You can use a simple adjustment model to factor out the usage of Prep in the outcomes rather than require that people avoid it. You just have to collect that in

You can use a simple adjustment model to factor out the usage of Prep in the outcomes rather than require that people avoid it. You just have to collect that information. Same with other measures like condom use. People can even lie about it randomly and you can still defuzz the outcomes.