Does the use of an automated tool for self-reporting mood by patients with bipolar disorder bias the collected data?