A Study Comparing Duloxetine to Other Antidepressants in the Treatment of Severe Depression

Visitwise probability of participants per treatment meeting remission criteria (QIDS-SR total score [TS]</=5 at week 12 endpoint) were estimated using a pseudolikelihood-based mixed-models repeated measures analysis for a categorical outcome, model included fixed, categorical effects of treatment group (duloxetine vs. SSRIs), visit, treatment group-by-visit & continuous, fixed covariate of baseline QIDS-SR TS, and random effect of participant. Primary analysis contrasted remission probability at week 12 endpoint between treatment groups.

The QIDS-SR is a 16-item, participant-rated short form of the Inventory of Depressive Symptomatology that assesses 9 domains: sad mood, concentration, self-outlook, suicidal ideation, involvement, energy/fatigability, sleep disturbance, appetite/weight increase/decrease and psychomotor agitation/retardation. Scores range from 0 (none) to 27 (very severe). The QIDS-SR total score was used to derive the mean change from baseline to endpoint depression.

Visitwise percentages of participants meeting remission criteria HAMD-17 total score [TS] </=7 at week 12 endpoint) were estimated using a categorical, pseudolike-lihood-based repeated measures approach, & included fixed, categorical effects of treatment group (duloxetine vs. SSRIs), visit, treatment group-by-visit interaction, & continuous, fixed covariate of baseline HAMD-17 TS. Primary analysis will be contrast of remission rates at week 12 endpoint between treatment groups, & represents estimated remission rates for each treatment group had all participants completed 12 weeks of therapy.

Visitwise percentages of participants meeting response criteria (50% reduction from baseline QIDS-SR total score at 12-week endpoint) were estimated using a categorical, pseudolikelihood-based repeated measures approach, & included fixed, categorical effects of treatment group, visit, treatment group-by-visit interaction, & continuous, fixed covariate of baseline QIDS-SR. The primary analysis will be the contrast of response rates at week 12 endpoint between treatment groups, and represents estimated response rates for each treatment group had all participants completed 12 weeks of therapy.

Visitwise percentages of participants meeting response criteria 50% reduction from baseline in HAMD-17 total score at 12-Week endpoint) were estimated using a categorical, pseudolike-lihood-based repeated measures approach, & included fixed, categorical effects of treatment group, visit, treatment group-by-visit interaction, & continuous, fixed covariate of baseline HAMD-17 TS. Primary analysis will be the contrast of response rates at week 12 endpoint between treatment groups, & represents estimated response rates for each treatment group had all participants completed 12 weeks of therapy.

The HAMD-17 is a rater-administered assessment of depression severity and improvement, with total score ranges from 0 (not at all depressed) to 52 (most severely depressed).

HAMD-17 subscale consists of items 10, 11, 12, 13, 15, and 17 evaluates agitation, and severity of psychic and somatic manifestations of anxiety. Total subscale scores range from 0 (normal) to 18 (severe). Mean change from baseline to endpoint.

HAMD-17 Maier Subscale consists of Items 1, 2, 7, 8, 9, 10 and represents the "core" symptoms of depression. Total subscale scores range from 0 (normal) to 24 (severe).

HAMD-17 Bech subscale consists of items 1, 2, 7, 8, 10, and 13 used to evaluate core symptoms of Major Depressive Disorder (MDD). Total subscale scores range from 0 (normal) to 22 (severe).

The HAMD-17 Retardation subscale consists of Items 1, 7, 8, 14 and evaluates dysfunction in mood, work, and sexual activity, as well as overall motor retardation. Total subscale scores range from 0 (normal) to 14 (severe).

The HAMD-17 Sleep Subscale consists of Items 4, 5, 6 and evaluates initial, middle, and late insomnia. Total subscale scores range from 0 (no difficulty) to 6 (difficulty).

The BPI is a self-reported scale measuring pain severity and pain-specific interference on function on a scale ranging from 0 (no pain) to 10 (pain as bad as you can imagine). The BPI average 24-hour pain measure was used to derive the overall mean change from baseline to endpoint, in those participants who had a BPI average 24-hour pain score of 3 or greater at baseline.

The BPI is a self-reported scale measuring pain severity and pain-specific interference on function, with scores ranging from 0 (does not interfere) to 10 (completely interferes). The BPI average 24-hour pain measure was used to derive the overall mean change from baseline to endpoint.

The SDS is a participant-rated anchored visual analog scale to assess disability across the three domains of work/school, social life, and family life, with each item scored from 0 (not at all) to 10 (very severely), with a summarization of the 3 items to evaluate global functioning. The Global Functional Impairment Score is a total score score that ranges from 0 (unimpaired) to 30 (highly impaired), and was used to derived the mean change from baseline to endpoint.

The SDS is completed by the participant and Item 1 is used to assess the effect of the participant's symptoms on their work/school schedule. Scores range from 0 to 10 with higher values indicating greater disruption in the participant's work/school life.

The SDS is completed by the participant and Item 3 is used to assess the effect of the participant's symptoms on their family life/home responsibilities. Scores range from 0 to 10 with higher values indicating greater disruption in the participant's family life/home responsibilities.

The SDS is completed by the participant and is used to assess the effect of the participant's symptoms on their work/social/family life. Total scores range from 0 to 30 with higher values indicating greater disruption in the participant's work/social/family life.

Mean change from baseline to endpoint in systolic blood pressure

Mean change from baseline to endpoint in diastolic blood pressure

Mean change from baseline to endpoint in pulse rate

Mean change from baseline to endpoint in weight

WP score was calculated by taking midpoint of annual before-tax income reported on HPQ. A multiplier of 1.25 produced estimated direct & indirect (i.e. benefits) income. Annual hours expected to work were calculated from expected daily work hours, multiplied by 236 days. Hourly, indirect income was total direct + indirect income, divided by # of expected annual work hours. Indirect hours lost annually for WP=hours expected to be worked annually times WP percent, times hourly rate=dollars earned, and then subtracted from total direct + indirect income=dollars lost annually due to WP.

Self-administered assessment used to determine a participant's work performance in terms of employment status, absenteeism if employed, productivity while at work, usual occupation, and annual income. Tool assesses the potential impact of change in depressive symptoms on work productivity and its associated employer costs. Scale ranges from 0 to 100% of work days in past 30 days. Absenteeism and presenteeism were combined into a measure of total lost work performance by adding absenteeism to the value ([100-absenteeism] × [100-presenteeism]). Mean change baseline to endpoint.

Self-administered assessment used to determine a subject's work performance in terms of employment status, absenteeism if employed, productivity while at work, usual occupation, and annual income. Tool assesses the potential impact of change in depressive symptoms on work productivity and its associated employer costs. Defined on a 0-100 scale for the percentage of work days the respondent missed in the past 30 days. Absolute absenteeism: actual hours worked minus expected hours equals number of missed work days. Mean change baseline to endpoint is reported.

Self-administered assessment used to determine a participant's work performance (employment status, absenteeism if employed, productivity while at work, usual occupation, & annual income). Tool assesses the potential impact of change in depressive symptoms on work productivity & its associated employer costs using a 0-100 scale in which 0 meant doing no work at all on days spent at work and 100 meant performing at the level of a top worker. Absolute presenteeism: difference between "score for self" and "score for average worker in same job". Mean change baseline to endpoint is reported.