again, according to a structured interview (if feasible). Alternatively, the person no longer meets the definition of remission/recovery (i.e., the person again scores 13 or above on the (C)Y-BOCS plus CGI-I rating of 6 ("much worse") or higher for at least one month, or needs to be withdrawn prematurely from the trial before one month has elapsed due to a severe worsening of OCD symptoms. Discontinuation of the trial due to reasons other than worsening in OCD symptoms (e.g. suicide risk) is not considered a relapse.Two comments are worth adding. First, in Round 1, to consider a patient a treatment responder or remitter, many experts (56% and 58%, respectively) thought that sustained improvement should be present for at least one month. However, this proposed duration clashes with the (C)Y-BOCS, which asks about symptoms during the "previous week". In addition, response has been defined in most prior OCD trials at the end of treatment. In Round 2, despite explicating this, only 64% and 46% of experts agreed with the proposal of "at least one week" for the duration of response and remission, respectively. To accommodate this disagreement in the field, the duration for response and remission above allows for "at least one week" and we recommend additional follow-up assessments where possible to assess whether response/remission status has been maintained over longer periods.Second, to judge that a patient relapsed, many experts (Round 1: 48%; Round 2: 87%) thought that worsening of symptoms should be present for at least one month to protect against transient flares in symptoms. However, some patients acutely deteriorate and require immediate clinical intervention 12 . For this reason, the relapse definition above indicates that patients who need to be removed from treatment protocols before one month because of worsening of OCD symptoms should also be considered to have relapsed.In summary, agreement was reached on how to define response, remission, recovery and relapse across a range of international professionals with expertise in OCD. We recommend that researchers report their results using these definitions whenever possible. As outlined by Frank et al 6 , doing so will lead to: a) improved design, interpretation and comparison of clinical trials of various modalities; b) improved communication of research findings between professionals and to the general public; c) improved guidelines for evaluation of clinical efficacy of various treatments by regulatory agencies; and d) development of improved treatment guidelines for clinical practice.