Land of Irrelevance #5 and #6

Time to hammer a couple of nails

Hello subscribers

Dogtensivists were so busy the last few weeks that I couldn't pace them.

Over the weekend, I worked three consecutive night shifts in the cardiac intermediate care unit, which allowed me to stray nerdily on the internet. My original purpose was to study polymixin toxicity, and before I noticed, I wormholed into the kidney's organic cation transporters. It's a cool topic that one day will be important (no kidding). In the process, I inevitably ran across a couple of wonderful pieces of Dogtensivism from Vanderbilt University and the Federal University of Sao Paulo (Unifesp), one of the leading research institutions in Brazil.

The studies are here and here.

The Dogtensivists (always capitalized) were easy to spot using the Nine Rules of Dogtensivism I posted a few weeks ago with The Manifesto.

Here it goes:

Rule #1 Theory-free science

Both studies were lazy enough to skip a serious reflection on the hypotheses. If one takes the paper at face value, it looks like there are doctors at Vanderbilt who believe that choosing Lactated Ringer's or Normosol-R may impact mortality in 30 days. There is no possible theory to explain that. Their other hypothesis, manifested in the primary outcome, is that the choice of balanced solutions could affect acid-base status. This is another theory-free hypothesis. Both solutions are buffered. So what are they measuring? Differences in the buffer effect? I must remember those patients had kidneys. Mother nature made these organs capable of keeping bicarb and other ions stable as long as people receive proper food and hydration. There is no excuse for such a bad, if not absent, reasoning. Please refer to my earlier posts about chloride and bicarb adventures in critical care.

On the Brazilian side, the ludicrous “ASA treats sepsis” hypothesis (May I say cringe? Like “it cringes me”?). The introduction section tries to justify the study, but stating doubts about preclinical research doesn't justify clinical research. This is not what Bradford Hill meant with RCTs. There must be a disease model behind the study question. Platelet aggregation is not a dominant cause of sepsis morbidity. Excess aggregation is neither necessary nor sufficient to kill a septic patient.

---

Rule #3 “Just do it” approach

Do you have enough aspirin in the hospital? How about trying it for sepsis? Just do it! Never mind all that was previously tested in coagulation/aggregation. Stating nonsense and platitudes in the introduction won't make a paper scientific but will make it publishable.

Out of money? Research crystalloids. All you need are the bags that you already have. Cheap and easy, harmless. Nobody dies because of a few grams of chloride or lactate. Just do it! There's no downside. If networking is good enough, you can make it to the Chest Journal.

---

Rule #4 Clinical constructs

Dogtensivists invented the clinical construct, a diagnostic that doesn't need biological reasoning. It exists only in the minds of theory-free scientists. The Brazilian study used the SOFA score for patient enrolment. How does a two-point rise in SOFA justify a better response to ASA, in human biology? Using a clinical construct like Sepsis-3 and its SOFA points instead of a disease model is the Petty and Bone mistake. I have written extensively about it in my substack. If you haven’t read it, please do.

The Vanderbilt study is so nonsensical that they didn't bother to have inclusion criteria, so I can't accuse them of using clinical constructs as nosologic surrogates.

---

Dr. Lynn about studies that use SOFA for both inclusion and outcome:

(full discussin here)

“This is a beautiful illustration of PettyBone science.

• Guessed triage set thresholds as the input measurement .

• Guessed set of thresholds as the primary output.

• The treated set derived from triage is disease and source agnostic.”

---

Rule #5 Shameless Terminal Bullshit

I was premonitory when I illustrated this rule using the normal saline x balanced solution case. Dogtensivists never fail me. They are not ashamed of BS hypotheses and the resulting irrelevant studies. Both groups did it, absent-minded and proudly.

---

Rule #8 Blindness to marginal utility

This one I won't stop repeating. Weren't they blind to marginal utility, they would try to describe a disease model where ASA or Normosol-R are dominant treatments. Being blind to it, they think one can infer causality in an RCT without having a strong case for biological plausibility before the RCT. This is not what Bradford Hill taught.

---

Well, they checked 5 out of 9 rules. This is a lot of Dogtensivism.

It is enough for today. If you want to dive deeper into those topics, I encourage you to upgrade to paid. The paywalled posts have a lot about what I call critical care epistemology. If you are short of cash or unsure about the value of the posts, I invite you to open a free trial and cancel it after interacting with the posts for a few days.

Thank you for reading The Thoughtful Intensivist.

Keep barking!