Search

Importance: Lung biopsies are sometimes performed in mechanically ventilated patients with acute hypoxemic respiratory failure (AHRF) of unknown etiology to guide patient management. While surgical lung biopsies (SLB) offer high diagnostic rates, they may also cause significant complications. Transbronchial forceps lung biopsies (TBLB) are less invasive but often produce non-contributive specimens. Transbronchial lung cryobiopsies (TBLC) yield specimens of potentially better quality than TBLB, but due to their novel implementation in the intensive care unit (ICU), their accuracy and safety are still unclear. Objectives: Our main objective was to evaluate the risk of adverse events in patients with AHRF following the three biopsy techniques. Our secondary objectives were to assess the diagnostic yield and associated modifications of patient management of each technique. Design, Settings and Participants: We conducted a retrospective cohort study comparing TBLC, TBLB, and SLB in mechanically ventilated patients with AHRF. Main Outcomes and Measures: The primary outcome was the proportion of patients with at least one complication, and secondary outcomes included complication rates, diagnostic yields, treatment modifications, and mortality. Results: Of the 26 patients who underwent lung biopsies from 2018 to 2022, all TBLC and SLB patients and 60% of TBLB patients had at least one complication. TBLC patients had higher unadjusted numbers of total and severe complications, but also worse Sequential Organ Failure Assessment scores and P/F ratios. A total of 25 biopsies (25/26, 96%) provided histopathological diagnoses, 88% (22/25) of which contributed to patient management. ICU mortality was high for all modalities (63% for TBLC, 60% for TBLB and 50% for SLB). Conclusions and Relevance: All biopsy methods had high diagnostic yields and the great majority contributed to patient management; however, complication rates were elevated. Further research is needed to determine which patients may benefit from lung biopsies and to determine the best biopsy modality. [ABSTRACT FROM AUTHOR]

Background: The CONvalescent Plasma for Hospitalized Adults With COVID‐19 Respiratory Illness (CONCOR‐1) trial was a multicenter randomized controlled trial assessing convalescent plasma in hospitalized COVID‐19 patients. This study evaluates the cost‐effectiveness of convalescent plasma and its impact on quality‐of‐life to provide insight into its potential as an alternative treatment in resource‐constrained settings. Methods: Individual patient data on health outcomes and resource utilization from the CONCOR‐1 trial were used to conduct the analysis from the Canadian public payer's perspective with a time horizon of 30 days post‐randomization. Baseline and 30‐day EQ‐5D‐5L were measured to calculate quality‐adjusted survival. All costs are presented in 2021 Canadian dollars. The base case assessed the EQ‐5D‐5L scores of hospitalized inpatients reporting at both timepoints, and a utility score of 0 was assigned for patients who died within 30 days. Costs for all patients enrolled were used. The sensitivity analysis utilizes EQ‐5D‐5L scores from the same population but only uses costs from this population. Results: 940 patients were randomized: 627 received CCP and 313 received standard care. The total costs were $28,716 (standard deviation, $25,380) and $24,258 ($22,939) for the convalescent plasma and standard care arms respectively. EQ‐5D‐5L scores were 0.61 in both arms (p =.85) at baseline. At 30 days, EQ‐5D‐5L scores were 0.63 and 0.64 for patients in the convalescent plasma and standard care arms, respectively (p =.46). The incremental cost was $4458 and the incremental quality‐adjusted life day was −0.078. Discussion: Convalescent plasma was less effective and more costly than standard care in treating hospitalized COVID‐19. [ABSTRACT FROM AUTHOR]

Objectives Distributed computations facilitate multi-institutional data analysis while avoiding the costs and complexity of data pooling. Existing approaches lack crucial features, such as built-in medical standards and terminologies, no-code data visualizations, explicit disclosure control mechanisms, and support for basic statistical computations, in addition to gradient-based optimization capabilities. Materials and methods We describe the development of the Collaborative Data Analysis (CODA) platform, and the design choices undertaken to address the key needs identified during our survey of stakeholders. We use a public dataset (MIMIC-IV) to demonstrate end-to-end multi-modal FL using CODA. We assessed the technical feasibility of deploying the CODA platform at 9 hospitals in Canada, describe implementation challenges, and evaluate its scalability on large patient populations. Results The CODA platform was designed, developed, and deployed between January 2020 and January 2023. Software code, documentation, and technical documents were released under an open-source license. Multi-modal federated averaging is illustrated using the MIMIC-IV and MIMIC-CXR datasets. To date, 8 out of the 9 participating sites have successfully deployed the platform, with a total enrolment of >1M patients. Mapping data from legacy systems to FHIR was the biggest barrier to implementation. Discussion and conclusion The CODA platform was developed and successfully deployed in a public healthcare setting in Canada, with heterogeneous information technology systems and capabilities. Ongoing efforts will use the platform to develop and prospectively validate models for risk assessment, proactive monitoring, and resource usage. Further work will also make tools available to facilitate migration from legacy formats to FHIR and DICOM. [ABSTRACT FROM AUTHOR]

We aimed to implement four data partitioning strategies evaluated with four federated learning (FL) algorithms and investigate the impact of data distribution on FL model performance in detecting steatosis using B-mode US images. A private dataset (153 patients; 1530 images) and a public dataset (55 patient; 550 images) were included in this retrospective study. The datasets contained patients with metabolic dysfunction-associated fatty liver disease (MAFLD) with biopsy-proven steatosis grades and control individuals without steatosis. We employed four data partitioning strategies to simulate FL scenarios and we assessed four FL algorithms. We investigated the impact of class imbalance and the mismatch between the global and local data distributions on the learning outcome. Classification performance was assessed with area under the receiver operating characteristic curve (AUC) on a separate test set. AUCs were 0.93 (95% CI 0.92, 0.94) for source-based partitioning scenario with FedAvg, 0.90 (95% CI 0.89, 0.91) for a centralized model, and 0.83 (95% CI 0.81, 0.85) for a model trained in a single-center scenario. When data was perfectly balanced on the global level and each site had an identical data distribution, the model yielded an AUC of 0.90 (95% CI 0.88, 0.92). When each site contained data exclusively from one single class, irrespective of the global data distribution, the AUC fell in the range of 0.34-0.70. FL applied to B-mode US images provide performance comparable to a centralized model and higher than single-center scenario. Global data imbalance and local data heterogeneity influenced the learning outcome., (© 2024. The Author(s).)