I do think there's an argument to be with, for example, needle exchange programs which arguably did reduce the spread of HIV. There's a weaker argument to be made with respect to methadone programs and injection sites. Harm reduction programs do have rational arguments behind them and a body of evidence, but even in the later two cases, where the policy objective is reducing drug use to zero but doing so minimising harms that do occur, there's a powerful argument to be made that those programs, in practice, service as secondary vectors for spreading drug abuse, which has a strong social component.
Given that, I would highly question a harm reduction model for administering hormones to minors.