| 1. |
- Bergman, Erik, et al.
(författare)
-
BERT based natural language processing for triage of adverse drug reaction reports shows close to human-level performance
- 2023
-
Ingår i: PLOS Digital Health. - : Public Library of Science (PLoS). - 2767-3170. ; 2:12
-
Tidskriftsartikel (refereegranskat)abstract
- Post-marketing reports of suspected adverse drug reactions are important for establishing the safety profile of a medicinal product. However, a high influx of reports poses a challenge for regulatory authorities as a delay in identification of previously unknown adverse drug reactions can potentially be harmful to patients. In this study, we use natural language processing (NLP) to predict whether a report is of serious nature based solely on the free-text fields and adverse event terms in the report, potentially allowing reports mislabelled at time of reporting to be detected and prioritized for assessment. We consider four different NLP models at various levels of complexity, bootstrap their train-validation data split to eliminate random effects in the performance estimates and conduct prospective testing to avoid the risk of data leakage. Using a Swedish BERT based language model, continued language pre-training and final classification training, we achieve close to human-level performance in this task. Model architectures based on less complex technical foundation such as bag-of-words approaches and LSTM neural networks trained with random initiation of weights appear to perform less well, likely due to the lack of robustness that a base of general language training provides.
|
|
| 2. |
- Geldsetzer, Pascal, et al.
(författare)
-
Healthcare provider-targeted mobile applications to diagnose, screen, or monitor communicable diseases of public health importance in low- and middle-income countries : A systematic review
- 2023
-
Ingår i: PLOS Digital Health. - : Public Library of Science (PLoS). - 2767-3170. ; 2:10
-
Tidskriftsartikel (refereegranskat)abstract
- Communicable diseases remain a leading cause of death and disability in low- and middle-income countries (LMICs). mHealth technologies carry considerable promise for managing these disorders within resource-poor settings, but many existing applications exclusively represent digital versions of existing guidelines or clinical calculators, communication facilitators, or patient self-management tools. We thus systematically searched PubMed, Web of Science, and Cochrane Central for studies published between January 2007 and October 2019 involving technologies that were mobile phone- or tablet-based; able to screen for, diagnose, or monitor a communicable disease of importance in LMICs; and targeted health professionals as primary users. We excluded technologies that digitized existing paper-based tools or facilitated communication (i.e., knowledge-based algorithms). Extracted data included disease category, pathogen type, diagnostic method, intervention purpose, study/target population, sample size, study methodology, development stage, accessory requirement, country of development, operating system, and cost. Given the search timeline, studies involving COVID-19 were not included in the analysis. Of 13,262 studies identified by the screen, 33 met inclusion criteria. 12% were randomized clinical trials (RCTs), with 58% of publications representing technical descriptions. 62% of studies had 100 or fewer subjects. All studied technologies involved diagnosis or screening steps; none addressed the monitoring of infections. 52% focused on priority diseases (HIV, malaria, tuberculosis), but only 12% addressed a neglected tropical disease. Although most reported studies were priced under 20USD at time of publication, two thirds of the records did not yet specify a cost for the study technology. We conclude that there are only a small number of mHealth technologies focusing on innovative methods of screening and diagnosing communicable diseases potentially of use in LMICs. Rigorous RCTs, analyses with large sample size, and technologies assisting in the monitoring of diseases are needed.
|
|
| 3. |
- Kong, So Yeon Joyce, et al.
(författare)
-
Mothers' acceptability of using novel technology with video and audio recording during newborn resuscitation: A cross-sectional survey.
- 2024
-
Ingår i: PLOS digital health. - 2767-3170. ; 3:4
-
Tidskriftsartikel (refereegranskat)abstract
- This study aims to assess the acceptability of a novel technology, MAchine Learning Application (MALA), among the mothers of newborns who required resuscitation.This study took place at Bharatpur Hospital, which is the second-largest public referral hospital with 13 000 deliveries per year in Nepal.This is a cross-sectional survey.Data collection took place from January 21 to February 13, 2022. Self-administered questionnaires on acceptability (ranged 1-5 scale) were collected from participating mothers. The acceptability of the MALA system, which included video and audio recordings of the newborn resuscitation, was examined among mothers according to their age, parity, education level and technology use status using a stratified analysis.The median age of 21 mothers who completed the survey was 25 years (range 18-37). Among them, 11 mothers (52.4%) completed their bachelor's or master's level of education, 13 (61.9%) delivered first child, 14 (66.7%) owned a computer and 16 (76.2%) carried a smartphone. Overall acceptability was high that all participating mothers positively perceived the novel technology with video and audio recordings of the infant's care during resuscitation. There was no statistical difference in mothers' acceptability of MALA system, when stratified by mothers' age, parity, or technology usage (p>0.05). When the acceptability of the technology was stratified by mothers' education level (up to higher secondary level vs. bachelor's level or higher), mothers with Bachelor's degree or higher more strongly felt that they were comfortable with the infant's care being video recorded (p = 0.026) and someone using a tablet when observing the infant's care (p = 0.046). Compared with those without a computer (n = 7), mothers who had a computer at home (n = 14) more strongly agreed that they were comfortable with someone observing the resuscitation activity of their newborns (71.4% vs. 14.3%) (p = 0.024).The novel technology using video and audio recordings for newborn resuscitation was accepted by mothers in this study. Its application has the potential to improve resuscitation quality in low-and-middle income settings, given proper informed consent and data protection measures are in place.
|
|
| 4. |
|
|
| 5. |
|
|
| 6. |
|
|
