Augmenting small tabular health data for training prognostic ensemble machine learning models using generative models
Liu D, Kababji SE, Mitsakakis N, Pilgram L, Walters TD, Clemons M, Pond GR, El-Hussuna A, Emam KE
Small datasets are common in health research. However, the generalization performance of machine learning models is suboptimal when the training datasets are small. To address this, data augmentation is one solution and is often used for imaging and time series data, but there are no evaluations on its potential benefits for tabular health data. Augmentation increases sample size and is seen as a form of regularization that increases the diversity of small datasets, leading them to perform better on unseen data.
Development of a Clinical Prediction Model for Anastomotic Leakage in Colorectal Surgery
Vincent Ochs, Stephanie Taha-Mehlitz, Alaa El-Hussuna Anas Taha et al
In this prognostic study of patients who underwent colon resection with anastomosis, a meta-model was found to improve predictive accuracy. Additional prospective studies are needed to assess the clinical utility of the model in real-time settings and its integration into surgical workflows
Safety of Prophylactic Ureteral Stents in Colorectal Surgery: A Systematic Review and Meta-Analysis
Mohammed Benhammou, Ghada Kareem, Muhammed Elhadi, Abdelaziz Salama, Sofian Zreigh, Rana Shembesh, Mohammad Hamza, Dina Elraggal, Sara Saleh, Hamza Hassan Abdalle, Mustafa Marzoug, Arowa Alansari, Dawood Alatefi, Alaa El-Hussuna, Steven Wexner
This study investigated the association between prophylactic ureteral stents (PUS) and postoperative complications was evaluated.
Enhancing Surgical Performance through Automated Video Analysis (ESPAVA) utilising computer vision and machine learning
Turkish Journal of Colorectal disease
Alaa El-Hussuna, Muhammed Elhadi1, Andreas Møgelmose, Hanan Aljuaid
This study aims to evaluate and enhance surgical performance through the use of automated performance metrics (APMs).
The study will enable efficient training programs within constrained working hours and address heightened ethical considerations regarding patient safety.
A systematic review on biomarkers implemented to measure surgical stress response in patients undergoing colorectal surgery
El-Hussuna A, Hussain MI, Cuk P, Ellebæk MB
Measuring the surgical stress response (SSR) involves assessing physiological and biochemical markers that indicate the extent of the body’s reaction to surgical trauma. The study aims to evaluate biomarkers used to measure (SSR) and to explore their potential clinical applicability. Mapping the existing literature may help address current limitations and inspire the development of new methods for measurement of SSR to measure SSR.
Synthetic Data Generation for Augmenting Small Samples
Dan Liu, Samer El Kababji, Nicholas Mitsakakis, Lisa Pilgram, Thomas Walters, Mark Clemons, Greg Pond, Alaa El-Hussuna, Khaled El Emam
Small datasets are common in health research. However, the generalization performance of machine learning models is suboptimal when the training datasets are small. To address this, data augmentation is one solution. Augmentation increases sample size and is seen as a form of regularization that increases the diversity of small datasets, leading them to perform better on unseen data.
Postoperative Outcomes in Tofacitinib-treated Patients with Acute Severe Ulcerative Colitis Undergoing Colectomy
Clinical Gastroenterology and Hepatology
Accepted for publication 2024
The aim of this study was to assess the intraoperative and post-operative complications with tofacitinib use for patients with acute severe ulcerative colitis (ASUC) undergoing semi-urgent colectomy.
This large multicenter adjusted case-control study found no different in postoperative complications in tofacitinib-treated patients with ASUC compared to patients with ASUC receiving usual care.
Small samples Retrospective studies with small sample size are still dominating colon cancer research: local recurrence in right sided colon cancer as a case in point
Turkish Journal of Gastroenterology
The aim of this study was to examine dominant study designs and investigate redundancy in surgical research.
The study highlights a prevalent challenge in cancer research that might limit the advancements in the field. Our findings advocate a strategic shift towards well-designed, larger-scale, and finely detailed multicentric studies, essential for fostering real progress in surgical research and patient care.
Automated Performance Metrics for Objective Surgical Skill Assessment in Laparoscopic Training
In Proceedings of VISIGRAPP 2025 – Volume 3 VISAPP: VISAPP, 824-830, 2025 , Porto, Portugal
The assessment of surgical skill is critical in advancing surgical training and enhancing the performance of surgeons. Traditional evaluation methods relying on human observation and checklists are often biased and inefficient, prompting the need for automated and objective systems. This study explores the use of Automated Performance Metrics (APMs) in laparoscopic surgeries, using video-based data and advanced object tracking techniques.
Synthetic Data Generation for Augmenting Small Samples
https://doi.org/10.1089/cyber.2023.0222
One of the biggest challenges with training machine learning models on health data is that many health datasets are small.
In this study we show how generative models can help solve that problem by simulating additional observations and this results in more diverse datasets that perform better on unseen data.
The improvement is non-trivial, and averages around an 15.5% increase in AUC.
The Association between Problematic Instagram Use, Psychological Distress, and Well-Being: A Systematic Review and Meta-Analysis.
Cyberpsychology, Behavior, and Social Networking
https://doi.org/10.1089/cyber.2023.0222
The study found significant conceptual heterogeneity across the studies included due to different study designs and population heterogeneity. The results show a positive association of PIU and IU with depression, anxiety, and other psychological distress(es). There is a need for robust designed longitudinal research to explore this association.
OpenSourceResearch Collaboration & ReplicaAnalytics
An evaluation of the replicability of analyses using synthetic health data
DOI:https://doi.org/10.1016/j.patter.2024.100946
The study considered two different types of generative models and also looked at the impact of data amplification.
Some interesting observations include that: when analyzed properly, data amplification has marginal value for improving the replicability of statistical analyses, replicability is maximized when ten synthetic datasets are generated and the resultant models are combined (as one does with multiple imputation), and that replicability does vary quite a bit with the type of generative model.
OpenSourceResearch Collaboration & ReplicaAnalytics
An evaluation of synthetic data augmentation for mitigating covariate bias in health data
DOI:https://doi.org/10.1016/j.patter.2024.100946
Important publication about synthetic data augmentation
The study compares the efficacy of bias-mitigating approaches in real-world health #data
Performance is evaluated based on predictive accuracy, parameter precision, and fairness
OpenSourceResearch Collaboration
Peak performance in surgery: how machine learning can improve surgical performance
DOI: https://doi.org/10.62463/surgery.15
Surgical research is associated with specific methodological and practical challenges related to the assessment of complex interventions and variation of clinical equipoise. Biases can, to some degree, be overcome by randomisation, blinding, and intervention standardisation, although randomised trials can be costly and time-consuming.
This poses a unique opportunity for innovation in surgical research design and evaluation trials and their assessment.
OpenSourceResearch Collaboration
Burnout among surgeons before and during the SARS-CoV-2 pandemic: an international survey
BMC Psychol 12, 48 (2024). https://doi.org/10.1186/s40359-023-01517-4
Members of OSRC participated in this important contribution. The article concludes that There was a significant burnout among trainees.
OpenSourceResearch Collaboration
Systematic Review of the Effect of Non-steroidal Anti-Inflammatory Drugs on the Exacerbation of Inflammatory Bowel Disease
Turk J Colorectal Dis 2023;33(4):92-102
Online publication
This article is part of investigator initiated research by members of OSRC. The aim of this systematic review is to investigate evidence on the role NSAIDs play in the exacerbation of IBD
OpenSourceResearch Collaboration
We need to rethink how to improve surgical research
Online publication
This article was written to encourage discussion of surgical research. Claims that surgical research are “comic operas” are not unsubstantiated as this article shows.
OpenSourceResearch Collaboration
How Computer Vision and Big Data Analytics are Revolutionizing Surgery
Online publication
This study was part of OpenSourceResearch Collaboration project
which aimed to explore The role of AI in assessment of surgical video films
OpenSourceResearch Collaboration
Preoperative assessment of blood supply and its role in predicting anastomotic leak
Online ahead of print
This study was part of OpenSourceResearch Collaboration project
which aimed to explore The role of AI in radiology
OpenSourceResearch Collaboration
Watchful waiting after radiological guided drainage of intra-abdominal abscess in patients with Crohn’s disease might be associated with increased rates of stoma construction
Online ahead of print
This study was part of OpenSourceResearch Collaboration project
which aimed to explore The need for AI research in IBD
OpenSourceResearch Collaboration
Risk of anorectal cancer in patients with Crohn’s disease and perianal fistula: a nationwide Danish cohort study
Online ahead of print.
This study was part of OpenSourceResearch Collaboration project
which aimed to explore register data as part of big data mining
OpenSourceResearch Collaboration
Algorithms to anonymize structured medical and healthcare data: A systematic review
Front. Bioinform., 22 December 2022
This study was part of OpenSourceResearch Collaboration project
which aimed to improve data management and sharing.
OpenSourceResearch Collaboration
Outcomes of emergency surgical interventions in right-sided colonic cancer: nationwide population-based study based on Danish Colorectal Cancer Group register
Volume 7, Issue 1, February 2023, zrac153
This study was part of OpenSourceResearch Collaboration project
which aimed to explore register data as part of big data mining
OpenSourceResearch Collaboration
Comparison of Microbiological Profile of Drained Intra-Abdominal Abscess in Patients with Crohn’s Disease and Colonic Diverticulitis
Turkish Journal of Colorectal diseases
Turk J Colorectal Dis 2022;32(4):252-257
This study was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in management of IBD optimize it.
Members of OpenSourceResearch Collaboration participated in ECCO research
ECCO Topical Review: Roadmap to optimal peri-operative care in IBD
This study was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in management of IBD optimize it.
OpenSourceResearch Collaboration
Inflammatory response in patients with Crohn’s Disease Compared to Ulcerative Colitis: secondary results of a prospective pilot study
DOI:doi.org/10.1093/crocol/otac047
This study was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in management of IBD optimize it.
OpenSourceResearch Collaboration
The use of datasets collected by surgical collaborating groups
British Journal of Surgery 2021 (ahead of print).
https://doi.org/10.1093/bjsopen/zrab075
This study was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
OpenSourceResearch Collaboration
Current status and future perspectives of collaboration in surgical research: A scoping review of the evidence
Surgery 2021 (ahead of print).
https://doi.org/10.1016/j.surg.2021.04.031
This study was part of OpenSourceResearch Collaboration project
which aimed to improve collaboration in research.
OpenSourceResearch Collaboration
Types, limitations, and possible alternatives of peer review based on the literature with the surgeons’ opinions via Twitter: a narrative review.
Science editing journal 2021 (ahead of print).
This study was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
OpenSourceResearch Collaboration
Postoperative complications and waiting time for surgical intervention after radiologically guided drainage of intra-abdominal abscess in patients with Crohn’s disease
British Journal of Surgery 2021 (ahead of print).
https://doi.org/10.1093/bjsopen/zrab075
This study was part of OpenSourceResearch Collaboration project
which aimed to investigate pre-operative optimization of patients with Crohn’s disease.
OpenSourceResearch Collaboration
Patient-reported outcome measures in colorectal surgery: Construction of core measures using open source research method
Surgical Innovation 2021 (ahead of print).
https://doi.org/10.1177/1553350621998871
This study was part of OpenSourceResearch Collaboration project
which aimed to explore optimization of pre-and postoperative course.
OpenSourceResearch Collaboration
Extended right sided colon resection does not reduce the risk of colon cancer local-regional recurrence: Nation-wide population-based study from Danish Colorectal Cancer Group Database
DOI: 10.1097/DCR.0000000000002358
This study was part of OpenSourceResearch Collaboration project
which aimed to explore local-regional recurrence in colon cancer.
Linda Tutty, Alaa El-Hussuna.
The power and potential of using social media to connect patients and health professionals
How social media can be used to empower patients, promote self management, and co-develop a “real world” core set of patient reported outcome measures for those undergoing colorectal surgery
This article was part of OpenSourceResearch Collaboration project which aimed to explore the optimization of pre-and post operative course.
OpenSourceResearch Collaboration
Biological treatment and the risk of unfavorable postoperative outcome in patients with inflammatory bowel disease: An open source conducted review of the current literature and future perspectives by an expert panel.
https://doi.org/10.1093/crocol/otz021
This study was part of OpenSourceResearch Collaboration project
which aimed to explore pre-operative optimization of patients with IBD.
Alaa El-Hussuna, Sameh H Emile, Steven D Wexner.
Comment on “Novel approaches to surgical trials and the assessment of new surgical technologies”: The OpenSourceResearch collaboration to conduct research using online tools and social media.
British Journal of Surgery 2020
https://doi.org/10.1093/bjs/znaa061
This article was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
Alaa El-Hussuna, Gaetano Gallo
Research as Part of Daily Clinical Life: Challenging But Possible
J Surg Res Prac. 2021;2(2):1-4.
http://dx.doi.org/10.46889/JSRP.2021.2203
This article was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
OpenSourceResearch Collaboration
Information Technologies to Support Research: Challenges and Action Plan.
Editorial by Alaa El-Hussuna
EC Gastroenterology and Digestive System 8.4 (2021).
This article was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
Ali Yalcinkaya, Alaa El-Hussuna.
#OpenSourceResearch: A novel medical research technique/style in terms of accessibility and way of work.
https://doi.org/10.1111/codi.15392
This article was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
Alaa El-Hussuna, Jeppe Lang
We need to rethink how to improve surgical research
Dan Med J 2022;69(6):A205119
This article was part of OpenSourceResearch Collaboration project
which aimed to explore collaboration in research and how to optimize it.
OpenSourceResearch in collaboration with EuroPower group
Association between use of enhanced recovery after surgery protocols and postoperative complications in colorectal surgery in Europe
2022 Apr 8;80:110752.
This article explores the optimization of surgical patients. Read more about OSRC optimization projects here
Khaled El Emam, Lucy Mosquera, Xi Fang & Alaa El-Hussuna
Utility Metrics for Evaluating Synthetic Health Data Generation Methods: A Validation Study
This article was part of OpenSourceResearch Collaboration project
which aimed to explore data de-identification
