Trusting AI made decisions in healthcare by making them explainable.

OBJECTIVES: In solving the trust issues surrounding machine learning algorithms whose reasoning cannot be understood, advancements can be made toward the integration of machine learning algorithms into mHealth applications. The aim of this paper is to provide a transparency layer to black-box machine learning algorithms and empower mHealth applications to maximize their efficiency. METHODS: Using a machine learning testing framework, we present the process of knowledge transfer between a white-box model and a black-box model and the evaluation process to validate the success of the knowledge transfer. RESULTS: The presentation layer of the final output of the base white-box model and the knowledge-infused white-box model shows clear differences in reasoning. The correlation between the base black-box model and the new knowledge-infused model is very high, indicating that the knowledge transfer was successful. CONCLUSION: There is a clear need for transparency in digital health and health in general. Adding solutions to the toolbox of explainable artificial intelligence is one way to gradually decrease the obscurity of black-box models.

Trends in Physiotherapy of Chronic Low Back Pain Research: Knowledge Synthesis Based on Bibliometric Analysis.

BACKGROUND: Physiotherapy and chronic low back pain (CLBP) form a broad and quickly developing research area. The aim of this article was to holistically, thematically and chronologically analyze and synthesize the literature production in this research area and identify the most prolific research entities and research themes. METHODS: This article quantitatively and qualitatively analyzed research literature production harvested from the Scopus bibliometric database, using a triangulation of bibliometric and thematic analysis. For this, Excel 2024, Bibliometrix Biblioshiny 4.1 and VOSviewer version 1.6.20 softwares were used. RESULTS: In the Scopus database, 2843 data sources were found, which were published between 1974 and 26 February 2024. The growth trend has been linearly positive since the beginning of publication, and after 2018 exponential growth began. A review of the most prolific entities showed that the most literature was published in America, Europe and Australasia. The thematic analysis of the information sources identified six main themes (pathophysiology of CLBP and the quantification assessment tools, diagnostics and CLBP treatment, CLBP questionnaires and surveys, quality of life, complementary methods in physiotherapy and psychosocioeconomic aspects), while the chronological analysis revealed three main areas of development: assessment tools, CLBP processing and study methodology. CONCLUSIONS: The results of this bibliometric study present a good starting point for further research, providing taxonomy and research landscapes as a holistic framework offering multidisciplinary knowledge about CLBP, while chronological analysis provides a basis for identifying prospective research trends. This article offers an interdisciplinary view of the current issue of public health. The results of this study provide a basis for the development of both the physiotherapy and epidemiological fields.

The role of correspondence analysis in medical research.

Correspondence analysis (CA) is a multivariate statistical and visualization technique. CA is extremely useful in analyzing either two- or multi-way contingency tables, representing some degree of correspondence between columns and rows. The CA results are visualized in easy-to-interpret "bi-plots," where the proximity of items (values of categorical variables) represents the degree of association between presented items. In other words, items positioned near each other are more associated than those located farther away. Each bi-plot has two dimensions, named during the analysis. The naming of dimensions adds a qualitative aspect to the analysis. Correspondence analysis may support medical professionals in finding answers to many important questions related to health, wellbeing, quality of life, and similar topics in a simpler but more informal way than by using more complex statistical or machine learning approaches. In that way, it can be used for dimension reduction and data simplification, clustering, classification, feature selection, knowledge extraction, visualization of adverse effects, or pattern detection.

Research Trends in Motivation and Weight Loss: A Bibliometric-Based Review.

Obesity is a complex disease that, like COVID-19, has reached pandemic proportions. Consequently, it has become a rapidly growing scientific field, represented by an extensive body of research publications. Therefore, the aim of this study was to present the research trends in the scientific literature on motivation and weight loss. Because traditional knowledge synthesis approaches are not appropriate for analyzing large corpora of research evidence, we utilized a novel knowledge synthesis approach called synthetic knowledge synthesis (SKS) to generate new holistic insights into obesity research focusing on motivation. SKS is a triangulation of bibliometric analysis, bibliometric mapping, and content analysis. Using it, we analyzed the corpus of publications retrieved from the Scopus database, using the search string TITLE-ABS-KEY((obesity or overweight) and "weight loss" and motiv*) in titles, keywords, and abstracts, without any additional inclusion or exclusion criteria. The search resulted in a corpus of 2301 publications. The United States of America, the United Kingdom, and Australia were the most productive countries. Four themes emerged, namely, weight loss and weight-loss maintenance through motivational interventions, lifestyle changes supported by smart ICT, maintaining sustainable weight with a healthier lifestyle, and weight management on the level of primary healthcare and bariatric surgery. Further, we established that the volume of research literature is growing, as is the scope of the research. However, we observed a regional concentration of research and its funding in developed countries and almost nonexistent research cooperation between developed and less-developed countries.

The Effects of Diabetes and Being Overweight on Patients with Post-COVID-19 Syndrome.

In the aftermath of the COVID-19 pandemic, post-COVID-19 syndrome (PCS) remains a challenge and may continue to pose a major health problem in the future. Moreover, the influences of type 2 diabetes and being overweight on PCS remain unclear. This study aimed to assess these influences. We performed an observational study from October 2020 to July 2022, which included 466 patients (269 males and 197 females) with a median age of 65. They were hospitalized due to COVID-19 pneumonia and had persistent symptoms after 1 month of COVID-19 infection. The patients were divided into four groups according to the study objectives: patients with type 2 diabetes, overweight patients, overweight patients with type 2 diabetes, and average-weight patients without type 2 diabetes. The clinical and demographic data collected during hospitalization and regular visits to the Community Healthcare Center dr. Adolf Drolc Maribor were analyzed. Our results showed that type 2 diabetes patients had more difficult courses of treatment and longer hospitalizations. Moreover, more type 2 diabetes patients underwent rehabilitation than the other study groups. The prevailing symptoms of our patients with PCS were dyspnea and fatigue, mostly among female patients with type 2 diabetes. Our study also showed that more women with type 2 diabetes and overweight women with type 2 diabetes suffered from secondary infections. Furthermore, more overweight patients were treated in the intensive care unit than patients from the other groups. However, our study showed an interesting result: patients with type 2 diabetes had the shortest PCS durations. Type 2 diabetes and being overweight are risk factors for PCS onset and prolonged duration. Therefore, our data that revealed a shorter duration of PCS in type 2 diabetes patients than the other investigated groups was unexpected. We believe that answering the questions arising from our unexpected results will improve PCS treatment in general.

Hybrid visualization-based framework for depressive state detection and characterization of atypical patients.

INTRODUCTION: Depression is a global concern, with a significant number of people affected worldwide, particularly in low- and middle-income countries. The rising prevalence of depression emphasizes the importance of early detection and understanding the origins of such conditions. OBJECTIVE: This paper proposes a framework for detecting depression using a hybrid visualization approach that combines local and global interpretation. This approach aims to assist in model adaptation, provide insights into patient characteristics, and evaluate prediction model suitability in a different environment. METHODS: This study utilizes R programming language with the Caret, ggplot2, Plotly, and Dalex libraries for model training, visualization, and interpretation. Data from the NHANES repository was used for secondary data analysis. The NHANES repository is a comprehensive source for examining health and nutrition of individuals in the United States, and covers demographic, dietary, medication use, lifestyle choices, reproductive and mental health data. Penalized logistic regression models were built using NHANES 2015-2018 data, while NHANES 2019-March 2020 data was used for evaluation at the global-specific and local level interpretation. RESULTS: The prediction model that supports this framework achieved an average AUC score of 0.748 (95% CI: 0.743-0.752), with minimal variability in sensitivity and specificity. CONCLUSION: The built-in prediction model highlights chest pain, the ratio of family income to poverty, and smoking status as crucial features for predicting depressive states in both the original and local environments.

Discrepancies among Scopus and Web of Science, coverage of funding information in medical journal articles: a follow-up study.

Objective: This follow-up study aims to determine if and how the coverage of funding information in Web of Science Core Collection (WoS) and Scopus changed from 2015 to 2021. Methods: The number of all funded articles published in 2021 was identified in WoS and Scopus bibliographic databases using bibliometric analysis on a sample of 52 prestigious medical journals. Results: The analysis of the number of funded articles with funding information showed statistically significant differences between Scopus and WoS due to substantial differences in the number of funded articles between some single journals. Conclusion: Due to significant differences in the number of funded articles indexed in WoS and Scopus, which might be attributed to the different protocols for handling funding data in WoS and Scopus, we would still advise using both databases to obtain and analyze funding information.

Profiling of patients with type 2 diabetes based on medication adherence data.

Introduction: Type 2 diabetes mellitus (T2DM) is a complex, chronic disease affecting multiple organs with varying symptoms and comorbidities. Profiling patients helps identify those with unfavorable disease progression, allowing for tailored therapy and addressing special needs. This study aims to uncover different T2DM profiles based on medication intake records and laboratory measurements, with a focus on how individuals with diabetes move through disease phases. Methods: We use medical records from databases of the last 20 years from the Department of Endocrinology and Diabetology of the University Medical Center in Maribor. Using the standard ATC medication classification system, we created a patient-specific drug profile, created using advanced natural language processing methods combined with data mining and hierarchical clustering. Results: Our results show a well-structured profile distribution characterizing different age groups of individuals with diabetes. Interestingly, only two main profiles characterize the early 40-50 age group, and the same is true for the last 80+ age group. One of these profiles includes individuals with diabetes with very low use of various medications, while the other profile includes individuals with diabetes with much higher use. The number in both groups is reciprocal. Conversely, the middle-aged groups are characterized by several distinct profiles with a wide range of medications that are associated with the distinct concomitant complications of T2DM. It is intuitive that the number of profiles increases in the later age groups, but it is not obvious why it is reduced later in the 80+ age group. In this context, further studies are needed to evaluate the contributions of a range of factors, such as drug development, drug adoption, and the impact of mortality associated with all T2DM-related diseases, which characterize these middle-aged groups, particularly those aged 55-75. Conclusion: Our approach aligns with existing studies and can be widely implemented without complex or expensive analyses. Treatment and drug use data are readily available in healthcare facilities worldwide, allowing for profiling insights into individuals with diabetes. Integrating data from other departments, such as cardiology and renal disease, may provide a more sophisticated understanding of T2DM patient profiles.

Lipoprotein(a) As a Risk Factor in a Cohort of Hospitalised Cardiovascular Patients: A Retrospective Clinical Routine Data Analysis.

INTRODUCTION: Lipoprotein(a) (Lp(a)) is a well-recognised risk factor for ischemic heart disease (IHD) and calcific aortic valve stenosis (AVS). METHODS: A retrospective observational study of Lp(a) levels (mg/dL) in patients hospitalised for cardiovascular diseases (CVD) in our clinical routine was performed. The Lp(a)-associated risk of hospitalisation for IHD, AVS, and concomitant IHD/AVS versus other non-ischemic CVDs (oCVD group) was assessed by means of logistic regression. RESULTS: In total of 11,767 adult patients, the association with Lp(a) was strongest in the IHD/AVS group (eß = 1.010, p < 0.001), followed by the IHD (eß = 1.008, p < 0.001) and AVS group (eß = 1.004, p < 0.001). With increasing Lp(a) levels, the risk of IHD hospitalisation was higher compared with oCVD in women across all ages and in men aged ≤75 years. The risk of AVS hospitalisation was higher only in women aged ≤75 years (eß = 1.010 in age < 60 years, eß = 1.005 in age 60-75 years, p < 0.05). CONCLUSIONS: The Lp(a)-associated risk was highest for concomitant IHD/AVS hospitalisations. The differential impact of sex and age was most pronounced in the AVS group with an increased risk only in women aged ≤75 years.

Lipoprotein(a) in Cardiovascular Diseases: Insight From a Bibliometric Study.

Lipoprotein(a) [Lp(a)] is a complex polymorphic lipoprotein comprised of a low-density lipoprotein particle with one molecule of apolipoprotein B100 and an additional apolipoprotein(a) connected through a disulfide bond. The serum concentration is mostly genetically determined and only modestly influenced by diet and other lifestyle modifications. In recent years it has garnered increasing attention due to its causal role in pre-mature atherosclerotic cardiovascular disease and calcific aortic valve stenosis, while novel effective therapeutic options are emerging [apolipoprotein(a) antisense oligonucleotides and ribonucleic acid interference therapy]. Bibliometric descriptive analysis and mapping of the research literature were made using Scopus built-in services. We focused on the distribution of documents, literature production dynamics, most prolific source titles, institutions, and countries. Additionally, we identified historical and influential papers using Reference Publication Year Spectrography (RPYS) and the CRExplorer software. An analysis of author keywords showed that Lp(a) was most intensively studied regarding inflammation, atherosclerosis, cardiovascular risk assessment, treatment options, and hormonal changes in post-menopausal women. The results provide a comprehensive view of the current Lp(a)-related literature with a specific interest in its role in calcific aortic valve stenosis and potential emerging pharmacological interventions. It will help the reader understand broader aspects of Lp(a) research and its translation into clinical practice.

Role of Agile in Digital Public Health Transformation.

The digitalisation of healthcare, fueled by advances in technology and the COVID-19 pandemic can not only empower equitable access to global expert-level healthcare but also make healthcare more patient-centric. Every digital health solution has one common fundamental component: they all run on computing platforms and are powered by complex software. Traditional software development life cycles have often failed in designing complex software; consequently, the agile approach was introduced. To assess the role of agile in digital public health transformation, we used the synthetic knowledge synthesis, a triangulation of bibliometric mapping, and thematic analysis to analyse the available literature harvested from PubMed. The analysis showed that the use of the agile approach is underutilised in developing digital health software. Moreover, the study revealed that health organisations did not yet embrace the agile culture and should adapt using innovative agile solutions to deliver clinical value to patients and public health systems. Following the software industry, where agile software development is becoming the mainstream approach also for sensitive and regulated software, it is becoming even more essential that the digital health software development process should be modernised. Furthermore, a shift to agile collaboration, agile decision-making, trial tolerance, active engagement, purposeful technology adoption, knowledge sharing, and an open agile innovation ecosystem must be achieved.

The quality of digital health software: Should we be concerned?

The digitalization of healthcare fuelled by advances in technology and the increased prevalence of mobile smart devices and health-related internet of things can offer equitable access to expert-level healthcare globally. Growing demand for telemedicine, mobile health apps, and advanced data analytics have further established their role in a modern information society during the Covid-19 crisis. Digital health is, in essence, powered by software (DHSW), which has to operate in the specific digital health environment characteristics and is therefore highly and intrinsically complex and prone to software defects and faults. Given the lack of standardization regarding DHSW quality, we explored the available reviewed research on this crucial topic in this brief paper, using a synthetic thematic analysis approach. We assert that neither the volume, distribution nor scope of the DHSW quality research content is satisfactory, and significant research gaps exist. Based on the presented evidence, we can only conclude that we should be concerned and that the time to act is now to ensure that the unavoidable increase of usage and prevalence of DHSW will not - in the end - reduce the quality of care due to subpar software and software-based digital health systems.

The Use of Machine Learning Techniques to Predict Diabetes in Patients with Cystic Fibrosis.

The accuracy of the prognosis of diabetes in patients with cystic fibrosis is crucial, as it highly connected with mortality and other complications. The prognosis of diabetes is a time-consuming process. Usually, it is performed by medical staff and can often lead to misdiagnosis. The aim of the study was to analyze and evaluate risk factors of developing diabetes in patients diagnosed with Cystic Fibrosis by using classification machine learning techniques. The ECFS data register was used to train and test the models. Visualization of our results using SHAP values highlights that most important features are age, antibiotic treatment, FEV1 value and lung transplant as risk predictors for diabetes.

Age-Related Changes in Lipid and Glucose Levels Associated with Drug Use and Mortality: An Observational Study.

BACKGROUND: The pathogenesis of type 2 diabetes mellitus is complex and still unclear in some details. The main feature of diabetes mellitus is high serum glucose, and the question arises of whether there are other statistically observable dysregulations in laboratory measurements before the state of hyperglycemia becomes severe. In the present study, we aim to examine glucose and lipid profiles in the context of age, sex, medication use, and mortality. METHODS: We conducted an observational study by analyzing laboratory data from 506,083 anonymized laboratory tests from 63,606 different patients performed by a regional laboratory in Slovenia between 2008 and 2019. Laboratory data-based results were evaluated in the context of medication use and mortality. The medication use database contains anonymized records of 1,632,441 patients from 2013 to 2018, and mortality data were obtained for the entire Slovenian population. RESULTS: We show that the highest percentage of the population with elevated glucose levels occurs approximately 20 years later than the highest percentage with lipid dysregulation. Remarkably, two distinct inflection points were observed in these laboratory results. The first inflection point occurs at ages 55 to 59 years, corresponding to the greatest increase in medication use, and the second coincides with the sharp increase in mortality at ages 75 to 79 years. CONCLUSIONS: Our results suggest that medications and mortality are important factors affecting population statistics and must be considered when studying metabolic disorders such as dyslipidemia and hyperglycemia using laboratory data.

Machine learning on small size samples: A synthetic knowledge synthesis.

Machine Learning is an increasingly important technology dealing with the growing complexity of the digitalised world. Despite the fact, that we live in a 'Big data' world where, almost 'everything' is digitally stored, there are many real-world situations, where researchers are still faced with small data samples. The present bibliometric knowledge synthesis study aims to answer the research question 'What is the small data problem in machine learning and how it is solved?' The analysis a positive trend in the number of research publications and substantial growth of the research community, indicating that the research field is reaching maturity. Most productive countries are China, United States and United Kingdom. Despite notable international cooperation, the regional concentration of research literature production in economically more developed countries was observed. Thematic analysis identified four research themes. The themes are concerned with to dimension reduction in complex big data analysis, data augmentation techniques in deep learning, data mining and statistical learning on small datasets.

Supporting Real World Decision Making in Coronary Diseases Using Machine Learning.

Cardiovascular diseases are one of the leading global causes of death. Following the positive experiences with machine learning in medicine we performed a study in which we assessed how machine learning can support decision making regarding coronary artery diseases. While a plethora of studies reported high accuracy rates of machine learning algorithms (MLA) in medical applications, the majority of the studies used the cleansed medical data bases without the presence of the "real world noise." Contrary, the aim of our study was to perform machine learning on the routinely collected Anonymous Cardiovascular Database (ACD), extracted directly from a hospital information system of the University Medical Centre Maribor). Many studies used tens of different machine learning approaches with substantially varying results regarding accuracy (ACU), hence they were not usable as a base to validate the results of our study. Thus, we decided, that our study will be performed in the 2 phases. During the first phase we trained the different MLAs on a comparable University of California Irvine UCI Heart Disease Dataset. The aim of this phase was first to define the "standard" ACU values and second to reduce the set of all MLAs to the most appropriate candidates to be used on the ACD, during the second phase. Seven MLAs were selected and the standard ACUs for the 2-class diagnosis were 0.85. Surprisingly, the same MLAs achieved the ACUs around 0.96 on the ACD. A general comparison of both databases revealed that different machine learning algorithms performance differ significantly. The accuracy on the ACD reached the highest levels using decision trees and neural networks while Liner regression and AdaBoost performed best in UCI database. This might indicate that decision trees based algorithms and neural networks are better in coping with real world not "noise free" clinical data and could successfully support decision making concerned with coronary diseasesmachine learning.

Meta approaches in knowledge synthesis in nursing: A bibliometric analysis.

OBJECTIVE: To analyze the bibliometric patterns of meta-approaches use in nursing research literature. METHODS: Descriptive, exploratory and historical bibliometrics analyses were used. The papers were harvested from the Web of Science Core Collection. FINDINGS: The search resulted in 2065 publications. The trends in using most individual meta approaches show that the use of meta-analysis is increasing exponentially, the use of meta-synthesis is increasing linearly, while the use of meta-ethnography is constant in last 6 years. Most productive countries were United States of America, United Kingdom and Peoples Republic of China. Most publications were published in the Journal of Advanced Nursing, International Journal of Nursing Studies, and Journal of Clinical Nursing. Twenty-seven percent of all publications were funded. Thirty-four meta approaches were identified. DISCUSSION: The study revealed that the trend in the literature production is positive. Research community use of meta-approaches in nursing exhibit considerable growth. Regional concentration of literature production was observed.

Application of bibliometrics in medicine: a historical bibliometrics analysis.

BACKGROUND: The application of bibliometrics in medicine enables one to analyse vast amounts of publications and their production patterns on macroscopic and microscopic levels. OBJECTIVES: The aim of the study was to analyse the historical perspective of research literature production regarding application of bibliometrics in medicine. METHODS: Publications related to application of bibliometrics in medicine from 1970 to 2018 were harvested from the Scopus bibliographic database. Reference Publication Year Spectroscopy was triangulated with the VOSViewer to identify historical roots and evolution of topics and clinical areas. RESULTS: The search resulted in 6557 publications. The literature production trend was positive. Historical roots analysis identified 33 historical roots and 16 clinical areas where bibliometrics was applied. DISCUSSION: The increase in productivity in application of bibliometrics in medicine might be attributed to increased use of quantitative metrics in research evaluation, publish or perish phenomenon and the increased use of evidence-based medicine. CONCLUSION: The trend of the literature production was positive. Medicine was in the forefront of knowledge development in bibliometrics. reference publication year spectroscopy proved to be an accurate method which was able to identify most of the historical roots.