Qin Guoqiang, Wei Jianxiang, Sun Yuehong, Du Wenwen
School of Management, Nanjing University of Posts and Telecommunications, Nanjing, People's Republic of China.
Library, Nanjing University of Posts and Telecommunications, Nanjing, People's Republic of China.
J Multidiscip Healthc. 2025 May 10;18:2603-2627. doi: 10.2147/JMDH.S516826. eCollection 2025.
Causal inference in clinical medicine provides scientific evidence for precision medicine and individualized treatment by revealing the true associations between interventions and health outcomes. This study aims to conduct a comprehensive bibliometric analysis to identify current research trends, primary themes, and future directions for the application of causal inference in clinical medicine.
We conducted a literature search in the Web of Science database using causal inference and medical terminology as subject keywords, covering the period from January 1986 to December 2024. After screening, we obtained 4,316 documents for analysis. Utilizing CiteSpace to generate network diagrams, we analyzed data related to the number of publications, citation analysis, collaboration relationships, keyword co-occurrence, and highlighted terms to illustrate the knowledge map and collaboration network in this field.
Publications on medical causal inference shows a fluctuating growth trend over time. The United States was the top contributors to this field. Harvard University is the leading research institution. George David Smith is the most prolific author, Robbins JM is the most cited scholar. The major research hotspots concentrated in fields such as epidemiology, coronary heart disease and health. Notably, marginal structural models, counterfactual forecasting, and Mendelian randomization have consistently been key methodologies in research. The burstness of keywords reveals that big data, DNA methylation, and robust estimation are emerging research directions.
In clinical research, counterfactual forecasting provides prospective guidance for optimizing clinical strategies; Mendelian randomization helps uncover potential therapeutic targets; and marginal structural models enhance the accuracy of causal effect estimation in clinical studies. The future integration of various data sources to improve causal inference methods is anticipated to enhance the sensitivity and specificity of trials, ultimately elucidating the complex mechanisms of diseases and drug effects. The literature retrieve strategy and the metrics of the tools adopted may have a certain impact on the results of this study.
临床医学中的因果推断通过揭示干预措施与健康结果之间的真实关联，为精准医学和个体化治疗提供科学依据。本研究旨在进行全面的文献计量分析，以确定临床医学中因果推断应用的当前研究趋势、主要主题和未来方向。
我们在科学网数据库中进行文献检索，使用因果推断和医学术语作为主题关键词，涵盖1986年1月至2024年12月期间。筛选后，我们获得4316篇文献进行分析。利用CiteSpace生成网络图，我们分析了与出版物数量、引文分析、合作关系、关键词共现和突出显示的术语相关的数据，以说明该领域的知识图谱和合作网络。
医学因果推断方面的出版物随时间呈波动增长趋势。美国是该领域的最大贡献者。哈佛大学是领先的研究机构。乔治·大卫·史密斯是最多产的作者，罗宾斯·J·M是被引用最多的学者。主要研究热点集中在流行病学、冠心病和健康等领域。值得注意的是，边际结构模型、反事实预测和孟德尔随机化一直是研究中的关键方法。关键词的突发性表明大数据、DNA甲基化和稳健估计是新兴的研究方向。
在临床研究中，反事实预测为优化临床策略提供前瞻性指导；孟德尔随机化有助于发现潜在的治疗靶点；边际结构模型提高了临床研究中因果效应估计的准确性。预计未来整合各种数据源以改进因果推断方法将提高试验的敏感性和特异性，最终阐明疾病和药物效应的复杂机制。文献检索策略和所采用工具的指标可能对本研究结果有一定影响。
