I was in the last group of interns selected by Dr George Thorn who was the Chief of Medicine at the Peter Bent Brigham Hospital from 1942 until 1972. I began my internship in Medicine on June 20, 1971. That night I was the admitting intern on the male ward, F Main. I can remember sitting at the nurse’s station on F Main at about five in the morning after that first night watching the daylight creep into the room while looking out over a dismal parking lot wondering what was to become of me. I kept falling asleep as I tried to write orders for the most recent of many admissions. I had not slept that night, and I had no idea how I was going to make it through the day, much less the next year. At that time all but two of our “rotations,” Neurology and the consult service, had every other night call schedules. Neurology and the consult service were light every third night “recovery rotations.” Eugene Braunwald came to the Brigham the next year as the Hersey Professor of the Theory and Practice of Physic, and the call schedule for interns changed to a merciful every third night. 

 

The Brigham has always been an interesting place to be a student of the practice of medicine. It was easy to feel that you were near the “source” of medical knowledge since George Thorn was one of the principal authors and editors of Harrison’s Principles of Internal Medicine. The Chief of Surgery was the famous Dr. Francis D. (Franny)  Moore, the author of the “bible of surgical residents,” The Metabolic Care of the Surgical Patient. Thorn and Moore headed the marquee, but the cast also included many other “greats” like Richard Gorlin, Louis Dexter, and Bernie Lown (whose politics were further left than Bernie Sanders’). Behind the hospital and across the alley that was called Shattuck Street were the Countway Library, The Harvard Medical School quadrangle, The Harvard School of Public Health, and Children’s Hospital. It was, and is, a medical “Mecca.” Walking down Longwood Avenue a couple of blocks from the quadrangle one comes to the intersection of Longwood and Brookline Avenues. Turn left to go to the New England Deaconess Hospital, Joslin Clinic, and at that time the Lahey Clinic, or turn right to the Beth Israel Hospital, and past that, the original offices of Harvard Community Health Plan near Fenway Park. For almost fifty years those few blocks were the center of my world. Things still don’t look much different now, although there are many new buildings for patient care and new corporate affiliations. The BI, New England Deaconess, and Lahey are now one huge enterprise, and The Brigham and The Massachusetts General Hospital, a few miles across Back Bay, form the core of Partners Healthcare. 

 

Sleep deprivation was part of the bizarre educational process that would go on for years that I began that first sleepless night “on call.” I was told that I should be happy not to have been an intern at Hopkins where the house staff worked five nights out of seven, and no one ever signed out their patients to a colleague. Dr. Thorn had come to the Brigham from Hopkins in 1942, and brought with him some of that culture. One had the idea that the sleep deprivation was part of a character building exercise. If it did not build character, it did separate the “wheat from the chaff.” Looking back through the haze of what I can remember, the impossible call schedule with long sleepless hours was essentially a right of passage. I have always thought that there must be a better way to train physicians, and I still have psychic scars and remnants of guilt associated with the thought of the damage I did to patients as I was earning my way into the profession that I loved, and that was so good to me. Perhaps that guilt explains some of my current interests in a transformation of healthcare. Healthcare would be much better for those who receive it, if the preparation of those who deliver it was more humane. You can get a more detailed example of my experience and the origin of my guilt by reading “Elizabeth McCarthy’s Story: The painful education of an intern,” which is posted on this website. The story begins:

 

I was an intern in the distant past when we worked every other night. The 48-hour cycle usually consisted of thirty six to forty hours at the hospital followed by eight to twelve hours of recovery before it started all over again. Those were the days when I took naps at stoplights and learned to love coffee. I thrived on coffee and beepers and was sustained at times by the adrenaline stimulated by the fear that the message announced by the beeper would be a challenge that exceeded my skill and knowledge.

 

Grand Rounds on Friday morning was a part of the hospital’s liturgical week. We all attended like we were attending church. The amphitheater had steep rows of seats. Along the walls and across the front of the large room there were pictures of famous Brigham physicians and surgeons. Soma Weiss, Henry Asbury Christain, Harvey Cushing, Sam Levine and others looked down on the weekly gathering, and never changed their expressions. The current crop of professors and visiting dignitaries occupied the front row. Some of them would eventually take their own place on the wall. 

 

I would usually hustle in at the latest possible moment, and try to find a comfortable seat near the top of the amphitheater on one of the back rows, preferably on an aisle so that I could stretch out my legs while I took a nap. It wasn’t that I was not interested in what was to be presented, it was just that I was sure that when the lights were turned off I would immediately fall asleep. These are the days of PowerPoint. Those were the days of lantern slides, and the darker it was, the better they were for viewing.  

 

I do remember one presentation when I did not fall asleep. Perhaps I was on one of the two every third night rotations, and maybe I was coming off two nights to recover from my last night on call. The presenter was a young visiting faculty member from Duke, and his subject was the use of relational databases in medical research. I was fascinated as he presented what they had accomplished at Duke, and predicted that in the not too distant future we would no longer need to do prospective double blind studies. We would be able to pose questions and then search data warehouses for the answers.  I remember that one of his examples was finding the best treatment for hypertension. In 1970 the famous VA Cooperative Study of Hypertension had been published by Fries and others. That study proved the benefit of treating men with a blood pressure of more than 140/90, so finding the best treatment was important. He imagined that rather than structuring studies like the VA study we could just ask the computer to review the records of all the men being treated across the country. Every doctor could prescribe what what he or she thought was best. If your new patient, say a sixty year old man with diabetes, was not responding to treatment, you could ask the computer what was working for other similar patients, and the computer could call up the results of similar patients to guide your management. Every doctor could use their best judgment, and as long as the computer could capture their choice, we would be advancing our knowledge of best practice by letting the computer collect the outcomes. 

 

Well, almost fifty years later we do use computers, and perhaps with Watson, or some other iteration of AI, we will soon make the prediction of that guest speaker at Friday Grand Rounds finally come true. The use of relational databases have already made a big difference in medicine, for example they have revealed drug/drug interactions that are now avoided. Databases have improved the management of chronic diseases and have facilitated politically controversial comparative effectiveness studies. Databases have been critical to the evolution of what we call population health. The speaker that Friday envisioned the use of computers to improve the care of the individual. I do not remember him talking about using his new tool to look at populations. In the early seventies we were not spending much time thinking about “populations” everything we did or were taught related to individuals. Doctors did not imagine that they had much reason to be concerned about the impact of the social environment on their patients. Germs were an external threat, but most of the other issues that occupied our thoughts were internal in nature. There was even controversy about whether or not there were agents like cigarette smoke that caused cancer. No one would have ever suggested that the education of children, the communities in which we live, the culture into which we were born, events that occurred a 150 years ago, or the “social capital” available to us, were determining factors in our current health, or might determine our life expectancy. 

 

We all exist in many different populations, and our nation gains its strength, as well as finds its most daunting challenges, in our diversity. We are simultaneously a part of a crowd like the one pictured in the header sitting in the end zone of a football game where our only health threats come from too many beers, too much sun, a tainted hot dog, or God forbid, the deranged action of some lunatic out to make a political statement or become an infamous mass murderer. Looking closer at any population with the tools that our powerful computers offer, that the man from Duke was offering as a future gift, we can now see that we are all parts of different populations that intersect. Indeed there is a new term, intersectionality. Here’s Mirriam-Webster’s definition:

 

 

My trip down memory lane this week to that lecture by the visiting professor from Duke, and on to musings about population heath and the complexities of populations, and how that evolving knowledge should impact how we think about the care we deliver was precipitated by listening to a fabulous conversation between Ezra Klein and Raj Chetty, a professor of economics at Harvard. In his introduction to the conversation, Klein caught my attention and stimulated my memory with his introduction:

 

Chetty is a Harvard economist who has been called “the most influential economist alive today.” He’s considered by his peers to be a shoo-in for the Nobel prize. He specializes in bringing massive amounts of data to bear on the question of social mobility: which communities have it, how they got it, and what we can learn from them.

What Chetty says in this conversation could power a decade of American social policy. It probably should.

 

I was hooked with:

….bringing massive amounts of data to bear on the question…. 

 

The conversation did reveal to me even more than Klein promised. Chetty is using massive amounts of data to do experiments in social policy that could have a tremendous impact on the social determinants of health. What he does for populations is what my Duke futurist predicted that we would someday do for individuals. A recent Vox article describes a study of social mobility that benefits poor children that Chetty has done that was a traditional double blind study, but the study concepts evolved from questioning large data sets over several years of study. I would suggest that you peruse a list of Chetty’s publications and ask yourself how these studies could impact policy development that might facilitate improving the health of the community, which is probably the most important leg of the Triple Aim. It is becoming increasingly clear to me that if we want to improve the health of every individual for a sustainable cost, the other two legs of the Triple Aim, we will fail until we direct more thought and attention to the social determinants of health, realizing that we all are a combination of multiple intersectional concerns. To work through that level of complexity to arrive at polices that can work efficiently for all of us we need the ability to capture and work with huge amounts of data. Chetty and his colleagues are giving us current examples of how to do that work. 

 

Chetty’s work is so important that The Atlantic has recently published a long article describing his impact on the evolution of our thinking, “The Economist Who Would Fix the American Dream.”  When you unpack his current work it goes like this: What we spend to improve the education of children, the communities where they live, and the social capital of their communities is returned many times over in their future health, the cost of care, and in a reduction in their need for social services. Chetty’s work identifies barriers that can be removed with cost effective measures that could gives us better outcomes for individuals while reducing the cost of care and improving the well being of everyone. 

 

Last year Chetty published data in JAMA on the declining life expectancy of subsets of our population that suffer from economic inequality. Wealthier men in America live 15 years longer than poorer men. Wealthier women live 10 years longer than poorer women. You can see slides demonstrating the findings by clicking on the last link. The origins of the differences are complex. The differences vary by  location. All men do better in New York than in Detroit. Chetty is quick to suggest that we hold our judgement on race and politics as the answer to these complex problems, but he does note that if we totally abolished deaths from cancer, we would improve life expectancy about four years while if we invest a fraction of what we spend on cancer, we would extend the life expectancy of everyone much more while moving everyone closer to the “American Dream.” It may be hard to believe that we could make a much bigger positive change than curing cancer if we had the focus and will to understand and correct the social issues that lead to the variation in life expectancy related to economic status and location. Chetty’s research gives us insight to the concerns people have about the future. He demonstrates that if you were born in the 40s or 50s you had a 90% chance of a better economic status than your parents. Children born in the 80s and 90s have only a 50% chance of a better life than their parents enjoyed.  That is not progress, and may be an important reason that so many Americans are angry as they feel their hopes and dreams are “Slip-slidin’ Away.” As Paul Simon sings near the end of his song:

 

God only knows
God makes his plan
The information’s unavailable
To the mortal man
We work our jobs
Collect our pay
Believe we’re gliding down the highway
When in fact we’re slip slidin’ away

 

Well, it turns out the information explaining why we are slip-slidin’ away is available, not only to God, but also to us through the work of social scientists like Chetty who are generating an understanding of large data sets that demonstrate that access to healthcare is insufficient to improve health. Their data can show us what else we might do to make life better for all of us. Local issues, economic issues, cultural issues, social mobility issues, and issues of race all have a tremendous impact on health and healthcare outcomes. The answers to what we might do to alleviate those disparities may flow from their work, if we take advantage of what is being discovered.

 

The Professor from Duke will eventually be right, we will improve the health of individuals with data from computers, but the old approaches to the care of the individual will not give us the insights we need to improve the health of the population, or the ability to lower the percentage of our collective economic assets that we spend on healthcare. To improve the health of the community and to substantially lower the cost of care while improving outcomes, we will need to manipulate and use the information and insights from very large data sets that social scientists like Raj Chetty are trying to offer us. He works like a physician. He does an exam and then formulates a plan of care making suggestions that could lead to improvement if we used his information to enable the writing of governmental and social policies that are like huge prescriptions that will improve the health of the nation. There is reason fo hope. Things could improve, if we have the will to fill the prescriptions and follow the treatment plans that the analysis of big data suggests.