The six-month test

Nearly five hundred dementia caregivers, eight online sessions, and a support worker who checked in but did not run the thing. Anxiety and depression came down at three months. The finding worth pausing on is that they were still down at six.

That is a low bar and almost nothing clears it. The category is full of tools that make a caregiver feel briefly seen: an app that names the exhaustion, a quiz that returns a score, a chatbot that says the right sentence at 2am. Feeling seen is real, and it is not the same as being different half a year later. Most of these studies stop measuring before the answer gets uncomfortable.

What the trial does not say is also worth holding. It is a press release, not the paper. No dropout rate, no effect size, nothing past month six. And the intervention was not clever: eight sessions and a human who occasionally checked in, delivered through a screen because a screen is what fits around a caregiving day.

The lesson is not that software fixed anything. It is that the reachable part of caregiver strain turned out to be reachable, and that the only honest measure of a caregiving tool is whether anyone is better six months later.