Definitions of major bleeding for predicting mortality in critically ill adult patients who survived 24 hours while supported with peripheral veno-arterial extracorporeal membrane oxygenation for cardiogenic shock: a comparative historical cohort study

PURPOSE

The severity of bleeding events is heterogeneously defined during peripheral veno-arterial extracorporeal membrane oxygenation (pVA-ECMO). We studied three bleeding definitions in pVA-ECMO: the Extracorporeal Life Support Organization (ELSO)-serious bleeding, the Bleeding Academic Research Consortium (BARC), and the universal definition of postoperative bleeding (UPDB) classifications.

METHODS

We included consecutive adult patients supported by pVA-ECMO for refractory cardiogenic shock admitted to Lille academic hospitals between January 2013 and December 2019. We assessed the association of bleeding definitions with the primary endpoint of 28-day all-cause mortality with the use of multivariate models accounting for time-dependent and competing variables. We compared models' performances using the Harrell's C-Index and the Akaike information criteria.

RESULTS

Twenty-eight-day mortality occurred in 128/308 (42%) 308 patients. The ELSO-serious bleeding (hazard ratio [HR], 1.67; 95% confidence interval [CI], 1.09 to 2.56) and BARC ≥ type 2 (HR, 1.55; 95% CI, 1.01 to 2.37) were associated with 28-day mortality (Harrell's C-index, 0.69; 95% CI, 0.63 to 0.74 for both). Predictors of ELSO-serious bleeding were postcardiotomy, body mass index, baseline platelets count, fibrinogen, and hemoglobin levels.

CONCLUSION

Extracorporeal Life Support Organization-serious bleeding and BARC ≥ type 2 are relevant definitions of major bleeding regarding their association with mortality in critically ill patients who survived the first 24 hr while supported with pVA-ECMO for cardiogenic shock.

STUDY REGISTRATION

CERAR (IRB 00010254-2022-050, Paris, France); first submitted on 18 April 2022.

Association between sarcopenia and risk of major adverse cardiac and cerebrovascular events-UK Biobank database

BACKGROUND

Few studies on the risk of incident major adverse cardiac and cerebrovascular events (MACCEs) in sarcopenia have been reported. The objective was to assess the association between presarcopenia and sarcopenia and a higher risk of MACCEs.

METHODS

This study on the UK Biobank prospective cohort, used data collected between 2006 and 2021. Community-dwelling Caucasian participants aged 37 to 73 years were included if values for Handgrip Strength (HGS) and Skeletal Muscle Index (SMI) were available and if no history of MACCEs was reported. Exposure was assessed using the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) criteria. Muscle strength was measured using HGS, and muscle mass using the SMI. Presarcopenia was defined through the two definitions available in the literature, as low HGS with normal SMI and as normal HGS with low SMI, whereas sarcopenia was defined as low HGS with low SMI. The main outcome was to determine whether presarcopenia and/or sarcopenia were predictors of MACCEs (composite events).

RESULTS

A total of 406,411 included participants (women: 55.7%) were included. At baseline, there were 18,257 (4.7%) presarcopenics-subgroup n°1 (low HGS only), 7940 (2.1%) presarcopenics-subgroup n°2 (low SMI only), and 1124 (0.3%) sarcopenics. Over a median follow-up of 12.1 years (IQR: [11.4; 12.8]), 28,300 participants (7.0%) were diagnosed with at least one event. Compared to NonSarc, presarcopenic (subgroups n°1 and n°2) and sarcopenic status were significantly associated with a higher risk of MACCEs (respectively fully adjusted HRs: HR = 1.25 [95% CI: 1.19; 1.31], HR = 1.33 [95% CI: 1.23; 1.45] and HR = 1.62 [95% CI: 1.34; 1.95]).

CONCLUSIONS

In a community-dwelling population, the risk of MACCEs was higher in both presarcopenic and sarcopenic participants.

Intraoperative Mechanical Power and Postoperative Pulmonary Complications in Noncardiothoracic Elective Surgery Patients: A 10-Year Retrospective Cohort Study

BACKGROUND

Postoperative pulmonary complications is a major issue that affects outcomes of surgical patients. The hypothesis was that the intraoperative ventilation parameters are associated with occurrence of postoperative pulmonary complications.

METHODS

A single-center retrospective cohort study was conducted at the Lille University Hospital, France. The study included 33,701 adults undergoing noncardiac, nonthoracic elective surgery requiring general anesthesia with tracheal intubation between January 2010 and December 2019. Intraoperative ventilation parameters were compared between patients with and without one or more postoperative pulmonary complications (respiratory infection, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm, and aspiration pneumonitis) within 7 days of surgery.

RESULTS

Among 33,701 patients, 2,033 (6.0%) had one or more postoperative pulmonary complications. The lower tidal volume to predicted body weight ratio (odds ratio per -1 ml·kgPBW-1, 1.08; 95% CI, 1.02 to 1.14; P < 0.001), higher mechanical power (odds ratio per 4 J·min-1, 1.37; 95% CI, 1.26 to 1.49; P < 0.001), dynamic respiratory system compliance less than 30 ml·cm H2O (1.30; 95% CI, 1.15 to 1.46; P < 0.001), oxygen saturation measured by pulse oximetry less than 96% (odds ratio, 2.42; 95% CI, 1.97 to 2.96; P < 0.001), and lower end-tidal carbon dioxide (odds ratio per -3 mmHg, 1.06; 95% CI, 1.00 to 1.13; P = 0.023) were independently associated with postoperative pulmonary complications. Patients with postoperative pulmonary complications were more likely to be admitted to the intensive care unit (odds ratio, 12.5; 95% CI, 6.6 to 10.1; P < 0.001), had longer hospital length of stay (subhazard ratio, 0.43; 95% CI, 0.40 to 0.45), and higher in-hospital (subhazard ratio, 6.0; 95% CI, 4.1 to 9.0; P < 0.001) and 1-yr mortality (subhazard ratio, 2.65; 95% CI, 2.33 to 3.02; P < 0.001).

CONCLUSIONS

In the study's population, decreased rather than increased tidal volume, decreased compliance, increased mechanical power, and decreased end-tidal carbon dioxide were independently associated with postoperative pulmonary complications.

EDITOR’S PERSPECTIVE

In France, the organization of perinatal care has a direct influence on the outcome of the mother and the newborn: Contribution from a French nationwide study

OBJECTIVE

To investigate maternal and neonatal outcomes after a delivery in France in 2019, according to hospital characteristics and the impact of distance and time of travel on mother and newborn.

METHODS

All parturients above 18 years of age who delivered in 2019 and were identified in the French health insurance database were included, with their newborns, in this retrospective cohort study. Main outcome measures were Severe Maternal Morbidity score and the Neonatal Adverse Outcome Indicator (NAOI).

RESULTS

Among the 733 052 pregnancies included, 10 829 presented a severe maternal morbidity (1.48%) and 77 237 had a neonatal adverse outcome (10.4%). Factors associated with an unfavorable maternal or neonatal outcome were Obstetric Comorbidity Index, primiparity, and cesarean or instrumental delivery. Prematurity was associated with less severe maternal morbidity but more neonatal adverse outcomes. Time of travel above 30 min was associated with a higher NAOI rate.

CONCLUSIONS

Results suggest the efficiency of regionalization of perinatal care in France, although a difference in both outcomes persists according to unit volume, suggesting the need for a further step in concentrating perinatal care. Perinatal care organization should focus on mapping the territory with high-level, high-volume maternity throughout the territory; this suggests closing down high-volume units and improving low-volume ones to maintain coherent mapping.

Exploring seasonality in catatonia diagnosis: Evidence from a large-scale population study

Catatonia is a severe psychomotor syndrome mainly associated with psychiatric disorders, such as mood disorders and schizophrenia. Seasonal patterns have been described for these psychiatric disorders, and a previous study conducted in South London showed for the first time a seasonal pattern in the onset of catatonia. In this study, we aim to extend those findings to a larger national sample of patients admitted to French metropolitan hospitals, between 2015 and 2022, and to perform subgroup analyses by the main associated psychiatric disorder. A total of 6225 patients diagnosed with catatonia were included. A seasonal pattern for catatonia diagnosis was described, using cosinor models. Two peaks of diagnoses for catatonic cases were described in March and around September-October. Depending on the associated psychiatric disorder, the seasonality of catatonia diagnosis differed. In patients suffering with mood disorders, peaks of catatonia diagnosis were found in March and July. For patients suffering with schizophrenia, no seasonal pattern was found.

Opening the black box of hospitalizations in French high-secure psychiatric forensic units

INTRODUCTION

Basic epidemiological data are rare concerning the activity of specialized forensic psychiatric facilities in France. Here, we investigated the activity of the ten (640 beds) French "units for difficult patients" (unités pour malades difficiles [UMDs]).

METHOD

We used the Programme de médicalisation des systèmes d'information (PMSI) database to describe the characteristics and evolution of psychiatric hospitalisations in UMDs between 2012 and 2021, as well as the age, sex, and principal diagnoses of the patients hospitalized in these facilities.

RESULTS

Between 2012 and 2021, 4857 patients were hospitalized in UMDs (6082 stays). Among them, 897 (18.5%) had more than one stay. The number of admissions ranged from a minimum of 434 to a maximum of 632 per year. The number of discharges ranged from a minimum of 473 to a maximum of 609 per year. The mean length of stay was 13.5 (SD: 22.64) months with a median of 7.3 months (IQR: 4.0-14.4). Among the 6082 stays, 5721 (94.1%) involved male patients. The median age was 33 (IQR: 26-41) years. The most frequent principal psychiatric diagnoses were psychotic disorders and personality disorders.

CONCLUSION

The number of individuals hospitalized in specialized forensic psychiatric facilities has been stable for 10 years in France and remains lower than in most European countries.

Organization of French outpatient psychiatric clinics and delay to appointment

OBJECTIVES

The goals of the study were to obtain a glimpse of the several types of organization of outpatient psychiatric clinics, as well as an overview of delays between a request for and the first appointment. We also look at geographical variability of appointment scheduling delays and assess the impact of the number of new applications on delays.

METHODS

We analyzed information collected from a phone survey conducted with the 103 adult outpatient psychiatric clinics of the French region Hauts-de-France. This survey had a one-week inclusion period in November 2022 and focused on the organization and delays before obtaining an appointment.

RESULTS

This study indicates that organizations seem to be homogeneous. Eighty-one outpatient psychiatric clinics (96.4%) receive with scheduled appointments. The initial evaluation by a nurse followed with a proposal for the organization of care seems to be the common practice. It also appears that primary psychiatric structures were capable of providing a response within a reasonable time frame in making a first appointment with a mental health professional. On the other hand, delays were much longer and heterogeneous for first appointments with psychiatrists and psychologists.

CONCLUSIONS

The organizations of the outpatient psychiatric clinics of the French region Hauts-de-France seem homogeneous. The evaluation by a nurse followed with a care proposal is the common practice. Primary psychiatric structures are able to provide a response within a reasonable time, with half the centers proposing an appointment with a nurse within 10.0 days.

Association Between Sarcopenia and Fracture Risk in a Population From the UK Biobank Database

Studies on the fracture risk in presarcopenic and sarcopenic patients report contradictory results. The objective was to assess whether presarcopenia and sarcopenia are associated with an increase in fracture risk. We conducted a retrospective study using the UK Biobank cohort and the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) criteria. Muscle strength was evaluated using hand-grip strength (HGS) and muscle mass using the skeletal muscle index (SMI; from bioimpedance analysis). Presarcopenia was defined through the two definitions available in the literature, as low HGS with normal SMI and as normal HGS with low SMI, and sarcopenia as low HGS and low SMI. Fracture events were recorded as "fracture" (location compatible with an osteoporotic origin) and "major osteoporotic fracture" (MOF), as listed in the FRAX tool. Associations were assessed using Cox proportional hazards models, adjusted for sarcopenia and osteoporosis risk factors. Adjusted hazard ratios (HRa ) and their 95% confidence intervals (CI) were reported. A total of 387,025 participants (women 54.4%; median age 58.0 years; interquartile range [IQR] 51.0-63.0 years) were included. At baseline, there were 18,257 (4.7%) presarcopenic participants-subgroup 1 (low HGS only), 7940 (2.1%) presarcopenic participants-subgroup 2 (low SMI only), and 1124 (0.3%) sarcopenic participants. Over a median follow-up of 12.0 years (IQR 11.4-12.6 years), 18,300 (4.7%) participants were diagnosed with at least one incident fracture. Presarcopenic (subgroups 1 and 2) and sarcopenic status were significantly associated with a higher risk of fracture (respectively adjusted HRs: HR = 1.26 [1.19-1.33], HR = 1.20 [1.11-1.30], HR = 1.30 [1.08-1.56]) and with a higher risk of MOF (respectively adjusted HRs: HR = 1.30 [1.21-1.40], HR = 1.19 [1.08-1.72], HR = 1.18 [0.93-1.49]). In a middle-aged population, the fracture and MOF risks were higher in both presarcopenic and sarcopenic participants compared with nonsarcopenic participants. © 2023 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Correction: Good practices for clinical data warehouse implementation: A case study in France

[This corrects the article DOI: 10.1371/journal.pdig.0000298.].

Good practices for clinical data warehouse implementation: A case study in France

Real-world data (RWD) bears great promises to improve the quality of care. However, specific infrastructures and methodologies are required to derive robust knowledge and brings innovations to the patient. Drawing upon the national case study of the 32 French regional and university hospitals governance, we highlight key aspects of modern clinical data warehouses (CDWs): governance, transparency, types of data, data reuse, technical tools, documentation, and data quality control processes. Semi-structured interviews as well as a review of reported studies on French CDWs were conducted in a semi-structured manner from March to November 2022. Out of 32 regional and university hospitals in France, 14 have a CDW in production, 5 are experimenting, 5 have a prospective CDW project, 8 did not have any CDW project at the time of writing. The implementation of CDW in France dates from 2011 and accelerated in the late 2020. From this case study, we draw some general guidelines for CDWs. The actual orientation of CDWs towards research requires efforts in governance stabilization, standardization of data schema, and development in data quality and data documentation. Particular attention must be paid to the sustainability of the warehouse teams and to the multilevel governance. The transparency of the studies and the tools of transformation of the data must improve to allow successful multicentric data reuses as well as innovations in routine care.

In France, distance from hospital and health care structure impact on outcome after arthroplasty of the hip for proximal fractures of the femur

BACKGROUND

Hip arthroplasty is a frequently performed procedure in orthopedic surgery, carried out in almost all health structures for two main issues: fracture and coxarthrosis. Even if volume-outcome relationship appeared associated in many surgeries recently, data provided are not sufficient to set surgical thresholds neither than closing down low-volumes centers.

QUESTION

With this study, we wanted to identify surgical, health care-related and territorial factors influencing patient' mortality and readmission after a HA for a femoral fracture in 2018 in France.

PATIENTS AND METHODS

Data were anonymously collected from French nationwide administrative databases. All patients who underwent a hip arthroplasty for a femoral fracture through 2018 were included. Patient outcome was 90-day mortality and 90-day readmission rate after surgery.

RESULTS

Of the 36,252 patients that underwent a HA for fracture in France in 2018, 0.7% died within 90-day year and 1.2% were readmitted. Male and Charlson comorbidity index were associated with a higher 90-day mortality and readmission rate in multivariate analysis. High volume was associated with a lower mortality rate. Neither time of travel nor distance upon health facility were associated with mortality nor with readmission rate in the analysis.

CONCLUSION

Even if volume appears to be associated with lower mortality rate even for longer distance and time of travel, the persistence of exogenous factors not documented in the French databases suggests that regionalization of hip arthroplasty should be organized with caution.

CLINICAL RELEVANCE

As volume-outcome relationship must be interpreted with caution, policy makers should not regionalize such surgery without further investigation.

Description of a French Population of Diabetics Treated Followed up by General Practitioners

In France, the prevalence of treated diabetes has been estimated at 4.6%, or more than 3 million people and 5.2% in Northern France. The reuse of primary care data allows to study outpatient clinical data such as laboratory results and drug prescriptions, which are not documented in claims and hospital databases. In this study, we selected the population of treated diabetics from the Wattrelos primary care data warehouse, in North of France. Firstly, we studied the laboratory results of diabetics by identifying whether the recommendations of the French National Authority for Health (HAS) were respected. In a second step, we studied the prescriptions of diabetics by identifying the oral hypoglycemic agents treatments and insulins treatments. The diabetic population represents 690 patients of the health care center. The recommendations on labortatory are respected for 84% of diabetics. The majority of diabetics are treated with oral hypoglycemic agents 68.6%. As recommended by the HAS, metformin is the first-line treatment in the diabetic population.

OpenDataPsy: An Open-Data Repository with Standardized Storage and Description for Research in Psychiatry

Sharing health data could avoid duplication of effort in data collection, reduce unnecessary costs in future studies, and encourage collaboration and data flow within the scientific community. Several repositories from national institutions or research teams have making their datasets available. These data are mainly aggregated at spatial or temporal level, or dedicated to a specific field. The objective of this work is to propose a standardized storage and description of open datasets for research purposes. For this, we selected 8 publicly accessible datasets, covering the fields of demographics, employment, education and psychiatry. Then, we studied the format, nomenclature (i.e., files and variables names, modalities of recurrent qualitative variables) and descriptions of these datasets and we proposed on common and standardized format and description. We made available these datasets in an open gitlab repository. For each dataset, we proposed the raw data file in its original format, the cleaned data file in csv format, the variables description, the data management script and the descriptive statistics. Statistics are generated according to the type of variables previously documented. After one year of use, we will evaluate with the users if the standardization of the data sets is relevant and how they use the dataset in real life.

Automated Monitoring Reports of the Activity of the French National Professional Suicide Prevention Helpline

The French Professional Suicidal Helpline 3114 was launched on October 1st, 2021. The objective of this study was to implement automated reports of the activity of the suicidal helpline. We developed automated reports and presentations with Rmarkdown. Two formats were developed, national reports to present for a funding agency and regional reports for each calling center. These reports fulfill a critical need to adjust call distribution patterns, identify problems, adjust communication across the territory and ensure that 3114 is delivering the service it is supposed to provide.

Publication Dynamics on Social Media During the Orpea Nursing Homes Scandal: A Twitter Analysis

The release of a book denouncing mistreatment in French nursing home triggered a scandal which was conveyed on social networks. The objectives of this study were to study the temporal trends and dynamics of publication on Twitter during the scandal as well as to identify the main topics of discussion.The first one is spontaneous and completely aligned with the actuality and fed by media and family of residents, while the second one is out of step with current events and fed by the company involved in the scandal.

Impact of COVID-19 pandemic and lockdowns on the consumption of anxiolytics, hypnotics and antidepressants according to age groups: a French nationwide study

BACKGROUND

Previous studies have shown a negative impact of the COVID-19 pandemic and its associated sanitary measures on mental health, especially among adolescents and young adults. Such a context may raise many concerns about the COVID-19 pandemic long-term psychological effects. An analysis of administrative databases could be an alternative and complementary approach to medical interview-based epidemiological surveys to monitor the mental health of the population. We conducted a nationwide study to describe the consumption of anxiolytics, antidepressants and hypnotics during the first year of the COVID-19 pandemic, compared to the five previous years.

METHODS

A historic cohort study was conducted by extracting and analysing data from the French health insurance database between 1 January 2015 and 28 February 2021. Individuals were classified into five age-based classes. Linear regression models were performed to assess the impact of the COVID-19 pandemic period on the number of drug consumers, in introducing an interaction term between time and COVID-19 period.

RESULTS

Since March 2020, in all five age groups and all three drug categories studied, the number of patients reimbursed weekly has increased compared to the period from January 2015 to February 2020. The youngest the patients, the more pronounced the magnitude.

CONCLUSIONS

Monitoring the consumption of psychiatric medications could be of great interest as reliable indicators are essential for planning public health strategies. A post-crisis policy including reliable monitoring of mental health must be anticipated.

Development and usage of an anesthesia data warehouse: lessons learnt from a 10-year project

This paper describes the development and implementation of an anesthesia data warehouse in the Lille University Hospital. We share the lessons learned from a ten-year project and provide guidance for the implementation of such a project. Our clinical data warehouse is mainly fed with data collected by the anesthesia information management system and hospital discharge reports. The data warehouse stores historical and accurate data with an accuracy level of the day for administrative data, and of the second for monitoring data. Datamarts complete the architecture and provide secondary computed data and indicators, in order to execute queries faster and easily. Between 2010 and 2021, 636 784 anesthesia records were integrated for 353 152 patients. We reported the main concerns and barriers during the development of this project and we provided 8 tips to handle them. We have implemented our data warehouse into the OMOP common data model as a complementary downstream data model. The next step of the project will be to disseminate the use of the OMOP data model for anesthesia and critical care, and drive the trend towards federated learning to enhance collaborations and multicenter studies.

The impact of including a medication review in an integrated care pathway: A pilot study

AIM

The objective of the present study was to measure the impact of the intervention of combining a medication review with an integrated care approach on potentially inappropriate medications (PIMs) and hospital readmissions in frail older adults.

METHODS

A cohort of hospitalized older adults enrolled in the French PAERPA integrated care pathway (the exposed cohort) was matched retrospectively with hospitalized older adults not enrolled in the pathway (unexposed cohort) between January 1st, 2015, and December 31st, 2018. The study was an analysis of French health administrative database. The inclusion criteria for exposed patients were admission to an acute care department in a general hospital, age 75 years or over, at least three comorbidities or the prescription of diuretics or oral anticoagulants, discharge alive and performance of a medication review.

RESULTS

For the study population (n = 582), the mean ± standard deviation age was 82.9 ± 4.9 years, and 380 (65.3%) were women. Depending on the definition used, the overall median number of PIMs ranged from 2 [0;3] on admission to 3 [0;3] at discharge. The intervention was not associated with a significant difference in the mean number of PIMs. Patients in the exposed cohort were half as likely to be readmitted to hospital within 30 days of discharge relative to patients in the unexposed cohort.

CONCLUSION

Our results show that a medication review was not associated with a decrease in the mean number of PIMs. However, an integrated care intervention including the medication review was associated with a reduction in the number of hospital readmissions at 30 days.

Definition of a Practical Taxonomy for Referencing Data Quality Problems in Health Care Databases

INTRODUCTION

 Health care information systems can generate and/or record huge volumes of data, some of which may be reused for research, clinical trials, or teaching. However, these databases can be affected by data quality problems; hence, an important step in the data reuse process consists in detecting and rectifying these issues. With a view to facilitating the assessment of data quality, we developed a taxonomy of data quality problems in operational databases.

MATERIAL

 We searched the literature for publications that mentioned "data quality problems," "data quality taxonomy," "data quality assessment," or "dirty data." The publications were then reviewed, compared, summarized, and structured using a bottom-up approach, to provide an operational taxonomy of data quality problems. The latter were illustrated with fictional examples (though based on reality) from clinical databases.

RESULTS

 Twelve publications were selected, and 286 instances of data quality problems were identified and were classified according to six distinct levels of granularity. We used the classification defined by Oliveira et al to structure our taxonomy. The extracted items were grouped into 53 data quality problems.

DISCUSSION

 This taxonomy facilitated the systematic assessment of data quality in databases by presenting the data's quality according to their granularity. The definition of this taxonomy is the first step in the data cleaning process. The subsequent steps include the definition of associated quality assessment methods and data cleaning methods.

CONCLUSION

 Our new taxonomy enabled the classification and illustration of 53 data quality problems found in hospital databases.

Identification of early symptoms of endometriosis through the analysis of online social networks: A social media study

Objective

Endometriosis is a complex full-body inflammation disease with an average time to diagnosis of 7-10 years. Social networks give opportunity to patient to openly discuss about their condition, share experiences, and seek advice. Thus, data from social media may provide insightful data about patient's experience. This study aimed at applying a text-mining approach to online social networks in order to identify early signs associated with endometriosis.

Methods

An automated exploration technique of online forums was performed to extract posts. After a cleaning step of the built corpus, we retrieved all symptoms evoked by women, and connected them to the MedDRA dictionary. Then, temporal markers allowed targeting only the earliest symptoms. The latter were those evoked near a marker of precocity. A co-occurrence approach was further applied to better account for the context of evocations.

Results

Results were visualised using the graph-oriented database Neo4j. We collected 7148 discussions threads and 78,905 posts from 10 French forums. We extracted 41 groups of contextualised symptoms, including 20 groups of early symptoms associated with endometriosis. Among these groups of early symptoms, 13 were found to portray already known signs of endometriosis. The remaining 7 clusters of early symptoms were limb oedema, muscle pain, neuralgia, haematuria, vaginal itching, altered general condition (i.e. dizziness, fatigue, nausea) and hot flush.

Conclusion

We pointed out some additional symptoms of endometriosis qualified as early symptoms, which can serve as a screening tool for prevention and/or treatment purpose. The present findings offer an opportunity for further exploration of early biological processes triggering this disease.

Mild pulmonary hemodynamic alterations in patients with systemic sclerosis: relevance of the new 2022 ESC/ERS definition of pulmonary hypertension and impact on mortality

Background and objective

The definition of pre-capillary pulmonary hypertension (PH) has been modified, with lowering of the mean pulmonary arterial pressure (mPAP) threshold from 25 to 20 mmHg and addition of a mandatory criterion of pulmonary vascular resistance (PVR) ≥ 2 Wood units (WU). Our objectives were: 1/ to estimate the proportion of patients reclassified as having pre-capillary PH when using the new 2022 ESC/ERS hemodynamic criteria (i.e. mPAP 21-24 mmHg and PVR ≥ 2 WU), and to describe their clinical characteristics and outcome; and 2/ to study the relationship between PVR and survival in patients with mPAP > 20 mmHg.

Methods

We retrospectively analyzed consecutive SSc patients included in our National Reference Center for a first right-heart catheterization between 2003 and 2018. The association between survival and PVR was studied using smoothing splines.

Results

We included 126 SSc patients with mPAP > 20 mmHg. Among them, 16 (13%) had a baseline mPAP value between 21 and 24 mmHg and PVR ≥ 2 mmHg and were reclassified as pre-capillary PH; 10 of which (62%) raised their mPAP ≥ 25 mmHg during follow-up. In patients with mPAP > 20 mmHg, we observed a linear relation between PVR and mortality for values < 6 WU.

Conclusion

A significant proportion of SSc patients is reclassified as having pre-capillary PH with the new 2022 ESC/ERS hemodynamic definition. Lowering the PVR threshold from 3 to 2 WU captures patients at risk of raising their mPAP > 25 mmHg, with a possibly less severe disease.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12931-022-02205-4.

Data Management for Health Data Reuse: Proposal of a Standard Workflow and a R Tutorial with Jupyter Notebook

The data collected in the clinical registries or by data reuse require some modifications in order to suit the research needs. Several common operations are frequently applied to select relevant patients across the cohort, combine data from multiple sources, add new variables if needed and create unique tables depending on the research purpose. We carried out a qualitative survey by conducting semi-structured interviews with 7 experts in data reuse and proposed a standard workflow for health data management. We implemented a R tutorial based on a synthetic data set using Jupyter Notebook for a better understanding of the data management workflow.

Master's Degree in Health Data Science: Implementation and Assessment After Five Years

Health data science is an emerging discipline that bridges computer science, statistics and health domain knowledge. This consists of taking advantage of the large volume of data, often complex, to extract information to improve decision-making. We have created a Master's degree in Health Data Science to meet the growing need for data scientists in companies and institutions. The training offers, over two years, courses covering computer science, mathematics and statistics, health and biology. With more than 60 professors and lecturers, a total of 835 hours of classes (not including the mandatory 5 months of internship per year), this curriculum has enrolled a total of 53 students today. The feedback from the students and alumni allowed us identifying new needs in terms of training, which may help us to adapt the program for the coming academic years. In particular, we will offer an additional module covering data management, from the edition of the clinical report form to the implementation of a data warehouse with an ETL process. Git and application lifecycle management will be included in programming courses or multidisciplinary projects.

Early hyperoxia and 28-day mortality in patients on venoarterial ECMO support for refractory cardiogenic shock: a bicenter retrospective propensity score-weighted analysis

Background

The mortality rate for a patient with a refractory cardiogenic shock on venoarterial (VA) extracorporeal membrane oxygenation (ECMO) remains high, and hyperoxia might worsen this prognosis. The objective of the present study was to evaluate the association between hyperoxia and 28-day mortality in this setting.

Methods

We conducted a retrospective bicenter study in two French academic centers. The study population comprised adult patients admitted for refractory cardiogenic shock. The following arterial partial pressure of oxygen (PaO₂) variables were recorded for 48 h following admission: the absolute peak PaO₂ (the single highest value measured during the 48 h), the mean daily peak PaO₂ (the mean of each day’s peak values), the overall mean PaO₂ (the mean of all values over 48 h), and the severity of hyperoxia (mild: PaO₂ < 200 mmHg, moderate: PaO₂ = 200–299 mmHg, severe: PaO₂ ≥ 300 mmHg). The main outcome was the 28-day all-cause mortality. Inverse probability weighting (IPW) derived from propensity scores was used to reduce imbalances in baseline characteristics.

Results

From January 2013 to January 2020, 430 patients were included and assessed. The 28-day mortality rate was 43%. The mean daily peak, absolute peak, and overall mean PaO₂ values were significantly higher in non-survivors than in survivors. In a multivariate logistic regression analysis, the mean daily peak PaO₂, absolute peak PaO₂, and overall mean PaO₂ were independent predictors of 28-day mortality (adjusted odds ratio [95% confidence interval per 10 mmHg increment: 2.65 [1.79–6.07], 2.36 [1.67–4.82], and 2.85 [1.12–7.37], respectively). After IPW, high level of oxygen remained significantly associated with 28-day mortality (OR = 1.41 [1.01–2.08]; P = 0.041).

Conclusions

High oxygen levels were associated with 28-day mortality in patients on VA-ECMO support for refractory cardiogenic shock. Our results confirm the need for large randomized controlled trials on this topic.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13054-022-04133-7.

"Book Music" Representation for Temporal Data, as a Part of the Feature Extraction Process: A Novel Approach to Improve the Handling of Time-Dependent Data in Secondary Use of Healthcare Structured Data

Book music is extensively used in street organs. It consists of thick cardboard, containing perforated holes specifying the musical notes. We propose to represent clinical time-dependent data in a tabular form inspired from this principle. The sheet represents a statistical individual, each row represents a binary time-dependent variable, and each hole denotes the "true" value. Data from electronic health records or nationwide medical-administrative databases can then be represented: demographics, patient flow, drugs, laboratory results, diagnoses, and procedures. This data representation is suitable for survival analysis (e.g., Cox model with repeated outcomes and changing covariates) and different types of temporal association rules. Quantitative continuous variables can be discretized, as in clinical studies. The "book music" approach could become an intermediary step in feature extraction from structured data. It would enable to better account for time in analyses, notably for historical cohort analyses based on healthcare data reuse.

Awareness and Perception of Google® Reviews Among French GPs

The objective of this study was to assess the perception of French general practitioners (GPs) and the impact of the Google® online physician rating system. We questioned the French GPs with a self-administered questionnaire. A total of 412 GPs had answered the complete questionnaire. 83.25% of respondents did not validate the relevance of the physician rating websites. The most decried limitations were the lack of validity of these opinions and the negative impact on physicians.

Implementation of a Data Warehouse in Primary Care: First Analyses with Elderly Patients

The implementation of clinical data warehouses has advanced in recent years. The standardization of clinical data in these warehouses has made it possible to carry out multicenter studies and to formalize the clinical vocabulary. However, there is limited insight into a patient's overall care pathway in the clinical domain. Regarding primary care data, the implementation of this type of warehouse in a routine way is hindered in particular by the analysis of textual data provided by general practitioners during patient consultations. In our study we collected primary care data for standardization in a data warehouse. The purpose of this analysis was to assess the feasibility of analyzing primary care data, and particularly to study the consultations and prescriptions of the elderly patient contained in our primary care data warehouse.

Goupile: A New Paradigm for the Development and Implementation of Clinical Report Forms

Despite the increasing computerization of hospital information systems, segments of patient care are still in paper format. Data extracted automatically from the hospital databases for one specific project are thus supplemented by data collected manually. Data collection tools are usually developed entirely, which requires computer knowledge and is tedious, or automatically from metadata or drag and drop controls, which is limiting in terms of functionality. To facilitate this manual collection, we developed a free and open-source tool for creating forms that does not require advanced computer skills, offers rich features, and is quickly implemented, tested and deployed. It was implemented for 15 projects and supported thousands of daily users for a complex interactive study at the national level.

Automated Twitter Extraction and Visual Analytics with Dashboards: Development and First Experimentations

Information found in the social media may help to set up infoveillance and track epidemics, identify high-risk behaviours, or assess trends or feelings about a subject or event. We developed a dashboard to enable novice users to easily and autonomously extract and analyze data from Twitter. Eleven users tested the dashboard and considered the tool to be highly usable and useful. They were able to conduct the research they wanted and appreciated being able to use this tool without having to program.

Jupyter Notebooks for Introducing Data Science to Novice Users

Data science is a bridge discipline involving computer science, statistics, and knowledge of the health field. We developed a Jupyter Notebook to enable novice users to easily and autonomously analyze data from social networks. We conducted an experimentation with non-programmer students. They had to adapt a R Notebook and complete 14 questions and to perform descriptive analyses. The average rate of correct answers was 90.7. Jupyter Notebook enabled novice users to easily and autonomously analyze data from Twitter.

Psychiatric Hospitalizations of People Found Not Criminally Responsible on Account of Mental Disorder in France: A Ten-Year Retrospective Study (2011-2020)

Background

Criminal responsibility is a key concept in the criminal sanctioning of people diagnosed with mental health disorders who have committed crimes. In France, based on the recommendations of one or more expert psychiatrists, a judge can declare a person not criminally responsible on account of mental disorder (NCRMD) if, at the time of the offense, the person was presenting a psychiatric disorder that abolished or altered his/her capacity for discernment and/or ability to control his/her actions. In such a case, the judge also generally orders an involuntary psychiatric hospitalization. The objectives of this study were to (1) describe longitudinal retrospective administrative data of psychiatric hospitalizations for people found NCRMD, (2) identify the age, sex, and principal diagnoses of these individuals, and (3) characterize the trajectories of their psychiatric care before and after NCRMD psychiatric hospitalization.

Methods

We used discharge reports from the French national hospital database called Programme de médicalisation des systèmes d'information (PMSI) to gather longitudinal data that describe psychiatric hospitalizations for people found NCRMD between 2011 and 2020, the age, sex, and principal diagnoses of these patients, the length of their hospitalization, and the trajectories of their psychiatric care before and after their NCRMD psychiatric hospitalization.

Results

We identified 3,020 patients who were hospitalized for psychiatric care after having been found NCRMD between 2011 and 2020. The number of admissions on these grounds has remained stable over this period, ranging from 263 in 2011 to 227 in 2021. They were mostly young men diagnosed with a psychotic disorder (62%). The majority (87%) were hospitalized in general psychiatric hospitals, and only 13% were admitted to maximum-security units (Unités pour malades difficiles, UMD). The median duration of hospitalization for these patients was 13 months. Our results show that 73% of the patients had already been hospitalized prior to their NRCMD hospitalization. The rehospitalization rate within 5 years of discharge from NCRMD psychiatric hospitalization was 62%.

Conclusion

We conducted the first study investigating the psychiatric hospital treatment of people declared NCRMD in France. There is an urgent need for further studies to investigate the clinical characteristics of these patients.

Psychiatric hospitalisations for people who are incarcerated, 2009-2019: An 11-year retrospective longitudinal study in France

BACKGROUND

Despite the poor mental health status of people who are incarcerated, few studies have examined the number of psychiatric hospitalisations in this population. Since 2010, France has progressively opened nine full-time inpatient psychiatric wards exclusively for people who are incarcerated, called "specially adapted hospital units" (unités hospitalières spécialement aménagées, UHSAs, 440 beds). This study aimed to present the annual rates of psychiatric hospitalisations and primary psychiatric diagnoses among people who are incarcerated in France from 2009 to 2019.

METHODS

We used discharge reports from the French national hospital database to describe longitudinal retrospective administrative data of psychiatric hospitalisations for people in jail and prison between 2009 and 2019, the age, sex, and principal diagnoses of these patients, the proportion of voluntary versus involuntary care, and the interactions between UHSAs and other facilities.

FINDINGS

Between Jan 1, 2009, and Dec 31, 2019, 32,228 (92.2% men, n = 29,721; 7.8% women, n = 2 507) incarcerated people were hospitalised for psychiatric care (64,481 stays). The main diagnoses were psychotic disorders (27.4%), personality disorders (23.2%), and stress-related disorders (20.2%). The annual number of incarcerated people hospitalised in psychiatric care increased from 3263 in 2009 to 4914 in 2019. The gradual increase in the activity of UHSAs (300 hospitalisations in 2010 versus 3252 in 2019) was not associated with a reduction in the rate of hospitalisation of incarcerated people in local psychiatric hospitals.

INTERPRETATION

The creation of psychiatric hospitals specifically dedicated to the prison population has not stopped the hospitalisation of people who are incarcerated at psychiatric hospitals. These findings suggest that access to psychiatric hospitalisation remains problematic for people who are incarcerated in France.

FUNDING

There was no funding source for this study.

Visualization of medical concepts represented using word embeddings: a scoping review

Background

Analyzing the unstructured textual data contained in electronic health records (EHRs) has always been a challenging task. Word embedding methods have become an essential foundation for neural network-based approaches in natural language processing (NLP), to learn dense and low-dimensional word representations from large unlabeled corpora that capture the implicit semantics of words. Models like Word2Vec, GloVe or FastText have been broadly applied and reviewed in the bioinformatics and healthcare fields, most often to embed clinical notes or activity and diagnostic codes. Visualization of the learned embeddings has been used in a subset of these works, whether for exploratory or evaluation purposes. However, visualization practices tend to be heterogeneous, and lack overall guidelines.

Objective

This scoping review aims to describe the methods and strategies used to visualize medical concepts represented using word embedding methods. We aim to understand the objectives of the visualizations and their limits.

Methods

This scoping review summarizes different methods used to visualize word embeddings in healthcare. We followed the methodology proposed by Arksey and O’Malley (Int J Soc Res Methodol 8:19–32, 2005) and by Levac et al. (Implement Sci 5:69, 2010) to better analyze the data and provide a synthesis of the literature on the matter.

Results

We first obtained 471 unique articles from a search conducted in PubMed, MedRxiv and arXiv databases. 30 of these were effectively reviewed, based on our inclusion and exclusion criteria. 23 articles were excluded in the full review stage, resulting in the analysis of 7 papers that fully correspond to our inclusion criteria. Included papers pursued a variety of objectives and used distinct methods to evaluate their embeddings and to visualize them. Visualization also served heterogeneous purposes, being alternatively used as a way to explore the embeddings, to evaluate them or to merely illustrate properties otherwise formally assessed.

Conclusions

Visualization helps to explore embedding results (further dimensionality reduction, synthetic representation). However, it does not exhaust the information conveyed by the embeddings nor constitute a self-sustaining evaluation method of their pertinence.

Exploration of the core protein network under endometriosis symptomatology using a computational approach

BACKGROUND

Endometriosis is defined by implantation and invasive growth of endometrial tissue in extra-uterine locations causing heterogeneous symptoms, and a unique clinical picture for each patient. Understanding the complex biological mechanisms underlying these symptoms and the protein networks involved may be useful for early diagnosis and identification of pharmacological targets.

METHODS

In the present study, we combined three approaches (i) a text-mining analysis to perform a systematic search of proteins over existing literature, (ii) a functional enrichment analysis to identify the biological pathways in which proteins are most involved, and (iii) a protein-protein interaction (PPI) network to identify which proteins modulate the most strongly the symptomatology of endometriosis.

RESULTS

Two hundred seventy-eight proteins associated with endometriosis symptomatology in the scientific literature were extracted. Thirty-five proteins were selected according to degree and betweenness scores criteria. The most enriched biological pathways associated with these symptoms were (i) Interleukin-4 and Interleukin-13 signaling (p = 1.11 x 10^-16), (ii) Signaling by Interleukins (p = 1.11 x 10^-16), (iii) Cytokine signaling in Immune system (p = 1.11 x 10^-16), and (iv) Interleukin-10 signaling (p = 5.66 x 10^-15).

CONCLUSION

Our study identified some key proteins with the ability to modulate endometriosis symptomatology. Our findings indicate that both pro- and anti-inflammatory biological pathways may play important roles in the symptomatology of endometriosis. This approach represents a genuine systemic method that may complement traditional experimental studies. The current data can be used to identify promising biomarkers for early diagnosis and potential therapeutic targets.

Transformation and Evaluation of the MIMIC Database in the OMOP Common Data Model: Development and Usability Study

Background

In the era of big data, the intensive care unit (ICU) is likely to benefit from real-time computer analysis and modeling based on close patient monitoring and electronic health record data. The Medical Information Mart for Intensive Care (MIMIC) is the first open access database in the ICU domain. Many studies have shown that common data models (CDMs) improve database searching by allowing code, tools, and experience to be shared. The Observational Medical Outcomes Partnership (OMOP) CDM is spreading all over the world.

Objective

The objective was to transform MIMIC into an OMOP database and to evaluate the benefits of this transformation for analysts.

Methods

We transformed MIMIC (version 1.4.21) into OMOP format (version 5.3.3.1) through semantic and structural mapping. The structural mapping aimed at moving the MIMIC data into the right place in OMOP, with some data transformations. The mapping was divided into 3 phases: conception, implementation, and evaluation. The conceptual mapping aimed at aligning the MIMIC local terminologies to OMOP's standard ones. It consisted of 3 phases: integration, alignment, and evaluation. A documented, tested, versioned, exemplified, and open repository was set up to support the transformation and improvement of the MIMIC community's source code. The resulting data set was evaluated over a 48-hour datathon.

Results

With an investment of 2 people for 500 hours, 64% of the data items of the 26 MIMIC tables were standardized into the OMOP CDM and 78% of the source concepts mapped to reference terminologies. The model proved its ability to support community contributions and was well received during the datathon, with 160 participants and 15,000 requests executed with a maximum duration of 1 minute.

Conclusions

The resulting MIMIC-OMOP data set is the first MIMIC-OMOP data set available free of charge with real disidentified data ready for replicable intensive care research. This approach can be generalized to any medical field.

Transforming Anesthesia Data Into the Observational Medical Outcomes Partnership Common Data Model: Development and Usability Study

BACKGROUND

Electronic health records (EHRs, such as those created by an anesthesia management system) generate a large amount of data that can notably be reused for clinical audits and scientific research. The sharing of these data and tools is generally affected by the lack of system interoperability. To overcome these issues, Observational Health Data Sciences and Informatics (OHDSI) developed the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) to standardize EHR data and promote large-scale observational and longitudinal research. Anesthesia data have not previously been mapped into the OMOP CDM.

OBJECTIVE

The primary objective was to transform anesthesia data into the OMOP CDM. The secondary objective was to provide vocabularies, queries, and dashboards that might promote the exploitation and sharing of anesthesia data through the CDM.

METHODS

Using our local anesthesia data warehouse, a group of 5 experts from 5 different medical centers identified local concepts related to anesthesia. The concepts were then matched with standard concepts in the OHDSI vocabularies. We performed structural mapping between the design of our local anesthesia data warehouse and the OMOP CDM tables and fields. To validate the implementation of anesthesia data into the OMOP CDM, we developed a set of queries and dashboards.

RESULTS

We identified 522 concepts related to anesthesia care. They were classified as demographics, units, measurements, operating room steps, drugs, periods of interest, and features. After semantic mapping, 353 (67.7%) of these anesthesia concepts were mapped to OHDSI concepts. Further, 169 (32.3%) concepts related to periods and features were added to the OHDSI vocabularies. Then, 8 OMOP CDM tables were implemented with anesthesia data and 2 new tables (EPISODE and FEATURE) were added to store secondarily computed data. We integrated data from 5,72,609 operations and provided the code for a set of 8 queries and 4 dashboards related to anesthesia care.

CONCLUSIONS

Generic data concerning demographics, drugs, units, measurements, and operating room steps were already available in OHDSI vocabularies. However, most of the intraoperative concepts (the duration of specific steps, an episode of hypotension, etc) were not present in OHDSI vocabularies. The OMOP mapping provided here enables anesthesia data reuse.

The Relationship between the Immigrant Rate and Health Status in the General Population in France

Mostly studied at the individual level, the analysis of immigrants' health status at a populational level may provide a different perspective to investigate, including social determinants as part of the explanation of the relationship between them and health status in France. We analyzed freely accessible databases curated by French public bodies. The dependent variables were death rate and mean age at death. Immigrant rate and covariates associated with either of the outcomes were explored in univariate and multivariate models. Linear models were used to explain the mean age at death, whereas tobit models were used to explain the death rate. The immigrant rate varied markedly from one department to another, as did healthcare accessibility, population's age profile, and economic covariates. Considering univariate models, almost all the studied covariates were significantly associated with comes. The immigrant rate was associated with a lower death rate and a lower age at death. In multivariate models, the immigrant rate was no longer associated with age at death but was still negatively associated with the death rate. In France, the departments with a higher proportion of immigrants were those with a lower death rate, possibly because immigrants are attracted to economically thriving areas.

Automated Generation of Individual and Population Clinical Pathways with the OMOP Common Data Model

INTRODUCTION

Clinical pathways represents the sequence of interventions from which the patients benefit during their encounters with health care structures. There are several complex issues which make it difficult to represent these pathways (e.g. high numbers of patients, heterogeneity of variables).

METHODS

We developed a tool to automate the representation of clinical pathways, from an individual and population points of view, and based on the OMOP CDM. The tool implemented the Sankey diagram in three stages: (i) data extraction, (ii) generation of individual sequence of steps and (iii) aggregation of sequence to obtain the population-level diagram. We tested the tool with three surgery procedures: the total hip replacement, the coronary bypass and the transcatheter aortic valve implantation.

RESULTS

The tool provided different ways of visualizing pathways depending on the question asked: a pathway before a surgery, the pathway of deceased patients or the complete pathway with different steps of interest.

DISCUSSION

We proposed a tool automating the representation of the clinical pathways, and reducing complexity of visualization. Representations are detailed from an individual and population points of view. It has been tested with three surgical procedures. The tool functionalities will be extended to cover a greater number of use cases.

Evaluation of Anti-Activated Factor X Activity and Activated Partial Thromboplastin Time Relations and Their Association with Bleeding and Thrombosis during Veno-Arterial ECMO Support: A Retrospective Study

BACKGROUND

We aimed to investigate the relationship between anti-activated Factor X (anti-FXa) and activated Partial Thromboplastin Time (aPTT), and its modulation by other haemostasis co-variables during veno-arterial extracorporeal membrane oxygenation (VA-ECMO) support. We further investigated their association with serious bleeding and thrombotic complications.

METHODS

This retrospective single-center study included 265 adults supported by VA-ECMO for refractory cardiogenic shock from January 2015 to June 2019. The concordance of anti-FXa and aPTT and their correlations were assessed in 1699 paired samples. Their independent associations with serious bleeding or thrombotic complications were also analysed in multivariate analysis.

RESULTS

The concordance rate of aPTT with anti-FXa values was 50.7%, with 39.3% subtherapeutic aPTT values. However, anti-FXa and aPTT remained associated (β = 0.43 (95% CI 0.4-0.45) 10^-2 IU/mL, p < 0.001), with a significant modulation by several biological co-variables. There was no association between anti-FXa nor aPTT values with serious bleeding or with thrombotic complications.

CONCLUSION

During VA-ECMO, although anti-FXa and aPTT were significantly associated, their values were highly discordant with marked sub-therapeutic aPTT values. These results should favour the use of anti-FXa. The effect of biological co-variables and the failure of anti-FXa and aPTT to predict bleeding and thrombotic complications underline the complexity of VA-ECMO-related coagulopathy.

Mid-Term Survival and Risk Factors Associated With Myocardial Injury After Fenestrated and/or Branched Endovascular Aortic Aneurysm Repair

OBJECTIVE

Myocardial injury after non-cardiac surgery (MINS) is an independent predictor of post-operative mortality in non-cardiac surgery patients and may increase health costs. Few data are available for MINS in vascular surgery patients, in general, and those undergoing fenestrated/branched endovascular aortic repairs (F/BEVAR), in particular. The incidence of MINS after F/BEVAR, the associated risk factors, and prognosis have not been determined. The aim of the present study was to help fill these knowledge gaps.

METHODS

A single centre, retrospective study was carried out at a high volume F/BEVAR centre in a university hospital. Adult patients who underwent F/BEVAR between October 2010 and December 2018 were included. A high sensitivity troponin T (HsTnT) assay was performed daily in the first few post-operative days. MINS was defined as a HsTnT level ≥ 14 ng/L (MINS₁₄) or ≥ 20 ng/L (MINS₂₀). After assessment of the incidence of MINS, survival up to two years was estimated in a Kaplan-Meier analysis and the groups were compared according to MINS status. A secondary aim was to identify predictors of MINS.

RESULTS

Of the 387 included patients, 240 (62.0%) had MINS₁₄ and 166 (42.9%) had MINS₂₀. In multivariable Cox models, both conditions were significantly associated with poor two year survival (MINS14: adjusted hazard ratio [aHR] 2.15, 95% confidence interval [CI] 1.10 - 4.19; MINS₂₀: aHR 2.43, 95% CI 1.36 - 4.34). In a multivariable logistic regression, age, revised cardiac risk index, duration of surgery, pre-operative estimated glomerular filtration rate (eGFR), and haemoglobin level were independent predictors of MINS.

CONCLUSION

After F/BEVAR surgery, the incidence of MINS was particularly high, regardless of the definition considered (MINS₁₄ or MINS₂₀). MINS was significantly associated with poor two year survival. The modifiable predictors identified were duration of surgery, eGFR, and haemoglobin level.

Importance of previous hospital stays on the risk of hospital re-admission in older adults: a real-life analysis of the PAERPA study population

BACKGROUND

consideration of the first hospital re-admission only and failure to take account of previous hospital stays, which are the two significant limitations when studying risk factors for hospital re-admission. The objective of the study was to use appropriate statistical models to analyse the impact of previous hospital stays on the risk of hospital re-admission among older patients.

METHODS

an exhaustive analysis of hospital discharge and health insurance data for a cohort of patients participating in the PAERPA ('Care Pathways for Elderly People at Risk of Loss of Personal Independence') project in the Hauts de France region of France. All patients aged 75 or over were included. All data on hospital re-admissions via the emergency department were extracted. The risk of unplanned hospital re-admission was estimated by applying a semiparametric frailty model, the risk of death by applying a time-dependent semiparametric Cox regression model.

RESULTS

a total of 24,500 patients (median [interquartile range] age: 81 [77-85]) were included between 1 January 2015 and 31 December 2017. In a multivariate analysis, the relative risk (95% confidence interval [CI]) of hospital re-admission rose progressively from 1.8 (1.7-1.9) after one previous hospital stay to 3.0 (2.6-3.5) after five previous hospital stays. The relative risk [95%CI] of death rose slowly from 1.1 (1.07-1.11) after one previous hospital stay to 1.3 (1.1-1.5) after five previous hospital stays.

CONCLUSION

analyses of the risk of hospital re-admission in older adults must take account of the number of previous hospital stays. The risk of death should also be analysed.

A Decentralized Framework for Biostatistics and Privacy Concerns

Biostatistics and machine learning have been the cornerstone of a variety of recent developments in medicine. In order to gather large enough datasets, it is often necessary to set up multi-centric studies; yet, centralization of measurements can be difficult, either for practical, legal or ethical reasons. As an alternative, federated learning enables leveraging multiple centers' data without actually collating them. While existing works generally require a center to act as a leader and coordinate computations, we propose a fully decentralized framework where each center plays the same role. In this paper, we apply this framework to logistic regression, including confidence intervals computation. We test our algorithm on two distinct clinical datasets split among different centers, and show that it matches results from the centralized framework. In addition, we discuss possible privacy leaks and potential protection mechanisms, paving the way towards further research.