1. |
- Dave, Nishi, et al.
(author)
-
Nosocomial SARS-CoV-2 infections and mortality during unique COVID-19 epidemic waves
- 2023
-
In: JAMA Network Open. - : American Medical Association (AMA). - 2574-3805. ; 6:11
-
Journal article (peer-reviewed)abstract
- Importance: Quantifying the burden of nosocomial SARS-CoV-2 infections and associated mortality is necessary to assess the need for infection prevention and control measures.Objective: To investigate the occurrence of nosocomial SARS-CoV-2 infections and associated 30-day mortality among patients admitted to hospitals in Region Stockholm, Sweden.Design, Setting, and Participants: A retrospective, matched cohort study divided the period from March 1, 2020, until September 15, 2022, into a prevaccination period, early vaccination and pre-Omicron (period 1), and late vaccination and Omicron (period 2). From among 303 898 patients 18 years or older living in Region Stockholm, 538 951 hospital admissions across all hospitals were included. Hospitalized admissions with nosocomial SARS-CoV-2 infections were matched to as many as 5 hospitalized admissions without nosocomial SARS-CoV-2 by age, sex, length of stay, admission time, and hospital unit.Exposure: Nosocomial SARS-CoV-2 infection defined as the first positive polymerase chain reaction test result at least 8 days after hospital admission or within 2 days after discharge.Main Outcomes and Measures: Primary outcome of 30-day mortality was analyzed using time-to-event analyses with a Cox proportional hazards regression model adjusted for age, sex, educational level, and comorbidities.Results: Among 2193 patients with SARS-CoV-2 infections or reinfections (1107 women [50.5%]; median age, 80 [IQR, 71-87] years), 2203 nosocomial SARS-CoV-2 infections were identified. The incidence rate of nosocomial SARS-CoV-2 infections was 1.57 (95% CI, 1.51-1.64) per 1000 patient-days. In the matched cohort, 1487 hospital admissions with nosocomial SARS-CoV-2 infections were matched to 5044 hospital admissions without nosocomial SARS-CoV-2 infections. Thirty-day mortality was higher in the prevaccination period (adjusted hazard ratio [AHR], 2.97 [95% CI, 2.50-3.53]) compared with period 1 (AHR, 2.08 [95% CI, 1.50-2.88]) or period 2 (AHR, 1.22 [95% CI, 0.92-1.60]). Among patients with nosocomial SARS-CoV-2 infections, 30-day AHR comparing those with 2 or more doses of SARS-CoV-2 vaccination and those with less than 2 doses was 0.64 (95% CI, 0.46-0.88).Conclusions and Relevance: In this matched cohort study, nosocomial SARS-CoV-2 infections were associated with higher 30-day mortality during the early phases of the pandemic and lower mortality during the Omicron variant wave and after the introduction of vaccinations. Mitigation of excess mortality risk from nosocomial transmission should be a strong focus when population immunity is low through implementation of adequate infection prevention and control measures.
|
|
2. |
- Henriksson, Aron, 1985-, et al.
(author)
-
Multimodal fine-tuning of clinical language models for predicting COVID-19 outcomes
- 2023
-
In: Artificial Intelligence in Medicine. - 0933-3657 .- 1873-2860. ; 146
-
Journal article (peer-reviewed)abstract
- Clinical prediction models tend only to incorporate structured healthcare data, ignoring information recorded in other data modalities, including free-text clinical notes. Here, we demonstrate how multimodal models that effectively leverage both structured and unstructured data can be developed for predicting COVID-19 outcomes. The models are trained end-to-end using a technique we refer to as multimodal fine-tuning, whereby a pre -trained language model is updated based on both structured and unstructured data. The multimodal models are trained and evaluated using a multicenter cohort of COVID-19 patients encompassing all encounters at the emergency department of six hospitals. Experimental results show that multimodal models, leveraging the notion of multimodal fine-tuning and trained to predict (i) 30-day mortality, (ii) safe discharge and (iii) readmission, outperform unimodal models trained using only structured or unstructured healthcare data on all three outcomes. Sensitivity analyses are performed to better understand how well the multimodal models perform on different patient groups, while an ablation study is conducted to investigate the impact of different types of clinical notes on model performance. We argue that multimodal models that make effective use of routinely collected healthcare data to predict COVID-19 outcomes may facilitate patient management and contribute to the effective use of limited healthcare resources.
|
|
3. |
- Pawar, Yash, et al.
(author)
-
Leveraging Clinical BERT in Multimodal Mortality Prediction Models for COVID-19
- 2022
-
In: In Proceedings of IEEE International Symposium on Computer-Based Medical Systems (CMBS 2022). - : IEEE. - 9781665467704 ; , s. 199-204
-
Conference paper (peer-reviewed)abstract
- Clinical prediction models are often based solely on the use of structured data in electronic health records, e.g. vital parameters and laboratory results, effectively ignoring potentially valuable information recorded in other modalities, such as free-text clinical notes. Here, we report on the development of a multimodal model that combines structured and unstructured data. In particular, we study how best to make use of a clinical language model in a multimodal setup for predicting 30-day all-cause mortality upon hospital admission in patients with COVID-19. We evaluate three strategies for incorporating a domain-specific clinical BERT model in multimodal prediction systems: (i) without fine-tuning, (ii) with unimodal fine-tuning, and (iii) with multimodal fine-tuning. The best-performing model leverages multimodal fine-tuning, in which the clinical BERT model is updated based also on the structured data. This multimodal mortality prediction model is shown to outperform unimodal models that are based on using either only structured data or only unstructured data. The experimental results indicate that clinical prediction models can be improved by including data in other modalities and that multimodal fine-tuning of a clinical language model is an effective strategy for incorporating information from clinical notes in multimodal prediction systems.
|
|
4. |
- Alam, Mahbub Ul, et al.
(author)
-
Deep Learning from Heterogeneous Sequences of Sparse Medical Data for Early Prediction of Sepsis
- 2020
-
In: Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies, Volume 5: HEALTHINF. - Setúbal : SciTePress. - 9789897583988 ; , s. 45-55
-
Conference paper (peer-reviewed)abstract
- Sepsis is a life-threatening complication to infections, and early treatment is key for survival. Symptoms of sepsis are difficult to recognize, but prediction models using data from electronic health records (EHRs) can facilitate early detection and intervention. Recently, deep learning architectures have been proposed for the early prediction of sepsis. However, most efforts rely on high-resolution data from intensive care units (ICUs). Prediction of sepsis in the non-ICU setting, where hospitalization periods vary greatly in length and data is more sparse, is not as well studied. It is also not clear how to learn effectively from longitudinal EHR data, which can be represented as a sequence of time windows. In this article, we evaluate the use of an LSTM network for early prediction of sepsis according to Sepsis-3 criteria in a general hospital population. An empirical investigation using six different time window sizes is conducted. The best model uses a two-hour window and assumes data is missing not at random, clearly outperforming scoring systems commonly used in healthcare today. It is concluded that the size of the time window has a considerable impact on predictive performance when learning from heterogeneous sequences of sparse medical data for early prediction of sepsis.
|
|
5. |
- Alam, Mahbub Ul, et al.
(author)
-
Terminology Expansion with Prototype Embeddings : Extracting Symptoms of Urinary Tract Infection from Clinical Text
- 2021
-
In: Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 5: HEALTHINF. - Setúbal : SciTePress. - 9789897584909 ; , s. 47-57
-
Conference paper (peer-reviewed)abstract
- Many natural language processing applications rely on the availability of domain-specific terminologies containing synonyms. To that end, semi-automatic methods for extracting additional synonyms of a given concept from corpora are useful, especially in low-resource domains and noisy genres such as clinical text, where nonstandard language use and misspellings are prevalent. In this study, prototype embeddings based on seed words were used to create representations for (i) specific urinary tract infection (UTI) symptoms and (ii) UTI symptoms in general. Four word embedding methods and two phrase detection methods were evaluated using clinical data from Karolinska University Hospital. It is shown that prototype embeddings can effectively capture semantic information related to UTI symptoms. Using prototype embeddings for specific UTI symptoms led to the extraction of more symptom terms compared to using prototype embeddings for UTI symptoms in general. Overall, 142 additional UTI symp tom terms were identified, yielding a more than 100% increment compared to the initial seed set. The mean average precision across all UTI symptoms was 0.51, and as high as 0.86 for one specific UTI symptom. This study provides an effective and cost-effective solution to terminology expansion with small amounts of labeled data.
|
|
6. |
- Alpkvist, Helena, et al.
(author)
-
Clinical and Microbiological Factors Associated with High Nasopharyngeal Pneumococcal Density in Patients with Pneumococcal Pneumonia
- 2015
-
In: PLOS ONE. - : Public Library Science. - 1932-6203. ; 10:10
-
Journal article (peer-reviewed)abstract
- Background: We aimed to study if certain clinical and/or microbiological factors are associated with a high nasopharyngeal (NP) density of Streptococcus pneumoniae in pneumococcal pneumonia. In addition, we aimed to study if a high NP pneumococcal density could be useful to detect severe pneumococcal pneumonia.Methods: Adult patients hospitalized for radiologically confirmed community-acquired pneumonia were included in a prospective study. NP aspirates were collected at admission and were subjected to quantitative PCR for pneumococcal DNA (Spn9802 DNA). Patients were considered to have pneumococcal etiology if S. pneumoniae was detected in blood culture and/ or culture of respiratory secretions and/or urinary antigen test.Results: Of 166 included patients, 68 patients had pneumococcal DNA detected in NP aspirate. Pneumococcal etiology was noted in 57 patients (84%) with positive and 8 patients (8.2%) with negative test for pneumococcal DNA (p<0.0001). The median NP pneumococcal density of DNA positive patients with pneumococcal etiology was 6.83 log(10) DNA copies/mL (range 1.79-9.50). In a multivariate analysis of patients with pneumococcal etiology, a high pneumococcal density was independently associated with severe pneumonia (Pneumonia Severity Index risk class IV-V), symptom duration >= 2 days prior to admission, and a medium/high serum immunoglobulin titer against the patient's own pneumococcal serotype. NP pneumococcal density was not associated with sex, age, smoking, co-morbidity, viral co-infection, pneumococcal serotype, or bacteremia. Severe pneumococcal pneumonia was noted in 28 study patients. When we studied the performance of PCR with different DNA cut-off levels for detection of severe pneumococcal pneumonia, we found sensitivities of 54-82% and positive predictive values of 37-56%, indicating suboptimal performance.Conclusions: Pneumonia severity, symptom duration similar to 2 days, and a medium/high serum immunoglobulin titer against the patient's own serotype were independently associated with a high NP pneumococcal density. NP pneumococcal density has limited value for detection of severe pneumococcal pneumonia.
|
|
7. |
|
|
8. |
- Athlin, Simon, 1971-, et al.
(author)
-
Management of community-acquired pneumonia in immunocompetent adults : updated Swedish guidelines 2017
- 2018
-
In: Infectious Diseases. - : Informa UK Limited. - 2374-4235 .- 2374-4243. ; 50:4, s. 247-272
-
Research review (peer-reviewed)abstract
- Based on expert group work, Swedish recommendations for the management of community-acquired pneumonia in adults are here updated. The management of sepsis-induced hypotension is addressed in detail, including monitoring and parenteral therapy. The importance of respiratory support in cases of acute respiratory failure is emphasized. Treatment with high-flow oxygen and non-invasive ventilation is recommended. The use of statins or steroids in general therapy is not found to be fully supported by evidence. In the management of pleural infection, new data show favourable effects of tissue plasminogen activator and deoxyribonuclease installation. Detailed recommendations for the vaccination of risk groups are afforded.
|
|
9. |
- Athlin, Simon, 1971-, et al.
(author)
-
Pneumococcal urinary antigen testing for antimicrobial guidance in community-acquired pneumonia : A register-based cohort study
- 2022
-
In: Journal of Infection. - : Elsevier. - 0163-4453 .- 1532-2742. ; 85:2, s. 167-173
-
Journal article (peer-reviewed)abstract
- OBJECTIVES: To evaluate the effect of pneumococcal urinary antigen test (UAT) usage on broad-spectrum antibiotic treatment in community-acquired pneumonia (CAP).METHODS: Patients admitted to 32 Swedish hospitals between 2011 and 2014 were retrospectively included from the Swedish National Quality Register of CAP. Using propensity score matched data, stratified by CRB-65 score, we studied the effect of performing UAT and of positive test results on treatment with broad-spectrum β-lactam monotherapy (BSBM) and antibiotics with coverage for atypical bacteria compared to narrow-spectrum β-lactam monotherapy (NSBM).RESULTS: UAT was performed for 4,995/14,590 (34.2%) patients, 603/4,995 (12.1%) of whom had positive test results. At day three, performing UAT was not associated with decreased use of BSBM (OR 1.07, 95% CI 0.94-1.23) but was associated with increased atypical coverage among patients with CRB-65 score 2 (OR 1.47, 95% CI 1.06-2.02). A positive UAT was associated with decreased BSBM use (OR 0.39, 95% CI 0.25-0.60) and decreased atypical coverage (OR 0.25, 95% CI 0.16-0.37), predominantly in non-severe CAP. At day one, performing UAT was associated with atypical coverage among patients with CRB-65 scores 2 (OR 2.60, 95% CI 1.69-3.98) and 3-4 (OR 3.69, 95% CI 1.55-8.79), and a positive test reduced the odds of BSBM treatment among CRB-65 score 3-4 patients (OR 3.49, 95% CI 1.02-12.0).CONCLUSIONS: Performing UAT had no overall effect on decreasing the use of BSBM treatment by day three of hospitalization, yet non-severely ill patients with positive UAT results were less likely to be treated with BSBM and antibiotics with atypical coverage.
|
|
10. |
|
|