Potti's claim was that he could predict which chemotherapy drugs would be most effective on a particular patient. The test patients were still getting good anti-cancer drugs. He wasn't giving them random drugs that had nothing to do with fighting cancer.

The most likely effect of the fake data on the test patients would be that their treatment was merely as effective as that of patients in the control group. It is very unlikely his methods produced worse outcomes--if they did, he would not have needed to fake anything! He could have published his methods as a way to identify drugs that would be ineffective for a particular patient, which would be about as big a breakthrough as what he was claiming.

do you know anything about chemotherapy? Taking these drugs is not just popping a pill and waking home like nothing happened.

In the trials, the control group and the trial group were being treated with the same sets of drugs. In the control group in one of the trials, for instance, patients were assigned to one of two sets of drugs randomly. In the test group, they were assigned to one of those same two sets of drugs using Potti's research.

Note that people in the test group received one of the treatments they could have received had they been in the control group. Since Potti's research was fake, there was effectively no difference between the test group and the control group.

no they wouldn't have. You don't go into cancer drug trials by taking just anyone who waltzes into the oncology ward.

I'm not sure what you are trying to say here. Yes, they don't just take random people. The breast cancer study, for example, wanted 18+ year old females, with a T1c to T3 invasive breast cancer, at any N stage, with no distant metastases, with a tumor size >1.5 cm in at least one dimension. They had to be a surgical candidate, with no chemotherapy, radiotherapy, or biologic/targeted therapy for that tumor or any other malignancy in the past 5 years. They had to have an ECOG status of 0-1, and adequate scores on several measures of organ function, and no significant heart problems or blood pressure problems.

The patients were randomly assigned to the control group or to the test group. Those in the control group were either treated with Doxorubicin/Cyclophosphamide or Docetaxel/Cyclophosphamide, randomly assigned.

Those in the test group were either treated with Doxorubicin/Cyclophosphamide or Docetaxel/Cyclophosphamide, assigned by the genomic data Potti claimed would predict which would work better for that patient.