Academic research

A study performed at Johns Hopkins Hospital incorporated preoperative risk factors associated with blood transfusion into a nomogram to accurately predict transfusion following hepatopancreaticobiliary or colorectal surgery. The nomogram was based on routinely collected preoperative data so as to maximize its clinical applicability and to ensure that it was generalizable and easy to use.  MetaVision was used to collect intraoperative fluid and transfusion data, which is maintained prospectively, updated monthly, and undergoes institutional quality review to verify accuracy. The authors point out that while the study provides a risk stratification tool to direct the development of improved decision-making pathways, based on risk management principles that allow redirection of resources to improve cost-effectiveness and patient outcomes, future studies are needed to validate  the proposed nomogram externally using an independent data set.

Read Abstract

A study performed at Institut Paoli-Calmettes in France evaluated the effects of ICU-related events, including the severity of acute respiratory distress syndrome (ARDS), on the long-term outcomes of haematology patients whose lungs required invasive mechanical ventilation.  The authors also wanted to identify early predictive factors (during the first 48 hours of ICU admission) for developing moderate to severe ARDS. They prospectively studied all consecutive patients with haematological malignancy who were admitted to their ICU and required invasive mechanical ventilation for more than 24 hours over a five-year period. All data was extracted  from MetaVision. This study is believed by the authors to be the first to evaluate long-term outcomes in patients with haematological malignancies who required ICU admission. The study results confirm that ICU mortality remains particularly high in the haematological population when invasive mechanical ventilation is required and that the extent of pulmonary infiltration observed on the first chest X-ray and the diagnosis of invasive fungal infection were the most relevant early predictive factors of the severity of ARDS.  The authors conclude that high-risk patients may be identified early, simply by performing a chest X-ray upon ICU admission, and that prompt microbiological identification and targeted therapy should be a major objective of physicians toward improving outcomes.

Read Abstract

The aim of a retrospective analysis performed at St. Antonius Hospital in the Netherlands was to investigate whether acute kidney injury (AKI) after coronary artery bypass grafting (CABG) can be attributed to intraoperative hypo-tension during cardiopulmonary bypass (IOH-CPB). The authors did not observe an association between IOH-CPB and AKI when they corrected this association for well-known confounders and suggest that future studies randomising patients to various IOH-CPB thresholds for CABG should be performed. Intra-operative and post-operative data were collected routinely in MetaVision, including the mean arterial pressure (MAP), which was monitored invasively and recorded every minute.

Read Abstract

The Johns Hopkins Hospital analysed the anaesthesia case start and case end times from paper anaesthesia records and compared them to case start and end times generated once they were fully operational on the MetaVision Anaesthesia Information Management System (AIMS). The authors hypothesised that with an AIMS, time capture would be more accurate, and therefore more randomly distributed, because it typically requires only a click of the computer mouse rather than manual entry after the fact. The authors found significant clustering around minutes ending in 0 and 5 whenever time capture was dependent on human judgement and not recorded in an automated manner. They demonstrated that the distribution of frequency of start and end times was more uniform with an AIMS than with paper records but point out that “Even with the AIMS, however, if data is entered retrospectively into the application, human bias will remain a factor. Thus, the design of the user interface of an AIMS for entering event times can affect the accuracy of the recorded information.”

Read Abstract

In a study done at Beth Israel Deaconess Medical Center, the authors developed novel deep learning models based on Gated Recurrent Units (GRU) to effectively exploit two types of informative missingness patterns, i.e., masking and time duration. They demonstrated the performance of their proposed models on one synthetic and two real-world healthcare datasets (MIMIC-III, PhysioNet) and compared them to several strong machine learning and deep learning approaches in classification tasks, achieving promising results.  In the supplementary section about MIMIC-III preprocessing details, the authors note that they used 19,714 admission records collected from 2008-2012 by MetaVision, which is still used at the hospital, and point out that “The data collection and organization in MetaVision system is much neater than the earlier Philips CareVue system [2001-2008].” Moving forward, the authors would like to explore deep learning approaches to characterise missing-not-at-random data and to conduct theoretical analysis to understand the behaviors of existing solutions to handle missing values.

Read Abstract

In order to address the ongoing controversy about the utility of electroencephalography guidance of anaesthesia and to potentially help to inform the standard of care going forward, the authors propose a block-randomised, double-blinded, comparative effectiveness trial which hypothesises that EEG-guidance of anesthetic administration prevents postoperative delirium and downstream sequelae, including falls and decreased quality of life. The ENGAGES study will enroll 1232 patients 60 years and older who will undergo elective major surgery at Barnes Jewish Hospital, St. Louis, Missouri, USA, an academic medical centre which is affiliated to Washington University School of Medicine. Perioperative data (including repeated measures data) will be retrieved from MetaVision.

Read Abstract