Special Issue "2nd Edition of Data Science for Health Services"

Dear Colleagues,

We are pleased to announce a Special Issue on “Data Science for Health Services II”. In recent years, health services have been transformed by the emergence and increased application of data science methods such as predictive modeling, visualization, and artificial intelligence. These methods are being used for service planning, service management, and the delivery of care, thus improving the health of individuals and communities. Research on data science methods for health services can be broadly grouped into three stages:

• At the collection stage, data need to be acquired, stored safely and effectively, and occasionally combined. Data may include demographic and clinical information obtained from electronic medical records (EMRs), insurance claims, and other administrative data, as well as data continuously flowing from devices grouped under the Internet of Things (IoT). Recent innovations include virtual hospitals, wearable biosensors, digital health apps, and smart monitors. New data warehouse designs are often sought after to handle constraints such as privacy preservation, the large scale of records, and the need to efficiently support various queries. Finally, data fusion is required to augment common sources with value-added information or derive comprehensive measures for health service performance (e.g., quality index).
• At the analysis and forecasting stage, artificial intelligence (AI) allows for the exploration of patterns or the assessment of possible future scenarios. Machine learning (ML) techniques can serve to predict healthcare outcomes such as quality, utilization, or cost. Modeling and simulation (M&S) provides estimates for scenarios, such as the impact of a vaccination scheme on the number of beds in intensive care units. ML and M&S both face challenges in terms of data (e.g., insufficient data for emerging problems and conflicting measures) and algorithmic efficiency (e.g., scaling to big data).
• The adoption of data science methods in health services sheds light on how to translate results into actions that improve care for individuals and better meet the health needs of communities. Such translational efforts include novel multidisciplinary initiatives which bridge academic or organizational silos such as, for example, when social scientists, epidemiologists, and modelers create joint frameworks. The adoption of these methods also needs to navigate regulatory and legal frameworks, particularly in a changing ecosystem (e.g., new laws on data protection) and, given the emergence of new approaches, to safely perform computations (e.g., federated learning and secure enclaves).

We solicit papers for this Special Issue that broadly deal with such challenges by addressing open questions, providing novel case studies, or encouraging interesting and challenging debates. Papers can be reviews, syntheses, viewpoints, meta-analyses, or original research articles.

Dr. Philippe J. Giabbanelli
Dr. Francisco Iacobelli
Dr. Charlotte James
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Information is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• clinical decision support
• clinical care models
• health informatics
• quality of care
• population health planning
• digital health

Title: Improving Cancer Prevention in Older Adults: A Machine Learning Perspective
Authors: Dr. Ashir Javeed; Prof. Johan Sanmartin Berglund; Prof. Peter Anderberg; Dr. Muhammad Asim Saleem
Affiliation: Karolinska Institutet
Abstract: Cancer is the second leading cause of death worldwide after cardiovascular disease. While cancer can affect people of any age, most cases occur in the fifth or sixth decade of life, highlighting the increased risk of cancer with age. Early detection of cancer predictors and associated risk factors is critical to improving survival rates. In light of this, we conducted a study in a sample of older adults in Sweden using a machine learning-based model to predict cancer and identify risk factors associated with cancer in this population. Our study provides valuable insights into cancer prevention and detection strategies in older adults and highlights the importance of early detection and intervention in improving health outcomes. The proposal is composed of two modules. The first module ranks the dataset's variables using a statistical F-score model that includes 74 variables. The second module is a classifier that uses the random forest algorithm (RF) with optimized hyperparameters using a genetic algorithm. The study found that the data set had severely imbalanced classes. To address this problem and avoid bias in the ML model, we used a random undersampling procedure to balance the classes. The model's components are integrated into a single "black box", we called this model F-RUS-RF. When only the six highest ranked variables are used, the F- RUS-RF model achieves the highest accuracy of 86.15%, with a sensitivity of 92.25% and a specificity of 85.14%. The F-RUS-RF model effectively predicts cancer and identifies risk factors in older adults. From the 74 variables in the data set, the F-RUS-RF model identified the six most significant variables contributing to cancer development in older adults. Addressing these risk factors may reduce cancer incidence in the older population.