Well, that depends a lot on what's wrong. If you've broken your leg, then believing a remedy is highly beneficial will hardly unbreak it. But on the other hand, broken legs heal. So the trick there is to persuade the patient to give the remedy credit for what would have happened anyway.
If you're suffering from a chronic, fluctuating illness, how you feel at any one time depends a lot on, well, how you feel. It can be quite psychological. So there, homoeopathy gets two bites at the cherry. It may be that the fluctuating condition was about to enter a period of improvement anyway, so hey, we'll take credit for that. Or else sometimes the "therapeutic consultation" can have such a positive psychological effect that the patient does indeed declare he feels batter even though the condition is objectively unchanged.
What has never been shown to occur is an objectively measurable improvement in an objective physical ailment due to "placebo". Yes, psychological things like anxiety and depression may genuinely improve because of positive reinforcement, but actual physical illness doesn't. In this case, "placebo" is either an improvement that would have happened anyway, being credited to the sugar pill, or a purely subjective report of "feeling better" which is not in fact a reflection of actual objective improvement.
Or, really, just a combination of coincidence and wishful thinking.
Rolfe.