Although I can imagine it tells you if your intervention/control hospitals have different patient populations, which could have influenced your results. But it does not tell you whether the specific baseline characteristic is relevant or not in explaining your results. Correct?

What would calculating a p-value tell you, when using it for differences in baseline characteristics in a cluster-randomised trial? (in this case each cluster is a different hospital) @DrToddLee I don’t see how it tells you anything but I’m curious about your ideas.

Regarding preferring oral treatment over iv, what is your opinion on valacyclovir @BradSpellberg ? In HSV encephalitis, would you treat orally? There seem to be differing opinions on bioavailability and adequate dosing of VACV.

Enjoyed your talk just now at ESCMID! @BradSpellberg

Wouldn’t a more accurate title be Diagnostic Testing Performance for Positive Urine Culture? Excluding the last part about imaging. Can’t find the source for these numbers in the table unfortunately.

Is this title really the best title for table 3 in the UTI wikiguideline @BradSpellberg ? Is spec for leukocytes in microscopy really that high? Usually these sens/spec numbers are in relation to predicting a positive urine culture, not predicting an actual uti.

https://t.co/migUSUjSdm A Longer Duration of Intravenous Antibiotic Treatment for Patients with Early Periprosthetic Joint Infections Is Not Associated with a Lower Failure Rate @ABsteward

Pristinamycin for oral treatment of E. faecium and VRE? I have never heard of this drug. But I see it’s used in France and even in Scotland. Can it really be used for E. faecium? I wonder why I’ve never heard of it. NHS Tayside even has guidance on it @IDiots_pod @BradSpellberg

Do you use oral flucloxacillin in Scotland @IDiots_pod ? Which dose do you use, and is this based on a specific guideline? I am reviewing the oral dose used here in The Netherlands, which is currently (usually) 500mg x4. Interesting to see what is done elsewhere.

Wouldn't you say that this study found short (6 days) inferior to long (12 days) treatment of severe cellulitis, since there was so much more relapse (6% vs 24%) @BradSpellberg ?

Simple question and I think I know the answer... in an RCT, does it make sense to calculate the p-value for differences in baseline characteristics? See screenshot for example. @DrToddLee

We aim for a vancomycin concentration of 20-24 during continuous infusion. I recall you and others being critical of targeting high vanc levels, is that right @BradSpellberg ? Is there literature discussing this?

We were discussing today, has amoxi+ceftriaxone synergy really been proven @BradSpellberg? The source that you refer to refers only to experimental models. I wonder if there is any good clinical data. Do you know?

Is there data on the diagnostic utility of cloudy or flaky urine for diagnosing UTI? It is often used by non-ID as an argument for starting treatment, but I have never heard a microbiologist/ID physician use it as for support for diagnosis. @BradSpellberg @DrToddLee @IDiots_pod

*By in stead of ‘vy’…

New guideline on CAP diagnostics and treatment @ABsteward Vy the Dutch Working Group on Antibiotic Policy (in English) https://t.co/CaNPMW0Vaf

"I suggest that Dr. Bik @MicrobiomDigest abandon this approach of hunting for mistakes in our or other people's work because it leads to unfounded accusations, but I suspect she and/or her assistants will find new arguments and tweet about extensively..."

Someone mentioned the quote by the US surgeon general that infections were a thing of the past today, so I googled it, and I was pleasantly surprised to find your myth-debunking article as the first result @BradSpellberg

Prosthetic valve endocarditis with viridans streptococcus with penicillin MIC of 0,75. eGFR of 35. So prefer not to use gentamicin. How would you treat? Penicillin 18 mil. U/day? Ceftriaxone 2g/day? Or different? #IDTwitter @BradSpellberg @TBrnaughtTB @gushamilton @MicroLukas

@gushamilton @DrToddLee @Josh_S_Davis Its true of many trials in ID that we haven't powered them to find small differences in mortality like 1-3% range (as opposed to cards which has for many drugs)