Scholarlink

<p>Process mining has become a crucial analytical approach for extracting data-driven insights from healthcare event logs to analyze and improve clinical workflows. Despite greater data availability, turning these insights into measurable improvements remains difficult. This systematic review (PRISMA 2020) synthesizes recent research to examine algorithmic trends, tool use, and healthcare-specific applications. Results confirm the widespread use of discovery techniques such as heuristic, inductive, and fuzzy miners, along with increasing adoption of hybrid and semantic-aware methods that incorporate machine learning for predictive analytics. Key applications include pathway analysis, conformance checking, and bottleneck detection. However, significant heterogeneity in study designs, contexts, and reported outcomes prevented a quantitative meta-analysis, highlighting the need for greater methodological standardization in the field. Ongoing obstacles, such as data heterogeneity, lack of standardization, privacy concerns, and limited validation, limit scalability. The review stresses the importance of FHIR-based standards, privacy-focused methods, comprehensive benchmarking, and closer collaboration between clinicians and data scientists.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Systematic review: a literature review of process mining in healthcare

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Aydin Fakhri Kaleybar,
  • Mohammadbagher Karimi,
  • Niloofar Naghshi,
  • Bahman Arasteh Abbasabad

摘要

Process mining has become a crucial analytical approach for extracting data-driven insights from healthcare event logs to analyze and improve clinical workflows. Despite greater data availability, turning these insights into measurable improvements remains difficult. This systematic review (PRISMA 2020) synthesizes recent research to examine algorithmic trends, tool use, and healthcare-specific applications. Results confirm the widespread use of discovery techniques such as heuristic, inductive, and fuzzy miners, along with increasing adoption of hybrid and semantic-aware methods that incorporate machine learning for predictive analytics. Key applications include pathway analysis, conformance checking, and bottleneck detection. However, significant heterogeneity in study designs, contexts, and reported outcomes prevented a quantitative meta-analysis, highlighting the need for greater methodological standardization in the field. Ongoing obstacles, such as data heterogeneity, lack of standardization, privacy concerns, and limited validation, limit scalability. The review stresses the importance of FHIR-based standards, privacy-focused methods, comprehensive benchmarking, and closer collaboration between clinicians and data scientists.