied a previously described population-based cohort of 358 patients with left-sided (N = 151) or pancolonic (N = 207) ulcerative colitis who received care for at least 2 years within an integrated health care system and were eligible for screening colonoscopy based on duration of disease > 8 years. All procedures and pathology records were available electronically as part of clinical care. We determined 1) What proportion of patients underwent at least one screening/surveillance colonoscopy between 2001–2005; 2) what proportion of patients who underwent surveillance colonoscopy underwent a second surveillance within the recommended 1–2 years. Results: Among 358 prevalent cases of UC eligible for screening/surveillance between 2001–2005, 123 (34%) underwent at least one surveillance colonoscopy during this time period. Among 123 patients who underwent one surveillance exam, 64 (52%) underwent an additional surveillance colonoscopy within the recommended 1–2 years. Overall, 64 of the 358 patients (18%) who were eligible for surveillance adhered to the recommended surveillance guidelines. Conclusion: In an integrated health care system with access to colonoscopy, few patients with UC adhered to the recommended colonoscopic surveillance guidelines. If these guidelines represent the current standard of care, reasons for such poor rates of adherence must be further investigated. Furthermore, strategies to improve participation in colonoscopic surveillance must be developed.

Colonoscopy practice for veterans within and outside the Veterans Affairs setting: a matched cohort study

SummaryObjectives: This survey aims to provide an overview of the current state of biomedical and clinical Natural Language Processing (NLP) research and practice in Languages other than English (LoE). We pay special attention to data resources, language models, and popular NLP downstream tasks.Methods: We explore the literature on clinical and biomedical NLP from the years 2020-2022, focusing on the challenges of multilinguality and LoE. We query online databases and manually select relevant publications. We also use recent NLP review papers to identify the possible information lacunae.Results: Our work confirms the recent trend towards the use of transformer-based language models for a variety of NLP tasks in medical domains. In addition, there has been an increase in the availability of annotated datasets for clinical NLP in LoE, particularly in European languages such as Spanish, German and French. Common NLP tasks addressed in medical NLP research in LoE include information extraction, named entity recognition, normalization, linking, and negation detection. However, there is still a need for the development of annotated datasets and models specifically tailored to the unique characteristics and challenges of medical text in some of these languages, especially low-resources ones. Lastly, this survey highlights the progress of medical NLP in LoE, and helps at identifying opportunities for future research and development in this field.