Treatments for bipolar disorder: can number needed to treat/harm help inform clinical decisions?

Objective: To compare bipolar treatment interventions, using number needed to treat (NNT) and number needed to harm (NNH). Method: Results of randomized controlled clinical trials were used to assess efficacy (NNT for response and relapse/recurrence prevention vs. placebo) and tolerability (e.g. NNH for weight gain and sedation vs. placebo). Results: United States Food and Drug Administration-approved bipolar disorder pharmacotherapies all have single-digit NNTs (i.e. > 10% advantage over placebo), but NNHs for adverse effects that vary widely. Some highly efficacious agents are as likely to yield adverse effects as therapeutic benefit, but may be interventions of choice in more acute severe illness. In contrast, some less efficacious agents with better tolerability may be interventions of choice in more chronic mild-moderate illness. Conclusion: Clinical trials can help inform clinical decision making by quantifying the likelihood of benefit vs. harm. Integrating such data with individual patient circumstances, values, and preferences can help optimize treatment choices.