Austin Robin R, Jantraporn Ratchada, Michalowski Martin, Marquard Jenna
School of Nursing, University of Minnesota, Minneapolis, Minnesota, USA.
Institute for Health Informatics, Minneapolis, Minnesota, USA.
J Nurs Scholarsh. 2025 Jan;57(1):72-81. doi: 10.1111/jnu.13025. Epub 2024 Sep 9.
A whole person approach to healthy aging can provide insight into social factors that may be critical. Digital technologies, such as mobile health (mHealth) applications, hold promise to provide novel insights for healthy aging and the ability to collect data between clinical care visits. Machine learning/artificial intelligence methods have the potential to uncover insights into healthy aging. Nurses and nurse informaticians have a unique lens to shape the future use of this technology.
The purpose of this research was to apply machine learning methods to MyStrengths+MyHealth de-identified data (N = 988) for adults 45 years of age and older. An exploratory data analysis process guided this work.
Overall (n = 988), the average Strength was 66.1% (SD = 5.1), average Challenges 66.5% (SD = 7.5), and average Needs 60.06% (SD = 3.1). There was a significant difference between Strengths and Needs (p < 0.001), between Challenges and Needs (p < 0.001), and no significant differences between average Strengths and Challenges. Four concept groups were identified from the data (Thinking, Moving, Emotions, and Sleeping). The Thinking group had the most statistically significant challenges (11) associated with having at least one Thinking Challenge and the highest average Strengths (66.5%) and Needs (83.6%) compared to the other groups.
This retrospective analysis applied machine learning methods to de-identified whole person health resilience data from the MSMH application. Adults 45 and older had many Strengths despite numerous Challenges and Needs. The Thinking group had the highest Strengths, Challenges, and Needs, which aligns with the literature and highlights the co-occurring health challenges experienced by this group. Machine learning methods applied to consumer health data identify unique insights applicable to specific conditions (e.g., cognitive) and healthy aging. The next steps involve testing personalized interventions with nurses leading artificial intelligence integration into clinical care.
采用全人方法促进健康老龄化能够深入了解可能至关重要的社会因素。数字技术，如移动健康（mHealth）应用程序，有望为健康老龄化提供新的见解，并具备在临床护理就诊之间收集数据的能力。机器学习/人工智能方法有潜力揭示健康老龄化的相关见解。护士和护理信息学家在塑造这项技术的未来应用方面具有独特视角。
本研究的目的是将机器学习方法应用于45岁及以上成年人的MyStrengths+MyHealth去识别化数据（N = 988）。一个探索性数据分析过程指导了这项工作。
总体而言（n = 988），平均优势为66.1%（标准差 = 5.1），平均挑战为66.5%（标准差 = 7.5），平均需求为60.06%（标准差 = 3.1）。优势与需求之间存在显著差异（p < 0.001），挑战与需求之间存在显著差异（p < 0.001），平均优势与挑战之间无显著差异。从数据中识别出四个概念组（思维、运动、情绪和睡眠）。与其他组相比，思维组在与至少一项思维挑战相关的统计上最显著的挑战（有十一项）、最高的平均优势（66.5%）和需求（83.6%）。
这项回顾性分析将机器学习方法应用于来自MSMH应用程序的去识别化全人健康恢复力数据。45岁及以上的成年人尽管面临诸多挑战和需求，但仍有许多优势。思维组的优势、挑战和需求最高，这与文献一致，并突出了该组共同面临的健康挑战。应用于消费者健康数据的机器学习方法识别出适用于特定状况（如认知）和健康老龄化的独特见解。下一步涉及测试由护士主导将人工智能整合到临床护理中的个性化干预措施。
