Genetic risks underlying all-cause dementia and overlap with vascular dementia
Most genome-wide association studies focus on Alzheimer's disease specifically and overlook other types of dementia. The Mega Vascular Cognitive Impairment and Dementia (MEGAVCID) (2024) consortium examined all-cause dementia, as well as the genetic overlap with vascular dementia in particular using a genome-wide association studies.
A dataset of 800 597 people across North America, Europe and Asia included 46 902 cases of all-cause dementia and 8702 cases of vascular dementia. This sample included individuals of European (98.5%), African (1.0%), Asian (0.4%) and Hispanic (0.1%) descent (mean age 54–80 years; 54–68% female).
The researchers sought to identify specific loci (positions on a chromosome where a particular gene or DNA sequence is located) associated with all-cause dementia and vascular dementia by initially replicating those known for Alzheimer's disease and prioritising genes that were likely to be functionally relevant with related traits and risk factors.
Genotyping was carr ied out using genotyping arrays specific to each cohort and genetic variants were imputed using the 1000 genomes project. The authors conducted study- and ethnicity-specific association analyses adjusting for age, sex, sites and population structure to test the association of each variant with vascular and all-cause dementias.
The novel loci identified for all-cause dementia were associated with the transportation of energy, the excitability of neurons, amyloid deposition in the brain and magnetic resonance imaging of small-vessel disease. For vascular dementia, novel loci were associated with hypertension, diabetes and neuron maintenance.
This study identified genetic risks underlying all-cause dementia and demonstrated their overlap with neurogenerative processes, vascular risk factors and cerebral small-vessel disease. It also identified likely genetic variants and biological pathways for all-cause and vascular dementias.
However, limitations included variation in diagnostic criteria for all-cause and vascular dementia, as well as a dataset of people consisting of participants mainly of European ancestry.
Improving dementia prediction using the cardiovascular risk factors, ageing and incidence of dementia risk score
Based on the Health 2000 Survey, Pietilä et al's (2025) study included a representative Finnish cohort of 5806 adults aged 30 years and above. Participants were observed for 19 years, during which 571 of them developed dementia; the majority of these cases being Alzheimer's disease. The cardiovascular risk factors, ageing and incidence of dementia (CAIDE) risk score predicted dementia with good accuracy and similar results were found when replacing body mass index with insulin resistance. When APOE ε4 (the strongest genetic risk factor for Alzheimer's disease) status was added to the model, predictive accuracy increased slightly.
Higher CAIDE scores were associated with significantly increased dementia risk (1.6% in the lowest category versus 26.3% in the highest). The study further confirmed that lifestyle-related factors such as physical inactivity, hypertension and high cholesterol were more prevalent in higher CAIDE categories. While insulin resistance and obesity were closely related, insulin resistance did not provide additional predictive benefits beyond that of body mass index. Limitations of the study included the use of registry-based diagnoses, which may under-report dementia, and the single-point collection of risk factor data. Although the inclusion of insulin resistance was methodologically sound, the authors noted that it lacked standardised cut-off points for clinical use and required blood tests, limiting the practicality of this approach.
Pietilä et al (2025) concluded that the CAIDE risk score was a useful, accessible tool for long-term dementia risk prediction within the general population. They emphasised the importance of targeting modifiable midlife risk factors and suggested that the CAIDE score could support preventive efforts in primary and community care settings.
Augmenting Israeli community nursing practice with artificial intelligence-based simulated cases
With the development of artificial intelligence (AI), generative AI (GenAI) in particular, the landscape of problem solving in healthcare has been transformed. These technologies show potential for rapid evolution in healthcare. However, research is still required before potential benefits can be fully realised.
Saad et al (2025) sought to compare community nurses against state-of-the-art GenAI in terms of their diagnostic accuracy and clinical decision making in simulated patient case scenarios using online surveys. This cross-sectional study, which took place between May and July 2024, evaluated four participant groups, one of which consisted of 114 Israeli human community nurses who were recruited using a snowball sampling technique via the Qualtrics XM online survey platform.
The other groups consisted of the following GenAI models: ChatGPT-4, Claude-3.0 and Gemini-1.5. The community nurses were drawn from primary care clinics, home healthcare services, professional community clinics with specialised services (such as diabetes or cardiology care) and community urgent care centres.
A clinical reasoning questionnaire was distributed to the human participants, featuring four clinical scenarios that they were required to assess, interpret related diagnostic tests for and determine strategies for appropriate patient management.
The GenAI models were given the same scenarios to assess and make treatment recommendations for, both with and without a word count restraint, in order for the researchers to evaluate the impact of their response length on the quality of their clinical reasoning and their conciseness.
Overall, the nurses in this study scored higher than the shortened AI responses from the GenAI models. The GenAI responses were faster and also more verbose, often containing unnecessary information. While ChatGPT was the first widely accessible large-language model in late 2022, this study demonstrated that the Claude-3.0 and Gemini-1.5 models (both full versions) achieved the highest accuracy score among the GenAI models included. The main limitations of this study were the small number of scenarios and their static nature, resulting in a lack of dynamism observed in practice.
According to this study, GenAI technology shows promise in terms of its potential to support community nursing practice. However, human nurses currently produce advantages in holistic clinical reasoning and the authors note that this skill requires not only experience, but also contextual knowledge and the ability to be concise and practical in responding. Real-world testing is required before GenAI models can be used to effectively support community nurses and other healthcare professionals to carry out various aspects of their work and to support patients (Figure 1).
Examining the impact of community nursing on health promotion in China
The role of community nurses in health promotion is well recognised in the UK, and is being increasingly acknowledged around the world. While community nurses’ interactions with patients and their impact on outcomes inevitably influence social power through their health promotion role, this aspect of community nursing is rarely examined or discussed.
In a Chinese study, Li (2025) uses an interpretive case study of hypertension management to examine the role of community nursing in health promotion in Shenzhen, China. The author draws on Michael Foucault's concept of biopower to enrich the analysis.
Real-world observations of nursing practices in hypertension management were combined with self-reported experiences gathered through semi-structured interviews with 22 community nurses via WeChat video calls in March 2023. Each call lasted between 80 and 110 minutes and included open-ended questions.
The study noted that since community nursing advances health promotion at both individual and population levels, it operates as a biopower mechanism. It uses three main biopower techniques that were examined in this study, which are persuasive, constructive and evidence-based.
These techniques are training, knowledge and social policy. Examples of such techniques include the nursing practices of patient follow ups, health education lectures and health statistics.
The author argued that community nurses went beyond providing care to individuals and actively shaped health governance. Li (2025) highlighted the contributions of community nursing to public health within Shenzhen's evolving healthcare landscape through an interpretive framework of the three key techniques mentioned, which are grounded in biopower.
Li (2025) concluded that the insights from this study offered implications for the development of policy and the optimisation of primary healthcare across diverse regions in China.