Publications

Improving the Conduct of Systematic Reviews: A Process Mining Perspective

Ba' Pham and Ebrahim Bagheri and Patricia Rios and Asef Pourmasoumi and Reid C. Robson and Jeremiah Hwee and Wanrudee Isaranuwatchai and Nazia Darvesh, Matthew Page and Andrea Tricco
Reference:
Ba' Pham; Ebrahim Bagheri; Patricia Rios; Asef Pourmasoumi; Reid C. Robson; Jeremiah Hwee; Wanrudee Isaranuwatchai; Nazia Darvesh, Matthew Page and Andrea Tricco Improving the Conduct of Systematic Reviews: A Process Mining Perspective. In Journal of Clinical Epidemiology, 2018.
Links to Publication: [www][pdf]
Abstract:
Objectives: To demonstrate the feasibility of using process mining concepts, techniques, and tools to examine and improve the systematic review process. Study Design and Setting: We conducted a simulation study evaluating a process used by one research team over one year. The process was characterized using an event log of review activities, start/end dates for review tasks, reviewers, and person-hours spent on tasks. We obtained process models from mining event logs for visual display/animation/replay of review activities. We analyzed the social networks of reviewer interactions to discern how reviewers worked together. Key outcomes included review timelines and person-time. Results: The 12 reviews included in the study included an average of 3831 titles and abstracts (range:1565-6368) and 20 studies (6-42). The average time was 463 days (range: 289-629) [881 person-hours (range: 243-1752)] per review. The average person-hours on each step were: study selection 26%, data abstraction 24%, report preparation 23%, and meta-analysis 17%. Social network analyses showed that the team handled tasks according to their expected roles (e.g., methodologists developed review questions, librarians conducted searches, and review coordinators coordinated tasks). Conclusion: Process mining is valuable for review teams interested in improving and modernizing the conduct of systematic reviews.
Bibtex Entry:
@article{jce2018, author = {Ba' Pham and Ebrahim Bagheri and Patricia Rios and Asef Pourmasoumi and Reid C. Robson and Jeremiah Hwee and Wanrudee Isaranuwatchai and Nazia Darvesh, Matthew Page and Andrea Tricco}, title = {Improving the Conduct of Systematic Reviews: A Process Mining Perspective}, journal = {Journal of Clinical Epidemiology}, url = {https://www.journals.elsevier.com/journal-of-clinical-epidemiology}, year = {2018}, abstract = {Objectives: To demonstrate the feasibility of using process mining concepts, techniques, and tools to examine and improve the systematic review process. Study Design and Setting: We conducted a simulation study evaluating a process used by one research team over one year. The process was characterized using an event log of review activities, start/end dates for review tasks, reviewers, and person-hours spent on tasks. We obtained process models from mining event logs for visual display/animation/replay of review activities. We analyzed the social networks of reviewer interactions to discern how reviewers worked together. Key outcomes included review timelines and person-time. Results: The 12 reviews included in the study included an average of 3831 titles and abstracts (range:1565-6368) and 20 studies (6-42). The average time was 463 days (range: 289-629) [881 person-hours (range: 243-1752)] per review. The average person-hours on each step were: study selection 26%, data abstraction 24%, report preparation 23%, and meta-analysis 17%. Social network analyses showed that the team handled tasks according to their expected roles (e.g., methodologists developed review questions, librarians conducted searches, and review coordinators coordinated tasks). Conclusion: Process mining is valuable for review teams interested in improving and modernizing the conduct of systematic reviews.}, webpdf = {http://ls3.rnet.ryerson.ca/wiki/images/f/f1/Jce2018.pdf} }




Powered by WordPress