By ANISH KOKA, MD
A number of politically tinged narratives have divided physicians during the pandemic. It would be unfortunate if politics obscured the major problem brought into stark relief by the pandemic: a system that marginalizes physicians and strips them of agency.
In practices big and small, hospital-employed or private practice, nursing homes or hospitals, there are serious issues raising their heads for doctors and their patients.
No masks for you
When I walked into my office Thursday, March 12th, I assembled the office staff for the first time to talk about COVID. The prior weekend had been awash with scenes of mayhem in Italy, and I had come away with the dawning realization that my wishful thinking on the virus from Wuhan skipping us was dead wrong. The US focus had been on travel from China and other Far East hotspots. There was no such limitation on travel from Europe. The virus had clearly seeded Italy and possibly other parts of Europe heavily, and now the US was faced with the very real possibility that there was significant community spread that had occurred from travelers from Europe and Italy over the last month. I had assumed that seeing no cases in our hospitals and ICUs by early March meant the virus had been contained in China. That was clearly not the case.
Our testing apparatus had also largely been limited in the US to symptomatic patients who had been to high-risk countries. If Europe was seeded, this meant we had not been screening nearly enough people. When I heard the first few cases pop up in my county, it was clear the jig was up. It was pandemic panic mode time. There was a chance that there were thousands of cases in the community we didn’t know about and that we were weeks away from the die-off happening in hospitals in China and Italy. So what I told the staff the morning of March 12th was that we needed to start acting now as if there was significant spread of COVID in the community. This meant canceling clinic visits for all but urgent patients, wearing masks, trying to buy masks, attention to hand hygiene, cleaning rooms between patients, screening everyone for flu-like symptoms before coming to the office, and moving to a skeleton staff in the office. I left the office that day wearing a mask as I headed to the ER.
Again, this was March 12th. Even the short walk from my office to the hospital ER resulted in quite a few stares. But I was going to the ER, after all, at a time when a possible surge was happening. I already knew the Chinese experience at this point. Health care workers got infected most likely because they didn’t have adequate Personal Protective Equipment (PPE) early on. Of course, I didn’t have any n95 respirators because I hadn’t prepared for this scenario. All I had was a limited quantity of surgical masks. I figured it was better than nothing, and at the very least, I would hopefully reduce the chance of transmitting the virus to patients I was going to come into contact with in case I picked it up in my travels. When I walked into the ER, I was surprised to find no separate area set up to handle COVID patients. There was one negative pressure room set up for COVID patients. That’s right, one room. The waiting room was a common area as well, which meant that people without COVID had a reasonable shot of acquiring COVID while in the waiting room. The initial Chinese data suggested 40% of COVID infections were acquired in the hospital.
I left the ER to travel to another part of the hospital, and in the lobby was stopped by a somewhat excitable administrator. Why was I wearing a mask, he wanted to know? Taken aback, I stammered something about trying to protect patients. Current guidelines, I was informed, were to only wear a mask if sick. Any other use was discouraged. We wouldn’t want to scare patients. Not wanting a scene, I took the mask off in the lobby to mollify the agitated administrator. Four days later, the same hospital had a universal mask policy for staff. Four weeks later, and the surgeon general was recommending universal masking for the entire population. Today, I am at risk of being dragged out of the local grocery store unless I’m wearing a mask.
PPE can be divided into plain old surgical facemasks, facemasks with shields, n95 respirators and Powered Air Purifying Respirators (PAPRs). The vast majority of US physicians spend little time thinking about PPE on a daily basis, because there has never been a widespread threat from a novel pathogen with no cure. Ebola and SARS would have qualified but American doctors domestically never faced these foes. Generally physicians let infection control teams in hospitals run the show. This translates into physicians putting on whatever PPE is sitting in the isolation cart that sits out of the patients rooms. The general theory that relates to the transmissibility of pathogens breaks down to four categories: Direct contact, indirect contact, droplet, and airborne.
Direct contact requires physical contact with some part of an infected person. Usually this happens when infected bodily fluids come into contact with the eyes, mouth, nose, genitals, or open wounds. Indirect contact is when an external surface like a doorknob serves as an intermediary for the communication of infected bodily fluids. Droplet transmission results from the expulsion of small particles of bodily fluids into the air through coughing, sneezing, or vomiting. These are thought to be heavier particles that don’t stay suspended in the air for very long, and so require close proximity for transmission to occur. Airborne transmission occurs via droplets that are small enough to float in the air and so can cause infection over long distances. Airborne particles are produced by coughing, sneezing, or certain aerosol-generating medical procedures that involve instrumenting the upper airways.
The type of transmission impacts the recommendations for PPE from organizations like the CDC. These guidelines provide the backbone for infection control in healthcare facilities. It’s an easy categorization scheme. Smallpox requires airborne and contact precautions, Ebola requires droplet and contact precautions, while HIV and staph require standard precautions. It should strike many how remarkable it is that millions of years of evolution has lead to infectious pathogens being neatly divided in a way that would make recommendations for PPE for healthcare workers so simple.
Of course biology isn’t this simple. A sneeze can aerosolize a pathogen. There isn’t a biological cliff at which point droplets suddenly become small enough to be airborne. If it’s not already obvious, the distinctions made have more to do with convenience than it has to do with the actual mechanics of disease transmission.
A beautiful review in the lay press by Ari Schulman notes that the current schema is rooted in distant history that has not been well updated. Charles V. Chapin, an eminent epidemiologist noted in 1910 that “there was no evidence that [infection by air] is an appreciable factor in the maintenance of most of our contagious diseases.” Therefore, he advised that attention be primarily paid “to the prevention of contact infection”. Not surprisingly, it was clinicians who began to question this paradigm. In 1919 George H. Weaver, a physician at Chicago’s Durand Hospital for the poor, subsequently observed “the ease with which infections may be transferred through mouth droplets when people are brought into intimate association”.
A famous paper by Mildred and William Wells in 1936 attempted to examine the variety of clinical reports of airborne infectious spread. The couple started with the assumption that the air was a conduit for transmission of bacteria based on the work of Louis Pasteur, whose classic experiments on spontaneous generation in 1861 proved that air was populated with microscopic germs which caused putrefaction and fermentation. The direct evidence of this proved difficult to produce in the decades to follow. The failure to produce nasopharyngeal organisms on plates exposed farther than a few feet from a person who was coughing and sneezing led to the conclusion that infectious pathogens settled out “in short distances and in brief time intervals”, and that airborne transmission just didn’t happen. The absence of evidence became evidence of absence.
Technical advances that allowed the observation of droplets were carried out by Wells to first demonstrate the plausibility of airborne spread. Animal experiments which involved inoculation of healthy ferrets with the air from an infected ferret with influenza suggested influenza could be airborne. Once again it was observations by clinicians that drove the research. Noting the recent reports of the spread of the psittacosis virus at a National Institutes of Health building via the ventilation system from the basement to the upper floors, the 1936 researchers attempted to mimic this by inoculating a culture of B.coli in the humidifying water of a one room air conditioner in the basement of the Harvard School of Public Health. They noted B. coli was recovered from the ends of every corridor up to the top floor of the three story building and concluded that “infected nuclei can therefore be dispersed by ventilating currents throughout a building.. “
The researchers next turned their attention to the effect of sneezing. Fifty sneezes were induced from a group of subjects in a standard air conditioned room. Bacterial samples collected in the air centrifuge on blood agar tubes revealed thousands of alpha streptococci and M. catarrhalis. Even after the subjects left the room, the air contained many hundreds per sample. Interestingly, in a nod to the current universal masking recommendations, an application of handkerchiefs to smother the sneeze showed a marked reduction in collected organisms.
The triumphant conclusion of this brilliant paper noted – “The burden of proof of air-borne infection, which had hitherto rested on bacteriology, has been lifted, and there now rests on epidemiology the burden of disproof of airborne disease”
To summarize, the world of science rejected a theory of airborne transmission based on clinical observations for decades because technical issues precluded organisms from being recovered from the air. Instead of holding on to the possibility that the clinical observations were ground truth, blind empiricism created a new dogma: Airborne transmission couldn’t happen. It would be convenient to assume the blind zeal attached to what can be measured is a product of a bygone era, but sad to say this is alive in well in a current society among an overly certain technocrat class that makes up a large portion of the intellectual elite.
In the hundred years since the Wells team published their conclusion, the debate on Personal Protective Equipment (PPE) to deal with infectious pathogens has been conveniently made simple by institutions that write the definitive guidelines on the subject. The neat categorization of PPE based on characteristics of pathogens never made biological sense. As was shown by the Wells team in one form or the other, infectious pathogens may be made airborne. We get away with artificial distinctions most of the time because the consequences of not having the perfect PPE are usually not apparent. This changes when humans encounter a lethal novel virus for which the population does not have immunity.
Consider Ebola. Establishing transmissibility of an infectious pathogen involves a demonstration of a variety of steps: Viability of virus in aerosolized particles, persistence in the environment, infective capacity, demonstration in animal models, epidemiology via contact tracing, and direct observation of healthcare workers wearing respiratory equipment. Ebola is able to reside within airborne sized particles, can stay persistent at certain temperatures for as long as 50 days, and animal models and epidemiologic data are mixed.
The conclusion from this mixed bag of data that does NOT rule out airborne transmission in a virus that has close to 50% mortality would seem to be to err on the side of recommending PPE assuming airborne transmission. Instead, the CDC and the WHO recommended against airborne precautions, in part because Ebola does not primarily affect the respiratory tract citing a lack of an evidence base to definitively prove transmissibility that would require more aggressive PPE. The official recommendation was instead for a plain surgical mask, unless health care workers were involved in medical procedures that could cause aerosolization.
In what seems to be reflective of the generally cavalier attitude of those in charge, an editorial published in the Lancet August 2014 argued against the use of more advanced PPE such as respirators despite noting that Ebola had been noted to “rarely be transmitted via an airborne route” because of the expense of complete respiratory protection. Even worse, “such an approach suggests that the only defense is individual protective equipment, which is inaccessible to the general population. Moreover, the image of workers with spectacular protective clothing might contribute to the panic in some communities. If this leads people to flee affected areas it could increase the spread of infection. It also reinforces the view that some lives are more valuable than others… ” The article concludes by emphasizing that the rational and efficient use of PPE may only be achieved by communicating a consistent message that the disease is essentially transmitted through direct contact. The overriding concern here in a document that is supposed to be about precautions for healthcare workers, is for the “public” psyche. The psyche of healthcare workers who are taking care of patients seems to be of minimal concern.
In another editorial, Harvard public health physicians scolded hospitals that were considering Ebola protocols that involved placing patients in negative pressure rooms, and compelling all personnel to wear full body hazmat suits, as well as requiring n95’s and powered air-purifying respirators. The public health officials pointed to “evidence-based” CDC guidelines that were really just cherry-picked case series and epidemiologic surveys to suggest contact with infectious body fluids was the primary mode of transmission. They did give a nod to potential airborne transmission by advising higher levels of protection for aerosol producing medical procedures.
But conveniently, the authors ignored the pragmatic consideration of natural aerosol producing procedures like coughing or sneezing, and go on to imagine the risks of self-contamination of taking off “unfamiliar PPE”. There is of course, no evidence provided to support the implication that wearing too much PPE when faced with a high mortality virus is more dangerous. The important thing is to not “inflate patients and caregivers anxiety levels, increase costs” and, as health care professionals, “strive to provide evidence-based care driven by science rather than by the media or mass hysteria.”
Never mind that the CDC’s guidelines for laboratory rather than health care settings puts Ebola on the list of pathogens requiring the most stringent possible level of protection. Ebola belongs to the Biosafety Level 4 club, and requires the use of powered respirators and other strict controls because :
The microbes in a BSL-4 lab are dangerous and exotic, posing a high risk of aerosol-transmitted infections. Infections caused by these microbes are frequently fatal and without treatment or vaccines.
Yet health care workers may work in close proximity with these same viruses in a sick patient that may be coughing, sneezing, and having loose stools wearing a gown, gloves and surgical mask, and then go home to their families, stopping to pick up a pork roast from the grocery store along the way.
The end result of all these machinations was to allow a natural experiment to play out demonstrating the folly of issuing recommendations that have more to do with calming the psyche of the public than with actual science. Low resource centers in Africa, driven in part by necessity, implemented less protection, while others did more. Doctors Without Borders sites used full body hazmat suits and respirators with a smattering of infections related to direct care of patients in facilities. Local hospitals used surgical masks with significantly different results. Even in the United States’ small experience of treating Ebola patients, two infected nurses in Dallas were not wearing respirators. Atlanta’s Emory Hospital, which required respirators, successfully treated four Ebola patients with no infections of health care workers. The CDC subsequently changed its guidelines quietly to recommend respirators for Ebola. The WHO still has not.
There has been evidence since at least 1936 that pathogens may transmit disease over long distances via the air, that coughing and sneezing can create aerosols, and that the simple handkerchief greatly reduces the transmission of aerosolized particles. Yet, it is March 12th, 2020, and faced with a potentially deadly pathogen with no cure, an excitable administrator wants to tell me about hospital guidelines that recommend against everyone wearing a mask so that the public doesn’t get scared.
This particular story of masks and PPE certainly isn’t a problem created by administrators. They simply are channeling the zeitgeist that permeates the world of public health and infection control for some time now. In doing so, they empower the cadre of non-clinician administrators who point to these ‘evidence-based’ proclamations when telling physicians what to do. It’s also very convenient for health systems to have infection control committees make rules that are less resource-intensive. This is the dirty secret of many a hospital committee – the goal is to rubber-stamp what the health system needs, rather than do what’s best for patients and physicians.
Don’t touch the ventilator
In another age, it was the clinicians that dictated process. Now processes are dictated to clinicians. In large part this relates to the incorporation of physicians as cogs of expansive, multibillion-dollar health systems. Just like decisions about PPE, the medical management of patients is now by protocol. The system is set up to thwart variance and dissuade individualized treatment.
This has made the traditional role of physician as team leader into a relic of long ago. Every trainee in medicine feels this to some degree when stepping into a hospital for the first time. There’s a way things are done that the nurses, respiratory therapists, physical therapists, charge nurses and case managers are anxious to imprint on the trainee. This is a necessary part of learning for all trainees. The more recent development is applying this imprinting process to trained physicians. The evolution was a byproduct of evidence-based medicine that rested on randomized control trials comparing new therapies or protocols to standard of care. It was a well-intentioned attempt to improve care by reducing variations in care that may be substandard. The roots of the focus on minimizing variation come from its successful application in industrial manufacturing in places like sausage factories.
Someone forgot to tell the intelligent designers that medicine isn’t a sausage factory.
Take the recent controversy that has emerged over mechanical ventilators and COVID. Since the coronavirus has a proclivity for lung tissue, the current conversation has focused on whether the nation has enough ventilators to support the sickest COVID patients. Managing patients on ventilators practically translates to assisting patients unable to breathe on their own. This has traditionally been done using machines that force air into the lungs in a rhythmic fashion to simulate the normal pattern of breathing. Ventilators don’t cure the underlying problem with the lungs, they simply help support patients until the underlying disease process affecting the lungs is resolved.
It became apparent decades ago that the process of ventilation was not a benign affair. The lungs are normally very compliant structures, which means relatively low pressures can be used to inflate lungs. In diseased lungs, where infection or inflammation affects the lungs in a non-uniform manner, it was recognized that pressures applied in a uniform manner may damage poorly compliant areas of the lung. A series of randomized controlled trials culminated in the famous ARDNet trial that examined lower administered volumes of air (a lung-protective strategy) than was standard at the time. The trial was a blockbuster at a time many were doubtful significant improvements were possible in critically ill patients on ventilators. Patients in the lung-protective, lower lung volume arm had 22% lower mortality.
Lower lung volumes meant less total ventilation, which meant letting surrogate markers that were dependent on ventilation to be less than perfect. The thought was that there must be damage happening at the higher lung volumes that was even worse for patients. As has been hammered home to many a medical trainee, patients don’t care about numbers, they care about living and getting off of the ventilator. And it sure seemed a simple adjustment to protocol could accomplish this.
The trial changed the practice of medicine. From that point on, every intensive care unit in the country made the new lung volume settings (6ml/kg) from ARDSNet standard. This was progress, of course, but it’s always the case that every new standard becomes a new dogma.
The rapid increase in trials that studied protocols has resulted in a rapid uptake of protocol-based management in practice. Hospitals today certainly feel like physicians oversee protocols rather than patients. A patient arrives in the intensive care unit on a ventilator, a templated order based on some guideline goes into effect, a respiratory therapist implements the order, and then checks daily via another templated order based on another guideline if a patient can be liberated from the ventilator. Management, by default, can become a technical affair that allows the physician to be minimally involved. The mantra from respiratory therapists – non-physician specialists who manage ventilators now – to physicians is : Don’t touch the ventilator.
Protocols dictate to nurses and therapists what parameters (laboratory values, patient metrics on the ventilator) need to be checked on a daily basis, and what path to follow with regard to management of the ventilator. Indeed studies have demonstrated that protocols may best the average physician in many cases because it saves on the time needed for scarce physician resources to get to the patient to make a decision.
This neat little system, however, runs into turbulent waters when patients don’t behave according to protocols. The COVID pandemic is a gigantic deviation from protocol. The disease appears to be marked by significant gas exchange abnormalities in the lung. Whether it’s due to coating of lungs with virus or an inflammatory reaction induced by the virus, the unit of the lungs that appears to be ground zero for COVID is the alveoli, the main unit of the lung involved in the exchange of oxygen and carbon dioxide. Generally, an important trigger for deciding on mechanical ventilation has been an inability to exchange these vital gases effectively, and an easy surrogate to follow has been the oxygenation level in the blood. This is easily done with the simple pulse oximeter which measures the changes in light absorption in oxygenated and deoxygenated blood to return a “percent saturation” on a monitor. The healthy human saturates close to 100%. As the number dips below 90% the anxiety of doctors and the health care team generally goes up exponentially. These low numbers are generally regarded as a warning that the patient may be approaching a cliff, and once over the cliff, resuscitation may prove to be very difficult. The approach has been to use the oxygen saturation as one data point that helps guide the timing of placement on a mechanical ventilator. This allows the administration of higher amounts of oxygen and may relieve the work of breathing of a patient in distress.
In the case of COVID, anecdotes of patients who were comfortably hypoxic began emerging: Some patients were being asked to get off their cell phone to be intubated. It has long been understood there is a downside to being placed on a ventilator. There is a cost that’s incurred by putting a tube down your throat, and blowing air through it. It’s generally uncomfortable, and often requires some type of sedation. This turns out to be especially true with COVID because patients are very hard to manage once on ventilators. Physicians with direct experience I’ve spoken to note that the patient’s numbers after intubation look terrible. Rapidly increasing levels of support on the ventilator are required in the form of fraction of oxygen, as well as the volume of air that requires to be moved per minute into the patient’s lungs. The patients often aren’t in-sync with the ventilator and require sedation along with muscular paralytics be administered to completely take over the work of breathing for the patient. One physician described the intensive care units they were managing as zombie units because of the significant portion of patients that were heavily sedated.
The need for significant sedation is complicating because of the downsides of sedation. Every day of complete bed rest results in muscle atrophy which may make it more difficult to ultimately get off the ventilator. Complete sedation puts the burden on the care team to do everything that the patient used to do for themselves. Fluid intake, shifting in bed to avoid sores and pressure ulcers, and caloric intake all need to be managed. In the best of circumstances, this is tough. In the midst of a pandemic with a potentially deadly virus where time spent in rooms is minimized, this is much harder.
As a result, some physicians wondered if it would be reasonable to tolerate more marginal numbers to minimize the intensity of sedation and ventilation that may be required. Interestingly, these were the type of steps taken that lead to the ARDSNet trial – tolerate worse gas exchange parameters that came along with lower delivered lung volumes so as to expose the lungs to less pressure. But the physicians ran into a bit of a problem. There was a new dogma. Experienced ICU nurses and respiratory therapists are wedded to the ARDSNet protocol. Deviation was seen as aberrant. I’m told physician orders, if not refused outright, were “slow played” if they ran counter to protocol. In a now-public video on Medscape, one physician had to leave the intensive care unit he was managing because he couldn’t ethically follow the protocols.
To be clear, this isn’t a story of miscreant physician trying to institute a brand new untested protocol broadly, it was physicians seeking to deviate from a standard for certain patients because they thought it best. The role of protocols should be to function as a basic framework. In practice, they can become rigid walls used by non-clinicians to veto doctors. ARDSNet showed that 12ml/kg was worse than 6ml/kg. It doesn’t say 8ml/kg is worse. The general consensus of lung protective ventilation that was introduced by ARDSNet has been to keep pressures within the lung below 30mmHg with every delivered breath. But what about 32mmHg? These are decisions that should be under the purview of the thoughtful clinician attempting to respond to the clinical scenario presented by the patient.
And they aren’t.
The desire to stamp out variance leaves little room for the nimble care that may be needed when faced with a novel disease. It’s likely too early to say what clinicians’ hunches will eventually prove to be salient. Many of them will no doubt be incorrect. The beauty of living in an interconnected world is that ideas are discarded by the community almost as quickly as they are generated. Multiple trials are, of course, underway to test a variety of these ideas and will help further inform the community. In the meantime, physicians are left to try to adjust their therapy to patients that are failing in front of their eyes. This doesn’t mean physicians have license to start offering spinal manipulations to treat COVID, but it does mean understanding that tidal volumes of 6ml/kg aren’t found in the Ten Commandments.
The consequences of being rigid in these circumstances runs the risk of making a bad situation worse. The end stages of any illness that involves the lungs to this degree is going to have a significant mortality independent of any interventions the medical community makes, but personal communications from experienced critical care physicians certainly make it plausible that following protocols blindly in this illness has a price.
Empowering physicians to help patients cannot be done without first freeing the thoughtful, experienced physician latitude to do what they believe is best.
Far from the ventilators in ICUs, even physician administrators find themselves swimming upstream. The COVID nursing home tragedy provides a good example of this.
Nursing homes and COVID: The disastrous shelter in place plan
Nursing homes that care for the elderly and infirm have been at the center of the COVID19 storm. The first US outbreak was in a Washington state nursing home, and the same scene has played out at multiple similar facilities across the country. It was always the case that the nursing home demographic would be hit especially hard with a virus like COVID that causes a lot of damage to the elderly with comorbid medical conditions, but this particular train wreck is almost definitely exacerbated by decision making taking place far from nursing homes.
Very early on, the medical caretakers of nursing home patients found themselves at odds with other parts of the health system, as well as local public health officials. Guidance was issued with regards to visitors to nursing homes early on, but there was precious little thought from central authorities about what to do when a resident at a nursing home got infected. As one of the medical caretakers at a nursing home, I thought it was obvious that patients who became infected needed to be rapidly removed from the nursing home. There was little doubt that the dormitory set up and significant health care needs of residents would make isolation almost impossible at the nursing home. If one person on a 40-bed unit was to get COVID, it was certain to spread like wildfire throughout the unit, and probably throughout the nursing home.
What I found was a local Department Of Health (DOH) that was laser-focused on keeping patients at nursing homes. This was understandable early on because the initial epidemiological modeling, and the scenes out of Italy and New York suggested local health systems could be overwhelmed in short order. Health systems were scrambling to build bed capacity and were worried about running out of ventilators. It made little sense to fill up beds in a system that was about to be overrun. The trajectory of the pandemic unfolded very differently across the country. There was the New York/New Jersey metropolitan area and there was everything else. Two weeks into the lockdown, Philadelphia hospitals had been emptied waiting for a New York-style surge that never came. But at this point the nursing homes unfortunately had started to see their first infections probably seeded from the nursing home staff. A nimble response at this point would have been to utilize the excess capacity of the hospitals and shelter the early positive nursing home patients. Hospitals had well trained, relatively highly paid staff, adequate PPE, and negative pressure rooms. Nursing homes had lowly paid workers with 10:1 patient ratios (ancillary staff) that were poorly equipped to effectively quarantine a patient with COVID. But nursing home patients were treated like patients from the community who were too well to be admitted to the hospital – they were sent home. The consequences of keeping these patients at the nursing home meant the health system had to eventually deal with the entire nursing home being infected.
We looked for help.
We asked the DOH to allow us to test everyone at the nursing home in order to effectively cluster everyone with COVID in one unit. They refused because the guidelines didn’t recommend this for those that were asymptomatic. We asked to utilize a large room to cohort patients with COVID. Nursing home administration and the DOH said this wasn’t possible. COVID tests were being sent to the DOH and had a turn around time of 2-4 days initially. I called the local large health system that had acquired a new rapid, same-day test to see if we could send them tests. No answer. I spoke to a nice hospitalist at another large health system who was very receptive to the idea of boarding COVID positive patients from the nursing home in the half-empty hospital to avoid the entire nursing home eventually being infected. An email chain followed to get permission from administrative units. Absolutely not was the answer. I was told the more fruitful endeavor was to discuss advance directives with the residents. Did they really want to be resuscitated if they got too sick? And if they didn’t want to be resuscitated did they really want to go to the hospital?
The implicit message: Keep the residents away from the pristine hospital. If they get too sick, hopefully they don’t need to be resuscitated. It’s sad. There’s a built-in bias society tends to have towards nursing homes that at times has reminded me of a leper colony. Out of sight, out of mind. In training, I’m ashamed to say I had the same biases. I groaned when I heard about the fever admission from the nursing home. The patients as I remembered them were non-verbal and almost always had a feeding tube. UTIs, meningitis, pneumonia. Usually, family members were scarce or unhelpful. Code status was never assessed.
As an attending physician, the view from the nursing home was wholly different. Everyone had a story, and at this nursing home at least, everyone had a personality even if the ability to communicate for a variety of reasons was unique. The first year I was there running a small cardiology clinic, I didn’t know that the holiday party was the same day. Mr. Rose was playing the guitar and wasn’t available. So I went to watch him play.
The system doesn’t know any of these things. The goal seemed to be to keep the health system free of these second class citizens.
Maddeningly, we physicians were supposed to be the advocates, and were powerless. This was always going to be hard. There was always going to be a death toll in this population. But with the limits placed on us, their medical guardians, that toll is bound to be much higher. Maybe there was a physician there somewhere behind the curtain who was running the show, but I never made it past the army of nurse administrators eager to tell me what the protocol was for my situation.
Nothing is easy about this. Mr. Rose won’t be playing in any more holiday shows. He wasn’t a well man. He had lived well beyond what someone with the injury that put him in the nursing home was expected to live. In some sense, he had been living on borrowed time. If not for COVID, perhaps a conventional pneumonia would have taken him. It doesn’t feel any better.
The traditional role of the physician is under siege precisely because it was felt the physician as owner lead to misaligned incentives. Third-party payment, embraced by physicians in the early years, has been steadily redirected to systems. The end result is the system employs physicians, and it is the system that directs medical care. The system will decide what PPE you will wear, what protocols are allowed on the ventilator, and how your nursing home residents are best managed. The intentions are good and based foundationally on the idea that decision making needs to be taken out of the hands of individuals as much as possible to reduce the variance in dealing with a problem. This schema was an attempt to apply what works well when working with mechanical systems to deliver mass production at scale with high quality. The innovation of the last few decades of health care system delivery has been to make hospitals more like factories with the aim to provide higher-value care with this approach. It would appear we have done anything but, and it’s possible the whole enterprise is based on rotten assumptions.
Gary Klein, a research psychologist who studies decision making and has written extensively on the topic based on an extensive historical review, notes that some of the best decisions are made by adapting to circumstances rather than being bound to protocols and checklists. The standard advice applies when everything is clear and situations are well ordered, but the tough decisions are made with imperfect information in complex and ambiguous situations.
The Miracle on the Hudson
Minutes after takeoff from LaGuardia, New York, US Airways Flight 1549 flew into a flock of large geese and lost both of its engines. The plane was at 3000 feet with no engine thrust and a rapidly descending flight trajectory. Turn back to LaGuardia? Try to make it to Teterboro Airport in New Jersey? Or land in the Hudson river? There was no protocol for this particular situation. There was only a well of experience to rely on. LaGuardia was quickly ruled out- it would require taking a rapidly descending commercial airliner over Manhattan. The initial plan was Teterboro, but the situation rapidly devolved. Teterboro was just too far to make. The next terse message to air traffic control — “We’re gonna be in the Hudson..” The rest is history. Captain “Sully” Sullenberger glided a fully loaded plane down into the Hudson – and everyone survived.
The established flight protocol on losing engines is to attempt to return to home base – LaGuardia, in this case. The National Transportation Safety Board (NTSB) had the job of deciding if Sully had done the right thing. Was Sully’s mental calculation correct? Was LaGuardia really not an option? Had the water landing been a reckless display of arrogance in breach of a safer protocol guided path to an airport? An initial flight simulation showed the plane could successfully return to LaGuardia after both engines were lost. But this didn’t incorporate the time taken to figure out that the engine’s were irretrievably lost. Delay the decision to return to LaGuardia by 35 seconds, and the simulation fails. The plane crashes. Everyone dies. Seventeen seconds after losing both engines, Captain Sully actually followed protocol by calling for the Engine Dual Failure checklist which the copilot began to go through to troubleshoot the engines. They never made it to the ‘ditching into the water’ part of the checklist because the checklist was three pages long, and the pilots ran out of time. Despite being unable to complete the checklist the crew did accomplish one critical item that the flight crew did not reach in the checklist: starting the Auxilliary Power Unit (APU). From the NTSB report:
“Starting the APU early in the accident sequence proved critical because it improved the outcome of ditching by ensuring electrical power was available to the airplane. Further, if the captain had not started the APU, the airplane would not have remained in normal law mode. This critical step would not have been completed if the flight crew had simply followed the order of the items in the checklist”
Captain Sully wasn’t perfect. He did not attain the required speed for his descent and landing, resulting in a high Angle of Attack (AOA). This resulted in significant structural damage to the fuselage on landing which caused water to enter the cabin. The directive on what speed to travel in case both engines failed was on page 2 of the checklist. The pilots only had time to get through page 1 before abandoning the checklist as time ran out. The checklist had been designed for problems at 30,000 feet. The aircraft lost both engines at 3,000 feet. The pilots only had 3 minutes to land the plane. Experience and intuition of a seasoned pilot carried the day. The protocols did not.
In the time of COVID, physicians more than ever find themselves as pretty baubles on display to reassure the public. The reality is that it’s the system that directs physicians and as a result it is the physicians who find themselves unable to protect themselves or their patients. The lessons from the pandemic, and yes, even the airline industry is loud and clear: Put the physicians back in charge.
Anish Koka is a cardiologist in private practice in Philadelphia. This post also appears at The Health Care Blog. Twitter: @anish_koka