A/HRC/38/43 provide the baseline that is indispensable to address the situation. The survey was the baseline for the plan of action adopted by the European Commission, a comprehensive programme that includes non-discrimination, education, employment, health, free movement, enlargement and foreign policy, asylum, and actions against hate speech and hate crime. This is clearly good practice on how data collection is the foundation of proper analysis of the causes, nature and scope of the challenges faced, and allows for the adoption of the measures necessary to eradicate violence and discrimination. 64. A major challenge in data collection is underreporting, 72 which stems from a multitude of reasons (A/HRC/38/43/Add.1, para. 52). In countries where same-sex sexual conduct is criminalized, where laws and policies are used to discriminate against lesbian, gay, bisexual, trans and gender non-conforming persons, or where stigma and prejudice are rampant, the probability that victims will dare to report abuses is very low, owing to fear of prosecution, stigma, reprisals or victimization, unwillingness to be “outed”, or lack of trust. Even in progressive environments, the worry of regression may lead to underreporting. In these and other contexts, there is also limited information available regarding violence against lesbian or bisexual persons in private settings; a monitoring exercise undertaken by the Inter-American Commission on Human Rights showed that underreporting of cases of violence against lesbian, gay, bisexual, trans and gender non-conforming persons was particularly apparent in cases of non-lethal attacks, since few of them were reported to authorities, monitored by non-governmental organizations or covered by the media, and even when they were reported, it was common for the victim’s sexual orientation and/or gender identity to be reported inaccurately.73 65. This lack of adequate data affects sectors other than justice; for example, the African Commission on Human and Peoples’ Rights found that only a limited amount of information was available on the impact of HIV on transgender women and men in Africa, and qualified them to be an “invisible” population in responses to the HIV epidemic. 74 Indeed, there is little research or interventions in relation to the health of trans men who have sex with men, and few studies have collected robust data on bisexuals. 75 VII. Support for effective State measures 66. At different levels, multiple sources support the effective implementation by States of measures to address violence and discrimination on the basis of sexual orientation or gender identity. All effective measures observed by the mandate holder share three fundamental traits: their point of departure is the acknowledgement of the problem to be addressed; they are based on evidence; and they have been designed and are implemented with the participation of the relevant communities, persons or populations. A. Global 67. During the initial stages of the third cycle of the universal periodic review, during which 28 States were reviewed, a significant number of recommendations relating to sexual orientation and gender identity were made. Even before the creation of the mandate, communications sent by special procedure mandate holders to States and non-State actors on issues pertaining to sexual orientation and gender identity, as well as the body of work of the treaty bodies, covered various areas addressed in the present report, including positive developments, such as the adoption of legal gender recognition bills, or regressive 72 73 74 75 African Commission on Human and Peoples’ Rights, “HIV, the Law and Human Rights” (see footnote 53), p. 11. Inter-American Commission on Human Rights, “An Overview of Violence” (see footnote 28). African Commission on Human and Peoples’ Rights, “HIV, the Law and Human Rights” (see footnote 53), p. 78. Harvey J. Makadon, Kenneth H. Mayer, Jennifer Potter and Hilary Goldhammer, Fenway Guide to Lesbian, Gay, Bisexual, and Transgender Health, American College of Physicians, 2nd edition, 2015, p. 29. 15