It's important to remember that the placebo effect only exists in a few specific symptoms that are known to have an important psychological component - pain, blood pressure, anxiety, nausea etc. In these cases, placebo interventions have a measurable positive effect (or negative, for nocebo) compared to no intervention.

But in the vast majority of other illnesses, placebo is exclusively a confounding factor. People's bones don't fuse from sham surgery. Sugar pills don't cure malaria or TB. Placebo chemotherapy doesn't kill cancer cells. The placebo effect does exist in these cases as well if not using double-blind trials, or if conducting patient subjective assessments, but it disappears if lab techs and objectively measurable data are investigated (things like viral load, not patient perception of their symptoms).