Medications for low bone density (bisphosphonates LIKE zoledronate)? Potential long term effects?

[DrFraser]

I used to work as a medical statistician before medical school … the incidence of the disease is critically important, a high incidence serious disease with an intervention that only makes a 10% rate reduction has a markedly bigger impact on all cause mortality than a low incidence less serious disease with an intervention that has a much bigger rate reduction. NNT is pretty important in any decision made to decide to treat. We also need to look at what outcome we are avoiding in the NNT and how this balances with the number of people needing to take the drug who don’t end up with benefit, but end up with risk/cost/side effects.
The extent of costs, side effects, alternatives must be factored before looking at whether Rx.
Naturally picking patients at high incidence of the given adverse outcome is important, and one’s ability to correctly pick those patients so that we treat patients at highest risk (which unfortunately we don’t do that great of a job predicting who will get spinal or hip fractures).
Irrespective, the decision to prescribe lipid medication is a more simple one than this - and I have very few patients who would meet any criteria to consider giving these medications.

[Davin8r]

@AnUser

He says “In other words, taking these drugs has not been convincingly shown to prevent fracturing your hip in the first place.”, when he is referencing this.

A meta analytic summation of 1 RCT and 5 observational trials.

Basically he decided to ruin a RCT although it was only trending towards significance, by adding a bunch of relative garbage studies, and create the equivalent of a mortgage-backed security.

Come on, this is just nit-picking. Neither the results of the RCT or the OS trials were statistically significant, per the diagrams you just noted. His statement “taking these drugs has not been convincingly shown to prevent fracturing your hip in the first place” remains true, per the data presented.

@AnUser

For the part of prior hip facture, he decides to cite the RR for the meta-analytic summation with 7 RCT’s and 9 observational trials:

Rather than the summation of the RCT’s which any evidence based practitioner would prioritize, and which decreases the RR more than the summation with the relatively garbage studies.

Wait, now you’re expecting him to selectively dismiss the RR for the summation of the RCT+OS studies and just present the RR for the RCTs alone because that suits your pro-pharmacy narrative marginally better? And HE is the one who is biased?

“The RCTs were strictly designed, conducted, and
reported, and the OSs employed multivariate statistical
analysis to minimize potential confounding. The quality of
all included studies was regarded as high.”

Is your strategy here to just get so far into the weeds that nobody will actually get in there to check your work? Actually reading through the review paper (which took some doing to find the full text), re-listening to Dr. Greger’s video, and taking into account Dr. Fraser’s comments above, I think Greger has it right.

[AnUser]

@Davin8r

Wait, now you’re expecting him to selectively dismiss the RR for the summation of the RCT+OS studies and just present the RR for the RCTs alone because that suits your pro-pharmacy narrative marginally better? And HE is the one who is biased?

Because a RCT is a better study design and doesn’t have confounding to a similar extent, so it is better able to detect causality.

@Davin8r

Is your strategy here to just get so far into the weeds that nobody will actually get in there to check your work? Actually reading through the review paper (which took some doing to find the full text), re-listening to Dr. Greger’s video, and taking into account Dr. Fraser’s comments above, I think Greger has it right.

Greger is wrong and knowingly so on so many topics, so it’s likely he is wrong on this one too.

[Davin8r]

@AnUser

Because a RCT is a better study design and doesn’t have confounding to a similar extent, so it is better able to detect causality.

Sure, but there’s also the “totality of evidence”. Even if RCTs (if done well) are weighted more heavily than observational trials, it doesn’t mean one has to completely throw out the results of well-done observational trials, especially when both showed significant results in secondary prevention of hip fractures.

@AnUser

Greger is wrong and knowingly so on so many topics, so it’s likely he is wrong on this one too.

We’ll have to agree to disagree, then. Nobody’s perfect, but I think he’s right on the vast, vast majority of topics. And the statement “knowingly” implies that you have psychic powers.

[DrFraser]

I love that you are on the board! You challenge me to think and evaluate what I do and recommend. Your points are well made. I’ll continue to refine my approach.