Forewarning of hypotensive events using a Bayesian artificial neural network in neurocritical care.

Traumatically brain injured (TBI) patients are at risk from secondary insults. Arterial hypotension, critically low blood pressure, is one of the most dangerous secondary insults and is related to poor outcome in patients. The overall aim of this study was to get proof of the concept that advanced statistical techniques (machine learning) are methods that are able to provide early warning of impending hypotensive events before they occur during neuro-critical care. A Bayesian artificial neural network (BANN) model predicting episodes of hypotension was developed using data from 104 patients selected from the BrainIT multi-center database. Arterial hypotension events were recorded and defined using the Edinburgh University Secondary Insult Grades (EUSIG) physiological adverse event scoring system. The BANN was trained on a random selection of 50% of the available patients (n = 52) and validated on the remaining cohort. A multi-center prospective pilot study (Phase 1, n = 30) was then conducted with the system running live in the clinical environment, followed by a second validation pilot study (Phase 2, n = 49). From these prospectively collected data, a final evaluation study was done on 69 of these patients with 10 patients excluded from the Phase 2 study because of insufficient or invalid data. Each data collection phase was a prospective non-interventional observational study conducted in a live clinical setting to test the data collection systems and the model performance. No prediction information was available to the clinical teams during a patient's stay in the ICU. The final cohort (n = 69), using a decision threshold of 0.4, and including false positive checks, gave a sensitivity of 39.3% (95% CI 32.9-46.1) and a specificity of 91.5% (95% CI 89.0-93.7). Using a decision threshold of 0.3, and false positive correction, gave a sensitivity of 46.6% (95% CI 40.1-53.2) and specificity of 85.6% (95% CI 82.3-88.8). With a decision threshold of 0.3, > 15 min warning of patient instability can be achieved. We have shown, using advanced machine learning techniques running in a live neuro-critical care environment, that it would be possible to give neurointensive teams early warning of potential hypotensive events before they emerge, allowing closer monitoring and earlier clinical assessment in an attempt to prevent the onset of hypotension. The multi-centre clinical infrastructure developed to support the clinical studies provides a solid base for further collaborative research on data quality, false positive correction and the display of early warning data in a clinical setting.

Cerebral Perfusion Pressure Variability Between Patients and Between Centres.

INTRODUCTION: The aim of this analysis was to investigate to what extent median cerebral perfusion pressure (CPP) differs between severe traumatic brain injury (TBI) patients and between centres, and whether the 2007 change in CPP threshold in the Brain Trauma Foundation guidelines is reflected in patient data collected at several centres over different time periods. METHODS: Data were collected from the Brain-IT database, a multi-centre project between 2003 and 2005, and from a recent project in four centres between 2009 and 2013. For patients nursed with their head up at 30° and with the blood pressure transducer at atrium level, CPP was corrected by 10 mmHg. Median CPP, interquartile ranges and total CPP ranges over the monitoring time were calculated per patient and per centre. RESULTS: Per-centre medians pre-2007 were situated between 50 and 70 mmHg in 6 out of 16 centres, while 10 centres had medians above 70 mmHg and 4 above 80 mmHg. Post-2007, three out of four centres had medians between 60 and 70 mmHg and one above 80 mmHg. One out of two centres with data pre- and post-2007 shifted from a median CPP of 76 mmHg to 60 mmHg, while the other remained at 68-67 mmHg. CONCLUSIONS: CPP data are characterised by a high inter-individual variability, but the data also suggest differences in CPP policies between centres. The 2007 guideline change may have affected policies towards lower CPP in some centres. Deviations from the guidelines occur in the direction of CPP > 70 mmHg.

Visualizing Cerebrovascular Autoregulation Insults and Their Association with Outcome in Adult and Paediatric Traumatic Brain Injury.

OBJECTIVE: The aim of this study is to assess visually the impact of duration and intensity of cerebrovascular autoregulation insults on 6-month neurological outcome in severe traumatic brain injury. MATERIAL AND METHODS: Retrospective analysis of prospectively collected minute-by-minute intracranial pressure (ICP) and mean arterial blood pressure data of 259 adult and 99 paediatric traumatic brain injury (TBI) patients from multiple European centres. The relationship of the 6-month Glasgow Outcome Scale with cerebrovascular autoregulation insults (defined as the low-frequency autoregulation index above a certain threshold during a certain time) was visualized in a colour-coded plot. The analysis was performed separately for autoregulation insults occurring with cerebral perfusion pressure (CPP) below 50 mmHg, with ICP above 25 mmHg and for the subset of adult patients that did not undergo decompressive craniectomy. RESULTS: The colour-coded plots showed a time-intensity-dependent association with outcome for cerebrovascular autoregulation insults in adult and paediatric TBI patients. Insults with a low-frequency autoregulation index above 0.2 were associated with worse outcomes and below -0.6 with better outcomes, with and approximately exponentially decreasing transition curve between the two intensity thresholds. All insults were associated with worse outcomes when CPP was below 50 mmHg or ICP was above 25 mmHg. CONCLUSIONS: The colour-coded plots indicate that cerebrovascular autoregulation is disturbed in a dynamic manner, such that duration and intensity play a role in the determination of a zone associated with better neurological outcome.

Early Detection of Increased Intracranial Pressure Episodes in Traumatic Brain Injury: External Validation in an Adult and in a Pediatric Cohort.

OBJECTIVE: A model for early detection of episodes of increased intracranial pressure in traumatic brain injury patients has been previously developed and validated based on retrospective adult patient data from the multicenter Brain-IT database. The purpose of the present study is to validate this early detection model in different cohorts of recently treated adult and pediatric traumatic brain injury patients. DESIGN: Prognostic modeling. Noninterventional, observational, retrospective study. SETTING AND PATIENTS: The adult validation cohort comprised recent traumatic brain injury patients from San Gerardo Hospital in Monza (n = 50), Leuven University Hospital (n = 26), Antwerp University Hospital (n = 19), Tübingen University Hospital (n = 18), and Southern General Hospital in Glasgow (n = 8). The pediatric validation cohort comprised patients from neurosurgical and intensive care centers in Edinburgh and Newcastle (n = 79). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The model's performance was evaluated with respect to discrimination, calibration, overall performance, and clinical usefulness. In the recent adult validation cohort, the model retained excellent performance as in the original study. In the pediatric validation cohort, the model retained good discrimination and a positive net benefit, albeit with a performance drop in the remaining criteria. CONCLUSIONS: The obtained external validation results confirm the robustness of the model to predict future increased intracranial pressure events 30 minutes in advance, in adult and pediatric traumatic brain injury patients. These results are a large step toward an early warning system for increased intracranial pressure that can be generally applied. Furthermore, the sparseness of this model that uses only two routinely monitored signals as inputs (intracranial pressure and mean arterial blood pressure) is an additional asset.

Artefact in Physiological Data Collected from Patients with Brain Injury: Quantifying the Problem and Providing a Solution Using a Factorial Switching Linear Dynamical Systems Approach.

INTRODUCTION: High-resolution, artefact-free and accurately annotated physiological data are desirable in patients with brain injury both to inform clinical decision-making and for intelligent analysis of the data in applications such as predictive modelling. We have quantified the quality of annotation surrounding artefactual events and propose a factorial switching linear dynamical systems (FSLDS) approach to automatically detect artefact in physiological data collected in the neurological intensive care unit (NICU). METHODS: Retrospective analysis of the BrainIT data set to discover potential hypotensive events corrupted by artefact and identify the annotation of associated clinical interventions. Training of an FSLDS model on clinician-annotated artefactual events in five patients with severe traumatic brain injury. RESULTS: In a subset of 187 patients in the BrainIT database, 26.5 % of potential hypotensive events were abandoned because of artefactual data. Only 30 % of these episodes could be attributed to an annotated clinical intervention. As assessed by the area under the receiver operating characteristic curve metric, FSLDS model performance in automatically identifying the events of blood sampling, arterial line damping and patient handling was 0.978, 0.987 and 0.765, respectively. DISCUSSION: The influence of artefact on physiological data collected in the NICU is a significant problem. This pilot study using an FSLDS approach shows real promise and is under further development.

Decompressive craniectomy as a second/third-tier intervention in traumatic brain injury: A multicenter observational study.

OBJECTIVES: RESCUEicp studied decompressive craniectomy (DC) applied as third-tier option in severe traumatic brain injury (TBI) patients in a randomized controlled setting and demonstrated a decrease in mortality with similar rates of favorable outcome in the DC group compared to the medical management group. In many centers, DC is being used in combination with other second/third-tier therapies. The aim of the present study is to investigate outcomes from DC in a prospective non-RCT context. METHODS: This is a prospective observational study of 2 patient cohorts: one from the University Hospitals Leuven (2008-2016) and one from the Brain-IT study, a European multicenter database (2003-2005). In thirty-seven patients with refractory elevated intracranial pressure who underwent DC as a second/third-tier intervention, patient, injury and management variables including physiological monitoring data and administration of thiopental were analysed, as well as Extended Glasgow Outcome score (GOSE) at 6 months. RESULTS: In the current cohorts, patients were older than in the surgical RESCUEicp cohort (mean 39.6 vs. 32.3; p < 0.001), had higher Glasgow Motor Score on admission (GMS < 3 in 24.3% vs. 53.0%; p = 0.003) and 37.8% received thiopental (vs. 9.4%; p < 0.001). Other variables were not significantly different. GOSE distribution was: death 24.3%; vegetative 2.7%; lower severe disability 10.8%; upper severe disability 13.5%; lower moderate disability 5.4%; upper moderate disability 2.7%, lower good recovery 35.1%; and upper good recovery 5.4%. The outcome was unfavorable in 51.4% and favorable in 48.6%, as opposed to 72.6% and 27.4% respectively in RESCUEicp (p = 0.02). CONCLUSION: Outcomes in DC patients from two prospective cohorts reflecting everyday practice were better than in RESCUEicp surgical patients. Mortality was similar, but fewer patients remained vegetative or severely disabled and more patients had a good recovery. Although patients were older and injury severity was lower, a potential partial explanation may be in the pragmatic use of DC in combination with other second/third-tier therapies in real-life cohorts. The findings underscore that DC maintains an important role in managing severe TBI.

Improving Health and Addressing Social Determinants of Health Through Hospital Partnerships.

Hospitals and health systems are forming partnerships to develop an integrated social network of services that better address the needs of their surrounding communities and their social determinants of health (SDOH). There is little research on the association of these partnered services with hospital outcomes. This study examined the association between hospital social need partnerships and activities to improve hospital and community outcomes. A secondary cross-sectional design to analyze 2021 census data of nonfederal short-term acute care hospitals in the United States was utilized. Data were obtained from the American Hospital Association. Four multilevel logistic regression models were used to analyze data from 1005 hospitals. The authors found that hospital partnership type differed in association to social need outcomes. They found that hospitals with a partnership with health insurance providers were more likely to have better health outcomes. Hospitals partnered with health insurance providers, local organizations addressing housing insecurity, local businesses, or chambers of commerce were more likely to have decreased health care costs. Hospitals partnered with health care providers, health insurance providers, local organizations providing legal assistance, or law enforcement/safety forces were more likely to have decreased utilization of hospital services. However, hospitals partnered with other local or state government or social service organizations were less likely to indicate decreased utilization of services. Many hospitals and health systems across the United States are screening for SDOH and are advancing health care delivery and improving the community's overall health and well-being by identifying unmet social needs and partnering with the community to address them.

Early warning of EUSIG-defined hypotensive events using a Bayesian Artificial Neural Network.

BACKGROUND: Hypotension is recognized as a potentially damaging secondary insult after traumatic brain injury. Systems to give clinical teams some early warning of likely hypotensive instability could be added to the range of existing techniques used in the management of this group of patients. By using the Edinburgh University Secondary Insult Grades (EUSIG) definitions for -hypotension (systolic arterial pressure <90 mmHg OR mean arterial -pressure <70 mmHg) we collected a group of ∼2,000 events by analyzing the Brain-IT database. We then constructed a Bayesian Artificial Neural Network (an advanced statistical modeling technique) that is able to provide some early warning when trained on this previously collected demographic and physiological data. MATERIALS AND METHODS: Using EUSIG defined event data from the Brain-IT database, we identified a Bayesian artificial neural network (BANN) topology and constructed a series of datasets using a group of clinically guided input variables. This allowed us to train a BANN, which was then tested on an unseen set of patients from the Brain-IT database. The initial tests used a particularly harsh assessment criterion whereby a true positive prediction was only allowed if the BANN predicted an upcoming event to the exact minute. We have now developed the system to the point where it is about to be used in a two-stage Phase II clinical trial and we are also researching a more realistic assessment technique. KEY RESULTS: We have constructed a BANN that is able to provide early warning to the clinicians based on a model that uses information from the physiological inputs; systolic and mean arterial pressure and heart rate; and demographic variables age and gender. We use 15-min SubWindows starting at 15 and 30 min before an event and process mean, slope and standard deviations. Based on 10 simulation runs, our current sensitivity is 36.25% (SE 1.31) with a specificity of 90.82% (SE 0.85). Initial results from a Phase I clinical study shows a model sensitivity of 40.95% (SE 6%) and specificity of 86.46% (SE 3%) Although this figure is low it is considered clinically useful for this dangerous condition, provided the false positive rate can be kept sufficiently low as to be practical in an intensive care environment. CONCLUSION: We have shown that using advanced statistical modeling techniques can provide clinical teams with useful information that will assist clinical care.

Trigger characteristics of EUSIG-defined hypotensive events.

BACKGROUND: Hypotension is a recognized -secondary insult after traumatic brain injury (TBI). There are many definitions of hypotension, an often cited example being the Brain Trauma Foundation's current (2007) "Guidelines for the Management of Severe Traumatic Brain Injury," which defines hypotension as systolic pressure <90 mmHg. However, this same document declares "The importance of mean arterial pressure, as opposed to systolic pressure should also be stressed, …." Our work shows that when using the Edinburgh University Secondary Insult Grades (EUSIG) definitions, which require monitoring of both systolic and mean arterial pressures, that most hypotensive events are in fact triggered by a breach of the mean arterial level of 70 mmHg. We suggest that close monitoring of mean arterial pressure would enable clinical teams to avoid more potentially damaging hypotensive events. MATERIALS AND METHODS: An analysis of 100 patients from the Brain-IT database was performed. Using the EUSIG definitions, 2,081 events can be obtained by analyzing the systolic and mean blood pressures on a minute by minute basis. A software program was written to identify and classify the trigger pattern for each event. A categorical analysis of these triggering patterns has been carried out. KEY RESULTS: Our analysis shows that most events are triggered by a drop in mean arterial pressure. In fact a large number of events (91%) occur where the mean arterial pressure is below the threshold limits whereas the systolic pressure does not cross the 90 mmHg limit at all. CONCLUSION: We suggest that more emphasis should be placed on closely monitoring mean arterial pressure as well as systolic pressure when trying to guard against hypotensive problems in traumatically brain injured patients. In future work we will study the underlying physiological mechanisms and attempt to further classify concomitant conditions that may be contributing to the onset of a hypotensive event.

Examining the Relationship Between Hospital-Community Partnerships and COVID-19 Case-Fatality Rates.

Abtract During the COVID-19 pandemic, hospitals across the United States were tasked to develop partnerships with other hospitals and community organizations to overcome the unexpected challenges. The aim of this study is to examine COVID-19 case-fatality rates and explore their relationship with hospital-community partnerships. This study employed a cross-sectional design using a multilevel generalized linear model with a Poisson regression distribution and publicly available COVID-19 mortality data from February to October 2020 across 2526 hospital service areas (HSAs). HSAs with a greater number of partnerships were found to have a reduced risk of higher case-fatality rates than those with fewer health system partnerships. The findings indicated the need for greater cooperation between individual health care systems, state and local governments, and community programs for better outcomes in the ongoing and evolving COVID-19 pandemic, and to be better prepared for future pandemics or large-scale public health crises. This study provides the necessary insights for policy makers, hospital administrators, and public health leaders to understand the critical importance of community partnerships and their influence on reducing the COVID-19 case-fatality rate, as well as their potential effects on improving the health of vulnerable populations as a means to achieve the Centers for Disease Control and Prevention's goal of achieving health equity. This research illustrates the need for further inquiries into the importance of these health care partnerships for positive health care outcomes.

Medicare and telehealth: The impact of COVID-19 pandemic.

STUDY RATIONALE: The swift progression of the COVID-19 pandemic appeared to facilitate the increase in telehealth utilization. However, it is clear neither how telehealth was offered by providers nor how it was used by patients during this time of unusual and rapid change within the health industry. AIM: To investigates the telehealth utilization patterns of Medicare beneficiaries during the height of the COVID-19 pandemic. METHODS AND MATERIALS: A cross-sectional study design was used to examine the responses of 9686 Medicare beneficiaries to the Centers for Medicare and Medicaid Services (CMS) Medicare Current Beneficiary Survey, Fall 2020 COVID-19 Supplement. Multiple logistic regression analyses were conducted to examine the relationship between telehealth offering and beneficiaries' sociodemographic variables. RESULTS: Over half (58%) of primary care providers provided telehealth services, while only 26%-28% of specialists did. Less than 8% of Medicare beneficiaries reported that they were unable to obtain care because of COVID-19. CONCLUSIONS: This research found that changes in Medicare policy, associated with CMS' declaration of telehealth waivers during the Public Health Emergency (PHE), likely increased the proliferation and utilization of telehealth services during the COVID-19 pandemic, providing important access to care for certain populations. With the impending conclusion of the PHE, policymakers must 1) ascertain which elements of the new telehealth landscape will be retained, 2) modernize the regulatory, accreditation and reimbursement framework to maintain pace with care model innovation and 3) address disparities in access to broadband connectivity with a particular focus on rural and underserved communities.

The Financial Burden of Opioid-Related Abuse among Surgical and Non-Surgical Patients in Florida: A Longitudinal Study.

Florida is one of the eight states labeled as a high-burden opioid abuse state and is an epicenter for opioid use and misuse. The aim of our study was to measure multi-year total room charges and costs billed for opioid abuse-related events and to compare the costs of inpatient opioid abusers and non-opioid abusers for Florida hospitals from 2011 to 2017. We constructed a retrospective case-control longitudinal study design on inpatient administrative discharge data across 173 hospitals. Opioid abuse was defined using both ICD-9-CM and ICD-10-CM systems. We found a statistically significant association between opioid abuse diagnosis and total room charge. On average, opioid abuse status increased the room charges by 8.1%. We also noticed year-to-year variations in opioid abuse had a remarkable influence on hospital finances. We showed that since 2015, the differences significantly increased from 4-5% to 13-14% for both room charges and cost, which indicates the financial burden due to opioid abuse becoming more frequent. These findings are important to policymakers and hospital administrators because they provide crucial insight into Florida's opioid crisis and its economic burden on hospitals.

The brain monitoring with Information Technology (BrainIT) collaborative network: EC feasibility study results and future direction.

BACKGROUND: The BrainIT group works collaboratively on developing standards for collection and analyses of data from brain-injured patients and to facilitate a more efficient infrastructure for assessing new health care technology with the primary objective of improving patient care. European Community (EC) funding supported meetings over a year to discuss and define a core dataset to be collected from patients with traumatic brain injury using IT-based methods. We now present the results of a subsequent EC-funded study with the aim of testing the feasibility of collecting this core dataset across a number of European sites and discuss the future direction of this research network. METHODS: Over a 3-year period, data collection client- and web-server-based tools were developed and core data (grouped into nine categories) were collected from 200 head-injured patients by local nursing staff in 22 European neuro-intensive care centres. Data were uploaded through the BrainIT website and random samples of received data were selected automatically by computer for validation by data validation staff against primary sources held in each local centre. Validated data were compared with originally transmitted data and percentage error rates calculated by data category. Feasibility was assessed in terms of the proportion of missing data, accuracy of data collected and limitations reported by users of the IT methods. FINDINGS: Thirteen percent of data files required cleaning. Thirty "one-off" demographic and clinical data elements had significant amounts of missing data (>15%). Validation staff conducted 19,461 comparisons between uploaded database data with local data sources and error rates were commonly less than or equal to 6%, the exception being the surgery data class where an unacceptably high error rate of 34% was found. Nearly 10,000 therapies were successfully recorded with start-times but approximately a third had inaccurate or missing "end-times" which limits the analysis of duration of therapy. Over 40,000 events and procedures were recorded but events with long durations (such as transfers) were more likely to have end-times missed. CONCLUSIONS: The BrainIT core dataset is a rich dataset for hypothesis generation and post hoc analyses, provided that studies avoid known limitations in the dataset. Limitations in the current IT-based data collection tools have been identified and have been addressed. In order for multi-centre data collection projects to be viable, the resource intensive validation procedures will require a more automated process and this may include direct electronic access to hospital-based clinical data sources for both validation purposes and for minimising the duplication of data entry. This type of infrastructure may foster and facilitate the remote monitoring of patient management and protocol adherence in future trials of patient management and monitoring.

Cerebral Perfusion Pressure Insults and Associations with Outcome in Adult Traumatic Brain Injury.

The definition of cerebral perfusion pressure (CPP) secondary insults in severe traumatic brain injury remains unclear. The purpose of the present study is to visualize the association of intensity and duration of episodes below or above CPP thresholds and outcome. The analysis was based on prospectively collected minute-by-minute intracranial pressure (ICP) and blood pressure data and outcome from 259 adult patients. The relationship of episodes of CPP below or above a certain threshold for certain duration with the 6-month Glasgow Outcome Score was visualized separately for episodes of active or deficient autoregulation (AR). In adults ≤65 years, an almost exponential transition curve separates the episodes of CPP associated with better outcomes from the episodes of low CPP associated with worse outcomes, indicating that lower CPP could only be tolerated for a brief time. Analysis of episodes of high CPP again showed a time-intensity dependent association with outcome. When combining the two plots, a safe CPP zone between 60 and 70 mm Hg could be delineated-however, only for AR active insults. The AR status predominantly affected the transition curve for insults of low CPP. Episodes with ICP >25 mm Hg were associated with poor outcome regardless of CPP. In the present study, the CPP pressure-time burden associated with poor outcome was visualized. A safe zone between 60 and 70 mm Hg could be identified for adults ≤65 years, provided AR was active and ICP was ≤25 mm Hg. Deficient AR reduces the tolerability for low CPP.

Visualizing the pressure and time burden of intracranial hypertension in adult and paediatric traumatic brain injury.

PURPOSE: To assess the impact of the duration and intensity of episodes of increased intracranial pressure on 6-month neurological outcome in adult and paediatric traumatic brain injury. METHODS: Analysis of prospectively collected minute-by-minute intracranial pressure and mean arterial blood pressure data of 261 adult and 99 paediatric traumatic brain injury patients from multiple European centres. The relationship of episodes of elevated intracranial pressure (defined as a pressure above a certain threshold during a certain time) with 6-month Glasgow Outcome Scale was visualized in a colour-coded plot. RESULTS: The colour-coded plot illustrates the intuitive concept that episodes of higher intracranial pressure can only be tolerated for shorter durations: the curve that delineates the duration and intensity of those intracranial pressure episodes associated with worse outcome is an approximately exponential decay curve. In children, the curve resembles that of adults, but the delineation between episodes associated with worse outcome occurs at lower intracranial pressure thresholds. Intracranial pressures above 20 mmHg lasting longer than 37 min in adults, and longer than 8 min in children, are associated with worse outcomes. In a multivariate model, together with known baseline risk factors for outcome in severe traumatic brain injury, the cumulative intracranial pressure-time burden is independently associated with mortality. When cerebrovascular autoregulation, assessed with the low-frequency autoregulation index, is impaired, the ability to tolerate elevated intracranial pressures is reduced. When the cerebral perfusion pressure is below 50 mmHg, all intracranial pressure insults, regardless of duration, are associated with worse outcome. CONCLUSIONS: The intracranial pressure-time burden associated with worse outcome is visualised in a colour-coded plot. In children, secondary injury occurs at lower intracranial pressure thresholds as compared to adults. Impaired cerebrovascular autoregulation reduces the ability to tolerate intracranial pressure insults. Thus, 50 mmHg might be the lower acceptable threshold for cerebral perfusion pressure.

Federating distributed clinical data for the prediction of adverse hypotensive events.

The ability to predict adverse hypotensive events, where a patient's arterial blood pressure drops to abnormally low (and dangerous) levels, would be of major benefit to the fields of primary and secondary health care, and especially to the traumatic brain injury domain. A wealth of data exist in health care systems providing information on the major health indicators of patients in hospitals (blood pressure, temperature, heart rate, etc.). It is believed that if enough of these data could be drawn together and analysed in a systematic way, then a system could be built that will trigger an alarm predicting the onset of a hypotensive event over a useful time scale, e.g. half an hour in advance. In such circumstances, avoidance measures can be taken to prevent such events arising. This is the basis for the Avert-IT project (http://www.avert-it.org), a collaborative EU-funded project involving the construction of a hypotension alarm system exploiting Bayesian neural networks using techniques of data federation to bring together the relevant information for study and system development.