Process mining of breast cancer care pathways from routine electronic medical records: coordination patterns, transition-time variation, and hospital-level differences

Abstract Background Breast cancer care involves multiple services over several months, yet most research has focused on individual treatment components. As a result, we still know relatively little about how clinical and coordination events are sequenced across services in routine practice. Linked electronic medical record (EMR) data and process mining provide an opportunity to reconstruct these patient-level pathways. This study aims to characterise breast cancer care pathways, identify recurring coordination patterns and transition-time variation, and examine their associations with measured clinical characteristics and treating hospital. Methods A retrospective cohort study was conducted using linked EMR data from 686 patients with invasive breast cancer who underwent surgery within the Illawarra Shoalhaven Local Health District, Australia, between 2010 and 2014. Patient-level pathways spanning diagnosis, care coordination and treatment initiation events during the initial treatment episode were reconstructed from three routine clinical information systems using process mining. Coordination patterns were classified as MDT-led or Referral-led based on the relative timing of multidisciplinary team (MDT) discussion and oncology referral. Transition times between care events were quantified, and multivariable logistic regression was used to identify factors associated with coordination pattern and pre-treatment oncology consultation. Results The event log contained 4,175 timestamped events and 143 distinct pathway variants. Among these variants, 81 (56.6%) were observed in only one patient each. Despite this patient-level heterogeneity, two recurring coordination patterns accounted for most patients: MDT-led, in which MDT discussion occurred before or on the same date as oncology referral, or Referral-led, in which referral occurred first. After adjustment, treating hospital was the only included model variable significantly associated with coordination pattern (Shoalhaven vs. Wollongong: OR 2.50, 95% CI 1.54–4.06). Molecular subtype, stage, grade, and age were not significantly associated. Timing also varied across selected transitions, with several intervals longer at Shoalhaven, including diagnosis to surgery and surgery to chemotherapy and radiotherapy. Conclusions Process mining of linked EMR data revealed substantial patient-level pathway heterogeneity alongside recurring patterns in the ordering of MDT discussion and oncology referral. In this historical cohort, coordination pattern was associated with treating hospital after adjustment for measured clinical characteristics, and hospital differences were also observed in selected transition times. These findings identify areas for service-level review, while the available data do not establish the organisational mechanisms underlying these differences or whether one coordination pattern leads to better care or outcomes.

Authors

Institutions

Publication Details

Journal
BMC Medical Informatics and Decision Making
Published
2026-11-14
DOI
https://doi.org/10.1186/s12911-026-03989-x
Primary Topic
Business Process Modeling and Analysis
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Process mining of breast cancer care pathways from routine electronic medical records: coordination patterns, transition-time variation, and hospital-level differences

Su C. Ang, Kimberley J. Davis, Ping Yu, Steven J. Craig et al.
3 links
BMC Medical Informatics and Decision Making
Business Process Modeling and Analysis
article

Process mining of breast cancer care pathways from routine electronic medical records: coordination patterns, transition-time variation, and hospital-level differences

Su C. Ang, Kimberley J. Davis, Ping Yu, Steven J. Craig, Rebekah Costelloe, Ting Song, Chantal Campbell
article en

Abstract

Abstract Background Breast cancer care involves multiple services over several months, yet most research has focused on individual treatment components. As a result, we still know relatively little about how clinical and coordination events are sequenced across services in routine practice. Linked electronic medical record (EMR) data and process mining provide an opportunity to reconstruct these patient-level pathways. This study aims to characterise breast cancer care pathways, identify recurring coordination patterns and transition-time variation, and examine their associations with measured clinical characteristics and treating hospital. Methods A retrospective cohort study was conducted using linked EMR data from 686 patients with invasive breast cancer who underwent surgery within the Illawarra Shoalhaven Local Health District, Australia, between 2010 and 2014. Patient-level pathways spanning diagnosis, care coordination and treatment initiation events during the initial treatment episode were reconstructed from three routine clinical information systems using process mining. Coordination patterns were classified as MDT-led or Referral-led based on the relative timing of multidisciplinary team (MDT) discussion and oncology referral. Transition times between care events were quantified, and multivariable logistic regression was used to identify factors associated with coordination pattern and pre-treatment oncology consultation. Results The event log contained 4,175 timestamped events and 143 distinct pathway variants. Among these variants, 81 (56.6%) were observed in only one patient each. Despite this patient-level heterogeneity, two recurring coordination patterns accounted for most patients: MDT-led, in which MDT discussion occurred before or on the same date as oncology referral, or Referral-led, in which referral occurred first. After adjustment, treating hospital was the only included model variable significantly associated with coordination pattern (Shoalhaven vs. Wollongong: OR 2.50, 95% CI 1.54–4.06). Molecular subtype, stage, grade, and age were not significantly associated. Timing also varied across selected transitions, with several intervals longer at Shoalhaven, including diagnosis to surgery and surgery to chemotherapy and radiotherapy. Conclusions Process mining of linked EMR data revealed substantial patient-level pathway heterogeneity alongside recurring patterns in the ordering of MDT discussion and oncology referral. In this historical cohort, coordination pattern was associated with treating hospital after adjustment for measured clinical characteristics, and hospital differences were also observed in selected transition times. These findings identify areas for service-level review, while the available data do not establish the organisational mechanisms underlying these differences or whether one coordination pattern leads to better care or outcomes.

BMC Medical Informatics and Decision Making
Griffith University (AU), University of Wollongong (AU), Shoalhaven District Memorial Hospital (AU), Illawarra Shoalhaven Local Health District (AU), Macquarie University (AU)
University of Wollongong
Good health and well-being
Business Process Modeling and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon