(4 papers in press)

Regular Issues

• A novel non-invasive technique for blood glucose level measurement  
  by Subramanya Bhat 
  Abstract: Diabetes mellitus (sugar) is one of the most common, incurable and critical health challenge to the medical field which can even cause lethal consequences. Controlling blood sugar levels is the sole effective treatment, to deter the further complications of the disease. To keep diabetes in check, a patient needs to track his or her blood glucose level frequently. The conventional methods available in the market are invasive, painful, time-consuming, expensive and often leads to fingertip infection in case of recurrent use. The complications aroused from conventional method have created the necessity for user-friendly non-invasive glucose monitoring device. In the proposed paper, a portable non-invasive device for glucose monitoring based on NIR spectroscopy using LED (940 nm to 1,300 nm) and a photo-detector, employing the principle of diffused reflectance to determine the blood glucose concentration is developed.
  Keywords: non-invasive; NIR spectroscopy; blood sugar level.
  DOI: 10.1504/IJAIH.2026.10075966
   
  
• Deep learning approaches for viral infection diagnosis via protein structure analysis  
  by Mueed Ahmed Mirza, Wasif Ali, Hafiz Haseeb Tasleem, Wasif Akbar, Muhammad Usman Karim, Muhammad Waqar Arshad 
  Abstract: Protein structural classes are highly essential in determining the function, facilitating genomics and enhancing drug discovery. This paper introduces a deep learning architecture of homology modeling with a fine-tuned CNN, VGG16, based on 2D images produced by the 3D structure of proteins to classify them. Protein models are translated into 2D models that represent alpha-helices, beta-sheets, surface exposure and amino acids properties. The model has an accuracy of 95.18, precision of 95.20, recall of 95.18 and F1-score of 95.17 which is better than the iProStruct2D baseline. These findings suggest valid and strong classification that can be used in bioinformatics.
  Keywords: protein structural class prediction; deep learning; homology modelling; computational proteomics; drug discovery; structural bioinformatics.
  DOI: 10.1504/IJAIH.2026.10076266
   
  
• Exploring the role of artificial intelligence in healthcare management in Iran: perspectives, challenges, and implementation strategies  
  by Abbas Sorkhi 
  Abstract: Artificial intelligence (AI) is increasingly recognised as a transformative force in healthcare. This qualitative study explored expectations, prerequisites, and barriers to AI adoption from the perspective of Iranian healthcare managers. Data were collected through focus group discussions with senior managers and analysed thematically. Findings revealed optimism regarding AIs potential to improve efficiency, decision-making, and resource allocation. However, successful implementation requires standardised data, skilled workforce training, robust IT infrastructure, sustainable funding, and ethicallegal frameworks. Context-specific, phased strategies were proposed, offering practical implications for policy and practice. These insights are also relevant for other low- and middle-income countries (LMICs).
  Keywords: artificial intelligence; healthcare management; decision-making; health information management; medical informatics; Iran.
  DOI: 10.1504/IJAIH.2026.10076669
   
  
• An investigation into the detection of paediatric obstructive sleep apnoea using polysomnography and machine learning  
  by Akhila Pockyarath Chathoth, Divya Sundarraj, Revathy Sekar 
  Abstract: Obstructive sleep apnoea (OSA) in children is frequently characterised by intermittent airway obstruction during sleep, a disease that can cause irregular sleep patterns and possibly long-term health issues. The gold standard for diagnosing OSA in children is polysomnography (PSG), which offers a thorough evaluation of respiratory episodes, sleep stages, and associated physiological markers. Machine learning (ML) approaches have demonstrated promise in recent years for increasing clinical efficiency, automating PSG signal interpretation, and boosting diagnostic precision. In order to diagnose paediatric sleep apnea, this study investigates the integration of PSG data with machine learning techniques.
  Keywords: paediatric obstructive sleep apnoea; POSA; polysomnography; machine learning; ML; convolutional neural networks; CNNs; supervised learning; SVMs; decision trees; apnoea episode categorisation; OSA severity determination; automated PSG signal interpretation.
  DOI: 10.1504/IJAIH.2026.10076917