Efficacy and risk of sexual orientation change efforts: a retrospective analysis of 125 exposed men [version 2; peer review: 2 approved]

Background: Voluntary therapeutic interventions to reduce unwanted same-sex sexuality are collectively known as sexual orientation change efforts (SOCE). Currently almost all evidence addressing the contested question whether SOCE is effective or safe consists of anecdotes or very small sample qualitative studies of persons who currently identify as sexual minority and thus by definition failed to change. We conducted this study to examine the efficacy and risk outcomes for a group of SOCE participants unbiased by current sexual orientation. Methods: We examined a convenience sample of 125 men who had undergone SOCE for homosexual-to-heterosexual change in sexual attraction, identity and behavior, and for positive and negative changes in psychosocial problem domains (depression, suicidality, self-harm, self-esteem, social function, and alcohol or substance abuse). Mean change was assessed by parametric (t-test) and nonparametric (Wilcoxon sign rank test) significance tests. Results: Exposure to SOCE was associated with significant declines in same-sex attraction (from 5.7 to 4.1 on the Kinsey scale, p <.000), identification (4.8 to 3.6, p < .000), and sexual activity (2.4 to 1.5 on a 4-point scale of frequency, p < .000). From 45% to 69% of SOCE participants achieved at least partial remission of unwanted same-sex sexuality; full remission was achieved by 14% for sexual attraction and identification, and 26% for sexual behavior. Rates were higher among married men, but 4-10% of participants experienced increased same-sex orientation after SOCE. From 0.8% to 4.8% of participants reported marked or severe negative psychosocial change following SOCE, but 12.1% to 61.3% reported marked or severe positive psychosocial change. Net change was significantly positive for all problem domains. Conclusion: SOCE was perceived as an effective and safe therapeutic practice by this sample of participants. We close by offering a unifying understanding of discrepant findings within this literature and caution against broad generalizations of our results