We included eight systematic reviews of high quality. We graded the quality of evidence for all relevant outcomes as moderate, low and very low. None of the studies were from Norway.
We included only Cochrane-reviews and there may exist other systematic reviews of high quality. Three systematic reviews had not conducted literature search after 2008. We do not know if the lack of updates is because of no longer published research in the field or other factors. If there are newer primary studies, we do not know whether these would change or confirm the conclusions from the systematic reviews.
The search for systematic reviews revealed more than 40 protocols for systematic reviews of people with dementia. Some of the protocols focused on diagnostic tests or medications, but many dealt with interventions such as dance and movement, art therapy, technological assistance measures and personally organized activity.
Considering a systematic review to be of high quality does not mean that all the included primary studies necessarily are of high quality. Many of the primary studies are unclear in reporting randomization sequence generation and allocation concealment, and there was a lack of blinding of the outcome assessor. For some of the comparisons and outcome the quality of the evidence was low and very low. This does not mean that these interventions do not work, but indicate that we have little confidence in the effect estimates.