The Petty and Bone paradigm is dead
Once more, top critical care researchers will reaffirm the death of the Petty and Bone paradigm - and refuse to see it.
Once more, top critical care researchers will reaffirm the death of the Petty and Bone paradigm - and refuse to see it.
THE DECEASED body lays pale, indifferent to the reverence of those who gathered to praise his life.
The Petty and Bone paradigm died old. Although there is no birth certificate, the older ones can tell he lasted 60 years. What a spoiled life! Many admirers of course, and a fair number of supporters. But no great achievements.
We are in Melbourne, at the Critical Care Reviews Down Under conference. We advanced one month into the future. It is December 10th, 2024. Top researchers are presenting the results of the latest clinical trials in critical care.
Notwithstanding the deceased inferior accomplishments, the night gathers a reasonable audience. Mostly, the ones who devoted their lives to him. They learned to pretend the devotion paid off. It does pay off, from a cynical perspective. Hindering the advance of critical care is a way of living.
The Petty and Bone paradigm wasn't born as the overwhelming paradigm in the field. It had fierce opponents. However, criticism subsided during the decades. Although incapable of providing efficient disease models, generations of physicians and researchers succeeded, and the mistake was normalized as a regular practice.
Recently, the Petty and Bone paradigm has been causing collective hallucinations. The lunacy is evident. Researchers discussed different SpO2 targets within the normal range in several clinical scenarios, tested the effects of sedatives on inflammation, hypothesized the mortality effect of grams of chloride, gave dapagliflosin to everyone just to see what happens, etc. All these extravagancies are only conceivable to researchers without a disease model in mind. Was it really about discovering something?
I cling to the collective hallucination thesis. Critical care research under the Petty and Bone paradigm is an unfruitful exercise of cognitive bias.
The origins of the Petty and Bone paradigm
The Petty and Bone paradigm was born a couple of decades after the most important advance in medical epistemology, the randomized clinical trial. By the 1960s, the RCT was gaining traction in the US as a revolution in medical science. Many leaders in critical care thought that that was the time to make the field more scientific by performing RCTs. That was the dawn of positive-pressure mechanical ventilation, a genuine revolution in patient care. It was natural to give it a scientific treatment.
By that time, medical researchers discovered that some preterm neonates could develop a severe respiratory syndrome, caused by the lack of pulmonary surfactant, termed the Respiratory Distress Syndrome (RDS). Pioneers of critical care in the military also realized that some severely injured soldiers who died on mechanical ventilation also lost the pulmonary surfactant. These soldiers presented with refractory hypoxia and chest infiltrates.
Soon, all adult patients in mechanical ventilation with hypoxemia and chest infiltrates were deemed to have the same condition, then called Adult RDS. This heuristic was called lumping. Patients were lumped if they shared the same clinical presentation, not the same disease.
Dr. Petty, a prominent critical care physician sixty years ago, decided to RCT the idea that lumped ARDS patients benefit from surfactant. Lumping was convenient for patient enrolling, of course, at the expense of causal reasoning. The RCT failed to prove a benefit for surfactant therapy, but it gave birth to the first ARDS definitions.
In 1989, Dr. Roger Bone lumped patients with complicated infections using arbitrary thresholds of physiologic variables to treat them with methylprednisolone or placebo. The clinical trial also failed to prove the benefits of corticosteroids, but Bone's study inclusion criteria became the first sepsis definition.
Like Petty's ARDS, Bone's sepsis is an arbitrary definition made up to enroll patients in clinical trials, not a disease model that points to a straightforward definition and treatable causes.
The Petty and Bone paradigm's complete story and ominous consequences are detailed in our book.
We must defeat the failed paradigm. Join us!
Here is what science looks like inside the Petty and Bone paradigm
For the sake of text length, I will only list the three main features of the failed paradigm and link useful resources.
Lack of a disease model
The Petty and Bone paradigm doesn't define diseases. It simply gives you a generic triage tool for a generic treatment or support strategy. Many people published ARDS and sepsis features and many tested treatments based on these features. All failed. They all failed because there is no disease model to articulate pre-clinical and clinical findings, and then to articulate translation from biological plausibility to clinical relevance.
My favorite disease model is peptic ulcer disease. It points to causes, diagnostic procedures, and treatments. It integrates anamnesis with pathophysiology, pharmacotherapy, and surgical intervention. Above all, the model elicits certain treatment strategies and is validated by the clinical result of such strategies. Any time a doctor successfully treats peptic disease anywhere in the world, the model is validated again.
Now, please compare it with sepsis research. A few researchers think that septic patients would die if receive normal saline instead of Ringer's, dexmedetomidine instead of propofol, or acetaminophen instead of placebo. Is it conceivable to fit these three hypotheses into a disease model? Of course not. Having no disease model means no limits to researchers’ imagination. And of course, JAMA will publish it.
Blindness to marginal utility
Without a disease model, researchers treat all proposed therapies as if they were equally capable of reducing mortality. This is absurd because it implies that if you add enough therapies you would reach zero mortality, which is impossible. That's a problem I treated in my first post and I strongly recommend it because it is one of the axes of my thinking.
A valid disease model presents a clear dominant cause, i.e., a treatable condition that contributes most to the disease development. Think of ST-elevation myocardial infarction. Whenever a doctor opens the culprit artery, the model is validated again. All other interventions, i.e. beta-blockers, are marginal. Once the artery is opened, the utility of additional interventions is marginal.
Inside the Petty and Bone paradigm, all things are equal. There is no dominance. No marginality. Hence they are blind to the marginal utility of their interventions. Expect nothing from Petty and Bone RCTs but marginal effects. Marginal effects are clinically irrelevant, but authors may use Creative Inferentials to find a nice p-value. It is happening now. Of course, JAMA will always publish it.
Irreproducibility
Petty and Bone defined the diseases by an arbitrary set of thresholds used as inclusion criteria. Without a disease model to inform the disease definition and the distribution of expected treatment effects, it is impossible to estimate the RCT case mix and the expected clinical effects. Any RCT is different from the next because of the different case mix. Indeed, a study found that only 6% of RCTs’ results were reproduced.
Dr. Lawrence Lynn has explained this point in a Datamethods forum discussion. The Petty and Bone definitions capture different disease sets every time they are applied. Let's consider we are participating in an ARDS RCT. All patients met current ARDS criteria for PaO2/FiO2 etc. We both have included 90 patients. Here is my case mix:1
10 cases of Influenza A pneumonia
5 cases of Aspiration pneumonia
25 cases of Covid pneumonia
5 Cases of severe burn-associated lung injury
5 cases of pancreatitis-associated lung injury
5 cases of lung injury due to major trauma with hypovolemic shock and transfusion
35 cases of lung injury due to systemic infection with shock
Now, look at your patients. Do our case-mix match? No, they don't. We treated diverse patients who fulfilled the inclusion criteria due to different pathophysiologic mechanisms that will respond differently to the study intervention. Forget about reproducibility.
The 2024 Melbourne Critical Care Review conference arrives at the end. Petty and Bone paradigm’s dead body remains unburied before the devotees. They won't let it rest.
The devotees will keep the unburied body for several years because they figured out how to carry on. They will parade the body worldwide at large and small critical care meetings. Before the idol, they will announce to the faithful meeting attendants that the Petty and Bone paradigm is alive. They will try to lure you using Bayesian Spins and Crystal-ball subgroup analysis. It is already happening.
We know what to do. Let's bury this dead paradigm.
They can not stop the rising sun.
Thanks for reading The Thoughtful Intensivist!
Further reading
From Dr. Lynn's post on the Datamethods forum.
Great piece. Thanks.