Differentiation between atypical anorexia nervosa and anorexia nervosa using machine learning

International Journal of Eating Disorders, Journal Year: 2024, Volume and Issue: 57(4), P. 937 - 950

Published: Feb. 14, 2024

Abstract Objective Body mass index (BMI) is the primary criterion differentiating anorexia nervosa (AN) and atypical despite prior literature indicating few differences between disorders. Machine learning (ML) classification provides us an efficient means of accurately distinguishing two meaningful classes given any number features. The aim present study was to determine if ML algorithms can distinguish AN ensemble features excluding BMI, not, inclusion BMI enables classify two. Methods Using aggregate sample from seven studies consisting individuals with who completed baseline questionnaires ( N = 448), we used logistic regression, decision tree, random forest models each trained on datasets, one containing demographic, eating disorder, comorbid without retaining all BMI. Results Model performance for as a feature deemed acceptable (mean accuracy 74.98%, mean area under receiving operating characteristics curve [AUC] 74.75%), whereas model diminished 59.37%, AUC 59.98%). Discussion acceptable, but not strong, included feature; no other meaningfully improved classification. When excluded, performed poorly at classifying cases when considering demographic clinical characteristics. suggest reconceptualization should be considered. Public Significance There growing debate about their diagnostic differentiation relies being similar otherwise. We aimed see machine could disorders found accurate only feature. This finding calls into question need differentiate

Language: Английский

---

Artificial Intelligence in the Management of Malnutrition in Cancer Patients: A Systematic Review

Advances in Nutrition, Journal Year: 2025, Volume and Issue: unknown, P. 100438 - 100438

Published: May 1, 2025

Malnutrition is a critical complication among cancer patients, affecting up to 80% of individuals depending on type, stage, and treatment. Artificial Intelligence (AI) has emerged as promising tool in healthcare, with potential applications nutritional management improve early detection, risk stratification, personalized interventions. This systematic review evaluates the role AI identifying managing malnutrition focusing its effectiveness status assessment, prediction, clinical outcomes, body composition monitoring. A search was conducted across PubMed, Cochrane Library, Cumulative Index Nursing Allied Health Literature, Excerpta Medica Database from June July 2024, following Preferred Reporting Items for Systematic Reviews Meta-Analyses guidelines. Quantitative primary studies investigating AI-based interventions analysis, optimization oncology were included. Study quality assessed using Joanna Briggs Institute (JBI) Critical Appraisal Tools, evidence certainty evaluated Oxford Centre Evidence-Based Medicine framework. Eleven (n=52,228 patients) met inclusion criteria categorized into three overarching domains: Nutritional Status Assessment Prediction, Clinical Functional Outcomes, Body Composition Cachexia Monitoring. models demonstrated high predictive accuracy detection (AUC >0.80). Machine learning algorithms, including decision trees, random forests, support vector machines, outperformed conventional screening tools. Deep applied medical imaging achieved segmentation (Dice Similarity Coefficient: 0.92-0.94), enabling cachexia detection. AI-driven virtual dietitian systems improved dietary adherence (84%) reduced unplanned hospitalizations. AI-enhanced workflows streamlined referrals, reducing referral times by 2.4 days. significant optimizing screening, monitoring, patients. Its integration nutrition care could enhance patient outcomes optimize healthcare resource allocation. Further research necessary standardize ensure applicability. 10.17605/OSF.IO/A259M STATEMENT OF SIGNIFICANCE: highlights transformative demonstrating superior Unlike previous studies, which focused isolated applications, this work comprehensively multiple domains, emphasizing their routine treatment personalization, overall outcomes.

Language: Английский

---

Examining the role of artificial intelligence to advance knowledge and address barriers to research in eating disorders

International Journal of Eating Disorders, Journal Year: 2024, Volume and Issue: 57(6), P. 1357 - 1368

Published: April 10, 2024

To provide a brief overview of artificial intelligence (AI) application within the field eating disorders (EDs) and propose focused solutions for research.

Language: Английский

---

On exploring features engineering and machine learning algorithms in the domain of clinical-psychological

Alexandria Engineering Journal, Journal Year: 2025, Volume and Issue: 123, P. 488 - 496

Published: March 29, 2025

Language: Английский

---

Integrating Artificial Intelligence and Smartphone Technology to Enhance Personalized Assessment and Treatment for Eating Disorders

International Journal of Eating Disorders, Journal Year: 2025, Volume and Issue: unknown

Published: May 21, 2025

ABSTRACT Objective Smartphone technology presents a promising path toward expanding access to evidence‐based eating disorder assessment and treatment. Despite rapid technological advances, research has yet harness these systems in ways that make personalized digital health care clinical reality. In this forum, we review extant testing smartphone intervention monitoring tools for disorders explore innovative integrating with AI can enhance assessment, symptom detection, efforts. Method We highlight three capabilities of smartphones hold promise delivering maximally effective tools: (1) passive sensing phenotyping; (2) natural language processing reflections from in‐app homework tasks; (3) closed‐loop adaptive interventions. discuss how augment current treatment efforts draw on literature other fields inform questions the field. Results Evidence demonstrates feasibility constructing data‐driven models sensor data textual input CBT activities predict outcomes. These may interventions, enabling apps deliver timely, support response real‐time changes user's needs. Conclusion The field lessons evaluate leverages personalization. Realizing potential will require addressing challenges related engagement, trust, governance, integration. testable presented here offer roadmap guide future large‐scale, collaborative aimed at transforming care.

Language: Английский

---

The Future of the Eating Disorder Field: Inclusive, Aware of Systems, and Personalized

Behaviour Research and Therapy, Journal Year: 2024, Volume and Issue: 183, P. 104648 - 104648

Published: Oct. 30, 2024

Language: Английский

---

Momentary physiological indices related to eating disorders: A systematic and methodological review

European Eating Disorders Review, Journal Year: 2024, Volume and Issue: 32(4), P. 700 - 717

Published: March 6, 2024

Abstract Eating disorders (ED) are serious psychiatric illnesses, with no everyday support to intervene on the high rates of relapse. Understanding physiological indices that can be measured by wearable sensor technologies may provide new momentary interventions for individuals ED. This systematic review, searching large databases, synthesises studies investigating peripheral (PP) commonly included in wristbands (heart rate [HR], heart variability [HRV], electrodermal activity [EDA], skin temperature [PST], and acceleration) Inclusion criteria included: (a) full peer‐reviewed empirical articles English; (b) human participants active ED; (c) containing one five measures. Kmet risk bias was assessed. Ninety‐four were (Anorexia nervosa [AN; N = 4418], bulimia [BN; 916], binge eating disorder [BED; 1604], other specified feeding [OSFED; 424], transdiagnostic [ 47]). Participants AN displayed lower HR EDA higher HRV compared healthy individuals. Those BN showed HRV, PST Other ED Transdiagnostic samples mixed results. PP differences indicated through various assessments ED, which suggest diagnostic associations, although more needed validate observed patterns. Results important therapeutic potential larger including diverse groups fully uncover their role

Language: Английский

---

Investigating Feature Selection for Physical Activity Prediction Based on Ecological Momentary Assessments: Towards Tailoring the Timing of Behaviour Change Support Messages (Preprint)

Published: Feb. 13, 2024

There has been a surge in the development of applications that aim to improve health, physical activity (PA), and well-being through behavior change. These apps often focus on creating long-term sustainable impact user. Just-in-time adaptive interventions (JITAIs) based passive sensing current user context (e.g., from smartphones wearables) have devised enhance effectiveness these foster PA. JITAIs provide personalized support interventions, such as encouraging messages, context-aware manner. However, limited range capabilities, getting timing right for delivering well accepted effective is challenging. Ecological Momentary Assessment (EMA) can personal by directly capturing assessments e.g. moods emotion. Thus, EMA might be useful complement determining when are triggered. Yet, extensive schedules need scrutinized they increase burden. Use machine learning (ML) balance feature set size questions with prediction accuracy regarding likelihood enacting A total 43 healthy participants (ages 19-67) completed four surveys daily three weeks. prospectively assessed different states including both motivational volitional variables PA preparation (e.g. intrinsic motivation, self-efficacy, perceived barriers) alongside stress mood/emotions. enactment was retrospectively via served outcome variable The best performing ML models predicted engagement an AUC score 0.87 ± 0.02 SD 5-fold cross validation test set. Particularly strong predictors included stress, planning, barriers, indicating small yield accurate prediction. features like planning barriers enough predict reasonably thus used meaningfully tailor sending well-timed aware messages.

Language: Английский

---

Investigating Feature Selection for Physical Activity Prediction Based on Ecological Momentary Assessments: Towards Tailoring the Timing of Behaviour Change Support Messages (Preprint)

JMIR mhealth and uhealth, Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 13, 2024

There has been a surge in the development of apps that aim to improve health, physical activity (PA), and well-being through behavior change. These often focus on creating long-term sustainable impact user. Just-in-time adaptive interventions (JITAIs) are based passive sensing user's current context (eg, via smartphones wearables) have devised enhance effectiveness these foster PA. JITAIs provide personalized support such as encouraging messages context-aware manner. However, limited range capabilities make it challenging determine timing for delivering well-accepted effective interventions. Ecological momentary assessment (EMA) can personal by directly capturing user assessments moods emotions). Thus, EMA might be useful complement determining when triggered. extensive schedules need scrutinized, they increase burden. The study was use machine learning balance feature set size questions with prediction accuracy regarding enacting A total 43 healthy participants (aged 19-67 years) completed 4 surveys daily over 3 weeks. prospectively assessed various states, including both motivational volitional variables related PA preparation intrinsic motivation, self-efficacy, perceived barriers) alongside stress mood or emotions. enactment retrospectively served outcome variable. best-performing models predicted engagement mean area under curve score 0.87 (SD 0.02) 5-fold cross-validation test set. Particularly strong predictors included stress, planning, barriers, indicating small yield accurate participants. EMA-based features like barriers enough predict reasonably well thus used meaningfully tailor sending well-timed messages.

Language: Английский

---

Using virtual reality in exploring the interaction between autonomous nervous system and eating stress disorder

˜The œRomanian Journal of Nutrition., Journal Year: 2024, Volume and Issue: 4(4), P. 11 - 11

Published: Jan. 1, 2024

Language: Английский

---