In our research with people with severe mental health problems, we see their fitbit data is very erratic, with lots of daytime naps to compensate for lack of nighttime sleeping.

Is there any literature about how to meaningfully analyse erratic Fitbit data? Many thanks for your suggestions.