Submission ID 130876
| Issue/Objective | Research in Canada by (Kelly, J.; Almukhtar, Z.; Eappen, P.; Fasanmi, A.) and academic studies (Fagan, et al. 2024; Dana et al., 2023) show strong association between housing, food security, and mental wellbeing among university students. This study analyses students' data and develops a mathematical framework to classify student mental health profiles and predict care needs based on social and environmental determinant. Primary outcomes show clear classification structures and combined determinants associated with higher needs. With parameter adaptation, the approach is transferable to both high resource, resource constraint and post conflict contexts. |
|---|---|
| Methodology/Approach | Together with descriptive analysis of impactful factors on student mental health, a machine learning model using (K Nearest Neighbor) KNN classification approach is applied. The relationship of mapping: Determinants → Classification → Level of service personalization → Outcomes. Pre- identified groups represent four degree structures of mental healthcare needs according to the WHO mental health pyramid. Suggestions for adaptation include incorporating variables such as exposure, availability of social support and re-calibration of thresholds to reflect elevated baseline distress and constrained resources. |
| Results | Initial analysis drawing on a dataset of over 1,500 university students, identified patterns across social, economic, and demographic dimension; mental wellbeing is positively associated with food security (r ≈ 0.34), and negatively associated with stress (r ≈ −0.52). Ethnicity shows a moderate association with perceived stigma highlighting the need for culturally responsive design. Results demonstrated clear segmentation into low, medium, upper medium and high need groups where higher risk profiles include multiple determinants. The complete model design includes variables such as help seeking behavior, perceived stigma, cultural background, alongside core socioeconomic indicators. |
| Discussion/Conclusion | The model can be translated into an implementable application to support decision making in universities, community health systems, and broader public health settings. Its adaptability across high resource, low resource, and post conflict settings supports global health efficiency and strengthens institutional leadership. The study contributes to the conference theme "Collectively reclaiming global health equity" by offering a scalable, evidence based mechanism to equitable, context responsive mental health service delivery. The framework supports governance, AI use and leadership to address systemic barriers aligning closely with Sub theme1 on upstream determinants of health equity. |
| Presenters and Affiliations | Zainab Almukhtar Cape Breton University Judith Kelly Cape Breton University Philip Eappen Cape Breton University Abidemi Fasanmi Cape Breton University |
