Opportunities and challenges when using record linkage of routinely collected electronic health care data to evaluate outcomes of systemic anti-cancer treatment in clinical practice

Mueller, T.; Laskey, J.; Baillie, K.; Clarke, J.; Crearie, C.; Kavanagh, K.; Graham, J.; Graham, K.; Waterston, A.; Jones, R.; Kurdi, A.; Morrison, D.; Bennie, M.

Objectives: To discuss the opportunities and challenges when applying an electronic record linkage methodology with respect to systemic anti-cancer therapy, and to highlight some of the potential pitfalls spanning the entire breadth and depth of the research process. Design: Retrospective cohort studies using routinely collected, administrative health data. Setting: Scotland Results: Studies conducted to-date have indicated that record linkage of routinely collected data to determine outcomes of treatment with cancer medicines is feasible, albeit currently within certain limits. While the general description of patient populations and the calculation of median overall survival are well supported, prevailing issues with combining data across regional boundaries and the limited availability of some variables (including molecular pathology data and information regarding toxicities) may restrict the extent of analyses feasible. Conclusion: There is scope to conduct large cohort studies to generate results from clinical practice using linkage of routinely collected health care data within a reasonable time frame; however, close collaboration between researchers, data controllers, and clinicians is required in order to obtain valid and meaningful results.