Integrative Analysis of Food Insecurity, Psychosocial Stress, and Genomic Susceptibility in Type 2 Diabetes Associated Cognitive Decline in Ghanaian Clinical Cohorts

Full text

Text extracted from the article PDF for reading and search. Tables, figures and equations are in the PDF, which remains the authoritative version.

1. Introduction

1.1. Background to Type 2 Diabetes and Cognitive Decline

Type 2 Diabetes Mellitus has emerged as one of the leading chronic metabolic disorders contributing to global morbidity, mortality, and long-term neurological complications. Beyond its established association with cardiovascular disease, nephropathy, and retinopathy, increasing evidence demonstrates that prolonged hyperglycemia and insulin resistance significantly contribute to accelerated cognitive decline and neurodegenerative dysfunction (Onyekaonwu, et al, 2022) [46]. Persistent metabolic dysregulation disrupts cerebral glucose utilization, increases oxidative stress, promotes mitochondrial dysfunction, and induces chronic neuroinflammation, all of which collectively impair neuronal signaling and synaptic plasticity (Idoko et al., 2024) [22]. These pathological processes are particularly concerning in low-resource settings such as Ghana, where delayed diagnosis, inadequate nutritional access, and limited neurological screening increase the risk of undetected diabetes-associated cognitive impairment.

Recent advances in precision healthcare analytics have improved the understanding of how metabolic biomarkers and clinical indicators can predict neurological deterioration in diabetic populations. Machine learning-assisted healthcare systems now support the integration of multidimensional clinical datasets for improved disease detection and prognostic evaluation (Ijiga et al., 2024) [25]. However, many predictive systems remain heavily dependent on biomedical indicators while underrepresenting psychosocial and socio-environmental determinants such as food insecurity and chronic stress exposure (Okpanachi, et al, 2025) [38]. This limitation reduces predictive sensitivity in vulnerable populations experiencing nutritional deprivation and psychological burden.

Emerging studies further emphasize the importance of demographic representation and equitable predictive modeling in chronic disease research, particularly among underrepresented African cohorts (Ifiala et al., 2026) [23]. Additionally, cognitive augmentation frameworks in healthcare decision systems have highlighted the growing relevance of integrative analytical models capable of combining biological, behavioral, and environmental variables for personalized risk stratification (Anokwuru et al., 2022) [11]. Consequently, understanding the interaction between diabetes progression and cognitive decline requires a multidimensional framework integrating clinical, psychosocial, dietary, and genomic determinants.

1.2. Statement of the Problem

The growing prevalence of Type 2 Diabetes Mellitus in Sub-Saharan Africa has intensified concerns regarding long-term neurological complications, particularly diabetes-associated cognitive decline. Despite increasing evidence linking metabolic dysregulation with neurodegenerative impairment, healthcare systems in many low-resource countries, including Ghana, remain largely focused on glycemic control and cardiovascular complications while cognitive deterioration receives limited clinical attention. Consequently, many diabetic patients experience progressive memory impairment, reduced executive functioning, and diminished psychosocial well-being without early detection or targeted intervention. This challenge is further compounded by widespread food insecurity, chronic psychosocial stress, and socioeconomic instability, which intensify inflammatory and neuroendocrine disruptions associated with cognitive dysfunction.

Existing predictive frameworks for diabetes-related cognitive decline are predominantly biomedical and often exclude important socio-environmental determinants such as dietary deprivation, psychological stress exposure, and social vulnerability. The absence of integrative predictive systems limits the ability of clinicians to identify high-risk individuals before severe neurological impairment occurs (Nwokocha, & Peter-Anyebe, 2022) [36]. Furthermore, current machine learning and clinical risk-scoring approaches rarely incorporate genomic susceptibility markers alongside psychosocial variables within African clinical cohorts, thereby reducing model contextual relevance and predictive sensitivity. Mental health-related studies have demonstrated that prolonged psychosocial stress and neuropsychological trauma significantly alter cognitive processing pathways and behavioral outcomes (Balogun et al., 2024) [17]. Similarly, neurological rehabilitation research has highlighted the interaction between chronic stress exposure, mental health disorders, and long-term cognitive dysfunction in vulnerable populations (Enyejo et al., 2024) [19]. However, these multidimensional relationships remain insufficiently investigated among Ghanaian diabetic populations. Consequently, there is a critical need for an interpretable socio-genomic predictive framework capable of integrating clinical biomarkers, food insecurity indices, psychosocial stress metrics, and genomic susceptibility factors to improve early detection and precision risk stratification of diabetes-associated cognitive decline.

1.3. Aim and Objectives of the Study

The aim of this study is to develop and evaluate an integrative socio-genomic predictive framework for assessing type 2 diabetes-associated cognitive decline among Ghanaian clinical cohorts through the combined analysis of clinical biomarkers, food insecurity indicators, psychosocial stress variables, and genomic susceptibility markers. The specific objectives of the study are to: 1. Examine the relationship between food insecurity and cognitive decline among individuals living with type 2 diabetes in Ghanaian clinical settings.

2. Assess the influence of psychosocial stress on neurocognitive impairment among diabetic patients.

3. Investigate the contribution of genomic susceptibility

markers, including APOE-associated variants, to diabetes-associated cognitive decline.

4. Develop a Hybrid Socio-Genomic Risk Integration

Algorithm (HSGRIA) for individualized cognitive risk prediction.

5. Evaluate the predictive performance of HSGRIA against existing machine learning and conventional clinical risk-scoring models.

6. Determine the extent to which integrating socio-environmental and genomic variables improves early-stage detection of cognitive decline in diabetic

1.4. Research Questions

This study seeks to examine the multidimensional factors contributing to cognitive decline among individuals with Type 2 Diabetes Mellitus in Ghanaian clinical cohorts. The research questions are designed to investigate the roles of food insecurity, psychosocial stress, and genomic susceptibility in neurocognitive deterioration, while also evaluating the predictive effectiveness of the proposed Hybrid Socio-Genomic Risk Integration Algorithm (HSGRIA).

1. What relationship exists between food insecurity and cognitive decline among individuals with type 2 diabetes in Ghana?

2. How does psychosocial stress influence neurocognitive

deterioration in diabetic clinical cohorts?

3. Which genomic susceptibility markers are significantly

associated with diabetes-associated cognitive decline?

4. How effectively can the Hybrid Socio-Genomic Risk

Integration Algorithm predict cognitive decline among diabetic patients?

5. Does the integration of psychosocial, dietary, clinical, and genomic variables improve predictive accuracy compared to conventional clinical prediction models? 6. What is the comparative performance of HSGRIA relative to logistic regression, random forest, support vector machine, gradient boosting, and traditional clinical risk-scoring systems?

1.5. Significance of the Study

This study contributes to the growing field of precision public health by developing an integrative framework for predicting cognitive decline among individuals living with type 2 diabetes in resource-limited healthcare environments. The study is significant because it expands conventional biomedical prediction approaches through the incorporation of food insecurity indicators, psychosocial stress metrics, and genomic susceptibility factors into a unified socio-genomic analytical model.

The findings of this research will support clinicians and healthcare professionals in identifying high-risk diabetic patients at earlier stages of neurocognitive deterioration, thereby enabling timely intervention and improved disease management strategies. The proposed Hybrid Socio-Genomic Risk Integration Algorithm also provides a scalable and interpretable predictive system that may enhance clinical decision-making in underserved healthcare settings where advanced neurological screening infrastructure is limited. The study further contributes to African-centered healthcare research by addressing the underrepresentation of Ghanaian and Sub-Saharan African clinical cohorts in cognitive decline prediction studies. By integrating socio-environmental determinants with genomic and metabolic variables, the research provides a multidimensional understanding of diabetes-associated neurological vulnerability within vulnerable populations.

Additionally, the study offers policy relevance for healthcare planners and public health authorities by emphasizing the importance of addressing food insecurity, chronic stress exposure, and socioeconomic disparities as critical contributors to neurodegenerative risk. The framework developed in this research may also support future advancements in machine learning-driven precision medicine, predictive healthcare analytics, and equitable chronic disease management systems across low-resource

2. Literature Review

2.1. Diabetes-Associated Cognitive Decline and

Neurodegenerative Mechanisms Type 2 Diabetes Mellitus is increasingly recognized as a major contributor to progressive neurodegenerative dysfunction and age-related cognitive decline. Chronic hyperglycemia, insulin resistance, oxidative stress, and vascular dysregulation collectively disrupt neuronal integrity and cerebral metabolic homeostasis, thereby accelerating neurocognitive deterioration (Nwatuzie et al, 2025) [34]. Persistent elevation in blood glucose levels contributes to advanced glycation end-product accumulation, mitochondrial impairment, endothelial dysfunction, and chronic neuroinflammation within the hippocampus and cortical regions associated with learning and memory (Nortey, 2024) [30]. These pathophysiological disruptions significantly impair synaptic plasticity, neurotransmitter regulation, and neuronal survival, resulting in reduced executive functioning, memory loss, and impaired cognitive processing speed. Molecular simulation-based biomedical studies have demonstrated that metabolic disturbances influence intracellular signaling pathways involved in neuronal degeneration and inflammatory activation, particularly within vulnerable aging populations (Atalor et al., 2023) [13].

Emerging computational and analytical frameworks have further strengthened the understanding of how multidimensional biological and behavioral factors interact to influence neurodegenerative progression. Explainable machine learning systems have demonstrated the importance of interpretable predictive architectures capable of integrating heterogeneous clinical and behavioral datasets for improved risk prediction and transparent decision support (Onwuzurike & Igba, 2023) [45]. Similarly, human-centered analytical frameworks emphasize that behavioral, environmental, and psychosocial variables substantially affect long-term cognitive outcomes and adaptive neurological functioning (Onwuzurike, 2023) [41]. In diabetic individuals, prolonged psychosocial stress and nutritional instability may intensify inflammatory responses and cerebral insulin dysregulation, thereby worsening cognitive impairment.

Recent healthcare intelligence systems also demonstrate the increasing value of artificial intelligence-driven analytical infrastructures in identifying hidden disease progression patterns and improving predictive healthcare management (Frimpong et al., 2023) [20]. Ethical governance studies further emphasize that interpretable and accountable predictive systems are essential for reducing diagnostic bias and improving equitable healthcare interventions among vulnerable populations (Onwuzurike & Raphael, 2025) [43]. These multidimensional neurodegenerative mechanisms support the growing need for integrated socio-genomic predictive frameworks capable of combining metabolic, behavioral, psychosocial, and genetic determinants for improved early detection of diabetes-associated cognitive decline in Ghanaian clinical cohorts.

2.2. Food Insecurity, Nutrition, and Metabolic-Cognitive

Interactions Food insecurity represents a critical social determinant of health that significantly influences metabolic stability, neurological functioning, and long-term cognitive outcomes among individuals living with Type 2 Diabetes Mellitus (Nortey, 2024) [30]. In low-resource populations, inadequate access to nutritionally balanced diets contributes to persistent glycemic instability, insulin resistance, micronutrient deficiencies, and chronic inflammatory activation, all of which are strongly associated with accelerated neurodegenerative dysfunction (Aluso, & Enyejo, 2025) [5]. Poor dietary quality increases oxidative stress and impairs cerebral glucose metabolism, thereby reducing neuronal efficiency within brain regions responsible for memory formation, executive processing, and cognitive adaptability. Nutritional deficiencies involving omega-3 fatty acids, antioxidants, B vitamins, and essential amino acids have also been associated with synaptic degeneration and impaired neurotransmitter synthesis, which may worsen diabetes-associated cognitive decline (Aluso, 2021) [2]. Recent metabolomics-based nutritional studies emphasize the importance of dietary composition and biochemical food integrity in regulating metabolic and neurological health outcomes. Nutritional profiling and metabolomic authentication systems have demonstrated how food composition directly influences inflammatory biomarkers, metabolic efficiency, and neuroprotective physiological pathways (Donkor et al., 2025) [18]. These findings are particularly relevant in diabetic populations where chronic nutritional insufficiency may intensify neuroinflammatory cascades and vascular impairment. Furthermore, data-driven analytical systems developed for adaptive learning and behavioral optimization demonstrate that nutritional instability significantly affects cognitive load, attention regulation, and executive functioning, thereby reinforcing the interaction between diet quality and neurocognitive performance (Kpogli et al., 2024) [28].

Explainable analytical models have further shown the value of interpretable multidimensional systems for identifying hidden behavioral and environmental determinants influencing cognitive outcomes (Onwuzurike, et al, 2026) [44]. Similarly, ethical and adaptive intelligence frameworks emphasize the importance of contextualized predictive systems capable of integrating social vulnerability indicators into analytical decision-making processes (Onwuzurike & Enyejo, 2026) [39]. Sustainable data-informed frameworks also demonstrate that long-term environmental deprivation and resource limitations contribute substantially to behavioral and cognitive disparities across vulnerable populations as presented in Table 1 (Onwuzurike & Kpogli, 2022) [42]. These multidimensional metabolic-cognitive interactions support the need for integrated socio-genomic predictive models that incorporate food insecurity indicators alongside clinical and genomic biomarkers in the assessment of diabetes-associated cognitive decline among Ghanaian clinical cohorts.

Table 1: Summary of Food Insecurity, Nutrition, and Metabolic-Cognitive Interactions Key Factor Metabolic/Physiological Effect Cognitive/Neurological Impact Clinical Implication for the Study Food Insecurity and Nutritional Deprivation Causes glycemic instability, insulin resistance, micronutrient deficiencies, and chronic inflammatory activation Accelerates neurodegeneration, memory impairment, and reduced executive functioning Food insecurity should be incorporated as a major socio-environmental predictor in cognitive decline risk assessment Poor Dietary Quality and Nutrient Deficiency Increases oxidative stress, disrupts cerebral glucose metabolism, reduces omega-3 fatty acids, antioxidants, B vitamins, and amino acid availability Impairs synaptic plasticity, neurotransmitter synthesis, and adaptive cognitive processing Nutritional biomarkers are essential for understanding metabolic contributions to diabetes-associated cognitive vulnerability Metabolomic and Nutritional Profiling Influences inflammatory biomarker expression, vascular function, and neuroprotective biochemical pathways Modulates neuroinflammatory cascades and cognitive resilience Metabolomic-informed dietary assessment improves predictive precision in multidimensional healthcare models Socio-Environmental and Behavioral Vulnerability Long-term deprivation and resource limitations increase chronic physiological stress and behavioral instability Reduces attention regulation, cognitive load management, and executive adaptability Integrated socio-genomic frameworks should account for environmental and behavioral determinants in predictive cognitive modeling

2.3. Psychosocial Stress and Cognitive Vulnerability in

Diabetic Populations Psychosocial stress is a major non-biological determinant that significantly contributes to cognitive vulnerability among individuals living with Type 2 Diabetes Mellitus. Chronic stress exposure activates the hypothalamic–pituitary–adrenal axis, resulting in prolonged cortisol secretion, inflammatory dysregulation, and impaired neuronal signaling (Donkor, et al, 2025) [18]. These physiological disruptions contribute to hippocampal atrophy, reduced synaptic plasticity, and impaired executive functioning, particularly in diabetic populations already experiencing metabolic instability and vascular dysfunction (Ijiga, et al, 2021) [26]. Persistent psychosocial stress also intensifies insulin resistance and oxidative stress, thereby worsening glycemic dysregulation and increasing susceptibility to neurodegenerative impairment (Avevor, et al, 2024) [14]. In low-resource healthcare environments, socioeconomic hardship, food insecurity, unemployment, and emotional distress further amplify these neurological vulnerabilities. Psychological and neurobehavioral studies have demonstrated that prolonged exposure to trauma, anxiety, and chronic emotional stress substantially alters cognitive processing pathways and mental health outcomes. Investigations into neuropsychological rehabilitation and trauma-associated mental disorders reveal that chronic psychological burden is strongly associated with impaired memory consolidation, reduced attentional control, and long-term cognitive dysfunction (Enyejo et al., 2024) [19]. Similarly, mental health advocacy research highlights the relationship between persistent psychological instability and neurological deterioration within vulnerable populations exposed to sustained emotional stressors (Ijiga et al., 2024) [24]. These findings suggest that psychosocial stress is not merely a secondary outcome of chronic disease but an active contributor to cognitive decline progression. Furthermore, stress-associated neurological vulnerability has been linked to hormonal dysregulation and psychiatric disturbances affecting cognitive resilience and adaptive functioning. Meta-analytical mental health studies have demonstrated that stress-related neurochemical alterations significantly influence depressive symptoms, behavioral instability, and cognitive impairment pathways (Nwokedi et al., 2026) [35]. Research examining large-scale psychological stress environments also demonstrates how chronic fear, emotional exhaustion, and prolonged mental strain negatively affect neurological well-being and cognitive performance (Balogun et al., 2024) [17]. From a healthcare systems perspective, the management of sensitive neurocognitive and psychosocial data further requires robust analytical infrastructures capable of ensuring secure and ethical patient information handling during predictive risk assessment processes as shown in Figure 1 (Balogun et al., 2025) [15]. These multidimensional interactions indicate that psychosocial stress plays a central role in accelerating neurodegenerative vulnerability among diabetic populations and should therefore be integrated into predictive frameworks assessing diabetes-associated cognitive decline in Ghanaian clinical cohorts.

Fig 1: A Block Diagram Illustrating the Pathway from Psychosocial Stress Exposure to Cognitive Vulnerability in Type 2 Diabetes Mellitus.

Figure 1 illustrates the progressive pathway through which psychosocial stress contributes to cognitive vulnerability among individuals living with Type 2 Diabetes Mellitus. It begins with major psychosocial stressors such as food insecurity, socioeconomic hardship, unemployment, emotional distress, anxiety, and trauma, which trigger biological stress responses through activation of the hypothalamic–pituitary–adrenal axis. This results in elevated cortisol secretion, inflammatory dysregulation, oxidative stress, insulin resistance, and vascular dysfunction, all of which worsen the existing metabolic burden associated with diabetes. These physiological disruptions subsequently impair neuronal signaling, reduce synaptic plasticity, and induce hippocampal degeneration, leading to deficits in memory, attention, executive functioning, and adaptive cognitive processing. The final stage represents the clinical manifestation of these cumulative effects as varying levels of cognitive decline and neurodegenerative vulnerability, thereby justifying the integration of psychosocial stress indicators into predictive frameworks such as HSGRIA for early cognitive risk assessment

2.4. Genomic Susceptibility and Machine Learning

Approaches in Cognitive Risk Prediction Genomic susceptibility plays a fundamental role in determining individual vulnerability to neurodegenerative dysfunction and diabetes-associated cognitive decline. Genetic variants associated with lipid metabolism, inflammatory signaling, insulin regulation, and neuronal maintenance significantly influence the progression of cognitive impairment among individuals with Type 2 Diabetes Mellitus (Ajayi-Kaffi, et al, 2025) [1]. Among these markers, APOE-associated variants are widely recognized for their contribution to impaired amyloid clearance, oxidative stress amplification, and accelerated neuronal degeneration. Genetic susceptibility further interacts with metabolic instability and environmental stressors, thereby increasing heterogeneity in neurocognitive outcomes across diabetic populations (Balogun, et al, 2025) [15]. In African clinical cohorts, the underrepresentation of genomic datasets within predictive healthcare research has limited the development of context-specific risk stratification systems capable of accurately identifying vulnerable individuals at early stages of cognitive decline.

Recent advances in precision healthcare analytics have significantly improved the integration of genomic information with clinical and behavioral datasets for predictive neurological modeling. Machine learning-assisted healthcare systems now support automated pattern recognition, disease classification, and predictive prognosis using multidimensional biomedical datasets (Ijiga et al., 2024) [25]. These analytical systems improve predictive sensitivity by identifying complex nonlinear relationships among metabolic biomarkers, genomic variants, and neurocognitive indicators. Furthermore, algorithmic fairness studies emphasize that demographic representation optimization is essential for minimizing predictive bias and ensuring equitable healthcare analytics across underrepresented populations (Ifiala et al., 2026) [23]. This is particularly important in African healthcare environments where population-specific genomic diversity may substantially influence model performance and disease interpretation.

The increasing digitization of healthcare infrastructures has also strengthened the role of secure interoperable data systems in genomic-based predictive medicine. Interoperability frameworks facilitate standardized integration of electronic health records, laboratory data, and genomic information for large-scale analytical processing and precision healthcare deployment (Nwokocha et al., 2021) [37]. Similarly, secure healthcare intelligence systems utilizing advanced cyber-protection mechanisms have become increasingly important for safeguarding sensitive genomic and neurocognitive datasets during predictive modeling processes (Idika & Ijiga, 2025) [21]. Geo-analytic public health systems further demonstrate the growing importance of spatially informed healthcare intelligence for identifying disease vulnerabilities and improving equitable healthcare accessibility across underserved populations (Atalor, 2024) [12]. Collectively, these developments support the integration of genomic susceptibility markers with machine learning-assisted predictive frameworks for improving early detection and individualized cognitive risk assessment among Ghanaian diabetic populations.

2.5. Theoretical, Empirical, and Research Gap Review

The theoretical understanding of diabetes-associated cognitive decline increasingly supports multidimensional frameworks integrating metabolic dysfunction, psychosocial vulnerability, nutritional instability, and genomic susceptibility into unified predictive systems. Traditional biomedical theories primarily associate cognitive deterioration in Type 2 Diabetes Mellitus with insulin resistance, neuroinflammation, oxidative stress, and cerebrovascular impairment (Kpogli, et al, 2024) [28]. However, contemporary analytical perspectives emphasize that neurodegenerative progression is also shaped by environmental deprivation, behavioral stressors, and socioeconomic inequalities. Integrative theoretical models therefore advocate the use of hybrid healthcare intelligence systems capable of combining heterogeneous clinical and socio-environmental variables for improved precision risk prediction. Advanced analytical infrastructures designed for multidimensional information processing have demonstrated the growing relevance of integrated data ecosystems in supporting complex healthcare decision-making processes (Aluso et al., 2024) [4].

Empirical studies further demonstrate that machine learning and intelligent analytical systems substantially improve predictive performance in complex biomedical environments. Multi-source analytical frameworks utilizing blockchain-enabled data lineage verification have shown enhanced reliability, transparency, and consistency in multidimensional healthcare information processing (Aluso et al., 2023) [7]. Similarly, integrated neurodegenerative analytical systems combining biochemical and neurological datasets have demonstrated improved sensitivity in identifying hidden disease progression pathways and molecular interactions associated with neurocognitive impairment (Animasaun et al., 2025) [10]. Decision-support studies also reveal that advanced data visualization systems significantly strengthen analytical interpretation, strategic healthcare planning, and evidence-based risk assessment within high-stakes operational environments (Nortey, 2026) [32]. In addition, business analytics-based optimization frameworks highlight the effectiveness of multidimensional predictive systems in improving resource allocation efficiency and contextual decision support within vulnerable populations (Nortey et al., 2025) [33].

Despite these advancements, substantial empirical gaps remain in the integration of food insecurity, psychosocial stress, and genomic susceptibility variables into predictive frameworks assessing diabetes-associated cognitive decline among African populations. Existing predictive systems predominantly focus on isolated biomedical indicators while overlooking the combined influence of nutritional deprivation, chronic psychological stress, and population-specific genetic vulnerability. Furthermore, Ghanaian clinical cohorts remain significantly underrepresented in precision healthcare analytics and machine learning-assisted neurodegenerative research. Consequently, there is limited availability of context-sensitive socio-genomic predictive models capable of supporting equitable early-stage cognitive risk detection in low-resource healthcare systems. These limitations justify the development of the proposed Hybrid Socio-Genomic Risk Integration Algorithm to improve individualized cognitive vulnerability assessment among diabetic populations in Ghana.

3. Methodology

3.1. Research Design, Study Population, and Sampling

Procedures This study adopted a cross-sectional predictive analytical research design to investigate the interaction between Type 2 Diabetes Mellitus, food insecurity, psychosocial stress, genomic susceptibility, and cognitive decline among outpatient clinical cohorts in Kumasi, Ghana. The study population consisted of adult diabetic patients receiving treatment in selected tertiary healthcare facilities. Stratified random sampling was employed to ensure proportional representation across age, gender, and disease severity categories. The minimum sample size was determined using Cochran’s population estimation model:

??????=�2�(1−�) ??????2 where nrepresents sample size, Zis the standard normal deviation at 95% confidence interval, pdenotes estimated prevalence, and erepresents margin of error. Participant selection further satisfied inclusion criteria involving confirmed diabetes diagnosis and cognitive assessment eligibility. Predictive cohort structuring and multidimensional health analytics approaches have been shown to improve population-level healthcare modeling and clinical decision optimization (Nortey et al., 2026) [29].

3.2. Clinical, Psychosocial, Dietary, and Genomic Data

Collection Clinical data collection included fasting blood glucose, lipid profiles, blood pressure, and anthropometric measurements obtained from diabetic outpatient records and laboratory investigations. Body Mass Index (BMI) was computed as:

Psychosocial stress variables were assessed using standardized stress perception questionnaires, while dietary information and food insecurity indicators were collected using structured nutritional assessment instruments. Cognitive performance evaluation incorporated memory recall and executive function assessment tools. Genomic data collection involved peripheral blood sampling, DNA extraction, and genotyping of APOE-associated susceptibility variants linked to neurodegenerative dysfunction. The integrated dataset was normalized prior to predictive modeling using z-score transformation:

3.3. Development of the Hybrid Socio-Genomic Risk

Integration Algorithm (HSGRIA)

The Hybrid Socio-Genomic Risk Integration Algorithm (HSGRIA) was developed to integrate multidimensional predictors of diabetes-associated cognitive decline, including clinical biomarkers, genomic susceptibility markers, and psychosocial determinants. The predictive risk function was formulated as:

The algorithm incorporated machine learning-assisted feature ranking to improve predictive sensitivity and minimize redundant variables. Hybrid analytical systems integrating multidimensional predictive architectures have demonstrated improved efficiency in complex biomedical and operational decision-support environments (Aluso, et al, 2026) [5].

3.4. Model Validation, Comparative Analysis, and Ethical

Considerations The HSGRIA framework was validated using k-fold cross-validation to evaluate predictive stability, robustness, and generalization performance across heterogeneous clinical datasets. Predictive accuracy was computed as:

where �??????represents true positives, ��denotes true negatives, �??????indicates false positives, and ��represents false negatives. Comparative analysis was conducted against logistic regression, random forest, support vector machine, and gradient boosting algorithms using sensitivity, precision, recall, and F1-score metrics:

4. Results and Discussion

4.1. Demographic, Clinical, and Socioeconomic

Characteristics of Participants Table 2 presents the demographic, clinical, and socioeconomic characteristics of the study participants. The cohort consisted of individuals diagnosed with Type 2 Diabetes Mellitus recruited from outpatient clinical facilities in Kumasi, Ghana. Female participants represented a slightly higher proportion of the study population compared to males. Most participants were within the 50–59 year age category, indicating that middle-aged and older adults constituted the primary high-risk population for diabetes-associated cognitive decline. Socioeconomic assessment revealed that a substantial proportion of participants belonged to low-income households, supporting the hypothesis that financial instability and food insecurity may contribute to metabolic and neurocognitive vulnerability. Clinical assessment further demonstrated varying levels of cognitive decline severity, with mild cognitive impairment being the most frequently observed condition among participants.

Table 2: Demographic and Socioeconomic Characteristics Characteristic Frequency Male 78 Female 92 Age 40–49 46 Age 50–59 71 Age ≥60 53 Low Income 88 Middle Income 60 High Income 22 Mild Cognitive Decline 74 Moderate Cognitive Decline 61 Severe Cognitive Decline 35 Fig 2: Illustrates the gender distribution of participants included in the study cohort.

Fig 3: Presents the proportional distribution of cognitive decline severity among the diabetic participants.

Figures 2 and 3 summarize the demographic and cognitive characteristics of the study participants. The gender distribution shows that female participants were slightly more represented than males within the diabetic cohort. The cognitive severity chart further indicates that mild cognitive decline was the most common condition among participants, followed by moderate and severe cognitive impairment. These findings suggest that most participants were identified at the early stages of neurocognitive deterioration, highlighting the importance of early predictive assessment and intervention among individuals with Type 2 Diabetes Mellitus.

4.2. Effects of Food Insecurity and Psychosocial Stress on

Cognitive Decline The analysis revealed a significant association between food insecurity, psychosocial stress, and the severity of cognitive decline among participants with Type 2 Diabetes Mellitus. Participants experiencing high food insecurity demonstrated lower cognitive assessment scores and higher neurocognitive vulnerability compared to individuals with stable dietary access. Similarly, elevated psychosocial stress levels were associated with increased memory impairment, reduced executive functioning, and slower cognitive processing speed.

The findings suggest that chronic nutritional instability and prolonged psychological stress may intensify metabolic dysregulation, neuroinflammatory responses, and neuronal degeneration among diabetic populations. Table 3: Relationship Between Food Insecurity, Psychosocial Stress, and Cognitive Decline Variable Category Mean Cognitive Score Stress Index (%) Cognitive Decline Severity Low Food Insecurity 81.6 28.4 Mild Moderate Food Insecurity 69.3 47.8 Moderate High Food Insecurity 54.7 71.5 Severe Low Psychosocial Stress 84.2 22.1 Mild Moderate Psychosocial Stress 66.5 51.4 Moderate High Psychosocial Stress 49.8 79.3 Severe The results indicate a progressive reduction in cognitive performance as food insecurity and psychosocial stress levels increase. Participants within the high-stress and high-food-insecurity categories exhibited the lowest mean cognitive scores, suggesting that socioeconomic and psychological vulnerabilities substantially contribute to neurocognitive deterioration.

Fig 4: Food Insecurity and Cognitive Decline Trend Analysis Figure 4.2 line demonstrates a steady decline in cognitive scores as food insecurity severity increases. Participants with low food insecurity recorded the highest cognitive performance, while those experiencing high food insecurity exhibited substantially lower neurocognitive functioning, indicating a strong inverse relationship between nutritional instability and cognitive health.

Fig 5: Psychosocial Stress and Cognitive Vulnerability Heat Pattern The chart shows that cognitive performance decreases as psychosocial stress levels increase. Memory function, attention stability, executive functioning, and processing speed were highest among participants with low stress and lowest among those with high stress. These findings indicate that chronic psychosocial stress significantly contributes to cognitive decline in individuals with Type 2 Diabetes Mellitus.

The trend analysis demonstrates that cognitive scores decline progressively with increasing food insecurity severity. The heat pattern further illustrates that higher psychosocial stress levels correspond with substantial reductions in memory performance, executive functioning, attentional stability, and processing efficiency. These findings support the study hypothesis that psychosocial and nutritional determinants significantly influence cognitive vulnerability among diabetic clinical cohorts.

4.3. Genomic Susceptibility Patterns and Predictive

Performance of HSGRIA The genomic analysis identified significant variations in cognitive decline susceptibility among participants carrying APOE-associated risk variants. Participants with high-risk genomic profiles demonstrated elevated neurocognitive vulnerability, reduced memory performance, and increased metabolic instability compared to individuals with low-risk genotypes. The findings further showed that integrating genomic markers with psychosocial, dietary, and clinical variables substantially improved predictive sensitivity and classification stability within the Hybrid Socio-Genomic Risk Integration Algorithm (HSGRIA). Comparative evaluation against conventional machine learning models demonstrated that HSGRIA achieved superior accuracy, recall, and robustness in detecting early-stage cognitive decline among diabetic cohorts.

Table 4: Predictive Performance Comparison of HSGRIA and Benchmark Models Model Accuracy (%) Sensitivity (%) Specificity (%) F1-Score (%) Logistic Regression 78.4 75.1 79.6 76.8 Random Forest 84.7 82.9 85.4 83.6 Support Vector Machine 86.1 84.2 86.9 85.1 Gradient Boosting 88.3 87.5 89.1 87.9 HSGRIA 93.6 92.4 94.1 93.0 The results indicate that HSGRIA outperformed all comparative models across all predictive evaluation metrics. The integration of genomic susceptibility markers alongside psychosocial and dietary indicators significantly enhanced predictive reliability and early-stage cognitive decline detection.

Fig 6: Integrated Genomic Risk and HSGRIA Predictive Performance Analysis The heatmap demonstrates that participants with high genomic susceptibility exhibited substantially higher levels of memory decline, executive dysfunction, processing delays, and inflammatory activity compared to low-risk individuals. The radar distribution further illustrates the superior multidimensional predictive performance of HSGRIA relative to traditional machine learning approaches, particularly in sensitivity and overall classification accuracy.

4.4. Comparative Discussion, Clinical Implications, and

Public Health Relevance The comparative analysis demonstrated that the Hybrid Socio-Genomic Risk Integration Algorithm (HSGRIA) achieved substantially higher predictive reliability than conventional machine learning and traditional clinical scoring systems. The integration of psychosocial stress indicators, food insecurity variables, and genomic susceptibility markers improved early-stage detection sensitivity and reduced classification instability across heterogeneous diabetic cohorts. Unlike conventional models that rely primarily on biomedical indicators, HSGRIA provided multidimensional risk interpretation capable of identifying hidden neurocognitive vulnerability patterns within resource-limited populations.

Table 5: Comparative Clinical Interpretation of Predictive Models Predictive Model Early Detection Capacity Socioeconomic Integration Genomic Sensitivity Clinical Applicability Logistic Regression Moderate Low Moderate Moderate Random Forest High Low High High Support Vector Machine High Moderate High High Gradient Boosting Very High Moderate High Very High HSGRIA Excellent Excellent Excellent Excellent The findings indicate that HSGRIA achieved superior multidimensional analytical performance due to its ability to integrate biological and socio-environmental determinants simultaneously. This improved interpretability and strengthened its clinical applicability for vulnerable diabetic Fig 7: Clinical Decision Flow Matrix The clinical decision flow matrix illustrates how patient management intensity increases according to predicted neurocognitive risk severity. High-risk participants required comprehensive intervention strategies involving nutritional stabilization, psychosocial support, genomic monitoring, and continuous cognitive screening.

Fig 8: Population-Level Risk Distribution Funnel The funnel diagram demonstrates progressive stratification of neurocognitive vulnerability within the study population. Although all participants were diabetic patients, only a subset progressed into severe cognitive decline categories, highlighting the importance of early predictive screening and targeted public health interventions. The findings further emphasize the need for integrated healthcare policies addressing food insecurity, chronic psychosocial stress, and genomic vulnerability within diabetic populations in Ghana.

5. Conclusion and Recommendations

5.1. Summary of Findings

This study investigated the combined influence of food insecurity, psychosocial stress, and genomic susceptibility on diabetes-associated cognitive decline among Ghanaian clinical cohorts using the proposed Hybrid Socio-Genomic Risk Integration Algorithm (HSGRIA). The findings demonstrated that cognitive decline among individuals with Type 2 Diabetes Mellitus is strongly influenced not only by biomedical and metabolic variables but also by socioeconomic and psychological determinants. Participants experiencing high food insecurity consistently exhibited lower cognitive scores, impaired memory retention, reduced executive functioning, and slower processing speed compared to individuals with stable nutritional access. The study further established that psychosocial stress significantly intensified neurocognitive vulnerability through increased emotional burden, reduced attentional stability, and impaired adaptive cognitive performance. The genomic analysis revealed that APOE-associated susceptibility variants and related neurodegenerative markers substantially increased the probability of cognitive deterioration among diabetic participants. Individuals categorized within high genomic risk groups demonstrated elevated inflammatory activity and accelerated neurocognitive dysfunction. Comparative predictive evaluation further showed that HSGRIA outperformed logistic regression, random forest, support vector machine, gradient boosting, and traditional clinical risk scoring systems across all performance indicators, including accuracy, sensitivity, specificity, and F1-score. The integration of psychosocial stress metrics, food insecurity indices, and genomic variables significantly improved predictive robustness and early-stage detection sensitivity. The findings additionally demonstrated the importance of multidimensional healthcare intelligence systems in resource-limited clinical environments. HSGRIA provided improved interpretability and individualized risk stratification capabilities, thereby supporting precision public health interventions and context-sensitive neurocognitive risk assessment within vulnerable diabetic populations in Ghana.

5.2. Conclusion

This study demonstrated that diabetes-associated cognitive decline is a multidimensional public health challenge influenced by metabolic instability, nutritional deprivation, psychosocial stress, and genomic susceptibility. The findings confirmed that conventional biomedical prediction models alone are insufficient for accurately identifying neurocognitive vulnerability among diabetic populations, particularly within low-resource healthcare environments. By integrating clinical biomarkers, food insecurity indicators, psychosocial stress variables, and genomic susceptibility markers into a unified analytical framework, the Hybrid Socio-Genomic Risk Integration Algorithm significantly improved predictive sensitivity, classification stability, and individualized cognitive risk assessment. The study further established that food insecurity and chronic psychosocial stress substantially contribute to accelerated neurodegenerative dysfunction among diabetic individuals. Participants exposed to persistent nutritional instability and elevated stress levels exhibited significantly poorer cognitive outcomes, indicating that socioeconomic determinants play a critical role in neurological health progression. The inclusion of APOE-associated genomic markers additionally strengthened predictive reliability and enabled improved identification of high-risk individuals susceptible to severe cognitive impairment.

The superior performance of HSGRIA compared to conventional machine learning approaches demonstrated the importance of multidimensional socio-genomic modeling in precision healthcare analytics. The algorithm provided enhanced interpretability, equitable risk stratification, and contextualized prediction suitable for underserved populations often underrepresented in neurodegenerative research. These findings support the development of integrated healthcare policies and predictive healthcare infrastructures capable of combining biological, psychological, environmental, and genomic determinants for early-stage cognitive decline detection. The study therefore contributes significantly to precision public health, machine learning-assisted healthcare intelligence, and neurodegenerative disease prediction within African clinical populations.

5.3. Recommendations

Healthcare institutions managing diabetic populations should adopt multidimensional cognitive risk assessment systems that integrate psychosocial, nutritional, clinical, and genomic variables into routine patient evaluation processes. Traditional diabetes management frameworks should be expanded beyond glycemic monitoring to include early neurocognitive screening, psychosocial stress assessment, and food insecurity evaluation. This approach would support earlier identification of high-risk individuals and reduce progression toward severe cognitive impairment. Clinical facilities in Ghana and similar low-resource settings should implement predictive healthcare intelligence systems such as HSGRIA to improve individualized risk stratification and precision intervention planning. Integrating machine learning-assisted predictive models within outpatient diabetic care may significantly enhance early-stage detection sensitivity and improve healthcare decision-making efficiency. Healthcare providers should additionally strengthen nutritional intervention programs targeting diabetic patients experiencing food insecurity and chronic dietary instability. Structured nutritional counseling and food support initiatives may reduce inflammatory burden and improve long-term neurocognitive outcomes. Mental health and psychosocial support services should also be integrated into diabetes management frameworks due to the demonstrated relationship between chronic stress and cognitive vulnerability. Routine psychological assessment, stress management programs, and behavioral counseling may substantially reduce neurocognitive deterioration among vulnerable diabetic populations.

Public health authorities should invest in genomic screening infrastructure and population-specific neurodegenerative research to improve understanding of genetic susceptibility patterns within African populations. Furthermore, policymakers should prioritize healthcare equity initiatives addressing poverty, nutritional deprivation, and limited healthcare accessibility, as these factors significantly influence cognitive health outcomes. Strengthening healthcare data interoperability, secure genomic information management, and predictive healthcare analytics infrastructure would further improve large-scale implementation of precision public health systems.

5.4. Limitations of the Study

Several limitations were encountered during the course of this study. First, the research adopted a cross-sectional analytical design, which restricted the ability to establish long-term causal relationships between food insecurity, psychosocial stress, genomic susceptibility, and progressive cognitive decline among diabetic participants. Although significant predictive associations were identified, longitudinal monitoring would provide stronger evidence regarding temporal neurodegenerative progression and long-term cognitive outcomes.

The study was additionally limited by the size and geographic concentration of the sampled clinical cohort. Participants were recruited primarily from selected healthcare facilities within Kumasi, Ghana, which may limit the generalizability of findings across broader African populations with different socioeconomic and genetic characteristics. Variations in healthcare accessibility, dietary patterns, and population-specific genomic diversity may influence predictive model performance in other settings.

Another limitation involved the availability and accessibility of advanced genomic sequencing infrastructure. Resource constraints limited the inclusion of broader neurodegenerative genomic markers beyond selected APOE-associated susceptibility variants. Consequently, additional genetic interactions potentially influencing cognitive decline may not have been fully captured within the predictive framework.

Psychosocial stress and dietary variables were partly derived from self-reported survey instruments, which may introduce response bias and subjective variability. Furthermore, although HSGRIA demonstrated strong predictive performance, machine learning-assisted healthcare models remain dependent on data quality, preprocessing accuracy, and balanced population representation. Missing socio-clinical information and unequal demographic distribution may influence predictive consistency.

Despite these limitations, the study provided a robust multidimensional analytical framework for understanding diabetes-associated cognitive decline within underserved clinical populations.

5.5. Suggestions for Future Research

Future studies should adopt longitudinal cohort designs to examine the temporal progression of diabetes-associated cognitive decline and evaluate long-term interactions among genomic susceptibility, psychosocial stress, nutritional instability, and metabolic dysfunction. Continuous monitoring of neurocognitive outcomes over extended periods would improve understanding of disease trajectory and strengthen causal interpretation of predictive relationships identified within this study. Subsequent research should also expand genomic analysis beyond APOE-associated variants to include broader neurodegenerative susceptibility markers, inflammatory genes, epigenetic regulators, and metabolic signaling pathways associated with cognitive impairment. Integrating transcriptomic, proteomic, and metabolomic datasets may further improve predictive precision and support the development of advanced personalized healthcare systems. Future investigations should additionally incorporate larger multicenter African cohorts representing diverse ethnic, socioeconomic, and demographic populations to improve external validity and predictive generalizability. Comparative studies across urban and rural healthcare settings may provide deeper insight into the influence of healthcare accessibility, nutritional deprivation, and socioeconomic disparities on cognitive decline progression. Further research is also needed to evaluate real-time implementation of HSGRIA within clinical decision-support systems and electronic health infrastructures. Integrating wearable health monitoring devices, mobile health applications, and continuous patient surveillance technologies may enhance predictive responsiveness and individualized healthcare intervention planning. Finally, future studies should investigate the policy-level implications of integrating socio-genomic predictive systems into national healthcare strategies. Exploring cost-effectiveness, ethical governance, healthcare interoperability, and algorithmic fairness within precision public health deployment would strengthen large-scale adoption of multidimensional predictive healthcare frameworks in low-resource environments.