We Put a Gorilla in the Hospital

Photo: We Put a Gorilla in the Hospital

In the locker room of the 1997 Chicago Bulls, while coach Phil Jackson was briefing the team, one player was always absent.

In a room of his own, Michael Jordan would excuse himself so as not to be distracted by the game plan.

Jordan was highly selective about the information he would take in.

While his peers would obsess over video footage of their adversaries, Jordan stated in a 1997 interview, “I don't really concern myself with other players.” His mentality was simple - He didn’t need to know about the other team. He was not competing with them. They were competing with him.

In a sport where the consensus was to study the opponent and identify weaknesses and responsive strategies, the Greatest Of All Time did not participate in the process. He had a firm belief that less information was better.

In the same city, during the same year, another top performer was excusing himself from the briefings of his peers.

In his book, Blink, Malcolm Gladwell tells the story of Dr. Brendan Reilly after taking the job of Chairman of Medicine at Cook Country Hospital in Chicago. In the 1990s, Cook County Hospital more closely resembled a war zone than a metropolitan hospital. Gang warfare was relentless. Gunshot victims, shackled to their hospital beds, were a regular occurrence, and police officers were often as present as doctors.

It was underfunded, overcrowded and chaotic. With no air conditioning, ward temperatures would skyrocket in the summer. During a 1995 heatwave, temperatures inside the hospital reached a suffocating 120 degrees F - patients got heatstroke, lying in their hospital beds.

For Reilly, who had transferred from the upscale hospital at Dartmouth College, Cook County was a hard landing, but it's what he was looking for. Like any profession, some are in it for the money, and some are in it for the impact. Reilly wanted to go where he was needed the most.

And where he was needed most was the triage section of Cook County’s Emergency Room. This chaotic and dangerous unit inspired the dramatic hit tv show E.R. But, partly because of Reilly’s contributions, Cook Country Hospital became one of the country's most innovative hospitals.

Cook Country would see a whopping 250,000 patients enter the ER doors each year, ranging from sick kids to street violence. But, like any hospital in the United States, the most resource-intensive patients came in with chest pain.

The US Medical community had been blitzing the public for years on recognizing early warning signals of heart attacks. The result was a surge of patients with mild and ambiguous symptoms showing up at the hospital daily - Cook County saw around 30 chest pain walk-ins per day.

A high-risk chest pain patient is typically kept under monitoring for three days at the cost of $6000. Doctors, familiar with the litigious nature of American patients, are generally hesitant to under-treat anyone for the threat of malpractice lawsuits. Consequently, Cook County was spending roughly $60 Million per year monitoring chest pain while subsequently discharging 90% of the patients as having no risk of a heart attack.

Reilly wanted to determine if improving the intake assessment of chest pain patients and reducing the millions of unnecessary care expenses was possible.

He began with an assessment. He had his ER doctors review hundreds of previous chest pain files to determine who, based on their symptoms, was actually under threat of cardiac arrest and who might be sent home with an antacid.

The results were deflating. His team of doctors, across all levels of expertise and experience, were given access to the same information - medical history, family history, blood results, ECG results, and the patients’ real-time physical symptoms. The result? Absolute zero consistency of diagnosis across hundreds of patients.

Three doctors would review the same file - one would send the patient home, one would recommend a day of monitoring, and another would rush the patient into urgent care.

Reilly determined that the inconsistency came from the complexity of variables.

A chest pain patient may have perfect ECG results and a healthy lifestyle but also have diabetes (strongly correlated to heart disease) and a family history of heart complications. They may have no nefarious medical history but be an overweight smoker with a lifetime of terrible eating. They might be young but also a drug user. They may walk in with a perfect bill of health and only mild discomfort but an ECG that sounded alarm bells.

There were endless combinations of inputs, none of which were conclusive.

(There is an enzyme test that can prove a heart attack definitively, but this test takes hours in a well-organized hospital. Cook County was far from organized and could not risk waiting a day to determine if a heart risk was serious.)

With nearly 700 patients coming through the ER daily, Cook County doctors had to complete a patient assessment and direct the next steps within minutes or face a traffic jam of gurneys at their door.

Reilly had read the work of a cardiologist named Lee Goldman. Goldman had been researching whether statistical analysis could determine heart attack risk better than doctors and believed it could, with four simple data points.

Goldman’s formula required an ECG test to be run, followed by three questions:

Is the patient experiencing unprovoked chest pain?

Is there fluid in the patient's lungs?

Is the patient's systolic blood pressure below 100?

He then drew a decision tree that algorithmically determined the severity of the risk. For example, a patient with a concerning ECG, but only one or fewer positive responses to the questions, was low risk. A patient with a healthy ECG but three positive answers was a medium risk. A patient with a concerning ECG and two or more positive answers should be rushed to urgent care.

But despite Goldman's findings, no hospitals or universities in the country were willing to test his findings in the real world.

American doctors typically go through 11-14 years of school and spend $300,000+ in tuition. After this, nobody wants to be replaced by an algorithm. In addition, intuition tells us that human bodies and medicine are complex. There are too many variables to reduce heart risk to four data points… or else… what did we do all of that school for?

But although a trained doctor, Brendan Reilly was an operator and had been hired to improve Cook County’s processes. So he began testing Goldman’s algorithm. For two years, he ran comparative tests - rotating back and forth between doctor discretion and algorithmic decision-making.

The results were astounding. When determining false symptoms of a heart attack, Goldmans’ algorithm was 70% more accurate than human doctors. When deciding on high-risk patients, the algorithm was 25% more accurate than doctors - with a never before seen 95% success rate.

Suddenly, if you were at high risk of a heart attack, Cook County Hospital was the safest place in America.

The lesson was simple - the human doctors wanted to know everything before deciding. But too much information was causing a wide range of discretion.

The algorithm cuts right to the signal.

We are an analytical species; we default to more information is better. The more consequential the decision, the more data we think we need.

Psychologist Jonathan Schooler has conducted studies showing that if we have seen a random stranger before, our ability to pick that stranger from a police lineup is quite good. However, we are far less successful if we are asked to describe the person verbally before we look at the lineup.

The addition of verbal descriptors creates more information to process when we look at the lineup, and the added variables cause more mistakes. Schooler describes a subconscious debate between our left hemisphere (thinks in words) and right hemisphere (thinks in pictures) of our brain, with one adding unnecessary information to the conversation.

This is interesting. In the case of both Michael Jordan and the Cook County doctors, a reduction in information led to a better outcome - but in each case, for very different reasons.

Michael Jordan’s competitive advantage was that he was the best on the court (maybe that sounds obvious) - but through obsessively redundant training, his instincts, agility, and positioning were superior to his competition. Consequently, anything distracting him from his performance decreased his competitive advantage. He recognized this: "I don't really concern myself with other players' games. I only worry about my own.He became so advanced in the game that intuition was the best strategy.

Doctors play a very different game - the human body has far more variables than a basketball court. In this case, leaning on intuition and gut feeling decreased the odds of success.

For Jordan, removing the noise allowed him to focus on himself. For doctors, removing the noise protected them from themselves.

Humans are endlessly fascinating.

If I were to speculate, the more primal the activity, the more advantageous our intuition is. For example, although a modern sport, basketball is very basic - your hands and feet compete against my hands and feet while trying to throw a ball in a hoop. On the other hand, modern medicine is chart displays of vital signs, blood data, and all varieties of human biology translated to numbers and percentages.

Give my ancestors from 5000 years ago an hour, and they would figure out basketball. But put them in an ER… and that, I would love to watch.

We run into trouble when we take our primal instincts into the modern world. The tricky part is accepting our limitations and acknowledging where our natural environment ends and the new world begins.

Maybe that is the stress test - what would happen if my Neanderthal cousins tried this?

23&Me tells me I have more Neanderthal DNA than 63% of their other customers. This may be why I keep my life simple with gold, steak and pushups.

I love writing my weekly essay. If you enjoy reading it, please send it to a friend who might like it too.

Love you all, and have an epic Sunday.


Now let's get down to business; How did we perform this week…...

Keep reading with a 7-day free trial

Subscribe to Jay’s Letter to keep reading this post and get 7 days of free access to the full post archives.

Start trial

A subscription gets you:

The Deep Dive - The Most Important Signals in Business, Finance and Culture.
Portfolio Allocation and Rationale
The Sovereign Life Blue Print