<![CDATA[Background: Mild cognitive impairment (MCI) is a critical transition stage toward Alzheimer’s disease (AD). While standard longitudinal and survival methods evaluate disease progression separately, joint modeling handles their complex interdependencies. However, no systematic review has exclusively evaluated joint modeling for predicting MCI-to-AD conversion. Objectives: To examine existing joint modeling approaches used to predict Alzheimer’s disease progression, with particular attention to modeling strategies, data sources, biomarkers, and methodological limitations. Methods: Scopus, Web of Science, and PubMed were searched from January 1999 to March 18, 2025. A total of 1614 records were identified, of which 891 duplicates were removed. After screening and assessing eligibility, 32 studies were included. Study selection was conducted in accordance with the PRISMA 2020 guidelines. The methodological quality of the included studies was assessed using a structured five-question quality assessment framework developed for this review. Eligible studies were original, peer-reviewed articles in English applying joint modeling for AD progression. A structured narrative synthesis was performed based on predefined categories, including modeling methodologies, data sources, biomarkers, and methodological limitations. Results: Among the 32 included studies, 50% (16/32) used a linear mixed-effects model combined with a Cox Proportional Hazards model, making this the most common approach. Most studies (60.5%, 20/32) used data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). The limited use of independent external cohorts was identified as a major constraint affecting the generalizability of the reported findings. Conclusion: This review provides a structured synthesis of joint modeling approaches for predicting AD progression and highlights key methodological gaps, particularly in MCI. Despite their growing use and methodological advantages, limited external validation restricts the generalizability of current findings. Clinically, joint modeling can help nurses stratify risk early, improve monitoring, and personalize care for patients at risk of Alzheimer’s disease. Dynamic, individualized predictions can guide cognitive assessment timing, prompt specialist referrals, inform caregiver education, and support proactive management and disease prevention. Registry: PROSPERO [CRD420261300219]]]>
Copyrights © 2026
