Lessons learned in a data linkage project on fatal drowning

Drowning is the leading cause of death in US children 1-4 years old. The epidemiology of drowning at a regional level is understudied because no single data source provides complete information on persons who drown. Probabilistic data linkage is a novel way of studying the epidemiology of drowning. This study aimed to document the lessons learned during the linkage process.

This was a cross-sectional study of persons of all ages who died from unintentional drowning in metropolitan Houston from 2016 to 2022. We describe the lessons learned during the project planning and execution phases which pertained to data curation, the regulatory aspects involved with obtaining data, data security, spatial identification, and the strengths and limitations of the different datasets.

Twelve datasets were reviewed; eight were successfully linked. During the planning phase, the key issues identified pertained to data ownership and governance and robustness of data which impacted the availability and quality of data, variation in the description of drowning location, and risk and protective factors which helped identify subpopulations at-risk for drowning. In the execution phase, the major issues included data security, data sharing, and dissemination of results.

There are a plethora of data sources for fatal drowning. The process of obtaining and analyzing data to describe the epidemiology of fatal drowning using probabilistic data linkage is complex, lengthy, and cumbersome. Documenting the process and lessons learned can support drowning research and inform regional drowning prevention strategies.