I'm in the final throws of my pulmonary and critical care fellowship at a "top-tiered" academic institution (oh how we love our "evidence"). As part of my training, I've done graduate work in biostatistics and epidemiology, which helped me tremendously to interpret studies for what they are actually worth. I absolutely agree with my colleague jdh's points.
Most of critical care medicine involves using available data about clinical scenarios and making bid decisions - often high stakes.
The evidence base for critical care is not worth bragging about. For example:
The overwhelming number of critical care trials are negative. So much so, that you could make the argument that with a collective p-value of 0.05 that several of the positive ones are by chance alone. This thinking is supported by the finding that many trials cannot be replicated (caveat: no two trials are completely the same).
Let's look at glucose control. Everyone got excited about tight glucose control and its potential to save lives. People rapidly implemented insulin drip protocols across many ICUs only to read follow up studies that spoke to the dangers of low blood sugar in critical illness (hypoglycemia is bad). Throw that on top of how commonly glucometers over-estimate (compared to glucose on a panel) by 30mg/dl or so and you have a recipe for potential harm. Tell me the right thing to do here? Prevent rapid swings in glucose and aim for 120-150. Evidence? Common sense.
What about activated protein C for septic shock? So many people go excited about this incredibly expensive drug. Threw it at people every chance they could get? Why? Because one study said it saved lives. Then, follow up study suggests harm. Removed from market. Now what? Back to salt water and pressors.
Speaking of sepsis management...The surviving sepsis campaign is a mindf*$k of guidelines based on pretty weak evidence. Take the time to read the 50 or so pages. The Rivers study was a single center study where the investigator himself managed many in the intervention arm. Despite its inherent limitations, rapidly adopted and implemented as a nationwide guideline. That point alone suggests that our evidence is pretty ****ty.
Ventilators? low-stretch mechanical ventilation. done. oh yeah, proning probably has an added benefit for severe hypoxemia but i'll be my measly salary that the magnitude of the benefit (e.g. the number of patients needed to treat to save one life) is grossly over-inflated in the study out of france in the recent NEJM. let's think before we flip everyone on their stomachs..
However, to confound results focused on short-term mortality (FACTT trial, ARDSnet) is new data suggesting that things that may save a life may create a new set of problems (post-ICU syndrome). For example, mortality benefit from a lower CVP target in FACTT trial appeared to lower mortality for ARDS. However, post-hoc analysis (quality of evidence not that great) suggests that these patients who survive go on to have higher rates of cognitive impairment compared to suvivors with a more liberalized CVP goal.
Oh god, show me the evidence that CVP is helpful? A MAP goal of 65? Rigidly adhering to these numbers harm patients on a daily basis.
Most of the evidence for critical care lies in prevention. Meaning, stop being stupid with your sedative and narcotic practices and minimize these drugs as much as possible while concomitantly assessing eligibility for extubation. It also means to hire crew of physical therapists who, on a daily basis, exercise patients with sepsis, respiratory failure, etc. Probably better than any drug. It means washing hands, elevating the head of the bed, considering proper patients for GI and DVT prophylaxis.
It also means, from my viewpoint, about being aggressive with patient-centered family communication so as not to ignore elephants in the room while you play around with fancy gadgets. A lot of the people speaking to critical care on this forum like to come off as cocky proceduralists, but I'd argue that an equally important skillset (that has some evidence to support its development) is communicating with patients and families.
So, JDH is right. We have little strong evidence to guide our treatment decisions - evidenced by the degree of heterogeneity in practice patterns across "centers of excellence"
How best to prepare, I think the Marino book stinks for the most part. If I had to pick a textbook I'd pick Hall Schmidt and Wood. But, who has time to read an entire textbook. Instead, read on ScvO2, sepsis, ARDS, COPD/Asthma on the vent, DKA, etc as they come up (or maybe on a few beforehand) in your patients. Reading about things while you have patients with those things solidified those things in your brain. I tried to offer advice as well in the link provided by JDH. The only thing I disagree with is buying your fellow coffee - unless you are truly buds - because I can't stand asskissers...