Doctoring Health Care Knowledge

Grander Lessons from a Failure of Robotic Surgery

This week on my podcast, I deviated briefly from cardiology to discuss a shocking and sad study highlighting the vital nature of doing randomized controlled trials in the practice of medicine.

The reason I mentioned a trial comparing 2 types of hysterectomy (removal of the uterus) in women with early stage cervical cancer was not to opine on matters of cancer, but because the tragic story shows the harm doctors can do if we incorporate therapies without proper testing.

The New England Journal of Medicine published a trial in which women with early cervical cancer were randomized to two types of hysterectomy:

  • One was the type you see on billboards: minimally invasive laparoscopic robotic surgery
  • The other group got traditional open surgery.

The investigators measured the rate of disease-free survival at 4.5 years. They put about 300 patients in each group.

The results were shocking: 86% of women in the minimal-invasive arm were free of disease vs 96.5% in the open surgery. The difference was a whopping 10.6% worse. In other words, the minimally invasive surgery vs the standard approach harmed one in ten women.

In the introduction of this paper, the authors write that current guidelines say either approach to this cancer is acceptable and “these recommendations have led to widespread use of a minimally invasive approach for radical hysterectomy, although there is a paucity of adequately powered RCTs.” Translation: doctors embraced this approach, hospitals advertised it and patients accepted it without proper evidence.

The message here transcends just one medical condition. It’s a message about doctors’ hubris–our excess confidence in accepting new therapies that have yet to be tested.

Cohort studies, small studies, retrospective studies should not lead to widespread acceptance of a new therapy. If you don’t randomize patients and blind the judges, you can’t be sure the treatment works. New approaches ought to be tested in this way.

Doctors need the courage to put our beloved therapies to the test. Patients need the knowledge to ask whether a proposed treatment has been tested in an RCT. Regulators could help by more often tying reimbursement to evidence generation. Viz: They could say, ok docs, you want that new device? We will pay for it if you enroll patients in a trial.

Finally: whenever you hear someone in medicine say we do something because it’s in the guidelines, I urge you to go to that document, find the paragraph and look at the citation. If it goes to anything besides a randomized controlled trial, be very skeptical.