In the ICU, it's the simple things that save lives
In Brazil, the U.S., and yes, your hospital too
“… And, pharmacy?”
For the overadrenalized, attentionally voracious intensivist, the enforced low-gear idle of interdisciplinary rounds may sometimes seem worse than Sisyphean in its repetitive monotony, a rock to be rolled up the mountain only to be found at the bottom again not tomorrow, but at the very next bed. The insipidity of the form, its evocations of apathetic classroom recitations, someone now with the rapid-fire monotone of an auctioneer, the next a low talker, each of their tics offering its own endearing evidence of a humanity resisting the homogenizing compulsory mode, with its enforced anti-rhetoric, its narrative flatness a constrictor snake asphyxiating each innocent wriggling datum as it struggles to escape the unfurling, featureless sonic tapestry entombing it, yearning instead to soar in the chaos of context, that jostling muddle where it can insist on its own truth and meaning, become real.
I had so much potential, the intensivist sighs invisibly. What am I doing here?
You really don’t know—you, who always had all the answers?
This one’s easy: saving lives.
Critical care researchers have spent decades searching for heroically transformative therapies in the ICU. N-of-1-omics. Stem cells. Brain-cooling. But in the ICU, as in many areas of medicine and surgery, the interventions with the largest impact have been the least glamorous: avoiding oversedation, liberating patients from ventilators, and removing unnecessary lines and tubes.
A new randomized trial in JAMA is the latest reminder that getting the basics of critical care consistently right matters more than any experimental therapy so far.
Researchers in Brazil launched a “tele-rehab” intervention at 20 public hospitals in random order (a stepped-wedge cluster-randomized trial, which is the right way to do this), ultimately enrolling 1,916 adults requiring invasive mechanical ventilation for acute hypoxemic respiratory failure in 2024 and 2025.
Hospitals sequentially crossed over from usual care to an integrated “telehealth rehabilitation program” spanning the full phase of care from the ICU to the hospital ward, to post-discharge follow-up for two months.
But It Wasn’t “Rehab”; It Was “Pre-hab”
The intervention basically amounted to increased diligence and accountability for consistently maintaining commonly accepted best practices in the ICU:
Reducing sedation and optimizing analgesia
Systematically performing spontaneous breathing trials
Delirium prevention efforts
Early mobilization and physical therapy
Removal of unnecessary invasive devices
The extra encouragement/pressure to adhere to these practices was delivered by trained local staff on multidisciplinary rounds that would probably be quite familiar to U.S. clinicians (if they spoke Portuguese).
A remote intensivist reviewed adherence weekly for three weeks in a one-hour videoconferencing session and provided targeted guidance and encouragement.
On the medical ward and after discharge, the focus shifted to true rehabilitation assessment and service delivery, provided via smartphone videoconferencing.
During usual care phases (in a variable lead-in period at each facility), the intervention was not performed.
Menos Pessoas Morreram
Fewer people died during the intervention periods overall: 71.8% all-cause mortality at 90 days vs. 78.3% during usual care periods (with a 95% confidence interval of −14.7% to −0.6%; P = .03).
They also had far shorter durations of mechanical ventilation: 9.9 days vs 15.5 days, adjusted difference, −6.2 days; 95% CI, −8.5 to −3.9; P < .001.
They had more days alive and out of the hospital (a mean of 17 vs. 12).
The primary outcome was health-related quality of life at 90 days after discharge using the EQ-5D-3L utility score, and this was minimally improved in the intervention group. The difference was entirely derived from the survival benefit. That is, patients who survived did not experience an improvement in self-reported quality of life as a result of the intervention.
Bundles: Baffling Black Boxes
Which component of the complex intervention caused the mortality difference—or was it due to something else entirely? There’s no way to know with clarity. The intervention spanned the entire care course and is impossible to unpack.
But it seems darned unlikely that one in every 13 patients’ lives was saved just by extra attention to their nutrition and PT plan after ICU discharge.
And we know that six fewer days on a ventilator will do wonders for anyone’s general health and life expectancy.
We can cautiously conclude that increased adherence with the ICU bundle’s multiple evidence-based practices was at least partly responsible for the improvements in survival and shortened duration of mechanical ventilation.
But in one of the biggest limitations of this paper, the authors didn’t record or report bundle compliance, so we don’t even know if the intervention increased it.
If Not the Bundle, What?
For system-level interventions like this, cluster randomization has major advantages over individually randomizing patients. But stepped-wedge cluster randomization is vulnerable to secular changes in care (e.g., unmeasured general improvements) over the course of the trial. Because all the units “flip” from usual care into the intervention (and never vice versa), albeit at varying times, generalized improvements unrelated to the intervention can spuriously favor the intervention in the results. Standard statistical adjustments are made for this, but they aren’t impervious to (e.g.) entire units becoming more engaged (or fearful, or both) and delivering better care overall during the intervention.
Clinicians were not blinded—obviously, as they were the ones performing rounds and were expected to change their behavior, by design. Surely their psychology and interpersonal interactions also changed, along with the local culture at each ICU, etc. Humans with other humans: the ultimate confounders.
Mortality approached a bone-shivering 75% at 90 days. In trials of patients requiring mechanical ventilation in the U.S. and Europe, mortality has generally been about 30-40%.
When mortality is that high, of course, there are often more opportunities to reduce it.
Conclusions
This trial’s results can’t be directly extrapolated from Brazil to health systems in wealthy nations like the U.S.
Because of the limitations in reporting, we don’t even have proof that ICU care bundles saved lives, or even reduced time on the ventilator.
But something did. Someone did.
Whether it was greater adherence to best practices in critical care, or the collaborative effort itself improved teamwork and care quality over time, doesn’t matter in the end.
We already have evidence that these measures help (albeit not nearly as impressively as here).
They help almost in direct proportion to how boring they are, how tedious and repetitive. And to then have to talk about them over and over again to say they were done, or notice they weren’t, can seem even more tedious and repetitive.
Yet it turns out that’s what good care is. It turns out tedium can be an act of love.
You will likely never see the difference you make, and after everything you’ve gone through to get where you are, that can be deeply disappointing and may seem unfair.
But every nth time you roll that rock up the hill, someone lives who would have died. You’ll never know who, and neither will they.
So let’s keep rolling the rock.
Reference
Rosa, R. G., et al (2026). Integrated Telehealth Rehabilitation and Quality of Life in Mechanically Ventilated Adults. JAMA. https://doi.org/10.1001/jama.2026.10617