DECIDE-AI: a new reporting guideline and its relevance to artificial intelligence studies in radiology.

DECIDE-AI is a new, stage-specific reporting guideline for the early and live clinical evaluation of decision-support systems based on artificial intelligence (AI). It answers a need for more attention to the human factors influencing clinical AI performance and more transparent reporting of clinical studies investigating AI systems. Given the rapid expansion of AI systems and the concentration of related studies in radiology, these new standards are likely to find a place in radiological literature in the near future. This review highlights some of the specificities of AI as complex intervention, why a new reporting guideline was needed for early stage, live evaluation of this technology, and how DECIDE-AI and other AI reporting guidelines can be useful to radiologists and researchers.

Failure to rescue following emergency surgery: A FRAM analysis of the management of the deteriorating patient.

BACKGROUND: Failure to rescue (FTR) denotes mortality from post-operative complications after surgery with curative intent. High-volume, low-mortality units have similar complication rates to others, but have lower FTR rates. Effective response to the deteriorating post-operative patient is therefore critical to reducing surgical mortality. Resilience Engineering might afford a useful perspective for studying how the management of deterioration usually succeeds and how resilience can be strengthened. METHODS: We studied the response to the deteriorating patient following emergency abdominal surgery in a large surgical emergency unit, using the Functional Resonance Analysis Method (FRAM). FRAM focuses on the conflicts and trade-offs inherent in the process of response, and how staff adapt to them, rather than on identifying and eliminating error. 31 semi-structured interviews and two workshops were used to construct a model of the response system from which conclusions could be drawn about possible ways to strengthen system resilience. RESULTS: The model identified 23 functions, grouped into five clusters, and their respective variability. The FRAM analysis highlighted trade-offs and conflicts which affected decisions over timing, as well as strategies used by staff to cope with these underlying tensions. Suggestions for improving system resilience centred on improving team communication, organisational learning and relationships, rather than identifying and fixing specific system faults. CONCLUSION: FRAM can be used for analysing surgical work systems in order to identify recommendations focused on strengthening organisational resilience. Its potential value should be explored by empirical evaluation of its use in systems improvement.

Improving the WHO Surgical Safety Checklist sign-out.

BACKGROUND: The WHO Surgical Safety Checklist has been shown to reduce perioperative morbidity and mortality worldwide. There is evidence to suggest that sign-out is the most poorly performed phase of the checklist as it coincides with a period of high workload for team members. This study aimed to see whether modification of this process might result in greater compliance. METHODS: A controlled longitudinal (before and after) study was performed to evaluate the effect of a modified checklist sign-out on compliance in a single surgical department. Checklist quality was evaluated by measurement of checklist completion, active participation, and team member presence. Workload assessment was performed to identify the optimal moment for the sign-out process. The sign-out process was modified through an iterative multidisciplinary approach, informed by results from the workload assessment. Feedback was obtained through staff surveys. RESULTS: A total of 185 operations were used, with an intervention group in vascular surgery and a control group in orthopaedics. The optimal timing for sign-out was identified as after final wound closure. The modified sign-out process improved active participation of team members (21 of 34 versus 31 of 34; P = 0.010). In the control group, complete compliance improved (48 of 76 versus 30 of 41; P = 0.041). However, active participation decreased (53 of 76 versus 19 of 41; P = 0.022). No differences were noted between groups in team member presence. Eighteen of 21 staff questioned viewed the modifications positively. CONCLUSION: The optimal sign-out timing was identified as immediately after final wound closure prior to undraping the patient.

Systematic review of methods for quantifying teamwork in the operating theatre.

BACKGROUND: Teamwork in the operating theatre is becoming increasingly recognized as a major factor in clinical outcomes. Many tools have been developed to measure teamwork. Most fall into two categories: self-assessment by theatre staff and assessment by observers. A critical and comparative analysis of the validity and reliability of these tools is lacking. METHODS: MEDLINE and Embase databases were searched following PRISMA guidelines. Content validity was assessed using measurements of inter-rater agreement, predictive validity and multisite reliability, and interobserver reliability using statistical measures of inter-rater agreement and reliability. Quantitative meta-analysis was deemed unsuitable. RESULTS: Forty-eight articles were selected for final inclusion; self-assessment tools were used in 18 and observational tools in 28, and there were two qualitative studies. Self-assessment of teamwork by profession varied with the profession of the assessor. The most robust self-assessment tool was the Safety Attitudes Questionnaire (SAQ), although this failed to demonstrate multisite reliability. The most robust observational tool was the Non-Technical Skills (NOTECHS) system, which demonstrated both test-retest reliability (P > 0·09) and interobserver reliability (Rwg = 0·96). CONCLUSION: Self-assessment of teamwork by the theatre team was influenced by professional differences. Observational tools, when used by trained observers, circumvented this.

Evaluation of high-intensity focused ultrasound ablation for uterine fibroids: an IDEAL prospective exploration study.

OBJECTIVE: To evaluate the clinical outcomes of high-intensity focused ultrasound (HIFU) and surgery in treating uterine fibroids, and prepare for a definitive randomised trial. DESIGN: Prospective multicentre patient choice cohort study (IDEAL Exploratory study) of HIFU, myomectomy or hysterectomy for treating symptomatic uterine fibroids. SETTING: 20 Chinese hospitals. POPULATION OR SAMPLE: 2411 Chinese women with symptomatic fibroids. METHODS: Prospective non-randomised cohort study with learning curve analysis (IDEAL Stage 2b Prospective Exploration Study). MAIN OUTCOME MEASURES: Complications, hospital stay, return to normal activities, and quality of life (measured with UFS-Qol and SF-36 at baseline, 6 and 12 months), and need for further treatment. Quality-of-life outcomes were adjusted using regression modelling. HIFU treatment quality was evaluated using LC-CUSUM to identify operator learning curves. A health economic analysis of costs was performed. RESULTS: 1353 women received HIFU, 472 hysterectomy and 586 myomectomy. HIFU patients were significantly younger (P < 0.001), slimmer (P < 0.001), better educated (P < 0.001), and wealthier (P = 0.002) than surgery patients. Both UFS and QoL improved more rapidly after HIFU than after surgery (P = 0.002 and P = 0.001, respectively at 6 months), but absolute differences were small. Major adverse events occurred in 3 (0.2%) of HIFU and in 133 (12.6%) of surgical cases (P < 0.001). Median time for hospital stay was 4 days (interquartile range, 0-5 days), 10 days (interquartile range, 8-12.5 days) and 8 days (interquartile range, 7-10 days). CONCLUSIONS: HIFU caused substantially less morbidity than surgery, with similar longer-term QoL. Despite group baseline differences and lack of blinding, these findings support the need for a randomised controlled trial (RCT) of HIFU treatment for fibroids. The IDEAL Exploratory design facilitated RCT protocol development. TWEETABLE ABSTRACT: HIFU had much better short-term outcomes than surgery for fibroids in 2411-patient Chinese IDEAL format study.

Does experience matter? A meta-analysis of physician rating websites of Orthopaedic Surgeons.

PURPOSE: To perform a systematic review evaluating online ratings of Orthopaedic Surgeons to determine: (1) the number of reviews per surgeon by website, (2) whether the number of reviews and rate of review acquisition correlated with years in practice, and (3) whether the use of ratings websites varied based on the surgeons' geographic region of practice. METHODS: The USA was divided into nine geographic regions, and the most populous city in each region was selected. HealthGrades and the American Board of Orthopaedic Surgery (ABOS) database were used to identify and screen (respectively) all Orthopaedic Surgeons within each of these nine cities. These surgeons were divided into three "age" groups by years since board certification (0-10, 10-20, and 20-30 years were assigned as Groups 1, 2, and 3, respectively). An equal number of surgeons were randomly selected from each region for final analysis. The online profiles for each surgeon were reviewed on four online physician rating websites (PRW, i.e. HealthGrades, Vitals, RateMDs, Yelp) for the number of available patient reviews. Descriptive statistics, analysis of variance (ANOVA), and Pearson correlations were used. RESULTS: Using HealthGrades, 2802 "Orthopaedic Surgeons" were identified in nine cities. However, 1271 (45%) of these were not found in the ABOS board certification database. After randomization, a total of 351 surgeons were included in the final analysis. For these 351 surgeons, the mean number of reviews per surgeon found on all four websites was 9.0 ± 14.8 (range 0-184). The mean number of reviews did not differ between the three age groups (p > 0.05) with 8.7 ± 14.4, (2) 10.3 ± 18.3, and (3) 8.0 ± 10.8 for Groups 1, 2, and 3, respectively. However, the rate that reviews were obtained (i.e. reviews per surgeon per year) was significantly higher (p < 0.001) in Group 1 (2.6 ± 7.7 reviews per year) compared to Group 2 (1.4 ± 2.4) and Group 3 (1.1 ± 1.4). There was no correlation between the number of reviews and years in practice (R < 0.001), and there was a poor correlation between number of reviews and regional population (R = 0.199). CONCLUSIONS: The number of reviews per surgeon did not differ significantly between the three defined age groups based on years in practice. However, surgeons with less than 10 years in practice were accumulating reviews at a significantly higher rate. Interestingly nearly half of "Orthopaedic Surgeons" listed were not found to be ABOS-certified Orthopaedic Surgeons.

Cross-sectional observational study of the availability of evidence supporting novel implantable devices used in gastrointestinal surgery.

BACKGROUND: Evidence supporting the implementation of novel surgical devices is unstandardized, despite recommendations for assessing novel innovations. This study aimed to determine the proportion of novel implantable devices used in gastrointestinal surgery that are supported by evidence from RCTs. METHODS: A list of novel implantable devices placed intra-abdominally during gastrointestinal surgery was produced. Systematic searches were performed for all devices via PubMed and clinical trial registries. The primary outcome measure was the availability of at least one published RCT for each device. Published RCTs were appraised using the Cochrane tool for assessing risk of bias. RESULTS: A total of 116 eligible devices were identified (implantable mesh 42, topical haemostatics 22, antiadhesion barriers 10, gastric bands 8, suture and staple-line reinforcement 7, artificial sphincters 5, other 22). One hundred and twenty-eight published RCTs were found for 33 of 116 devices (28·4 per cent). Most were assessed as having a high risk of bias, with only 12 of 116 devices (10·3 per cent) supported by a published RCT considered to be low risk. A further 95 ongoing and 23 unpublished RCTs were identified for 42 of 116 devices (36·2 per cent), but many (64 of 116, 55·2 per cent) had no evidence from published, ongoing or unpublished RCTs. The highest stage of innovation according to the IDEAL Framework was stage 1 for 11 devices, stage 2a for 23 devices, stage 2b for one device and stage 3 for 33 devices. The remaining 48 devices had no relevant clinical evidence. CONCLUSION: Only one in ten novel implantable devices available for use in gastrointestinal surgical practice is supported by high-quality RCT evidence.

Practical guide to the Idea, Development and Exploration stages of the IDEAL Framework and Recommendations.

BACKGROUND: Evaluation of new surgical procedures is a complex process challenged by evolution of technique, operator learning curves, the possibility of variable procedural quality, and strong treatment preferences among patients and clinicians. Preliminary studies that address these issues are needed to prepare for a successful randomized trial. The IDEAL (Idea, Development, Exploration, Assessment and Long-term follow-up) Framework and Recommendations provide an integrated step-by-step evaluation pathway that can help investigators achieve this. METHODS: A practical guide was developed for investigators evaluating new surgical interventions in the earlier phases before a randomized trial (corresponding to stages 1, 2a and 2b of the IDEAL Framework). The examples and practical tips included were chosen and agreed upon by consensus among authors with experience either in designing and conducting IDEAL format studies, or in helping others to design such studies. They address the most common challenges encountered by authors attempting to follow the IDEAL Recommendations. RESULTS: A decision aid has been created to help identify the IDEAL stage of an innovation from literature reports, with advice on how to design and report the IDEAL study formats discussed, along with the ethical and scientific rationale for specific recommendations. CONCLUSION: The guide helps readers and researchers to understand and implement the IDEAL Framework and Recommendations to improve the quality of evidence supporting surgical innovation.

[Better Reporting of Interventions: Template for Intervention Description and Replication (TIDieR) Checklist and Guide]. / Die TIDieR Checkliste und Anleitung - ein Instrument für eine verbesserte Interventionsbeschreibung und Replikation.

Without a complete published description of interventions, clinicians and patients cannot reliably implement interventions that are shown to be useful, and other researchers cannot replicate or build on research findings. The quality of description of interventions in publications, however, is remarkably poor. To improve the completeness of reporting, and ultimately the replicability, of interventions, an international group of experts and stakeholders developed the Template for Intervention Description and Replication (TIDieR) checklist and guide. The process involved a literature review for relevant checklists and research, a Delphi survey of an international panel of experts to guide item selection, and a face-to-face panel meeting. The resultant 12-item TIDieR checklist (brief name, why, what (materials), what (procedure), who intervened, how, where, when and how much, tailoring, modifications, how well (planned), how well (actually carried out)) is an extension of the CONSORT 2010 statement (item 5) and the SPIRIT 2013 statement (item 11). While the emphasis of the checklist is on trials, the guidance is intended to apply across all evaluative study designs. This paper presents the TIDieR checklist and guide, with a detailed explanation of each item, and examples of good reporting. The TIDieR checklist and guide should improve the reporting of interventions and make it easier for authors to structure the accounts of their interventions, reviewers and editors to assess the descriptions, and readers to use the information.

Systematic review of methodological quality of individual performance measurement in surgery.

BACKGROUND: Calls for greater transparency with improved quality, safety and outcomes have led to performance tracking of individual surgeons. This study evaluated the methodology of studies investigating individual performance in surgery. METHODS: MEDLINE, Embase, PsycINFO, AMED and the Cochrane Database of Systematic Reviews (from their inception to July 2014) were searched. Two authors independently reviewed citations using predetermined inclusion and exclusion criteria; 91 data points per study were extracted. RESULTS: The search strategy yielded 8514 citations; 101 were eligible, comprising 1 006 037 procedures by 14 455 surgeons. Thirty-four studies were prospective and 66 were retrospective. The aim of the studies was either to assess individual performance and describe the learning curve of a procedure, to describe factors influencing performance, or to describe methods for routine performance monitoring. Some 51·5 per cent of the studies investigated 500 or fewer procedures. Most (77 of 101) were single-centre studies. Less than half of the studies (42, 41·6 per cent) employed statistical modelling or stratification to adjust performance measures. Forty studies (39·6 per cent) adjusted outcomes for case mix. Seventeen (16·8 per cent) adjusted metrics for surgeon-specific factors. Thirteen studies (12·9 per cent) considered clustering in their analyses. The most frequent outcome studied was duration of operation (59·4 per cent), followed by complication rate (45·5 per cent) and reoperation rate (29·7 per cent); 15·8 per cent of studies recorded mortality, and 4·0 per cent explored patient satisfaction. Only 48·5 per cent of studies displayed procedural learning curves using a graph. CONCLUSION: There exist substantial shortcomings in methodological quality, outcome measurements and quality improvement evaluation among current studies of individual surgical performance. Methodological guidelines should be established to ensure that assessments are valid.

Compliance and use of the World Health Organization checklist in U.K. operating theatres.

BACKGROUND: The World Health Organization (WHO) Surgical Safety Checklist is reported to reduce surgical morbidity and mortality, and is mandatory in the U.K. National Health Service. Hospital audit data show high compliance rates, but direct observation suggests that actual performance may be suboptimal. METHODS: For each observed operation, WHO time-out and sign-out attempts were recorded, and the quality of the time-out was evaluated using three measures: all information points communicated, all personnel present and active participation. RESULTS: Observation of WHO checklist performance was conducted for 294 operations, in five hospitals and four surgical specialties. Time-out was attempted in 257 operations (87.4 per cent) and sign-out in 26 (8.8 per cent). Within time-out, all information was communicated in 141 (54.9 per cent), the whole team was present in 199 (77.4 per cent) and active participation was observed in 187 (72.8 per cent) operations. Surgical specialty did not affect time-out or sign-out attempt frequency (P = 0.453). Time-out attempt frequency (range 42-100 per cent) as well as all information communicated (15-83 per cent), all team present (35-90 per cent) and active participation (15-93 per cent) varied between hospitals (P < 0.001 for all). CONCLUSION: Meaningful compliance with the WHO Surgical Safety Checklist is much lower than indicated by administrative data. Sign-out compliance is generally poor, suggesting incompatibility with normal theatre work practices. There is variation between hospitals, but consistency across studied specialties, suggesting a need to address organizational culture issues.

Trigeminal Medullary Dorsal Horn Neurons Activated by Nasal Stimulation Coexpress AMPA, NMDA, and NK1 Receptors.

Afferent information initiating the cardiorespiratory responses during nasal stimulation projects from the nasal passages to neurons within the trigeminal medullary dorsal horn (MDH) via the anterior ethmoidal nerve (AEN). Central AEN terminals are thought to release glutamate to activate the MDH neurons. This study was designed to determine which neurotransmitter receptors (AMPA, kainate, or NMDA glutamate receptor subtypes or the Substance P receptor NK1) are expressed by these activated MDH neurons. Fos was used as a neuronal marker of activated neurons, and immunohistochemistry combined with epifluorescent microscopy was used to determine which neurotransmitter receptor subunits were coexpressed by activated MDH neurons. Results indicate that, during nasal stimulation with ammonia vapors in urethane-anesthetized Sprague-Dawley rats, activated neurons within the superficial MDH coexpress the AMPA glutamate receptor subunits GluA1 (95.8%) and GluA2/3 (88.2%), the NMDA glutamate receptor subunits GluN1 (89.1%) and GluN2A (41.4%), and NK1 receptors (64.0%). It is therefore likely that during nasal stimulation the central terminals of the AEN release glutamate and substance P that then produces activation of these MDH neurons. The involvement of AMPA and NMDA receptors may mediate fast and slow neurotransmission, respectively, while NK1 receptor involvement may indicate activation of a nociceptive pathway.

Ghrelin does not orchestrate the metabolic changes seen in fasting but has significant effects on lipid mobilisation and substrate utilisation.

OBJECTIVE: Short-term fasting is associated with increased GH pulsatility and mobilisation of fats, but underlying mechanisms are unclear. We studied ghrelin's role during fasting and the effects of exogenous ghrelin on lipid mobilisation. DESIGN: Randomised placebo-controlled study. METHODS: In this study, ten controls (body mass index (BMI) 23.3±3.2), ten morbidly obese subjects (BMI 50.1±10.6) and six post-gastrectomy subjects (BMI 25.2±1.0) were fasted for 36  h undergoing regular blood sampling. On a separate occasion, subjects were infused with either i.v. ghrelin (5  pmol/kg per min) or saline over 270  min. RESULTS: Obese and post-gastrectomy subjects had lower ghrelin compared with controls (ANOVA, P=0.02) during the fast. Controls and gastrectomy subjects showed a similar increase in GH pulsatility, circulating non-esterified fatty acids (NEFA) and 3ß-hydroxybutyrate (3 HB). Obese subjects had an impaired GH response (P<0.001), reduced excursions of 3 HB (P=0.01) but no change in NEFA excursions (P=0.09) compared with controls. Ghrelin infusion increased GH, NEFA and ketone bodies (ANOVA, P<0.0001) in all the three groups, but GH response was impaired in the obese subjects (P=0.001). Ghrelin also induced a significant (ANOVA, P=0.004) biphasic NEFA response to meals in all the subjects. CONCLUSIONS: Despite low circulating ghrelin, gastrectomy subjects maintain a normal metabolic response to fasting, implying that ghrelin plays a minimal role. In contrast, infused ghrelin has significant effects on lipid mobilisation and induces a marked biphasic NEFA response to meals. Hence, ghrelin may play a significant role in meal-related substrate utilisation and metabolic flexibility.

Interventions to improve teamwork and communications among healthcare staff.

BACKGROUND: Concern over the frequency of unintended harm to patients has focused attention on the importance of teamwork and communication in avoiding errors. This has led to experiments with teamwork training programmes for clinical staff, mostly based on aviation models. These are widely assumed to be effective in improving patient safety, but the extent to which this assumption is justified by evidence remains unclear. METHODS: A systematic literature review on the effects of teamwork training for clinical staff was performed. Information was sought on outcomes including staff attitudes, teamwork skills, technical performance, efficiency and clinical outcomes. RESULTS: Of 1036 relevant abstracts identified, 14 articles were analysed in detail: four randomized trials and ten non-randomized studies. Overall study quality was poor, with particular problems over blinding, subjective measures and Hawthorne effects. Few studies reported on every outcome category. Most reported improved staff attitudes, and six of eight reported significantly better teamwork after training. Five of eight studies reported improved technical performance, improved efficiency or reduced errors. Three studies reported evidence of clinical benefit, but this was modest or of borderline significance in each case. Studies with a stronger intervention were more likely to report benefits than those providing less training. None of the randomized trials found evidence of technical or clinical benefit. CONCLUSION: The evidence for technical or clinical benefit from teamwork training in medicine is weak. There is some evidence of benefit from studies with more intensive training programmes, but better quality research and cost-benefit analysis are needed.

Control chart methods for monitoring surgical performance: a case study from gastro-oesophageal surgery.

Graphical methods are becoming increasingly used to monitor adverse outcomes from surgical interventions. However, uptake of such methods has largely been in the area of cardiothoracic surgery or in transplants with relatively little impact made in surgical oncology. A number of the more commonly used graphical methods including the Cumulative Mortality plot, Variable Life-Adjusted Display, Cumulative Sum (CUSUM) and funnel plots will be described. Accounting for heterogeneity in case-mix will be discussed and how ignoring case-mix can have considerable consequences. All methods will be illustrated using data from the Scottish Audit of Gastro-Oesophageal Cancer services (SAGOCS) data set.