Does the Use of an Automated Tool for Self-Reporting Mood by Patients With Bipolar Disorder Bias the Collected Data?

Michael Bauer, MD, PhD; Natalie Rasgon, MD, PhD; Paul Grof, MD, PhD; Laszlo Gyulai, MD; Tasha Glenn; Peter C. Whybrow, MD

In This Article


For the validation study, 96 patients were consecutively recruited from the mood disorders clinics at UCLA, the University of Ottawa, and the University of Pennsylvania. After a complete explanation of the study, all patients signed the written consent form approved by the Institutional Review Board at their respective institutions. All patients met the DSM-IV criteria for bipolar disorder, diagnosed by clinical interview and confirmed with the MINI International Neuropsychiatric Interview.[26] Other inclusion criteria were age 18 years or older, daily access to a personal computer and the skills to use it, and the ability to read and write English. Patients with antisocial personality disorder or dementia were excluded from the study. To avoid selection bias, no other inclusion or exclusion criteria were used.

The ChronoRecord data collection software presents the patient with large colorful icons for mood, medication, and sleep to facilitate data entry. For recording mood, ChronoRecord uses a 100-unit visual analogue scale between the extremes of mania and depression that the patient marks proportionately. During training, patients set personal anchor points describing the most depressed and most manic states they ever experienced. They were told to describe the predominant features of the extreme state in addition to mood and often depicted a mixed episode as the most manic experience. After one-half hour of training, patients were given ChronoRecord software to install on their home computers. For a 3-month period, the patients entered mood, sleep, menstrual data, psychiatric medications, and life events daily. Weight was entered weekly. Patients returned data by e-mail or diskette.

During the validation study, observer ratings were obtained at 4 visits over the 3-month period (HAMD, n=281; YMRS, n=283). For the current analysis, patient data were grouped by severity of illness using HAMD scores of 7 or below (n=169), 8 to 13 (n=48), and 14 or greater (n=64), and YMRS scores of 7 or below (n=253), 8 to 12 (n=20), and 13 or greater (n=10). The ChronoRecord scores for these groups were evaluated to see if the following demographic variables had any effect on patient self-reporting: gender, site, ethnicity, diagnosis of bipolar I or II, age, disability status, and years of education. The 2-way analysis of variance (ANOVA) was used to compare the effect of the demographic variable and the interaction between the HAMD or YMRS score and the demographic variable. Effect significance was measured by the F statistic. A separate analysis using the general linear mixed model was done to account for within-patient variability using ChronoRecord mood as the dependent variable. Estimates were made for models with HAMD and each of the demographic characteristics, and for YMRS and each of the demographic characteristics as fixed factors. The individual patient code was included as a random factor. The mixed-model analysis did not include interaction effects due to the small number of observations for each patient. SPSS Version 11.5 was used to perform all statistical computations.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.