Background: Pharmacological and non-pharmacological approaches are commonly used to treat patients' institutionalised for nighttime wandering. Actigraphy and other scales have been used to evaluate the efficacy of these treatments. However, in clinical settings, nursing records are often the sole source of daily observation of nighttime wandering. Thus, physicians rely on nursing records to evaluate pharmacological and non-pharmacological treatments. This study examined nighttime movements of patients with dementia, comparing the results of integrated circuit tag monitoring with hourly nighttime nursing records. We tested which factors were associated with agreement rates between the two data sources. Methods: The study hospital was a general hospital in Osaka, Japan. Monitoring was conducted in a closed 60-bed dementia care unit. An integrated circuit tag monitoring system was used to monitor the movement of institutionalised dementia patients for over half a year. The distance moved per hour by subjects was measured using the monitoring system, and the data were compared with hourly nighttime nursing records. Agreement rates were calculated between the two data sources. Results: Thirty-five patients were monitored. Between 10-30% of subjects moved at any given hour during the night. The overall agreement rate between sources of movement data was 39%. Agreement rates were significantly correlated with the interquartile of the distance moved, changes in medication, physical conditions, and behavioural and psychological symptoms of dementia measured during the day. Conclusions: Although the agreement rate was low, staff appeared to pay more attention to patients associated with notable events during the day and patients exhibiting variability in distance moved.