Random forest of perfect trees: concept, performance, applications and perspectives.

MOTIVATION: The principle of Breiman's random forest (RF) is to build and assemble complementary classification trees in a way that maximizes their variability. We propose a new type of random forest that disobeys Breiman's principles and involves building trees with no classification errors in very large quantities. We used a new type of decision tree that uses a neuron at each node as well as an in-innovative half Christmas tree structure. With these new RFs, we developed a score, based on a family of ten new statistical information criteria, called Nguyen information criteria (NICs), to evaluate the predictive qualities of features in three dimensions. RESULTS: The first NIC allowed the Akaike information criterion to be minimized more quickly than data obtained with the Gini index when the features were introduced in a logistic regression model. The selected features based on the NICScore showed a slight advantage compared to the support vector machines-recursive feature elimination (SVM-RFE) method. We demonstrate that the inclusion of artificial neurons in tree nodes allows a large number of classifiers in the same node to be taken into account simultaneously and results in perfect trees without classification errors. AVAILABILITY AND IMPLEMENTATION: The methods used to build the perfect trees in this article were implemented in the 'ROP' R package, archived at https://cran.r-project.org/web/packages/ROP/index.html. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Electronic Medication Regimen Complexity Index at admission and complications during hospitalization in medical wards: a tool to improve quality of care?

OBJECTIVE: Adverse events during hospitalization are a major worry considering their frequency and their burden. Many could be avoided by immediate identification of at-risk patients at admission and adapted prevention. The complexity of a patient's medication regimen immediately available at admission is a good indicator of the complexity of the patient's condition. This study aims to determine whether the electronic Medication Regimen Complexity Index (MRCI) at admission is associated with complications during hospitalization. DESIGN: We performed a multilevel logistic regression model, adjusted for age and sex. SETTING: Premier Perspective™ database, a clinical and financial information system from 417 US hospitals. PARTICIPANTS: Adults hospitalized for more than 3 days in a medical ward and included in Premier's Perspective™ database for 2006. INTERVENTION(S): Multilevel logistic regression. MAIN OUTCOME MEASURE: Association of the MRCI and complications during hospitalization, defined as in-hospital death, hospital-acquired infection, pressure ulcers; and need for highly technical healthcare, identified as the secondary introduction of catecholamines. RESULTS: In total, 1 592 383 admissions were included. The median MRCI at admission was 13 [interquartile range: 9-19]. The higher the MRCI, the higher the adjusted odds ratio of the following: in-hospital mortality, hospital-acquired infections, pressure ulcers and the secondary introduction of catecholamines. CONCLUSIONS: Our results suggested that the MRCI at admission was correlated with patient complexity, independent of age. Considering that patients with complex conditions pose a heavier workload for staff, measuring MRCI at admission could be used to allocate resources in medical wards at an institutional level. The MRCI might be a useful tool to assess the management of care.

[Alternatives to heparin and protamine anticoagulation for cardiopulmonary bypass in cardiac surgery]. / Les différentes alternatives d'anticoagulation au couple héparine/protamine en chirurgie cardiaque sous circulation extra-corporelle.

PURPOSE: Heparin anticoagulation followed by protamine reversal is commonly used in cardiopulmonary bypass (CPB) cardiac procedures, but this strategy has some limitations. The primary objective of this study was to determine the reliable alternatives for anticoagulation during CPB for cardiac surgery. For each drug proposed, the secondary objectives were to outline the main advantages and disadvantages, to propose a therapeutic protocol, and to provide a cost-benefit analysis. SOURCE: A systematic review of the literature was performed between September 2012 and December 2013. It was based on the protocol established by the "Cochrane collaboration Handbook". Twenty articles were analyzed. The Thériaque database from the University Hospital of Grenoble made the economic analysis possible. PRINCIPAL FINDINGS: Seven alternative anticoagulation strategies were considered: danaparoid sodium, lepirudin, argatroban, bivalirudin, ancrod, idraparinux, and EP217609. Danaparoid sodium has issues with individual variability. Several studies (EVOLUTION-ON, CHOOSE-ON) proposed a reliable therapeutic protocol for bivalirudin. Ancrod resulted in an increase in the transfusion of blood products. Direct thrombin inhibitors offer a promising alternative. EP217609 is a synthetic anticoagulant currently undergoing Phase IIa clinical trials. It is an indirect inhibitor of factor Xa, a direct inhibitor of free and bound thrombin, and can be neutralized by avidin. CONCLUSIONS: The ideal anticoagulation strategy for cardiac surgery with CPB does not exist. Heparin and protamine remain the gold standard for anticoagulation therapy. To date, bivalirudin is the most promising molecule despite its high cost and lack of a readily available antagonist.

Analysis of screening for neonatal hypoglycemia in large-for-gestational-age newborns without risk factors, and proposed changes in practice at Grenoble University Hospital.

INTRODUCTION: The aim of this study was to evaluate the relevance of screening for neonatal hypoglycemia as it is currently performed, in order to improve the comfort of newborns by reducing the number of painful procedures such as venipunctures or capillary punctures. The primary objective was to determine the prevalence of neonatal hypoglycemia in large-for-gestational-age newborns. The secondary objective was to determine a threshold percentile of birth weight for optimal screening for hypoglycemia. METHODS: We performed a descriptive, cross-sectional, single-center study, based on a structured review of obstetrical records from 11 January 2017 to 21 January 2020, from the maternity department of the University Hospital of Grenoble. Eligible neonates were large-for-gestational-age (birth weight >90th percentile) at term (37-42 weeks) without other risk factors for hypoglycemia. The primary outcome was the prevalence of neonates with capillary or venous glucose levels <2.2 mmol/L in the first 48 hours of life. We performed a sensitivity and specificity analysis of the birth weight percentile as a determinant of the threshold for hypoglycemia detection (ROC curve, area under the curve, Youden index, Brier score, Hosmer-Lemeshow test). RESULTS: In all, 19.2% of the newborns presented at least one hypoglycemic episode during the first 48 hours of life, and 75.7% of the hypoglycemic episodes occurred at 1 hour of life. The cut-off percentile that seemed most appropriate for screening was determined to be the 97th percentile of birth weight (AUC=0.64; 95% CI: 0.52-0.75). CONCLUSIONS: Our statistical model is robust and allows us to state that the currently used birth weight percentile threshold can be revised upwards. Thus, the protocol for neonatal hypoglycemia screening can be updated to improve the comfort of newborns at risk of hypoglycemia.

Active older adults goal setting outcomes for engaging in a physical activity app and the motivation characteristics of these goals (MOVEAGE-ACT).

Approximately 70% of older adults do not meet physical activity (PA) guidelines. While many interventions, are used in promoting PA, few target older adults or include substantial behavioural change techniques. Setting PA goals is often used but there is less research on goal setting outcomes, like improving health, preventing age effects, improving flexibility, goals that have been associated with increased likelihood of maintenance of PA. To understand the concept more fully in this cohort, the aim of this study was to identify older adults' goal setting outcomes - the purpose of engaging in a PA app and through analysis determine the motivation characteristics of these. A cross-sectional, qualitative online survey was completed by 24, 60+, community dwelling, mostly active, French and Irish older adults. Thematic template analysis was used, and the motivation of these outcomes was assessed using the Self-Determination Theory of Motivation. The themes were: improving/staying healthy or physically active, maintaining functional aspects of physical health, continuing to do the things I want, sustaining mental wellbeing, and preventing disease and aging. Individuals cited goal setting outcomes that were generic, specific or both, and goals related to maintenance of PA and prevention of aging decline, were cited most. The motivation characteristics of these goals in mostly active older adults were autonomous and internally driven. Interventions, including apps, for older adults that encourage them to set specific goal setting outcomes/purposes for PA, are likely to generate stronger internally driven motivation, enhance ownership and participation, and may therefore increase effectiveness.

Learning from informed consent litigation to improve practices: A systematic review.

OBJECTIVE: To describe the reasons that lead judges to qualify malpractice as a lack of information, then rule in favour or not of the health professional (HP). METHODS: We conducted a systematic review of case law relating to the breach of disclosure obligations over a ten-year period from 2010 to 2020. We used 3 legal databases: Légifrance, Dalloz and Lexis 360, all identified as the most exhaustive. RESULTS: Of the 514 law cases included: judges found malpractice owing to lack of information in 377 (73.3%) cases. Among the latter, malpractices were lack of risk information (N = 257, 68.2%), lack of proof of information (N = 243, 64.5%) and/or lack of information on therapeutic alternatives (N = 49, 13.0%). These malpractices resulted in a conviction of the HP in 268 (71.1%) of the cases. CONCLUSION: Case law is an important source of information for improving the quality of HP, lawyers, and judges' practices. PRACTICE IMPLICATIONS: This review suggests that.

Comparative evaluation of socioeconomic insecurity in peripheral and coronary artery disease patients.

BACKGROUND: Cardiovascular disease represents the leading cause of death worldwide. Socioeconomic deprivation is a risk factor for cardiovascular disease. We have previously shown that precariousness was more frequent in symptomatic peripheral artery disease (PAD) patients than in the general population. According to a previous study, coronary artery disease (CAD) patients have a higher level of education than CAD with PAD, but no study directly compared the level of precariousness in PAD and CAD patients. AIM: To measure and compare the level of socioeconomic insecurity in patients suffering from symptomatic PAD with those suffering from isolated CAD, i.e without symptomatic PAD. METHODS: We conducted an observational, cohort, prospective, multicenter study. Patients suffering from symptomatic PAD or CAD were recruited through the medical or surgical vascular or cardiology departments, or the vascular rehabilitation center. The EPICES score and the INSEE parameters were used for analysis. The individual is considered precarious when his or her score is greater than or equal to 30. Cardiovascular risk factors and peripheral arterial disease stages were also collected. RESULTS: In total, 230 patients were included. According to the EPICES score, 47.8% [95%CI, 38.7-56.7] of patients with symptomatic PAD were in a precarious situation compared to 17.4% [95%CI, 10.5-24.3] of patients suffering from isolated CAD (P<0.001). The mean EPICES score was 33.3 (SD 22.5) in the PAD and 16.9 (SD 17.02) in the CAD population, respectively (P<0.001). In the PAD population, the level of education was low, with an under-representation of patients with a baccalaureate or higher education degree: 21.7% [95%CI, 14.2-29.3] vs. 41.7% [95%CI, 32.7-50.7] in the PAD and CAD populations, respectively. There was also an under-representation of executives and intellectual and intermediate professions in the PAD population, 18.3% [95%CI, 11.2-25.3], compared to the CAD population, 31.3% [95%CI, 22.8-39.8]. CONCLUSION: PAD patients are more precarious than patients suffering from CAD. A better detection of socioeconomic deprivation in patients suffering from peripheral arterial disease could allow comprehensive care and thus hope for an improvement in terms of morbidity and mortality.

Freedom of Master's Degree Students to Study in Health Curricula: Switching to Optimized Blended Learning as a Solution!

OBJECTIVES: The Grenoble (France) Master's degree in health includes 17 sub-specialty programs, 120 separate teaching units (TUs) and caters for up to 400 students per year. We present the pedagogical transition to blended learning based on flipped classroom initiated in 2010 to overcome the pedagogical limitations of classical lectures. METHODS: The pedagogical organization of each TU is based on the weekly and sequential implementation of five sequences. The first three sequences comprise the learning stages of (1) self-learning on knowledge capsules, (2) interactive on-line questions and votes of interest, and (3) interactive on-site training and explanation meetings. The last two sequences include the evaluation stages with (4) positioning tests, and (5) an anonymous evaluation of the TU allowing access to personalized follow-ups. This pedagogical sequence is completed with a final certification on a tablet computer. RESULTS: The systematic evaluation and debriefing sessions of TUs gave us a clear SWOT vision of the revised Master's degree in health. The feedback was very positive from students, teachers, and the institution, which encourages us to move forward in this transition. Nonetheless, some of this positive feedback was unexpected, such as the ease of managing mobile learners (e.g. Erasmus, International internship) or personalized reinforcement. CONCLUSION: Our results indicate that a switch to blended learning is feasible in a large Master program, with improvements on student/teacher equity and for the institution.

Is virtual reality effective to teach prevention of surgical site infections in the operating room? study protocol for a randomised controlled multicentre trial entitled VIP Room study.

INTRODUCTION: Some surgical site infections (SSI) could be prevented by following adequate infection prevention and control (IPC) measures. Poor compliance with IPC measures often occurs due to knowledge gaps and insufficient education of healthcare professionals. The education and training of SSI preventive measures does not usually take place in the operating room (OR), due to safety, and organisational and logistic issues. The proposed study aims to compare virtual reality (VR) as a tool for medical students to learn the SSI prevention measures and adequate behaviours (eg, limit movements…) in the OR, to conventional teaching. METHODS AND ANALYSIS: This protocol describes a randomised controlled multicentre trial comparing an educational intervention based on VR simulation to routine education. This multicentre study will be performed in three universities: Grenoble Alpes University (France), Imperial College London (UK) and University of Heidelberg (Germany). Third-year medical students of each university will be randomised in two groups. The students randomised in the intervention group will follow VR teaching. The students randomised in the control group will follow a conventional education programme. Primary outcome will be the difference between scores obtained at the IPC exam at the end of the year between the two groups. The written exam will be the same in the three countries. Secondary outcomes will be satisfaction and students' progression for the VR group. The data will be analysed with intention-to-treat and per protocol. ETHICS AND DISSEMINATION: This study has been approved by the Medical Education Ethics Committee of the London Imperial College (MEEC1920-172), by the Ethical Committee for the Research of Grenoble Alpes University (CER Grenoble Alpes-Avis-2019-099-24-2) and by the Ethics Committee of the Medical Faculty of Heidelberg University (S-765/2019). Results will be published in peer-reviewed medical journals, communicated to participants, general public and all relevant stakeholders.

Biostatistics Disruptive Acculturation Through Serious Gaming: A New Hope.

Biostatistics is one of the transversal subjects that all future doctors must acquire and master. Nonetheless, it is a subject that has the reputation of being difficult, which has not been able to be corrected even with the application of new pedagogical methods such as blended learning. We address this problem with our acculturative and disruptive approach in the form of a serious game scenario in clinical research that integrates biostatistics with our R4Web adapted tools. Our approach was launched in 2008 for the second year of medical school. Here we describe this LOE scenario for serious game including the biostatistics disruptive acculturation task and present its new international version.

Contact allergy in chronic leg ulcers: results of a multicentre study carried out in 423 patients and proposal for an updated series of patch tests.

BACKGROUND: There is a lack of prospective studies investigating contact sensitization in patients with chronic leg ulcers. OBJECTIVES: To determine the frequency of contact sensitization in patients with chronic leg ulcers using a special series of patch tests and to determine whether the number of sensitizations was correlated with the duration of the chronic leg ulcers. PATIENTS/METHODS: Multicentre study carried out in patients with chronic leg ulcers; patch tests with the European baseline series and with an additional 34 individual allergens or mixes and 3 commercial products. RESULTS: Of the 423 patients (301 women, 122 men, mean age 68.5 years) with chronic leg ulcers, 308 (73%) had at least one positive patch test with 3.65 positive patch tests per patient. The main allergens were Myroxylon pereirae (41%), fragrance mix I (26.5%), antiseptics (20%), and corticosteroids (8%). The number of positive tests per patient was not correlated with the cause of ulcer but was increased with the duration of the ulcer with a statistical difference between the group of the <1 year compared with the group >10 years duration. CONCLUSIONS: From this large prospective multicentre study, polysensitization is frequent in patients with chronic leg ulcers, increasing with the duration of the ulcer. We propose avoidance of topical antiseptics and ointments containing perfumes in patients with chronic leg ulcers and an updated patch test series for investigating these patients.

Robustness in regulatory interaction networks. A generic approach with applications at different levels: physiologic, metabolic and genetic.

Regulatory interaction networks are often studied on their dynamical side (existence of attractors, study of their stability). We focus here also on their robustness, that is their ability to offer the same spatiotemporal patterns and to resist to external perturbations such as losses of nodes or edges in the networks interactions architecture, changes in their environmental boundary conditions as well as changes in the update schedule (or updating mode) of the states of their elements (e.g., if these elements are genes, their synchronous coexpression mode versus their sequential expression). We define the generic notions of boundary, core, and critical vertex or edge of the underlying interaction graph of the regulatory network, whose disappearance causes dramatic changes in the number and nature of attractors (e.g., passage from a bistable behaviour to a unique periodic regime) or in the range of their basins of stability. The dynamic transition of states will be presented in the framework of threshold Boolean automata rules. A panorama of applications at different levels will be given: brain and plant morphogenesis, bulbar cardio-respiratory regulation, glycolytic/oxidative metabolic coupling, and eventually cell cycle and feather morphogenesis genetic control.

Disruptive Analysis of Closed Questions Assessments at Medical School, Interest of Massive Multi-Choice Tests.

Descriptive statistics, Classical Test Theory (CTT), and Item Response Theory (IRT) are used to analyze learning. These results are a powerful synthesis for students to evaluate their progression into their personal pathway of learning, for teachers to improve the quality of different multiple-choice questions (MCQ) series, and for administration to assist in making pedagogical decisions. Our study explores tools to evaluate MCQ tests, and MCQs. Answers to tutorials and final examination of students of the first year of health contest were collected and analyzed.

Blended Learning for French Health Students: Does Acceptance of a Learning Management System Influence Students' Self-Efficacy?

A learning management system (LMS) used for the initial training of health professionals has been rejected by students. Our study aimed to explain the reason of this rejection. We performed this evaluation on a sample of health students in 2012 and 2017 (n = 144). We used scales from the literature (Technology Acceptance Model, General Self-Efficacy Scale, LMS-Self-Efficacy Scale), and studied the social representation of the LMS. The system seemed accessible and useful, but unfortunately with similarities to the system used in a traumatic student environment. Health students using the system did not have a lower self-efficacy. Although the LMS seemed relevant to students, its initial rejection might have been due to a confounding context that created confusion in the acceptability of the tool. To conclude, there is a need to create new dematerialized course formats but with strong tutorship to improve the usage of the technologies by students.

Consequences of Switching to Blended Learning: The Grenoble Medical School Key Elements.

INTRODUCTION: In 2006, the Grenoble-Alpes University Medical School decided to switch the learning paradigm of the first year to a blended learning model based on a flipped classroom with a continuous dual assessment system providing personal follow-up. We report a descriptive analysis of two pedagogical models. METHODS: The innovative blended learning model is divided into 5 week-sequences of learning, starting with a series of knowledge capsules, following with Interactive On Line Questions, Interactive On Site Training and an Explanation Meeting. The fourth and final steps are the dual assessment system that prepares for the final contest and the personal weekly follow-up. The data were extracted from the information systems over 17 years, during which the same learning model was applied. RESULTS: With the same student workload, the hourly knowledge/skills ratio decreased to approximately 50% with the blended learning model. The teachers' workload increased significantly in the first year (+70%), and then decreased each year (reaching -20%). Furthermore, the type of education has also changed for the teacher, from an initial hourly knowledge/skill ratio of 3, to a ratio of 1/3 with the new model after a few years. The institution also needed to resize the classroom from a large amphitheatre to small interactive learning spaces. DISCUSSION: There is a significant initial effort required to establish this model both for the teachers and for the institution, which have different needs and costs However, the satisfaction rates and the demand for extension to the other curriculums from medics and paramedics learners indicate that this model provides the enhanced learning paradigm of the future.

Does the acceptance of hybrid learning affect learning approaches in France?

PURPOSE: Acceptance of a learning technology affects students' intention to use that technology, but the influence of the acceptance of a learning technology on learning approaches has not been investigated in the literature. A deep learning approach is important in the field of health, where links must be created between skills, knowledge, and habits. Our hypothesis was that acceptance of a hybrid learning model would affect students' way of learning. METHODS: We analysed these concepts, and their correlations, in the context of a flipped classroom method using a local learning management system. In a sample of all students within a single year of study in the midwifery program (n= 38), we used 3 validated scales to evaluate these concepts (the Study Process Questionnaire, My Intellectual Work Tools, and the Hybrid E-Learning Acceptance Model: Learner Perceptions). RESULTS: Our sample had a positive acceptance of the learning model, but a neutral intention to use it. Students reported that they were distractible during distance learning. They presented a better mean score for the deep approach than for the superficial approach (P< 0.001), which is consistent with their declared learning strategies (personal reorganization of information; search and use of examples). There was no correlation between poor acceptance of the learning model and inadequate learning approaches. The strategy of using deep learning techniques was moderately correlated with acceptance of the learning model (rs= 0.42, P= 0.03). CONCLUSION: Learning approaches were not affected by acceptance of a hybrid learning model, due to the flexibility of the tool. However, we identified problems in the students' time utilization, which explains their neutral intention to use the system.

Comparison of the impact of cardiovascular guidelines on a working population.

UNLABELLED: The influence of guidelines is more and more important in the medicine using standardised knowledge as Clinical Practice Guideline (CPG). OBJECTIVES: 1) Determine the impact of different CPGs on a working population 2) Using the Framingham risk calculation as gold standard to check the CPG classifications. METHODS: This work is done in the context of two hypertension management CPGs published in 1997 and 2000, and one dyslipidemia CPG by the French agency HAS. A French regional study sample of 2817 patients was selected. The results of the classification system by CPG were compared to those given by the chosen gold standard: Framingham cardiovascular risk calculation. RESULTS: The HBP CPG concerns more of the quarter of the working population. The mean of the cardiovascular risk calculation for the patients with drug treatment is higher than for the patients with no drug treatment and for those not screened by the CPG. Some patients with a very high calculated risk, are not selected by the screening CPG based on too few or on not adequate variables. CONCLUSIONS: For the two High Blood Pressure Management CPGs, the selected patients are not distributed the same way in the two CPGs. Observed agreement is poor between the two hypertension CPG versions. Screening people for specific health care should be based on risk calculation not on few variables. Feeding back our results to the CPG creator group should reduce the variability of recommendations.

Prediction of intra-hospital mortality after severe trauma: which pre-hospital score is the most accurate?

PURPOSE: Computing trauma scores in the field allows immediate severity assessment for appropriate triage. Two pre-hospital scores can be useful in this context: the Triage-Revised Trauma Score (T-RTS) and the Mechanism, Glasgow, Age and arterial Pressure (MGAP) score. The Trauma Revised Injury Severity Score (TRISS), not applicable in the pre-hospital setting, is the reference score to predict in-hospital mortality after severe trauma. The aim of this study was to compare T-RTS, MGAP and TRISS in a cohort of consecutive patients admitted in the Trauma system of the Northern French Alps(TRENAU). MATERIALS AND METHODS: From 2009 to 2011, 3260 patients with suspected severe trauma according to the Vittel criteria were included in the TRENAU registry. All data necessary to compute T-RTS, MGAP and TRISS were collected in patients admitted to one level-I, two level-II and ten level-III trauma centers. The primary endpoint was death from any cause during hospital stay. Discriminative power of each score to predict mortality was measured using receiver operating curve (ROC) analysis. To test the relevancy of each score for triage, we also tested their sensitivity at usual cut-offs. We expected a sensitivity higher than 95% to limit undertriage. RESULTS: The TRISS score showed the highest area under the ROC curve (0.95 [CI 95% 0.94-0.97], p<0.01). Pre-hospital MGAP score had significantly higher AUC compared to T-RTS (0.93 [CI 95% 0.91-0.95] vs 0.86 [CI 95% 0.83-0.89], respectively, p<0.01). MGAP score<23 had a sensitivity of 88% to detect mortality. Sensitivities of T-RTS<12 and TRISS<0.91 were 79% and 87%, respectively. DISCUSSION/CONCLUSION: Pre-hospital calculation of the MGAP score appeared superior to T-RTS score in predicting intra-hospital mortality in a cohort of trauma patients. Although TRISS had the highest AUC, this score can only be available after hospital admission. These findings suggest that the MGAP score could be of interest in the pre-hospital setting to assess patients' severity. However, its lack of sensitivity indicates that MGAP should not replace the decision scheme to direct the most severe patients to level-I trauma center.

PROK1 Level in the Follicular Microenvironment: A New Noninvasive Predictive Biomarker of Embryo Implantation.

CONTEXT: Prokineticin 1 (PROK1), also called endocrine gland-derived vascular endothelial growth factor, is a well-established regulator of endometrial receptivity and placental development. However, its clinical usefulness as a noninvasive predictive biomarker of embryo implantation is yet to be validated. OBJECTIVE: The main objective of this article was to determine the relationship between PROK1 levels in the follicular fluid (FF) and fertilization culture media (FCM) and the reproductive outcome in patients who received a first conventional in vitro fertilization-embryo transfer. The secondary objective was to characterize the expression of PROK1 and its receptors (PROKRs) in the human follicular microenvironment. DESIGN AND SETTING: We conducted a prospective study between January 2013 and June 2015 at the University Hospital of Grenoble. PATIENTS: A total of 135 infertile in vitro fertilization patients and 10 women undergoing ovarian tissue cryopreservation were included. INTERVENTIONS: The PROK1 concentration was measured by ELISA in FF and FCM collected on the day of oocyte retrieval and the day of the oocyte denudation step, respectively. Follicular expression of the PROK1/PROKR system was determined by immunohistochemistry, RT-quantitative PCR, and ELISA. MAIN OUTCOME MEASURE: Assessment of the clinical pregnancy rates was the main outcome. RESULTS: FF and FCM PROK1 levels were significantly higher in the embryo implantation group (P < .001) and were predictive of subsequent embryo implantation (area under the receiver operating characteristic curve, 0.91 [95% confidence interval, 0.81-1.00], P = .001; and 0.88 [0.72-1.00], P = .001, respectively). FF and FCM PROK1 levels remain similar irrespective of the embryo morphokinetic parameters (P = .71 and P = .83, respectively). The PROK1/PROKR system is expressed during human folliculogenesis. CONCLUSIONS: PROK1 levels in FF and FCM could constitute new predictive noninvasive markers of successful embryo implantation in conventional in vitro fertilization-embryo transfer.

[The keys to success in French Medical National Ranking Examination: Integrated training activities in teaching hospital and medical school]. / Les clefs de la réussite à Grenoble aux Épreuves classantes nationales : les formations intégrées hospitalo-universitaire.

INTRODUCTION: The National Ranking Examination (NRE) is the key to the choice of career and specialty for future physicians; it lets them choose their place of employment in a specialty and an hospital for their internship. It seems interesting to model the success factors to this exam for the medical students from Grenoble University. METHODS: For each of the medical students at Grenoble University who did apply to the NRE in 2012, data have been collected about their academic background and personal details from the administration of the University. A simple logistic regression with success set as being ranked in the first 2000 students, then a polytomous logistic regression, have been performed. RESULTS: The 191 students in the models are 59% female, 25 years old in average (SD 1.8). The factors associated to a ranking in the first 2000 are: not repeating the PCEM1 class (odds ratio [OR] 2.63, CI95: [1.26; 5.56]), performing nurse practice during internships (OR=1.27 [1.00; 1.62]), being ranked in the first half of the class for S3 pole (OR=6.04 [1.21; 30.20] for the first quarter, OR=5.65 [1.15; 27.74] for the second quarter) and being in the first quarter at T5 pole (OR=3.42 [1.08; 10.82]). CONCLUSION: Our study finds four factors independently contributing to the success at NRE: not repeating PCEM1, performing nurse practice and being ranked in the top of the class at certain academic fields. The AUC is 0.76 and student accuracy is more than 80%. However, some items, for example repeating DCEM4 or participating in NRE mock exams, have no influence on success. A different motivation should be a part of the explanation… As these analysed data are mainly institutional, they are accurate and reliable. The polytomic logistic model, sharing 3 factors with the simple logistic model, replace a performing nurse practice factor's by a grant recipient factor.