BackgroundAt some point in the disease process many persons with dementia (PWD) will have a missing incident and be unable to safely return to their care setting. In previous research studies, researchers have begun to question whether this phenomenon should continue to be called wandering since the antecedents and characteristics of a missing incident are dissimilar to accepted definitions of wandering in dementia. The purpose of this study was to confirm previous findings regarding the antecedents and characteristics of missing incidents, understand the differences between those found dead and alive, and compare the characteristics of a missing incident to that of wandering.MethodsA retrospective design was used to analyse 325 newspaper reports of PWD missing in the community.ResultsThe primary antecedent to a missing incident, particularly in community-dwelling PWD, was becoming lost while conducting a normal and permitted activity alone in the community. The other common antecedent was a lapse in supervision with the expectation that the PWD would remain in a safe location but did not. Deaths most commonly occurred in unpopulated areas due to exposure and drowning. Those who died were found closer to the place last seen and took longer to find, but there were no significant differences in gender or age. The key characteristics of a missing incident were: unpredictable, non-repetitive, temporally appropriate but spatially-disordered, and while using multiple means of movement (walking, car, public transportation). Missing incidents occurred without the discernible pattern present in wandering such as lapping or pacing, repetitive and temporally-disordered.ConclusionsThis research supports the mounting evidence that the concept of wandering, in its formal sense, and missing incidents are two distinct concepts. It will be important to further develop the concept of missing incidents by identifying the differences and similarities from wandering. This will allow a more targeted assessment and intervention strategy for each problem.