Someone in my family currently is having a 24 hour electrocardiography (ECG). Basically, the heart region is fully wired and connected to a "wearable" mini-computer/database and data is collected over a 24 hour period. That part of the 24 h ECG is fully automated. The person in question gets wired, the computer set up, and everything is ready to go. The more complicated part is keeping an activity log. Think of it as a silly form of Twitter.
You write down the time and the start of activities:
16:05 running
16:15 sitting and talking
16:40 reading
...
18:18 eating & drinking
...
x:yy sleeping
After the recording period is over, the data is evaluated manually. Someone will take a look at the plotted graphs, look for "interesting" peaks, gaps, dense graphs, etc. and correlate that with the activity log.
Wouldn't that be something that could be done by a database application and mostly automated? Looking for patterns, for irregularities, correlating timestamped data, and eventually reporting back a list of events to take a closer look? Instead of keeping a paper-based log, timestamped activities could be stored with the mini-computer/database (enter it with a wireless device or have that person wear a USB connector... ;-) or correlated with the twitter account...