Background and Aims: Deviating growth patterns in children can be a symptom of coeliac disease (CD). Knowledge of growth of a total real world data childhood population is scarce. The aim of the study was to investigate the occurrence of potential growth deviations, associated with CD, at what age deviations occurred related to date of diagnosis and whether there were differences between the sexes in preschool Swedish children.

Methods: Weight and height were gathered from the Regional Healthcare Information Platform (RHIP) in the Region Halland, Sweden, including children registered January 2009 to December 2022, with a diagnosis of CD. Descriptive statistics were performed.

Results: 121 (65.4%) of the children with CD were girls and 64 (34.6%) boys. The negative growth deviations were highest at diagnosis; mean -0.54(±1.2) weight standard deviation score (weight-SDS), -0.48(±1.1) height standard deviation score (height-SDS), with no significant differences between the sexes. More pronounced growth deviations were seen in children diagnosed at 36-48 months of age compared to other age groups. Significant improvement in growth for weight-SDS and height-SDS were seen between diagnosis and 2 years after diagnosis.

Conclusions: The significant differences at the measurement points in proximity to diagnosis indicated differences in growth depending on age of CD diagnosis. Catch up growth was seen 2 years after diagnosis. Practical implications are that it is possible to observe negative growth deviations years before CD diagnosis, by growth-monitoring.

Background Associations between growth deviations and coeliac disease (CD) in children have been documented, but longitudinal evidence-particularly concerning when such deviations first emerge-is limited.The aim of this study was to investigate the occurrence and timing of growth deviations in relation to CD diagnosis, and to assess whether these patterns differed between boys and girls in a Swedish preschool population. Methods This retrospective longitudinal study was conducted as part of the project Evidence based knowledge about deviant growth in children 0-6 years. A total of 185 children with CD were identified through the Regional Healthcare Information Platform. Background characteristics and growth deviations relative to the time of CD diagnosis were analysed using descriptive statistics and tests, including chi square, Mann-Whitney U, one way ANOVA, Kruskal-Wallis, paired t tests, and Wilcoxon signed rank tests. Results Of the 185 participating children, 64 (34.6%) were boys and 121 (65.4%) were girls. Growth deviations were most pronounced at the time of diagnosis, with mean weight SDS of-0.54 (+/- 1.2) and mean height SDS of-0.48 (+/- 1.1). No significant sex differences were observed. Children diagnosed between 36 and 48 months exhibited significantly greater negative height SDS deviations at one year and six months prior to diagnosis compared with younger age groups. Significant improvements in both weight SDS and height SDS were observed during the two years following diagnosis and initiation of treatment. Conclusion Growth deviations varied according to age at CD diagnosis, with the most pronounced deviations occurring in proximity to the diagnostic timepoint. Growth improved significantly during the two years following diagnosis.These findings suggest that subtle growth deviations may be detectable years before diagnosis, supporting the clinical value of careful growth monitoring in paediatric care. Further research on the implementation of growth based screening parameters in routine practice is warranted. © The Author(s) 2026. 

Objective: Half of all adult emergency department (ED) visits with a complaint of dyspnea involve acute heart failure (AHF), exacerbation of chronic obstructive pulmonary disease (eCOPD), or pneumonia, which are often misdiagnosed. We aimed to create an artificial intelligence (AI) diagnostic decision support tool to detect patients with AHF, eCOPD, and pneumonia among dyspneic adults at the beginning of their ED visit. Methods: In this cross-sectional study, we included all ED visits of patients 18 years or older with dyspnea at two regional Swedish EDs 07/01/2017–12/31/2019. In-hospital or ED discharge notes were used as outcome labels, with a subset manually reviewed by experts. We analyzed data from a complete regional healthcare system, along with socioeconomic factors, using Hierarchical Attention Networks. Each patient displayed a unique set of variables important for diagnosing dyspnea. All patients’ unique variable sets were aggregated into a variable list. The top 100, 50, and 20 variables were tested in a simpler CatBoost model. Finally, performance was compared after adding medical expertise to the AI model. Results: We included 10,869 visits, with 15.1% having AHF, 13.6% eCOPD, and 13.1% pneumonia. The median number of variables per unique ED visit was 187 (IQR 111–307). Aggregating the unique sets of variables resulted in a cohort list of 2,064 variables. The median micro AUROC was 87.8% (2.5–97.5 percentile; 86.4–89.3%). Age, ECGs, previous diagnoses, and medication were considered important by the AI model, while sex, vital signs, and socioeconomic factors were deemed almost non-predictive. Using the top 20 AI-selected variables, the AUROC was 86.6% (85.1–88.1%). Adding human medical expertise did not significantly change the AUROC. Conclusion: Based on the analysis of a high-dimensional dataset, we designed a lightweight 20-variable machine learning model that can early and effectively diagnose AHF, eCOPD, and pneumonia among ED patients with dyspnea. © 2025 The Authors. Published by Elsevier B.V.

Background: Dyspnoea is one of the emergency department’s (ED) most common and deadly chief complaints, but frequently misdiagnosed and mistreated. We aimed to design a diagnostic decision support which classifies dyspnoeic ED visits into acute heart failure (AHF), exacerbation of chronic obstructive pulmonary disease (eCOPD), pneumonia and “other diagnoses” by using deep learning and complete, unselected data from an entire regional health care system.

Methods: In this cross-sectional study, we included all dyspnoeic ED visits of patients ≥ 18 years of age at the two EDs in the region of Halland, Sweden, 07/01/2017–12/31/2019. Data from the complete regional health care system within five years prior to the ED visit were analysed. Gold standard diagnoses were defined as the subsequent in-hospital or ED discharge notes, and a subsample was manually reviewed by emergency medicine experts. A novel deep learning model, the clinical attention-based recurrent encoder network (CareNet), was developed. Cohort performance was compared to a simpler CatBoost model. A list of all variables and their importance for diagnosis was created. For each unique patient visit, the model selected the most important variables, analysed them and presented them to the clinician interpretably by taking event time and clinical context into account. AUROC, sensitivity and specificity were compared.

Findings: The most prevalent diagnoses among the 10,315 dyspnoeic ED visits were AHF (15.5%), eCOPD (14.0%) and pneumonia (13.3%). Median number of unique events, i.e., registered clinical data with time stamps, per ED visit was 1,095 (IQR 459–2,310). CareNet median AUROC was 87.0%, substantially higher than the CatBoost model´s (81.4%). CareNet median sensitivity for AHF, eCOPD, and pneumonia was 74.5%, 92.6%, and 54.1%, respectively, with a specificity set above 75.0, slightly inferior to that of the CatBoost baseline model. The model assembled a list of 1,596 variables by importance for diagnosis, on top were prior diagnoses of heart failure or COPD, daily smoking, atrial fibrillation/flutter, life management difficulties and maternity care. Each patient visit received their own unique attention plot, graphically displaying important clinical events for the diagnosis.

Interpretation: We designed a novel interpretable deep learning model for diagnosis in emergency department dyspnoea patients by analysing unselected data from a complete regional health care system. © 2024 Heyman et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Objectives: To describe chronic kidney disease (CKD) regarding treatment rates, comorbidities, usage of CKD International Classification of Diseases (ICD) diagnosis, mortality, hospitalisation, evaluate healthcare utilisation and screening for CKD in relation to new nationwide CKD guidelines.

Design: Population-based observational study.

Setting: Healthcare registry data of patients in Southwest Sweden.

Participants: A total cohort of 65 959 individuals aged >18 years of which 20 488 met the criteria for CKD (cohort 1) and 45 470 at risk of CKD (cohort 2).

Primary and secondary outcome measures: Data were analysed with regards to prevalence, screening rates of blood pressure, glucose, estimated glomerular filtration rate (eGFR), Urinary-albumin-creatinine ratio (UACR) and usage of ICD-codes for CKD. Mortality and hospitalisation were analysed with logistic regression models.

Results: Of the CKD cohort, 18% had CKD ICD-diagnosis and were followed annually for blood pressure (79%), glucose testing (76%), eGFR (65%), UACR (24%). UACR follow-up was two times as common in hypertensive and cardiovascular versus diabetes patients with CKD with a similar pattern in those at risk of CKD. Statin and renin-angiotensin-aldosterone inhibitor appeared in 34% and 43%, respectively. Mortality OR at CKD stage 5 was 1.23 (CI 0.68 to 0.87), diabetes 1.20 (CI 1.04 to 1.38), hypertension 1.63 (CI 1.42 to 1.88), atherosclerotic cardiovascular disease (ASCVD) 1.84 (CI 1.62 to 2.09) associated with highest mortality risk. Hospitalisation OR in CKD stage 5 was 1.96 (CI 1.40 to 2.76), diabetes 1.15 (CI 1.06 to 1.25), hypertension 1.23 (CI 1.13 to 1.33) and ASCVD 1.52 (CI 1.41 to 1.64).

Conclusions: The gap between patients with CKD by definition versus those diagnosed as such was large. Compared with recommendations patients with CKD have suboptimal follow-up and treatment with renin-angiotensin-aldosterone system inhibitor and statins. Hypertension, diabetes and ASCVD were associated with increased mortality and hospitalisation. Improved screening and diagnosis of CKD, identification and management of risk factors and kidney protective treatment could affect clinical and economic outcomes. © Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Background: After a heart failure (HF) hospital discharge, the risk of a cardiovascular (CV) related event is highest in the following 100 days. It is important to identify factors associated with increased risk of readmission. Method: This retrospective, population-based study examined HF patients in Region Halland (RH), Sweden, hospitalized with a HF diagnosis between 2017 and 2019. Data regarding patient clinical characteristics were retrieved from the Regional healthcare Information Platform from admission until 100 days post-discharge. Primary outcome was readmission due to a CV related event within 100 days. Results: There were 5029 included patients being admitted for HF and discharged and 1966 (39%) were newly diagnosed. Echocardiography was available for 3034 (60%) patients and 1644 (33%) had their first echocardiography while admitted. The distribution of HF-phenotypes was 33% HF with reduced ejection fraction (EF), 29% HF with mildly reduced EF and 38% HF with preserved EF. Within 100 days, 1586 (33%) patients were readmitted, and 614 (12%) died. A Cox regression model showed that advanced age, longer hospital length of stay, renal impairment, high heart rate and elevated NT-proBNP were associated with an increased risk of readmission regardless of HF-phenotype. Women and increased blood pressure are associated with a reduced risk of readmission. Conclusions: One third had a CV-readmission within 100 days. This study found clinical factors already present at discharge that are associated with increased risk of readmission which should be considered at discharge. © 2023 The Authors

As Deep Neural Networks (DNN) make their way into safety-critical decision processes, it becomes imperative to have robust and reliable uncertainty estimates for their predictions for both in-distribution and out-of-distribution (OOD) examples. This is particularly important in real-life high-risk settings such as healthcare, where OOD examples (e.g., patients with previously unseen or rare labels, i.e., diagnoses) are frequent, and an incorrect clinical decision might put human life in danger, in addition to having severe ethical and financial costs. While evidential uncertainty estimates for deep learning have been studied for multi-class problems, research in multi-label settings remains untapped. In this paper, we propose a DEep Evidential Doctor (DEED), which is a novel deterministic approach to estimate multi-label targets along with uncertainty. We achieve this by placing evidential priors over the original likelihood functions and directly estimating the parameters of the evidential distribution using a novel loss function. Additionally, we build a redundancy layer (particularly for high uncertainty and OOD examples) to minimize the risk associated with erroneous decisions based on dubious predictions. We achieve this by learning the mapping between the evidential space and a continuous semantic label embedding space via a recurrent decoder. Thereby inferring, even in the case of OOD examples, reasonably close predictions to avoid catastrophic consequences. We demonstrate the effectiveness of DEED on a digit classification task based on a modified multi-label MNIST dataset, and further evaluate it on a diagnosis prediction task from a real-life electronic health record dataset. We highlight that in terms of prediction scores, our approach is on par with the existing state-of-the-art having a clear advantage of generating reliable, memory and time-efficient uncertainty estimates with minimal changes to any multi-label DNN classifier. © 2023 The Author(s)

OBJECTIVES: To describe clinical characteristics and prognosis related to heart failure (HF) phenotypes in a community-based population by applying a novel algorithm to obtain ejection fractions (EF) from electronic medical records. DESIGN: Retrospective population-based cohort study. SETTING: Data were collected for all patients with HF in Southwest Sweden. The region consists of three acute care hospitals, 40 inpatient wards, 2 emergency departments, 30 outpatient specialty clinics and 48 primary healthcare. PARTICIPANTS: 8902 patients had an HF diagnosis based on the International Classification of Diseases, Tenth Revision during the study period. Patients <18 years as well as patients declining to participate were excluded resulting in a study population of 8775 patients. PRIMARY AND SECONDARY OUTCOME MEASURES: The primary outcome measure was distribution of HF phenotypes by echocardiography. The secondary outcome measures were 1 year all-cause mortality and HR for all-cause mortality using Cox regression models. RESULTS: Out of 8775 patients with HF, 5023 (57%) had a conclusive echocardiography distributed into HF with reduced EF (35%), HF with mildly reduced EF (27%) and HF with preserved EF (38%). A total of 43% of the cohort did not have a conclusive echocardiography, and therefore no defined phenotype (HF-NDP). One-year all-cause mortality was 42% within the HF-NDP group and 30% among those with a conclusive EF. The HR of all-cause mortality in the HF-NDP group was 1.27 (95% CI 1.17 to 1.37) when compared with the confirmed EF group. There was no significant difference in survival within the HF phenotypes. CONCLUSIONS: This population-based study showed a distribution of HF phenotypes that varies from those in selected HF registries, with fewer patients with HF with reduced EF and more patients with HF with preserved EF. Furthermore, 1-year all-cause mortality was significantly higher among patients with HF who had not undergone a conclusive echocardiography at diagnosis, highlighting the importance of correct diagnostic procedure to improve treatment strategies and outcomes. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Recent years have witnessed an unparalleled surge in deep neural networks (DNNs) research, surpassing traditional machine learning (ML) and statistical methods on benchmark datasets in computer vision, audio processing and natural language processing (NLP). Much of this success can be attributed to the availability of numerous open-source datasets, advanced computational resources and algorithms. These algorithms learn multiple levels of simple to complex abstractions (or representations) of data resulting in superior performances on downstream applications. This has led to an increasing interest in reaping the potential of DNNs in real-life safety-critical domains such as autonomous driving, security systems and healthcare. Each of them comes with their own set of complexities and requirements, thereby necessitating the development of new approaches to address domain-specific problems, even if building on common foundations.

In this thesis, we address data science related challenges involved in learning effective prediction models from structured electronic health records (EHRs). In particular, questions related to numerical representation of complex and heterogeneous clinical concepts, modelling the sequential structure of EHRs and quantifying prediction uncertainties are studied. From a clinical perspective, the question of predicting onset of adverse outcomes for individual patients is considered to enable early interventions, improve patient outcomes, curb unnecessary expenditures and expand clinical knowledge.

This is a compilation thesis including five articles. It begins by describing a healthcare information platform that encapsulates clinical, operational and financial data of patients across all public care delivery units in Halland, Sweden. Thus, the platform overcomes the technical and legislative data-related challenges inherent to the modern era's complex and fragmented healthcare sector. The thesis presents evidence that expert clinical features are powerful predictors of adverse patient outcomes. However, they are well complemented by clinical concept embeddings; gleaned via NLP inspired language models. In particular, a novel representation learning framework (KAFE: Knowledge And Frequency adapted Embeddings) has been proposed that leverages medical knowledge schema and adversarial principles to learn high quality embeddings of both frequent and rare clinical concepts. In the context of sequential EHR modelling, benchmark experiments on cost-sensitive recurrent nets have shown significant improvements compared to non-sequential networks. In particular, an attention based hierarchical recurrent net is proposed that represents individual patients as weighted sums of ordered visits, where visits are, in turn, represented as weighted sums of unordered clinical concepts. In the context of uncertainty quantification and building trust in models, the field of deep evidential learning has been extended. In particular for multi-label tasks, simple extensions to current neural network architecture are proposed, coupled with a novel loss criterion to infer prediction uncertainties without compromising on accuracy. Moreover, a qualitative assessment of the model behaviour has also been an important part of the research articles, to analyse the correlations learned by the model in relation to established clinical science.

Put together, we develop DEep Evidential Doctor (DEED). DEED is a generic predictive model that learns efficient representations of patients and clinical concepts from EHRs and quantifies its confidence in individual predictions. It is also equipped to infer unseen labels.

Overall, this thesis presents a few small steps towards solving the bigger goal of artificial intelligence (AI) in healthcare. The research has consistently shown impressive prediction performance for multiple adverse outcomes. However, we believe that there are numerous emerging challenges to be addressed in order to reap the full benefits of data and AI in healthcare. For future works, we aim to extend the DEED framework to incorporate wider data modalities such as clinical notes, signals and daily lifestyle information. We will also work to equip DEED with explainability features.