Jared M Wohlgemut’s research while affiliated with Queen Mary University of London and other places

Proposals of artificial intelligence (AI) solutions based on more complex and accurate predictive models are becoming ubiquitous across many disciplines. As the complexity of these models increases, there is a tendency for transparency and users’ understanding to decrease. This means accurate prediction alone is insufficient to make an AI-based solution truly useful. For the development of healthcare systems, this raises new issues for accountability and safety. How and why an AI system made a recommendation may necessitate complex explanations of the inner workings and reasoning processes. While research on explainable AI (XAI) has grown significantly in recent years, and the demand for XAI in medicine is high, determining what constitutes a good explanation is ad hoc and providing adequate explanations remains a challenge. To realise the potential of AI, it is critical to shed light on two fundamental questions of explanation for safety–critical AI such as health-AI that remain unanswered: (1) What is an explanation in health-AI? And (2) What are the attributes of a good explanation in health-AI? In this study and possibly for the first time we studied published literature, and expert opinions from a diverse group of professionals reported from a two-round Delphi study. The research outputs include (1) a proposed definition of explanation in health-AI, and (2) a comprehensive set of attributes that characterize a good explanation in health-AI.

Introduction Traumatic injuries are responsible for a huge amount of mortality, morbidity, and disability globally. Within global surgery, Bellwether procedures have previously been identified to measure the surgical proficiency of a hospital or a region, however traumatic injuries often have distinct epidemiological and demographic patterns, compared to routine surgical care. Using a focused set of procedures or processes to measure trauma performance could allow for improved trauma assessments and outcomes on a global scale. Methodology An international Delphi study will be conducted in attempt gain consensus on the optimal procedures and processes that can be used to assess the performance of trauma care within any region or hospital worldwide. Recognised guidelines for conducting the Delphi process will be followed, comprising of 3 separate rounds and participation from a public and patient involvement (PPI) group. Respondents will be identified through pre-existing collaborative networks and research partners, to ensure adequate global representation from across the trauma care pathway. The final Bellwether procedures will be those that have shown consensus in agreement and in stability throughout successive rounds Discussion This Delphi process aims to identify the optimal Bellwether metrics that could be used to assess trauma care worldwide. Using a focused set of procedures or processes to assess trauma performance globally will reduce complexity and improve ease of use compared to current methods.

Background The primary aim of this scoping review was to synthesise key domains and sub-domains described in existing clinical decision support systems (CDSS) implementation frameworks into a novel taxonomy and demonstrate most-studied and least-studied areas. Secondary objectives were to evaluate the frequency and manner of use of each framework, and catalogue frameworks by implementation stage. Methods A scoping review of Pubmed, Scopus, Web of Science, PsychInfo and Embase was conducted on 12/01/2022, limited to English language, including 2000–2021. Each framework was categorised as addressing one or multiple stages of implementation: design and development, evaluation, acceptance and integration, and adoption and maintenance. Key parts of each framework were grouped into domains and sub-domains. Results Of 3550 titles identified, 58 papers were included. The most-studied implementation stage was acceptance and integration, while the least-studied was design and development. The three main framework uses were: for evaluating adoption, for understanding attitudes toward implementation, and for framework validation. The most frequently used framework was the Consolidated Framework for Implementation Research. Conclusions Many frameworks have been published to overcome barriers to CDSS implementation and offer guidance towards successful adoption. However, for co-developers, choosing relevant frameworks may be a challenge. A taxonomy of domains addressed by CDSS implementation frameworks is provided, as well as a description of their use, and a catalogue of frameworks listed by the implementation stages they address. Future work should ensure best practices for CDSS design are adequately described, and existing frameworks are well-validated. An emphasis on collaboration between clinician and non-clinician affected parties may help advance the field.

Proposals of artificial intelligence (AI) solutions based on increasingly complex and accurate predictive models are becoming ubiquitous across many disciplines. As the complexity of these models grows, transparency and users' understanding often diminish. This suggests that accurate prediction alone is insufficient for making an AI-based solution truly useful. In the development of healthcare systems, this introduces new issues related to accountability and safety. Understanding how and why an AI system makes a recommendation may require complex explanations of its inner workings and reasoning processes. Although research on explainable AI (XAI) has significantly increased in recent years and there is high demand for XAI in medicine, defining what constitutes a good explanation remains ad hoc, and providing adequate explanations continues to be challenging. To fully realize the potential of AI, it is critical to address two fundamental questions about explanations for safety-critical AI applications, such as health-AI: (1) What is an explanation in health-AI? and (2) What are the attributes of a good explanation in health-AI? In this study, we examined published literature and gathered expert opinions through a two-round Delphi study. The research outputs include (1) a definition of what constitutes an explanation in health-AI and (2) a comprehensive list of attributes that characterize a good explanation in health-AI.

Aims The primary aim was to determine the volume of emergency robotic general surgery performed in the UK. Secondary aims were to determine compatible operations with opportunities for training and the availability of robotic-trained staff. Methods Freedom of Information (FOI) requests were sent to 122 UK Acute Trusts, asking for data on general surgery procedures (open, laparoscopic, robotic) from January 2019 to January 2023. Data on common emergency procedures, in-patient length of stay, robotic complications, and availability of robotically trained staff was requested. Results Of the 122 Acute Trusts, 22 did not respond (82% response rate), 32 had no general surgery robotic access, 16 could not comply due to cost (under section 12 of FOI Act), and 52 provided data. Fifteen Trusts performed emergency robotic surgeries: cholecystectomy (262 cases), hernia repair (140), and appendectomy (3). 86% of these Trusts utilized the DaVinci system, while the CMR Versius system was employed by the remaining 14%. There was a median of 12.5 (range 2-40) robotic-trained staff in-hours, versus a median of 0 (range 0-4) out-of-hours. Conclusions Emergency robotic general surgery in the UK is limited by financial constraints, staff availability, and selective procedural use. Compatible emergency operations that have been performed were cholecystectomies, hernias, and appendectomies. To expand its role, there's a need for training, infrastructure investment, and further research into the cost-effectiveness and outcomes of robotic surgery in emergency cases.

Aims Determine the desired future operative role of dedicated Emergency General Surgery (EGS) surgeons in the United Kingdom (UK) and the Republic of Ireland (ROI), and how to improve the attractiveness of an EGS post. Methods A live interactive online survey, designed and endorsed by ASGBI, was completed by attendees at two ASGBI conferences in 2023. The study was conducted according to CHERRIES and exempt from REC according to HRA. Consent was obtained from study participants, who were asked using a Likert scale which operations a dedicated EGS surgeon should perform and how an EGS post could be more attractive. Results The survey was completed by 156/371 eligible participants, of which 132 responses were analysed once duplicate responses (n=16), and non-UK/ROI respondents (n=8), were excluded. Participants included consultants (34%), surgical trainees (42%), and other doctors (23%), 61% were male and 39% were female, aged between 25 and 69. Respondents stated the desired future operative role of EGS surgeons included most EGS operations, including for torso trauma. Few believed EGS surgeons should perform emergency colonic resection for cancer (49%) or inflammatory bowel disease (37%), gastric volvulus (34%), bariatric complications (26%), or Boerhaaves (28%). The vast majority agreed that EGS posts could be improved by including elective general and subspecialty operating lists, continuing professional development, minimally-invasive (laparoscopic/robotic) opportunities, rest days post-on-call, family-friendly hours, EGS ambulatory and follow-up clinics, and major trauma responsibility. Conclusions EGS surgeons should operate on the majority of EGS patients. Job plans that include elective operating, professional development, and patient follow-up may improve recruitment and retention.

Aims Determine human factors implicated in cases submitted to the Confidential Reporting System in Surgery (CORESS), using the non-technical skills for surgeons (NOTSS) framework. Methods Consecutive cases submitted to CORESS, between 2005 and 2011 were evaluated as an interim analysis. The type of case, whether patient harm occurred, Clavien-Dindo (CD) classification, and the location of the incident were collected. Demographic data was not collected because they are often changed to protect patient identification. Each case was reviewed according to the non-technical skills for surgeons (NOTSS) taxonomy. Results There were 100 cases included in this interim analysis. The most common types of cases were delayed diagnosis (n=19), known complications (n=10), near misses (n=10), poor surgical technique (n=9), misuse of equipment/device (n=8), and wrong site surgery (n=5). Two-thirds of patients were harmed, with complications ranging from minor to life-threatening (9 CD I, 10 CD II, 32 CD III, 12 CD IV), and 4 resulting in death (CD V). Incidents occurred most frequently in the operating theatre (n=54), the ward (n=16), or the emergency department (n=9). There were 71 incidents that endangered patient safety related to deficiencies in situational awareness, 59 to poor communication and teamwork, 48 to poor decision-making, and 48 related to leadership failures. Conclusions The main human factors implicated in surgical patient safety incidents related to errors in situational awareness, communication and teamwork, and to a lesser extent, decision-making, and leadership. Efforts to prevent medical errors should be focused on clinician human factors training.