Maybe some back pain is caused by low-grade bacterial infection and therefore treatable with antibiotics. I have been following the story of this hypothesis since 2013. A new test (Bråten et al) has more or less slammed the door shut on it: antibiotics don’t make enough of a difference to take seriously. This attracted some amusing snark from people puzzled by the need for such a study:
Important RCT for anyone who’s been treating low back pain with <checks notes> amoxicillin.
— David Juurlink (@DavidJuurlink) October 19, 2019
But the study was charmingly defended by Dr. Paul Sax (also just a solid review of the topic):
Allow me to defend the people who did the study, and go even further — this is exactly the sort of practical, hypothesis-testing trial I wish we’d see more often. … Even negative studies are important. It could have been an H. pylori and peptic ulcer disease.
Imagine if it had been! Checking is good. I’d even like to remove “even” from that statement: studies with negative results are not second class citizens in any way. We need them. In some contexts, they are disastrously missing (see file drawer effect and the AllTrials.net Campaign)
I’ve updated my back pain book with the new reference.