Should We Treat Depression with drugs or psychological interventions? A Reply to Ioannidis

Table 1 Response to antidepressants in the treatment of an episode.

	Percent Response			Drug/Placebo
Initial Antidepressant Studies 1957 ~1990:	Drug	Placebo	Difference	Ratio	References
Very early 1957-1974 Imipramine	70%	39%	31%	1.8	[19]
1953-1990 All older antidepressant (TCA class)	63%	36%	27%	1.7	[20]
1953-1990 All older antidepressant (MAO class)	66%	32%	35%	2.1	[20]
1973-1980 Trazodone	61%	29%	32%	2.1	[21]
Early fluoxetine (Prozac) studies	64%	32%	32%	2.0	[77]
More Recent Studies to Present:
Antidepressant (TCA's ~1979-91))	46%	31%	15%	1.5	[78]
Venlafaxine (a newer antidepressant)	45%	25%	20%	1.8	[79]
Severe outpatient depression (Duloxetine, a newer antidepressant)	42%	21%	21%	2.0	[35]
Duloxetine (all patients)	48%	35%	12%	1.3	[35]
Paroxetine (a newer antidepressant) Ioannidis example of no efficacy	53%	43%	11%	1.2	[28]

The table summarizes meta-analyses of antidepressant efficacy including both studies done from 1957 up until the 1980s or 1990s and studies done from roughly the 1990s to the present. There is excellent agreement that all the antidepressants have roughly the same efficacy, although there may be some minor differences [18]. We would caution against making causative inferences between the earlier and the later studies, as cause should not be attributed to inferences from correlational or observational comparison, though they can support inferences. The study Ioannidis cites is in the bottom row.

ISSN: 1747-5341