The internet offers a powerful means of gathering information about the lived experience of dementia, especially with patients and caregivers increasingly using online tools or apps for symptom management. A recent study by our group (Rockwood et al., Reference Rockwood2019) employed such a tool to explore agitation from user-completed online reports (n = 2,577) on a dementia symptom tracking website, www.dementiaguide.com. Website users (97% of whom are caregivers) selected symptoms to track to better understand the evolution of dementia, and/or assess response to treatment in the person for whom they care. The study compared the proportion of people with neuropsychiatric symptoms and agitation, and the number of such symptoms, across groups with mild cognitive impairment (MCI), and mild, moderate, and severe dementia. Important to readers of International Psychogeriatrics, agitation was operationalized using terms consistent with the International Psychogeriatric Association (IPA) consensus definition (Cummings et al., Reference Cummings2015) and the Neuropsychiatric Inventory-Questionnaire (NPI-Q) agitation/aggression domain (Cummings Reference Cummings1997). This facilitated comparison between the agitation profiles found on our website and those described in other settings. We found that the proportion with neuropsychiatric symptoms and agitation symptoms increased with increasing cognitive impairment, as did the number of neuropsychiatric symptoms, including symptoms associated with agitation as per the IPA consensus definition (Cummings et al., Reference Cummings2015).

Of broader interest is the notable similarity between our estimates of the frequency of neuropsychiatric symptoms generally and agitation symptoms specifically and those from clinically adjudicated reports that are acquired using traditional data collection approaches. For example, Siafarikas et al. (Reference Siafarikas, Selbaek, Fladby, Benth, Auning and Aarsland2018), using data on 2,141 patients in the Norwegian national registry of memory clinics, investigated neuropsychiatric symptoms in relation to dementia. As part of our validation study, we compared trends in the two reports. We found that estimates of the proportions of people with at least one neuropsychiatric symptom were consistent (Table 1). Similar comparability was observed with regard to agitation. Further support for the validity of the IPA consensus and of using online data comes from a study by Peters et al. (Reference Peters2006) that employed data from A Canadian Cohort Study of Cognitive Impairment and Related Dementias (ACCORD) (Feldman et al., Reference Feldman2003) to characterize neuropsychiatric symptoms (n = 769) (Table 1). Despite differences in selection and reporting procedures between the studies, it appears that once a sample of people in whom a high proportion have dementia is assembled, comparable associations between neuropsychiatric symptoms and estimates can be observed. The similarity in the trends observed in our online database and estimates from clinically adjudicated reports suggests both validity of the IPA definition and the merit in exploring the internet as a means to capture the lived experience of people with dementia.

Table 1. Proportions of people with neuropsychiatric symptoms and agitation symptoms in samples with a high proportion of people living with cognitive impairment

* For Rockwood et al. (Reference Rockwood2019) this figure reflects the proportion in those with severe dementia, and for Siafarikas et al. (Reference Siafarikas, Selbaek, Fladby, Benth, Auning and Aarsland2018) this figure reflects the proportion in those with AD.

^† Peters et al. This figure reflects the proportion in those that were deemed cognitively impaired-not demented (CIND).