The Prevalence of mRNA Related Discussions During the Post-COVID-19 Era.

Vaccinations are one of the most significant interventions to public health, but vaccine hesitancy and skepticism are raising serious concerns for a portion of the population in many countries, including Sweden. In this study, we use Swedish social media data and structural topic modeling to automatically identify mRNA-vaccine related discussion themes and gain deeper insights into how people's refusal or acceptance of the mRNA technology affects vaccine uptake. Our point of departure is a scientific study published in February 2022, which seems to once again sparked further suspicion and concern and highlight the necessity to focus on issues about the nature and trustworthiness in vaccine safety. Structural topic modelling is a statistical method that facilitates the study of topic prevalence, temporal topic evolution, and topic correlation automatically. Using such a method, our research goal is to identify the current understanding of the mechanisms on how the public perceives the mRNA vaccine in the light of new experimental findings.

Semantic relations of binary compounds annotated with SNOMED CT.

The objective of this paper is to investigate whether a meaningful interpretation can be easily assigned to compound medical terms that have been assigned two distinct concepts taken from one of the most comprehensive, clinical healthcare terminologies, the (Swedish) SNOMED CT®. A binary compound term is a union of two terms to construct a complex term of special meaning that is quickly conveyed by not using separated (multiword) terms but rather solid or closed ones (not separated by space or hyphen). This is a case when the vocabulary lacks a single code for such concept; at the same time, solid compounds is the norm for expressing compounds in Germanic languages, such as Swedish. It is therefore useful and challenging to both identify such compounds and also determine the semantic relation that holds between the compound's constituents. The hypothesis we explore is that, if there are two or more noun compounds in which the head and modifier of each one belong to the same semantic type respectively, then the same relation probably holds in each of them. The juxtaposition of the concepts' membership within the SNOMED CT is used for determining an approximation of the semantic relation between head and modifier.

What is the coverage of SNOMED CT®on scientific medical corpora?

This paper reports on the results of a large scale mapping of SNOMED CT on scientific medical corpora. The aim is to automatically access the validity, reliability and coverage of the Swedish SNOMED-CT translation, the largest, most extensive available resource of medical terminology. The method described here is based on the generation of predominantly safe harbor term variants which together with simple linguistic processing and the already available SNOMED term content are mapped to large corpora. The results show that term variations are very frequent and this may have implication on technological applications (such as indexing and information retrieval, decision support systems, text mining) using SNOMED CT. Naïve approaches to terminology mapping and indexing would critically affect the performance, success and results of such applications. SNOMED CT appears not well-suited for automatically capturing the enormous variety of concepts in scientific corpora (only 6,3% of all SNOMED terms could be directly matched to the corpus) unless extensive variant forms are generated and fuzzy and partial matching techniques are applied with the risk of allowing the recognition of a large number of false positives and spurious results.

Voice quality and speech fluency distinguish individuals with Mild Cognitive Impairment from Healthy Controls.

Mild Cognitive Impairment (MCI) is a syndrome characterized by cognitive decline greater than expected for an individual's age and education level. This study aims to determine whether voice quality and speech fluency distinguish patients with MCI from healthy individuals to improve diagnosis of patients with MCI. We analyzed recordings of the Cookie Theft picture description task produced by 26 patients with MCI and 29 healthy controls from Sweden and calculated measures of voice quality and speech fluency. The results show that patients with MCI differ significantly from HC with respect to acoustic aspects of voice quality, namely H1-A3, cepstral peak prominence, center of gravity, and shimmer; and speech fluency, namely articulation rate and averaged speaking time. The method proposed along with the obtainability of connected speech productions can enable quick and easy analysis of speech fluency and voice quality, providing accessible and objective diagnostic markers of patients with MCI.

Using a Discourse Task to Explore Semantic Ability in Persons With Cognitive Impairment.

This paper uses a discourse task to explore aspects of semantic production in persons with various degree of cognitive impairment and healthy controls. The purpose of the study was to test if an in-depth semantic analysis of a cognitive-linguistic challenging discourse task could differentiate persons with a cognitive decline from those with a stable cognitive impairment. Both quantitative measures of semantic ability, using tests of oral lexical retrieval, and qualitative analysis of a narrative were used to detect semantic difficulties. Besides group comparisons a classification experiment was performed to investigate if the discourse features could be used to improve classification of the participants who had a stable cognitive impairment from those who had cognitively declined. In sum, both types of assessment methods captured difficulties between the groups, but tests of oral lexical retrieval most successfully differentiated between the cognitively stable and the cognitively declined group. Discourse features improved classification accuracy and the best combination of features discriminated between participants with a stable cognitive impairment and those who had cognitively declined with an area under the curve (AUC) of 0.93.

Predicting MCI Status From Multimodal Language Data Using Cascaded Classifiers.

Recent work has indicated the potential utility of automated language analysis for the detection of mild cognitive impairment (MCI). Most studies combining language processing and machine learning for the prediction of MCI focus on a single language task; here, we consider a cascaded approach to combine data from multiple language tasks. A cohort of 26 MCI participants and 29 healthy controls completed three language tasks: picture description, reading silently, and reading aloud. Information from each task is captured through different modes (audio, text, eye-tracking, and comprehension questions). Features are extracted from each mode, and used to train a series of cascaded classifiers which output predictions at the level of features, modes, tasks, and finally at the overall session level. The best classification result is achieved through combining the data at the task level (AUC = 0.88, accuracy = 0.83). This outperforms a classifier trained on neuropsychological test scores (AUC = 0.75, accuracy = 0.65) as well as the "early fusion" approach to multimodal classification (AUC = 0.79, accuracy = 0.70). By combining the predictions from the multimodal language classifier and the neuropsychological classifier, this result can be further improved to AUC = 0.90 and accuracy = 0.84. In a correlation analysis, language classifier predictions are found to be moderately correlated (ρ = 0.42) with participant scores on the Rey Auditory Verbal Learning Test (RAVLT). The cascaded approach for multimodal classification improves both system performance and interpretability. This modular architecture can be easily generalized to incorporate different types of classifiers as well as other heterogeneous sources of data (imaging, metabolic, etc.).

Semantic relation mining of solid compounds in medical corpora.

In the context of scientific and technical texts, meaning is usually embedded in noun compounds and the semantic interpretation of these compounds deals with the detection and semantic classification of the relation that holds between the compound's constituents. Semantic relation mining, the technology applied for marking up, interpreting, extracting and classifying relations that hold between pairs of words, is an important enterprise that contribute to deeper means of enhancing document understanding technologies, such as Information Extraction, Question Answering, Summarization, Paraphrasing, Ontology Building and Textual Entailment. This paper explores the application of assigning semantic descriptors taken from a multilingual medical thesaurus to a large sample of solid (closed form) compounds taken from large Swedish medical corpora, and determining the relation(s) that may hold between the compound constituents. Our work is inspired by previous research in the area of using lexical hierarchies for identifying relations between two-word noun compounds in the medical domain. In contrast to previous research, Swedish, as other Germanic languages, require further means of analysis, since compounds are written as one sequence with no white space between the words, e.g. virus diseases vs. virussjukdomar, which makes the problem more challenging, since solid compounds are harder to identify and segment.

Identification of Mild Cognitive Impairment From Speech in Swedish Using Deep Sequential Neural Networks.

While people with mild cognitive impairment (MCI) portray noticeably incipient memory difficulty in remembering events and situations along with problems in decision making, planning, and finding their way in familiar environments, detailed neuropsychological assessments also indicate deficits in language performance. To this day, there is no cure for dementia but early-stage treatment can delay the progression of MCI; thus, the development of valid tools for identifying early cognitive changes is of great importance. In this study, we provide an automated machine learning method, using Deep Neural Network Architectures, that aims to identify MCI. Speech materials were obtained using a reading task during evaluation sessions, as part of the Gothenburg MCI research study. Measures of vowel duration, vowel formants (F1 to F5), and fundamental frequency were calculated from speech signals. To learn the acoustic characteristics associated with MCI vs. healthy controls, we have trained and evaluated ten Deep Neural Network Architectures and measured how accurately they can diagnose participants that are unknown to the model. We evaluated the models using two evaluation tasks: a 5-fold crossvalidation and by splitting the data into 90% training and 10% evaluation set. The findings suggest first, that the acoustic features provide significant information for the identification of MCI; second, the best Deep Neural Network Architectures can classify MCI and healthy controls with high classification accuracy (M = 83%); and third, the model has the potential to offer higher accuracy than 84% if trained with more data (cf., SD≈15%). The Deep Neural Network Architecture proposed here constitutes a method that contributes to the early diagnosis of cognitive decline, quantify the progression of the condition, and enable suitable therapeutics.

Automated Syntactic Analysis of Language Abilities in Persons with Mild and Subjective Cognitive Impairment.

In this work we analyze the syntactic complexity of transcribed Swedish-language picture descriptions using a variety of automated syntactic features, and investigate the features' predictive power in classifying narratives from people with subjective and mild cognitive impairment and healthy controls. Our results indicate that while there are no statistically significant differences, syntactic features can still be moderately successful at distinguishing the participant groups when used in a machine learning framework.

Readability, suitability and comprehensibility in patient education materials for Swedish patients with colorectal cancer undergoing elective surgery: a mixed method design.

OBJECTIVE: To characterize education materials provided to patients undergoing colorectal cancer surgery to gain a better understanding of how to design readable, suitable, comprehensible materials. METHOD: Mixed method design. Deductive quantitative analysis using a validated suitability and comprehensibility assessment instrument (SAM+CAM) was applied to patient education materials from 27 Swedish hospitals, supplemented by language technology analysis and deductive and inductive analysis of data from focus groups involving 15 former patients. RESULTS: Of 125 patient education materials used during the colorectal cancer surgery process, 13.6% were rated 'not suitable', 76.8% 'adequate' and 9.6% 'superior'. Professionally developed stoma care brochures were rated 'superior' and 44% of discharge brochures were 'not suitable'. Language technology analysis showed that up to 29% of materials were difficult to comprehend. Focus group analysis revealed additional areas that needed to be included in patient education materials: general and personal care, personal implications, internet, significant others, accessibility to healthcare, usability, trustworthiness and patient support groups. CONCLUSION: Most of the patient education materials were rated 'adequate' but did not meet the information needs of patients entirely. Discharge brochures particularly require improvement. PRACTICE IMPLICATIONS: Using patients' knowledge and integrating manual and automated methods could result in more appropriate patient education materials.

Medication extraction and guessing in Swedish, French and English.

Extraction of information related to the medication is an important task within the biomedical area. Our method is applied to different types of documents in three languages. The results indicate that our approach can efficiently update and enrich the existing drug vocabularies.

Characteristics of Finnish and Swedish intensive care nursing narratives: a comparative analysis to support the development of clinical language technologies.

BACKGROUND: Free text is helpful for entering information into electronic health records, but reusing it is a challenge. The need for language technology for processing Finnish and Swedish healthcare text is therefore evident; however, Finnish and Swedish are linguistically very dissimilar. In this paper we present a comparison of characteristics in Finnish and Swedish free-text nursing narratives from intensive care. This creates a framework for characterising and comparing clinical text and lays the groundwork for developing clinical language technologies. METHODS: Our material included daily nursing narratives from one intensive care unit in Finland and one in Sweden. Inclusion criteria for patients were an inpatient period of least five days and an age of at least 16 years. We performed a comparative analysis as part of a collaborative effort between Finnish- and Swedish-speaking healthcare and language technology professionals that included both qualitative and quantitative aspects. The qualitative analysis addressed the content and structure of three average-sized health records from each country. In the quantitative analysis 514 Finnish and 379 Swedish health records were studied using various language technology tools. RESULTS: Although the two languages are not closely related, nursing narratives in Finland and Sweden had many properties in common. Both made use of specialised jargon and their content was very similar. However, many of these characteristics were challenging regarding development of language technology to support producing and using clinical documentation. CONCLUSIONS: The way Finnish and Swedish intensive care nursing was documented, was not country or language dependent, but shared a common context, principles and structural features and even similar vocabulary elements. Technology solutions are therefore likely to be applicable to a wider range of natural languages, but they need linguistic tailoring. AVAILABILITY: The Finnish and Swedish data can be found at: http://www.dsv.su.se/hexanord/data/.