Typically, you can have more valid and reliable results if you use a scale that is longer than only 10 items. I might advise using the full BFI, rather than a ten item version.
You could also check out the BFAS, another nice Big-5 measure.
If you already have data collected, that's a different question, and we would need to hear a bit more about what other measures/scales are in your dataset.
Since the Big 5 are traits, they wouldn't normally change except over very long periods of time (i.e. many years). They might be dependent variables if you're using an extended time lag of many years, or measuring some kind of traumatic event, but it would otherwise be an uphill battle to model them as outcomes.