Controlled trials in bipolar I depression: focus on switch rates and efficacy