Volume 11 · Issue 01 · 2025

Chronic diseases such as heart disease, cancer, and diabetes are leading drivers of mortality worldwide, underscoring the need for improved efforts around early detection and prediction. The pathophysiology and management of chronic diseases have benefitted from emerging fields in molecular biology like genomics, transcriptomics, proteomics, glycomics, and lipidomics. The complex biomarker and mechanistic data from these "omics" studies present analytical and interpretive challenges, especially for traditional statistical methods. Machine learning (ML) techniques offer considerable promise in unlocking new pathways for data-driven chronic disease risk assessment and prognosis. This review provides a comprehensive overview of state-of-the-art applications of ML algorithms for chronic disease detection and prediction across datasets, including medical imaging, genomics, wearables, and electronic health records. Specifically, we review and synthesize key studies leveraging major ML approaches ranging from traditional techniques such as logistic regression and random forests to modern deep learning neural network architectures. We consolidate existing literature to date around ML for chronic disease prediction to synthesize major trends and trajectories that may inform both future research and clinical translation efforts in this growing field. While highlighting the critical innovations and successes emerging in this space, we identify the key challenges and limitations that remain to be addressed. Finally, we discuss pathways forward toward scalable, equitable, and clinically implementable ML solutions for transforming chronic disease screening and prevention.

The rising incidence and death rates linked to Alzheimer's disease (AD) highlight an urgent issue. Genetic screening is celebrated as a significant advancement for its early detection capabilities, pinpointing those at risk before the emergence of symptoms. Yet, the limited availability of these technologies highlights a critical gap in widespread application. This review pivots to the potential of presymptomatic clinical assessments as a readily available, economical, and simple strategy for early detection. Traditionally, AD diagnosis relies on the late-stage identification of cognitive deterioration, functional impairments, and neuropsychiatric symptoms, coinciding with advanced brain degeneration. Conversely, emerging research identifies early indicators preceding significant degeneration, manifesting years before clinical symptoms. We introduce a mnemonic, MEMORIES, to categorize these prodromal: Metabolism changes, Eye/visual impairments, March (refer to gait disturbances), Olfactory dysfunction, Rhythm (blood pressure and heart rate), Insensitivity of the tongue, Ears (hearing loss), and Stool alterations. Recognizing these prodromal through clinical examinations provides a valuable strategy for initiating preventative actions against brain degeneration. This approach advocates for broadening the screening lens beyond genetic screening to encompass clinical evaluations, enhancing early detection and intervention opportunities for AD.

During antenatal care, gestational diabetes mellitus (GDM) screening is crucial for early diagnosis and treatment to ameliorate clinical outcomes and limit health care expenses. Dietary management and physical activity are central to GDM treatment, however, adherence is often influenced by personal preferences, socioeconomic barriers, and psychological stress. Pharmacologically, insulin and oral hypoglycemic medications, are the main GDM treatment that can be subject to patients' resistance due to fears of needles and side effects. Metformin is increasingly preferred for its ease of administration and lower cost. In the postpartum stage, regular screening for type 2 diabetes mellitus (T2DM) should always be considered despite the possible limitations that could arise, including communication gaps, lack of long-term focus, and personal barriers. Overall, women with GDM prefer personalized, flexible management plans that consider their lifestyle, support from health care professionals (HCPs), and family involvement. Addressing psychological and socioeconomic barriers through education, counseling, and support networks is crucial for improving adherence and health outcomes. Enhancing patient-centered care and shared decision-making can empower women with GDM to manage their condition effectively and maintain lifestyle changes postpartum. Therefore, this review aimed to assess pregnant women's preferences in GDM management, focusing on screening, dietary recommendations, physical activity, and treatment. Additionally, this review examined GDM care in terms of these patients' quality of life and postpartum experiences.

The prevalence of type 2 diabetes has been growing among younger and middle-aged adults in the United States. A portion of this increase for this age group may be attributable to shared type 2 diabetes risks with family members. How family history of type 2 diabetes history is associated with type 2 diabetes risk among younger and middle-aged adults is not well understood.

This population-based retrospective cohort study uses administrative, genealogical, and electronic medical records from the Utah Population Database. The study population comprises offspring born between 1970 and 1990 and living in the four urban Utah counties in the United States between 1990 and 2015. The sample comprises 360,907 individuals without a type 2 diabetes diagnosis and 14,817 with a diagnosis. Using multivariate logistic regressions, we estimate the relative risk (RR) of type 2 diabetes associated with the number of affected first- (FDRs), second- (SDRs), and third-degree (first cousin) relatives for the full sample and for Hispanic-specific and sex-specific subsets.

Individuals with 2+ FDRs with type 2 diabetes have a significant risk of type 2 diabetes in relation to those with no affected FDRs (RR = 3.31 [3.16, 3.48]). Individuals with 2+ versus no SDRs with type 2 diabetes have significant but lower risks (RR = 1.32 [1.25, 1.39]). Those with 2+ versus no affected first cousins have a similarly low risk (RR= 1.28 [1.21, 1.35]). Larger RRs are experienced by males (2+ vs. 0 FDRs, RR = 3.55) than females (2+ vs. 0 FDRs, RR = 3.18) (p < 0.05 for the interaction). These familial associations are partly mediated by the individual's own obesity.

The risks of type 2 diabetes are significantly associated with having affected first-, second-, and third-degree relatives, especially for men. One of the forces contributing to the rising patterns of type 2 diabetes among young and middle-aged adults is their connection to affected, often older, kin.

The patterns of dual antiplatelet therapy (DAPT) use and the associated clinical outcomes in current practice remain limited. This study evaluates DAPT regimen patterns and clinical outcomes among acute coronary syndrome (ACS) patients undergoing percutaneous coronary intervention (PCI).

This multicenter retrospective cohort study included ACS patients treated with PCI from January 2017 to February 2022 at five tertiary hospitals in Thailand. DAPT was categorized as nonpotent (NP-DAPT) or potent (P-DAPT). We described DAPT trends, with major adverse cardiovascular events (MACEs) and major bleeding, as primary efficacy and safety outcomes. Outcomes were assessed using inverse probability treatment weighting (IPTW) with Cox's proportional hazards model.

The study included 1877 patients with ACS undergoing PCI. The mean age was 64.51 years (standard deviation 11.34), with 639 (34.04%) female patients and 1159 (61.75%) presenting ST-elevation myocardial infarction (STEMI). Of these, 924 (49.23%) received NP-DAPT, and 953 (50.77%) were prescribed P-DAPT. Crude MACE incidence was lower in the P-DAPT compared to the NP-DAPT group (6.82% vs. 10.28%). After applying IPTW and conducting Cox's proportional hazard analysis, no significant differences in MACE were observed between groups (hazard ratio [HR]: 0.85, 95% confidence interval [CI]: 0.58-1.25, p = 0.408), nor in major bleeding (HR: 0.80, 95% CI: 0.37-1.70, p = 0.555). P-DAPT was associated with any higher bleeding risk (HR: 1.52, 95% CI: 1.13-2.03, p = 0.005).

Standard DAPT remains predominant among Thai ACS patients, with NP-DAPT prescriptions approaching those of P-DAPT. Despite similar rates of MACE and major bleeding between the groups, P-DAPT was associated with a higher risk of any bleeding.

The prevalence of type 2 diabetes mellitus (T2DM) has been rapidly growing in Chinese populations in recent decades, and the shift in eating habits is a key contributing factor to this increase. Eating out of home (EOH) is one of the major shifts in eating habits during this period. However, the influence of EOH on the incidence of T2DM among Chinese urban workers is unknown.

The cross-sectional study involved an analysis of 13,904 urban workers recruited from 11 health examination centers in the major cities of China to explore the relationship between EOH and T2DM between 2013 September and 2016 March.

Average weekly EOH frequency ≥10 times was positively associated with increased incidence of T2DM in the sampled population (OR: 1.31 [1.11-1.54], p < 0.01), most notably in participants ≤45 years old (OR: 1.41[1.11-1.80], p < 0.01]) and in males (OR:1.26 [1.06-1.51], p < 0.01). An EOH frequency of 5 times/week appears as a threshold for a significant increase in the odds of T2DM. Weekly EOH frequency ≥5 times was associated with increased odds of T2DM in a dose-response manner in the total population and almost all subgroups (p_{overall association} < 0.05 and p_nonlinearity ≤ 0.05).

This study showed that a frequency of EOH (≥5 times/week) was associated with a frequency-dependent increase in the odds of T2DM urban workers in China. More nutrition promotion is needed to improve the eating behavior of Chinese urban workers to reduce T2DM risk.

对小编来说，