Tommy

In September of 1953, my mother entered Muhlenberg Hospital in Plainfield, NJ to deliver my brother Tommy, her third son in four years. By all accounts, it was a normal delivery. Unfortunately, while in the hospital she was infected with the polio virus. (Vaccine deniers, take note.) She spent most of the following year in the hospital, and Tommy was raised by her brother’s sister, a taciturn emotionally bankrupt woman who rarely spoke and never smiled. As Tommy grew, he had a dark soul. He was often angry and in trouble. He was arrested for drinking in junior high and high school. He dropped out of college after one year and bounced from job to job before getting married and settling down to work intermittently as a cable man in North Carolina.

But Tommy’s demons continued to haunt him. He drank constantly and died in his early 50s of alcoholic liver disease. Unlike Tommy, his other three siblings have been relatively well-adjusted and successful in both their family lives and careers.

Did Tommy’s separation from our mother during the first, most influential year of his life result in his having lifelong mental illness? One anecdote does not constitute proof, but there is strong evidence that it could have. In a landmark article published in 1998 in the American Journal of Preventive Medicine, Vincent Felitti reported the results of the Adverse Childhood Experiences (ACE) Study. They surveyed more than 9500 adults about their history of experiences in seven categories: psychological, physical, or sexual abuse; violence against mother; living with household members who were substance abusers, mentally ill or suicidal; or ever imprisoned. More than half of those surveyed had at least one ACE, and 499 (6.2%) had four or more. Compared to adults with no ACEs, those with four or more were at 4- to 12-fold increased risks for alcoholism, drug abuse, depression, and suicide attempts. They also had 60% more diabetes, more than twice the risk of heart disease or stroke, and four times the risk of chronic lung disease.

What does this tell us about what is happening on the U.S.-Mexico border where children were being separated from their parents? Similar to the long-term psychological and physical health effects noted among Japanese families locked into internment camps during World War II, we can expect there to be downstream consequences for many years among the thousands of children recently separated from their detained parents.

What to do? I spent more than 25 years working in an inner-city emergency room, and when there is major trauma, the first step is to stop the bleeding, stop ongoing damage. We can only hope that President Trump’s recent executive order forbidding further separation of families will actually stop this practice and that currently separated families will be quickly reunited. However, locking children away with their families for months or longer into concentration camps while awaiting hearings may itself cause acute psychological trauma, further increasing the children’s likelihood of long-term poor physical and mental health. One key is rapid adjudication of their cases and settling families into stable situations either in the U.S. or their countries of origin.

But what about the damage that has already occurred among thousands of immigrant children? To me, the rule should be: You broke it, so now you own it. The U.S. government should take responsibility for treating these children now to lessen the effects of the trauma they have already experienced and establish a fund for longer-term treatment once they are settled, wherever they are settled. If we don’t do this, what does that say about the moral character of the American people and the government that represents them?

My brother Tommy got no treatment for the demons that haunted and tortured him. I can’t say with any confidence that it would have helped. But getting no treatment certainly did not help. We are at risk of creating an army of young Tommies through our misguided, even mean and purposefully hurtful policies. The long-term impact on them, their families, their communities, and on us will be substantial if we don’t stop the bleeding and start the healing now.

© 2018 by William M. Tierney, MD

All rights reserved

Chris’ Tears

As soon as I arrived for morning rounds, the resident pulled me aside and said, “You need to talk to Chris.”

“Why?”

“Mr. Simons died last night, and Chris took it pretty hard. He was crying.”

I was attending on the inpatient medicine service of our VA hospital and had come in for Saturday morning rounds. Chris was a medical student on the first clinical rotation of his junior year. When he started his rotation, one of the patients he picked up was Mr. Simons. At 6 feet, 5 inches, Mr. Simons had been a big, strong farmer. Even into his 80s, he had been working every day and reigning as the patriarch of a large, close-knit rural Indiana family until, over a period of several years, a series of events had turned him into a dependent invalid. Renal cancer required a nephrectomy. Recurrence in the other kidney required a second nephrectomy and chronic hemodialysis. A stroke left him hemiparetic and aphasic but cognitively intact. He became frustrated and depressed at his inability to communicate or care for himself, let alone lead his family. His world was crashing around him, slowly, inexorably.

Hailing from rural Indiana himself, Chris became engrossed in Mr. Simons’ care. He seemed to understand and resonate with the old man and his family. He fussed over everyday medical details before and after our rounds, and it was clear that the Simonses saw Chris as “The Doctor.” Although clearly not fully comfortable in that role, Chris accepted it and worked hard to maximize Mr. Simons’ care and minimize his discomfort.

Mr. Simons had extensive arteriosclerosis which resulted not only in his stroke but in multiple small heart attacks and intermittent, severe bouts of angina. On two occasions, episodes of chest pain resulted in his being transferred to intensive care where, for a few days, a different set of doctors and nurses would care for him. Although he always returned to our care after his ICU stays, such interruptions were upsetting to Mr. Simons and his family. One day, Chris came to me and asked, “Why do we keep doing this?”

“Doing what?”

“Moving Mr. Simons to the ICU when he gets chest pain.”

“To watch him closely, treat him more aggressively, hopefully prevent a heart attack, monitor for arrhythmias, etc.”

“To what end? He is going to die, sooner rather than later, and all we are doing is prolonging his suffering. Or even making it worse: it really upsets him when he goes into the ICU. They don’t treat him well.”

“How so?”

“They don’t spend time with him. They don’t try to understand him when he can’t get the words out. They do the work, but they don’t seem to care for him.”

“So what do you want to do?”

“Make him a DNR. Don’t minimize his care on the wards, but don’t sent him to the unit if he gets chest pain. Don’t resuscitate him if he codes.”

We agreed that we would treat his chest pain with morphine and other medications as needed, but we would not move him into intensive care, nor would we resuscitate him. Mr. Simons and his family also agreed with this plan and seemed relieved.

It was a couple of nights later when, while our team was on call, Mr. Simons died. Chris was one of the first there and was the primary communicant with the family. He brought the family into the room. He then stood quietly in the corner, shedding silent tears.

I spoke to Chris after rounds the next morning and asked him what had happened. He described the event and was satisfied that we had treated Mr. Simons well, and that his death was as peaceful as possible. He thought the family was satisfied with his care.

“How about you? Are you OK?”

“Yeah,” he said. Clearly embarrassed, he added, “I kinda got emotional, too.”

Later that day, I called Mr. Simons’ daughter, who had served as the main spokesperson for his large, extended family. I told her I was sorry he had passed and hoped she and the family were doing OK. Was there anything they wanted to know from us?

“You know,” she told me, “we are a very religious family. We believe that God takes an active hand in our lives. We believe that everything that happens to us is done by God, for a reason. I must say that for years, we had questioned why God was making Dad suffer so much. The cancers, the stroke, the heart attacks, his inability to care for himself, to even talk – God was doing this to Dad, and we couldn’t figure out why. At times, we were angry at God for making Dad suffer. But now we understand why God was doing this. He did it to help make Chris a better doctor.”

In the dark of night in a small room in a VA hospital in the heartland, a family gathered to say good-bye to the man who had been its anchor for more than 60 years. There was weeping, hugs, and whispers of comfort. And there was Chris, in the corner, whose tears had given meaning to a family’s years of cruel suffering and reaffirmed their faith in God and our healing profession.

[Originally published in the Annals of Internal Medicine, 2011. All rights reserved.]

Miles

I grin at the thought of how my habit of putting several miles of road beneath my feet every couple of days must appear to others, especially now as I pull on a pair of running shorts and a light jersey to face the first freezing weather of the winter season. Under a gray sky promising snow, I face into the wind as it whispers and growls across the Indiana farmland and turn on to a deserted country road.

In reality, I run in a cloud of memories of failures, of lost lives and abandoned hopes. As I run, I remember too well …

I remember Bradley, a young leukemic patient who had failed all standard and experimental chemotherapy, whose parents had lost their only other child to another malignancy, hating having her die in the confines of a sterile hospital, now only wanting their hopelessly ill son to die at home. I remember him in the hospital admitting room, the small purple lesions with black, necrotic centers on his extremities, his impossible fever, the obvious pseudomonal sepsis, his skin wet, pale, his eyes open too wide, the pupils too large, begging for release. His parents only wanted some relief, not a cure-we decided to give him a quick liter of saline, some narcotics, and whisked him home to die. Suddenly my heart feels heavy as I run, remembering the “stat” page to the admitting room, arriving there breathless, as I am now, seeing in the eyes of his parents that that death, too, had been stolen from them. Their kindness and thankfulness gall me still. And I run.

I remember Judy, the second-grade schoolteacher in her 20s, pretty and ill, cyanotic in the ravages of postinfluenzal staphylococcal pneumonia. Day after day, disaster after disaster, I fought for each organ as it failed: the septic shock, the renal failure, the gastrointestinal bleeding, the gradual decline into delirium, the cardiac arrest, the resuscitation, the return of vital signs. Then came the months, the cachexia, the contractures. I remember the accusations of her parents with less pain and bitterness than Bradley’s parents’ understanding and forgiveness.

Then Bob, a kindly 55-year-old man whom I met in the coronary care unit of the Veterans Administration hospital, his third visit there for angina in two weeks. He had suffered three myocardial infarctions in two years. He showed me his work: he would hollow out blocks of Lucite with a motorized tool, delicate and patient work, and then paint the inside of the block with a minute brush to make roses so real as to astonish, frozen forever in plastic.

As I run, now sweating despite the biting of the wind as it swirls over the fence rows, I remember his acceptance of his death, foreseeing it unafraid, as I admonished him not to give up, that there was always hope; and his indulgent smile, without words. I was speaking to him of inconsequentials when he died, falling suddenly back against the pillows, surprising me by the speed with which he turned that sickly blue-gray. I remember the resuscitation as each piece of equipment failed, as I failed, and the long walk down the short corridor to tell his wife of his death. She knew, and she accepted. As I whipped myself before her, expounding my failures, her face mirrored those of Bradley’s parents, showing kindness and forgiveness that hurt more than denigration.

I gather my dead about me as I run on this sunless November day across the Indiana farmland. I am running hard now, allowing the thoughts that all physicians cage deep within themselves to bubble to the surface and linger. I try to gain a hold on my place in all of this, and I find myself a part of the countryside, dead and barren, but within, beneath the frozen furrows, there lies a future, the next planting, the toil from which springs the next harvest. As from the windswept fields of November comes future plenty, so from the ashes of my memory, the losses and failures, come compassion for the suffering, passion for the living, and striving for what is right for my patient.

I head home now. The raw, wintry air roughens my throat as I run, but I run free. I have called up my shades once more and put them to rest. I am a man. I cannot place my talents above the lives of those whom I serve. I can only offer what I can, never enough, but all I have. I have neither beauty nor wisdom nor strength, only a heart that beats achingly against a frail chest, both on the road and in the passions of my art. Bradley, Judy, Bob, and all the others live on in me, making their way into my consciousness when they may, either on the back roads of Indiana or in the critical care waiting room. They are there, always there. Their lesson has been sobering, disquieting, and real. I listen when I can and run when I must.

[Originally printed in the Journal of the American Medical Association, March 1, 1985. All rights reserved.]

Patient names

All patient names in my blogs are fictitious. The narratives are based on real patients for whom I have cared, but I always change the names.

D5 Half-Normal

The ringing of the telephone in a dark on-call room always startled me, even far into my internship (after I had stopped sleeping with my shoes on).

“Hello?” I answered dully, my heart pounding.

“Dr. Tierney?”

“Yeah.”

“I have an IV that needs re-starting.”

“Who?”

“24B – Bridgewater.”

“The quad?”

“Yeah.”

My heart sank. Donny Bridgewater had a C6 transection that he’d acquired in an automobile accident. His arms were of little use except as IV sites for treating his innumerable urinary tract infections and bedsores. Every intern and resident on campus had treated him at some time or another.

“What’s he getting?”

“D5 half-normal.”

Hope! If he were only getting fluids, perhaps the IV could wait until morning!

“…and tobramycin.”

Groan! No way out. “Okay. I’m coming.” I kicked off my blanket, sat up, and pulled on my shoes. My roommate rolled over.

“Bilgewater?” [I always wondered if the cynical residents and students even recognized the reference to Huckleberry Finn.]

“Yeah.”

“Good luck.” He meant it.

The walk to the ward was punctuated by yawns. After picking up an IV tray at the nurse’s station, I walked to his room and turned on the grossly inadequate light, a single yellowed bulb at the head of his bed.

“Hi, Donny. I’m Dr. Tierney. I’m here to re-start your IV.” He glanced up, lanced me with an expression that was a mixture of disdain and indifference, then closed his eyes and ignored me. As I cleaned his atrophic, contracted arm, I noticed a wire contraption attached to his hand. After a few moments, I realized that it must hold a pen or pencil. A look to his bedside tray-table revealed the pencil and a sheet of paper covered with barely legible scrawl:

I was shaken. I could not reconcile his physical incapacity with the breadth of his vision and the depth of his feeling. Of the two of us, who was more handicapped? Whose mobility was more limited? Which one of us was mired in the mud? Whose numbness was greater? In a small, dark room amidst terror, hope, and pain, his angry spark of defiance and feeling had blinded me in my weary performance of routine tasks. I realized that I was a nameless, annoying, meaningless detail in his life.

Later, as I sat on my on-call bed and pulled off my shoes, my roommate rolled over again.

“Get it?”

“First stick.”

“Lucky.”

“Yeah.”

[Copyright © 2015 by William M. Tierney. All rights reserved.]

Running, Geology, and Academia

I was on Spring Break in Tucson, Arizona, running up Sabino Canyon. The canyon cut almost due north into the Catalina Mountains, its four mile access road climbing 1500 feet from the entrance of the park to the rock wall at the end of the canyon. The afternoon air was cooling and the sun was about a hand’s breadth from the horizon as I began my run up the canyon, the surroundings so foreign to this Hoosier flatlander.

I was younger then (45) and in decent shape. The first part was quite flat, and I settled into a comfortable pace. It had been an unusually wet winter, and the Spring colors of the Sonoran Desert wildflowers were shocking. The land I passed was humming with life. Literally. The bees were so plentiful that the entire valley had a background hum like some huge, restless underground machine. You could almost feel the ground vibrating with life.

With this backdrop, I thought about my professional life and how it, and I, had changed since my first faculty appointment in 1980. I used to go to professional meetings excited and self-conscious. I was jealous of anyone with a “presenter” ribbon on his badge. I remember sitting in small workshops and larger lectures given by the luminaries in my field. I soaked it all up. Over the years my self-consciousness has faded but has never left completely. The number of faces I meet at each meeting that I categorize as “friend” has grown steadily until it is difficult to traverse a hall at professional meetings in my fields without stopping to greet a friend.

I pondered my career’s progress as my course began to climb up the canyon. I was getting hot as I ran hard in the afternoon sun. But the valley narrowed, the stratified limestone walls closed in above me, and the road fell into shadow and cooled. Running comfortably again, I let my eyes trace the lines of sedimentary rock. The remains of tiny, delicate invertebrates had drifted down onto the floor of an ancient ocean for millennia and had been turned to stone by the slow, relentless, irresistible high-pressure dance of the continents. The stone had then been elevated more than a mile and squeezed into undulations not unlike tousled bedding. The weathering of the calcium carbonate rock fed the alluvium that anchored the stately saguaro. These enormous succulents, found nowhere else, grow and flourish in the harsh high desert where most plants whither.

Like these stately plants, my career had slowly advanced. I thought about the balance between the time I spent serving each of the academic missions: clinical care, teaching, and research. They are considered separate, but in reality they are each ways that academic clinician-teacher-researchers like me provide care to patients: directly as a provider and indirectly by educating the next generation of clinicians and researching ways of providing more efficient and effective care.

As the canyon narrowed further, the road began following the course of a lively stream, cold and swollen with the abundant snowmelt from high in the Catalinas. I marveled at the contrast of such noisy crystalline water passing so close to baking desert flora that would never drink it. I passed a youth group carrying bamboo rods, most of which were too short and too big around to be walking sticks. I wondered briefly about them before turning my attention to climbing the increasingly steep roadway. However, a minute or two later, I heard a haunting chorus of notes rising out of the darkening valley behind me. The bamboo rods had been the pipes of a mobile organ, with a person blowing into each different length rod louder and then softer, producing an eerie melody. I got gooseflesh despite my exertions in the afternoon desert heat.

I found a parallel to this song in my academic life in the sudden spark of an idea that leads to a successful project, the flush of excitement in a young colleague’s eyes when her work is recognized as important by someone else, the dawn of understanding in a student, the relaxing of an anxious patient. There was always something both new and familiar at each turn.

Tired physically yet mentally exhilarated, I reached the cul de sac at the end of the canyon road, but I did not stop. I continued another 500 feet up the canyon wall along a steep path that switched back and forth around boulders and patches of wildflowers. When I was as high as the remaining daylight and the terrain would allow me to go, I turned and faced the Sabino Valley and Tucson below it. Lit tangentially by the setting sun, the scene felt paradoxically idyllic and surreal. Having no one to share it with, I silently stood and panted from my exertions, not wanting to leave despite the increasingly chilly darkness and the protestations of innumerable bees.

Finally, I turned and bounced quickly down the path. I hardly remember the run back down the canyon road. I rarely looked at the road or the darkening rock walls, just at the ever widening view of the valley below, burnished in the orange twilight. But I ran hard and fast, sweating, wanting to get the most out of the 2000 feet of elevation that I had gained on my upward trek. As I ran, I felt the 15 years of my career behind me and the decades ahead, both existing simultaneously within that day as my life’s fulcrum. I was at peace with the world and what I had accomplished to date. Not satisfied, but accepting of the grace that my life’s opportunities had presented me. Like the darkening valley, my future played out before me, ever hopeful, pregnant with possibilities.

My Sabino run is etched in my mind. I’ve come far over the intervening 20 years, have learned much, and realize how little I understand, how little we all understand. A wise Pakistani professor of mine stated that our lives are a circle of knowing surrounded by an infinite plain of the unknown. The more we learn—the more we move from outside to inside the circle—the larger the circle becomes and the more contact it makes with what we don’t know. Our knowledge demonstrates our ignorance: the more we know, the more we know how truly ignorant we are. I have always found that thought to be both humbling and energizing, feeding my hunger for knowledge and my need to impact the world. My journey has been punctuated by moments of clarity like Sabino nestled among accomplishments and failures. There is no end in sight; there can be no end. The journey is the thing.

[Copyright © 2015 by William M. Tierney. All rights reserved.]

Miss Ellie’s Story

I had run into a hornet’s nest. It was my first day of attending on the inpatient general medicine wards of our inner-city county teaching hospital. Miss Ellie was my only ICU patient. She was 87 years old, an African-American woman who had sustained an acute intracranial hemorrhage that had rendered her irrevocably comatose. The hospital’s standard Brain Death Protocol had been applied, and she failed to show any responses to any stimuli. Miss Ellie was brain dead, but the ventilator and artificial nutrition and hydration were keeping her alive. Her body was strong, so this could go on for some time.

She had nine kids, seven of whom were in town and actively interacting with her providers. There was great disagreement among her family about whether she should be maintained on the ventilator or whether “nature should take its course.” The health care team felt we should withdraw life-sustaining interventions, but members of her family disagreed, at times vehemently. It seemed that those who lived the furthest away and had been the least engaged with her over the years were the most insistent that everything be done to keep Miss Ellie alive. They couldn’t allow us to let her die “of neglect” when the lay media was filled with stories of miraculous recoveries of patients who’d spent years in a persistent vegetative state.

The attending internist I had replaced was a long-term friend of mine who had never been comfortable withdrawing therapy or even discussing such issues. So he took the stance that as long as patients were alive, we should do everything to support them. I had inherited his ward service in the past, and each time I was met by the housestaff with a list of patients for whom they wanted to withdraw therapies. Realizing that our interventions frequently do more to patients than for them, I was always willing to entertain such discussions, painful though they may be.

This one was tough. The family members who frequented her bedside were unyielding in their individual opinions which tended more towards doing everything than limiting care. So I called a family conference and invited every family member who wanted to participate, regardless of where they were. The conference was delayed a couple of days to accommodate the arrival of a couple of her remote children. When it was finally held, more than 20 people crowded into the room: 15 family members (her children and a few grandchildren), her pastor, and our team of physicians and medical students. The family was all black, the health care team all white. Our differences were stark, and obvious.

I introduced our team and asked the family to do the same. I summarized her recent events as I understood them, asking for clarifications from our team and the family. Following the guidance of our hospital’s palliative care providers, I then made a specific suggestion that, because her coma was irreversible and she would regain no semblance of human interactions, we should disconnect her ventilator and let her die peacefully. I reassured them that our repeated tests had shown that her brain was so damaged that she would feel no pain or breathlessness. We were keeping her alive, but for all human intents and purposes, she was dead.

The family would have none of it. Several held out hope that our tests were wrong—no test was ever 100 percent right, was it? They hoped for a miracle. They wanted anything but to face her passing. But a few agreed we were only keeping her body alive, that there was no reason for this to continue. After 15 minutes of back and forth discussion, we were no closer to consensus. If anything, the family was more polarized than when we had begun the conference.

I then said, “This is my second day taking over for the other hospital doctor leading this team, and I don’t really know Miss Ellie. Can you tell me about her?” The gates opened. The family members smiled and then regaled me with the story of a remarkable woman. She had been raised in an incredibly poor share cropper family in Alabama. As a young woman, she had migrated north, started a successful business, raised and became the matriarch of a large and highly successful family. She founded her church and led it for decades as a deaconess and was a community leader in many other ways. Hers was truly a remarkable life. The stories and anecdotes flowed for many minutes.

As the stories of Miss Ellie’s life gradually abated, silence filled the room and slowly enveloped us all. Then, in a voice barely above a whisper, her eldest son said, “She wouldn’t want this.” Heads nodded, and with that, consensus was reached. By telling her story and embracing her life, her family realized that prolonging it was unnecessary and even selfish. She had attained her goals, and this was the end to her story, an acceptable end.

We arranged for her beautician to come in and cut and fix her hair. Her daughter applied makeup and put on her best Sunday dress. Gathering around her bed in the ICU, her family had pictures taken while she was still alive on the ventilator. The pastor held a brief service after which I disconnected the ventilator, disconnected her IV lines, tucked her in so that only the tip of the endotracheal tube was showing, and removed the medical equipment, leaving only Miss Ellie in bed surrounded by her family. I held her hand for 3 minutes and felt her pulse slow and then stop. She made no moves, no gasps for breath. She was at peace. I listened to her heart, heard no sounds, nodded, and backed away. No one moved, so I said, “It’s okay, you can touch her.” Slowly they moved in, embracing her and cherishing her passing. A few embraced me and other members of our health care team. The family then took more pictures amid tears and smiles and more stories as our team slowly, quietly exited the room.

Miss Ellie had a dignified, peaceful end because her family told her life’s story. The struggles and distrust within her family and between her family and our health care team melted in the warm fullness of her life. The coda of her song had been reached, and any prolongation was sacrilege. Her powerful story yielded the right end for her and for us.

[Copyright © 2015 by William M. Tierney. All rights reserved.]

The Boxer

I’m usually doing something trivial—riding my bike, mowing the grass—when a certain song from that late ‘60s comes onto my MP3 player, and it never fails to pull me into a state of reverie. I first heard the song in the Spring of my senior year of high school while driving to the YMCA to play basketball. I recall the day because I had not gone for months, and when I hit the court, Greg was there, alone, practicing his jump shot.

Greg looked at me and asked, “Have you been coming here a lot recently?”

“No, first time in months,” I answered.

“Weird! First time in months for me too.” I wondered how closely Greg’s and my minds were linked, what subconscious event had sent us both to the “Y” to shoot hoops on that same day. That day and that song remain linked in my memory. We both shot well that day.

Two years previously, at the beginning of my junior year of high school, I had moved from the New Jersey shore to a mid-sized industrial town in Indiana. I was a stranger in a strange land, a surfer banished to the cornfields, and had little hope of making friends until I met Greg who lived across the street. On my first day in Indiana, he was in his driveway, shooting baskets. He and I were very much alike: same size, same build, same smart-ass personality. I was a decent athlete, but coming from a state where baseball and football were king, I had no game. Greg took me to the local YMCA and taught me basketball. He asserted that basketball and jazz were the only two uniquely American art forms.

Greg taught me the fundamentals: the pick and roll, give and go, the turnaround jumper, blocking out, and always taking the open shot from the free throw line. Being new and having few friends and little else to do, I played a lot of “ball.” And I got good. Together, Greg and I played at the “Y” almost every week in both pickup games and leagues. We joined a team of guys from his old (rough) neighborhood and would play other neighborhood teams in games that at times resembled rugby more than basketball. We won more than we lost, and we loved it.

Greg and I were also close in school where we were both in honors classes in English, Math, and Physics. We often double dated. We went to high school basketball games together: a big deal in Indiana where, at that time in the 1960s, basketball was like religion (think “Hoosiers”). My high school gym seated 9100 people and was always packed with howling fans.

After high school, we both went to the same state university, but we parted ways. He lived in a dorm and studied business while I moved into a fraternity and was pre-med. We lost touch, both at school and even in the summertime when we both came home. We still lived across the street from each other, but I worked in factories and on construction crews while he worked on his family farm outside of town. We rarely saw each other.

Right before I started medical school, I got married and Greg came to my wedding. We spoke only briefly, awkwardly. We had little to say to each other. And then I didn’t see him again until my junior year of med school. I had begun a month-long rotation in Psychiatry at the state mental hospital, and as I walked through the cafeteria on the first day, Greg was sitting there, thin, disheveled, unwashed, and smoking. I learned that beginning in his last two years of college, he had progressively fallen into psychosis. He had begun to lose touch with the reality that he and I had shared, and no treatment seemed to work. Despite intensive inpatient care huge doses of antipsychotics, and even ECT, he was still nonfunctional, paranoid, and lost.

During our month together on the Psych Ward, Greg taught me about mental illness, especially schizophrenia. I was shocked to learn that he knew he was crazy. He knew that the things he was thinking and the voices he was hearing weren’t real. He just couldn’t ignore them. He couldn’t seem to find his way back to the reality he knew was out there somewhere, the reality he and I had once shared. He was lost, and neither he nor I could bring him back. God knows, I tried. I worked hard with him during that month, trying to fathom his illness and help him find his way out. I figured that nobody caring for him could understand him and help him as well as I. But I was ignorant and impotent, and I failed. Worse than failing, I represented the reality, the functional and successful existence, that was permanently closed to him. No matter how hard he tried, his illness bludgeoned him into submission, defeated and hopeless.

My month on Adult Psychiatry ended. On my last day, I wished Greg luck and left. I’d like to say that I thought about Greg a lot over the ensuing weeks, but this was my Junior Year of med school. It was intense. I was immersed in the anxieties and excitement of my work, and I had no extra cerebral energy to ponder Greg and his demons. Then, two years later, during my internship, I got a call from Greg’s former girlfriend. He had gone into his father’s barn, squirted a ring of gasoline around himself, and consumed himself in fire. Unfortunately, he didn’t die immediately. He lived two days after being helicoptered to the Burn Center in my hospital. He died there, alone, without family or friends. I didn’t even know he was in the same building in which I was putting in 18-hour days as an intern.

Since then, I have occasionally thought about Greg and the lessons he taught me, especially when working in my inner-city hospital’s emergency department where it was my responsibility to screen all psychotic patients to rule out a medical cause for their behavior before calling Psychiatry. Greg taught me that each of them had a past when they were not crazy, not screaming or spitting on me. Greg was always in the back of my mind when I worked with patients in the hospital or my primary care practice who were depressed, or anxious, or schizophrenic, and lost. He taught me patience and empathy. I try to recall Greg and his struggles when faced with a “difficult patient.” And his memory also comes to me unbidden when, mowing my lawn or riding my bike, that song from our carefree high school days takes me back to a day at the “Y” and basketball, jump shots, and the demons that killed Greg and motivate me.

[Copyright © 2015 by William M. Tierney. All rights reserved.]

Testimony to ONC on the JASON Report

On July 31, 2014 I was asked to testify at a “listening session” to ONC concerning the JASON Report, “A Robust Health Data Infrastructure.” I was asked to participate in their panel of researchers and respond to specific questions. I also made some generic comments about the JASON Report. I also submitted the following written testimony.

**************************************************

Testimony by William M. Tierney, MD, FACMI

President and CEO, Regenstrief Institute, Inc.

Associate Dean for Clinical Effectiveness Research, Indiana University School of Medicine

Ré: JASON Report, “A Robust Health Data Infrastructure”

 

Dear Members of ONC’s Health Information Policy Committee:

I am pleased to offer the following testimony concerning the JASON report that was commissioned by ONC and the Agency for Healthcare Research and Quality. I represent the Regenstrief Institute, Inc. as its current President and CEO. The Regenstrief Institute was established in 1969 and, at 45 years of age, is the oldest research institution in the country dedicated to improving health systems. Its Center for Biomedical Informatics has been a pioneer in building electronic health record (EHR) systems and health information exchange (HIE). I am also a member of the EHR 2020 Task Force of the American Medical Informatics Association. Although my testimony does not necessarily represent the opinions of AMIA or its EHR Task Force, I have discussed the JASON report with them and have included selected responses in my final testimony. That said, my testimony reflects my own opinions and do not necessarily reflect the opinions of the Regenstrief Institute, AMIA, or the Indiana University School of Medicine, for which I serve as Associate Dean for Clinical Effectiveness Research.

On the following pages, my comments are divided into three parts: (1) general comments, (2) responses to specific questions posed to the Research Panel, and (3) responses to additional questions.

General Comments

Overall, I found the JASON report to be surprisingly naïve, outdated, unrealistic, and not responding to its charge to recommend how to collate data from disparate health information systems for analysis. Instead, it recommends a radically different broad architecture for EHRs and their interchange of data in the U.S. that is impractical and unworkable at this time. Smaller increments using current systems and standards would have been more helpful.

Section 3.1 lists 15 problems with the current state of EHRs and HIEs in the U.S. I agree with all of them, and they are the main reasons why what JASON recommends can’t happen in the near future. They are all very thorny issues that are deeply embedded in our national fragmented health care (non)system that must be overcome before making radical changes to the EHR and HIE infrastructure. Overcoming them all right now in one fell swoop is simply not possible. Each needs substantial work, and overcoming them is outside of JASON’s charge. JASON should have provided a practical way of moving the needle forward now to enhance the creation of analyzable datasets of clinical data in the near term. The increasing number of EHR implementations are coming from a decreasing number of large vendors such as Epic and Cerner which creates opportunities for dialog and movement. JASON suggests that ONC and CMS can solve these problems by forcing the EHR industry to accept common EHR standards and APIs on which vendors can build their systems and communicate data between systems. That is simply not going to happen anytime soon, so this report as it stands provides little practical advice. And again, it doesn’t respond to JASON’s charge by ONC and AHRQ on pages 11 and 12, specifically:

• How can complex data handling techniques and Internet-based technologies be applied to health care to promote the development of real-time integrated datasets at a scale seen in other industries?
• How can the various users of health data in the clinical research and public health communities be presented with tailored and highly specific data views in near real time based on routinely collected health data?
• As health data grows from megabits to gigabits per individual, what fine-grained analytics should be made available to patients and health care providers to guide health care decisions?
• What fundamental data management capabilities are needed to support potential future requirements in an open-ended manner?

[The entire JASON report seems to respond to this one bullet item, but it does so by starting over with a new national EHR/HIE infrastructure.]

• What are the national security consequences of not addressing comprehensive health data opportunities in clinical research and public health?

Our experience with the Indiana Network for Patient Care (INPC) has shown that collating data from dozens of commercial EHRs can be done with today’s technology.[1]^,[2] It is difficult, to be sure—interfaces change constantly, standards evolve, and interpretations and implementations vary. But JASON could have made recommendations on how ONC and CMS might expand and support existing HIEs and encourage their use of existing technologies to capture discrete data for clinical care and research. JASON decided that the U.S. health information infrastructure needed a complete reboot with a new conceptual model. I don’t disagree that a more standardized and open model would provide robustness, but retooling hundreds of health systems’ EHR implementations just won’t happen in the near term.

JASON naively thinks that if HIT is using the standard architecture it suggests, the data will just come together:

• A new software architecture will make aggregated health care data available to all biomedical researchers, not just those who happen to work at a large academic center with strength in a particular specialty. The federated database will provide large effective sample sizes, both to support statistical significance and to identify statistical outliers.

Establishing a federation of data providers is hard work—it took the Mini-Sentinel initiative years to develop the organizational infrastructure to share data in a federated manner, and it is highly entropy-prone. The sociological and organizational barriers to maintaining such federations are much higher than the technical ones.

JASON also seems to make the assumption that notes and encoded data being entered into EHRs are accurate and worthy of extraction for secondary use, including research. Currently this is not a safe assumption: many notes containing history, physical exam findings, assessments, and plans are inaccurate because of copy/paste, template use, time pressure and weakness of EHR functionality. Encoded data are often entered by busy providers thinking about something else, or they are entered before complete information is available. For example, a recent study by Regenstrief investigators showed that the diagnosis of heart failure from two different health systems was proven wrong by imaging data (mainly from echocardiograms) 57% of the time.[3] Physicians entered heart failure as the most likely diagnosis for a set of non-specific symptoms and findings, and more often than not subsequent studies showed that the patient was suffering from something else, most often chronic obstructive pulmonary disease, which presents with shortness of breath, abnormal lung sounds, and a large heart shadow on chest xray, similar to heart failure.

Hence JASON’s assumption that problem lists, medication lists, pain scores, and other codified data would be highly valuable to a learning healthcare system may be inaccurate. Knowing what data are useful—separating the signal from the noise—is an ongoing, active area of developmental research within clinical database epidemiology community and critical to JASON’s charge. (For example, adding prescriptions for ACE inhibitors or beta-blockers to the diagnosis of heart failure lowers the false-positive rate significantly.) Yet it is not discussed.

Moreover, JASON spends a lot of time describing a radical approach to EHR and HIE architecture but little or no effort describing the functions of such architecture and their attendant policies, procedures, and expectations. Such functions include basic operations (entering and retrieving notes, orders, and other information) and enhancements to health care delivery, such as e-prescribing and clinical decision support. These things don’t happen and won’t if the architecture and its attendant policies and procedures don’t support them.

JASON makes two fundamental “findings”:

• The current lack of interoperability among data resources for EHRs is a major impediment to the unencumbered exchange of health information and the development of a robust health data Interoperability issues can be resolved only by establishing a comprehensive, transparent, and overarching software architecture for health information.
• The twin goals of improved health care and lowered health care costs will be realized only if health-related data can be explored and exploited in the public interest, for both clinical practice and biomedical research. That will require implementing technical solutions that both protect patient privacy and enable data integration across patients.

I disagree with the first finding: as described below, the Indiana Network for Patient Care and other hits have been able to acquire and merge data from a variety of existing health care information systems using existing mature health information standards for data coding and transmission. Other HIEs and research networks have also had success. Therefore linking health information systems and data to perform clinical operations or research is not impossible. It is simply extremely challenging and success has been limited to a few places. This more informed finding suggests that there exist workarounds and opportunities to incrementally improve how systems interoperate to achieve the goal of collating disparate data for analysis in a learning health system.

Today’s lack of interoperability between EHRs not a problem of lacking interoperability standards. It’s the lack of fully implementing them mainly due to the reticence of health systems and EHR vendors to share data. Lack of EHR data sharing is therefore more a problem of sociology than technology. JASON’s prescription for a standard architecture and APIs may work in countries with strong central ministries of health and without pre-existing electronic health record systems, such as Rwanda, Kenya, and the Philippines (where the open-source systems OpenMRS and OpenHIE are being implemented nationwide), but it can’t work in short order in the U.S., in my opinion.

I wholly agree with and support the second finding.

JASON states that “Interoperability is best achieved through the development of a comprehensive, open architecture.” I agree with this statement, but it is naïve to think that this will happen without dramatic changes in federal regulations and subsequent overhauls of existing EHRs and interface technologies, which would take 5-10 years. This report is “pie in the sky” and does not give practical advice in the current reality. JASON repeatedly states that merging data from disparate EHR systems will require APIs and semantic interoperability standards without any mention of the enormous challenges of establishing these latter, as recently outlined by D’Amore et al.[4]

JASON also asserts: “This lack of evidence [that EHRs and HIEs improve quality, safety, or efficiency of care] is partly attributable to slower-than-anticipated adoption rates of computerized HIT systems, especially among small health care organizations and individual providers [2,3].” These references are woefully out of date (2006 and 2009). ONC’s own website provides information showing that by 2012, 80% of non-federal hospitals[5] and half of physician practices in the U.S.[6] have either implemented a certified EHR or have contracted to do so. Furthermore, a recent survey of HIE initiatives[7] found that 61% more HIEs were actively exchanging clinical data in 2012 when compared to an earlier survey in 2010. And a very recent data brief from ONC reveals that the ability to electronically report laboratory results to public health departments grew dramatically from 8 states in 2005 to 48 states in 2014.[8]

“Even with the emerging evidence of benefit, implementing the useful exchange of EHRs across the entirety of the US presents an enormous challenge.” Agreed, so don’t try to do it all at once! Support large demonstration HIE projects like the Indiana Network for Patient Care^1,2 and New York’s State Health Information Exchange[9] to show what can be done and improve on current interoperability policies and procedures for health information exchange (HIE) and the organizational constructs and business models to support HIE along with the evidence that HIE works and is cost-effective. Then support other HIEs to implement these technologies.

JASON states that, “Evidence that the widespread use of EHRs and HIEs actually improves the quality, safety, or efficiency of health care in the US has been slow to accumulate.” It appears that the authors of the JASON report accepted this statement and did not attempt a relevant literature review. Yet in Indiana alone, use of INPC data has been shown to improve diabetes care and colorectal cancer screening,[10] increase HIV testing among emergency department patients,[11] lower the costs of emergency department care,[12] and enhance reporting of notifiable conditions to the State Department of Health.[13] Many similar studies from other HIEs have been reported in the peer-reviewed literature.

In summary, JASON fails to meet its original charge by ONC and AHRQ and instead proposes a radically different national EHR architecture. The latter is needed, but in the meantime there are important uses to which EHR data can be carefully applied. NIH recently created and generously funded the BD2K (Big Knowledge to Data) Initiative to develop tools, approaches, and methods to doing just that.[14] Perhaps the results if the BD2K Initiative will provide ONC with the guidance towards the collation and use of EHR data that JASON failed to provide.

Research Panel Questions

1. Could you address the tension in the JASON report around consumer control of data sharing versus unfiltered data?

In 2011, the Regenstrief Institute received an ONC Challenge Grant (with me as the Principal Investigator) to design a system that would (a) capture patients’ preferences for granular control over access to their EHRs and then (b) allow patients’ preferences to determine what INPC data would be displayed to different categories of users. The goal of this project was to work through the technical and organizational challenges of creating and implementing a patient-controlled data segmentation system. This project included a demonstration study where this data segmentation program was fully implemented in one primary care clinic affiliated with Eskenazi Health, Indianapolis’ urban public teaching health system affiliated with the Regenstrief Institute and the Indiana University School of Medicine. At the end of this six month demonstration study, we surveyed patients and providers opinions of data segmentation and its effects on patient care and communication. The comments below reflect what we learned as part of this three-year project for which a series of articles will be published in a special dedicated supplement to the Journal of General Internal Medicine in late 2014.

In our study, we tagged individual data items as being in selected categories of “highly sensitive” data as defined by the National Committee on Vital and Health Statistics.[15] Although meta-data tagging was not difficult for discrete data (with numeric or coded results), it was not possible for information embedded in free-text notes and reports, where a lot of patient information is recorded. For example, a primary care physician’s note could say, “The patient had an abortion at age 18 and post-partum depression after her first child at age 23.” This sentence contains two categories of sensitive data—mental health and reproductive health—and yet the note would be tagged as “primary care” based on who wrote it or where it was written. Granular control of EHR data will not be possible until natural language processing (NLP) can tag every important concept within a free-text note and then redact text based on patients preferences.[16] Currently, this is not possible, but we are working on it.[17]

Even discrete data can cause problems. For example, tricyclic antidepressant medications are often used for treating neuropathic pain and preventing headaches, making it difficult to know how to tag them. Similarly, some antipsychotic medications are also used to treat pain and nausea as well as schizophrenia. Until NLP matures and is widely implemented to extract data from free-text to clarify conditions for which drugs are prescribed or electronic prescribing systems force physicians to say what conditions they are treating with each prescription, classifying and controlling access to medication information will prove difficult. Yet JASON makes no mention of these difficulties in meta-data tagging. I believe this is an important area of development, and ONC should liberally support research and development of NLP as a core technology for EHRs and health information exchange.

Our research suggests that patients have widely different opinions of what data they would like visible to which type of health care provider and others who might have access to information in their EHRs.[18]^,[19] The JASON report overlooks patients’ desires to provide very different levels of access to different types of health care providers and non-provider users of EHR information. For example, Caine et al. found that virtually 100% of patients would allow full access to their EHR data for their primary care physicians while only 15% would provide full access to researchers.¹⁸

Moreover, our research showed that health care providers were highly concerned with the adverse effects on the quality of care and provider-patient relationships if patients were allowed to restrict access to their EHRs. [20] Yet the JASON report is silent on this issue. For example, on page 33, JASON states, “It is anticipated that different patients would opt for different levels of assumed risk associated with sharing their personal data, in return for different perceived benefits that may accrue from that sharing, both for themselves and for society.” This statement and the entire JASON report completely ignore the risks associated with patients hiding their health care data from their clinicians. Health care is an information-intensive business: most of the business of health care involves the collection, storage, reading, processing, and transmitting of information. Restricting provider access to patient information will undoubtedly pose risks to patient safety, resulting in errors of omission and commission, and could increase costs through duplicate testing and treatment.

Moreover, health care providers have a solemn responsibility to deliver the highest quality of care, often in the fact of life-threatening conditions. Would it be fair to expect them to do so without the full complement of patient data? Balanced solutions that would give patients granular control over access to data in their EHRs must weigh their interests in controlling EHR access and how their data are used (“nothing about me without me”) against their health care providers’ responsibility to delivery the highest quality, most efficient care. JASON completely ignores this issue.

In summary, allowing patients to have effective granular control over access to data in their EHRs—determining what information they would hide from whom—requires a level of NLP that is currently lacking. Hiding and redacting data from health care providers could have adverse effects on the quality, safety, costs, and outcomes of care and on provider-patient relationships.

2. What do you need in terms of unbiased participation?

Biased data yields biased results that cannot be relied upon to tell the truth about the topic being studied. Truly unbiased studies must have full access to all available data, and only unbiased results should be used for clinical decision-making.

3. How you feel about opt in or opt out option (for patients)?

The question about opting in and opting out can be asked of both data segmentation and access to data for research. I will discuss each of these separately.

In our ONC-funded data segmentation study, we took an opt-out approach. We did this for two reasons. First, our qualitative research with patients showed that patients expected health care providers to have access to their EHRs unless the patient desires otherwise. And in health care, providers have had such access throughout history. Hence, both patients and providers believe the status quo is opting out. Suddenly changing to an opt-in approach would be much more difficult to implement from both cultural and technical reasons. Moreover, the INPC uses an opt-out approach to capturing patient data into the HIE, and it has functioned well for 25 years.

As for opting in and out of using data for research, first we have to define research. Generally, there are two categories of investigations that use patients’ EHR data: little ‘r’ research (where analyses are performed, usually by provider organizations, to improve quality and safety and to manage practices more efficiently) and big ‘R’ research (where analyses attempt to generate new knowledge, usually by academic investigators). I believe health systems should have full access to patient data for quality improvement research unless patients opt out. Such studies are necessary and patients directly benefit. Academic researchers using identified data should be required to get consent (opting in). Researchers should be able to use de-identified data unless patients opt-out. In both cases, Institutional Review Board review and approval is required.

Let’s talk about de-identification. On page 50, JASON states that de-identification is impossible. That is patently false, and JASON provides no references. I have found no case were data completely lacking any of the 18 HIPAA personal health identifiers could be linked back to individual patients unless one has access to the health system’s data in the first place. All individuals who have such EHR access are governed by enforceable data use agreements. If they are going to breach patient confidentiality, they don’t need de-identified data to do it.

Because truly de-identified data as defined by HIPAA cannot be traced back to individual patients, patients are at very low (if any) risk of harm from their de-identified data being used for big ‘R’ research aimed at generating new knowledge. For patients whose data are being analyzed, such research could be of direct benefit to them. Hence, an opt-out approach to research using de-identified data is appropriate and will generate unfiltered, complete data since our experience has shown that very few Eskenazi Health patients choose to opt out of having their data used for research (and during each encounter with Eskenazi Health, patients are given the option of opting out).

4. What is your position on consent? JASON report argues that de-identification techniques are not robust for ensuring privacy so they advocate the need for informed consent. Federal research rules does include a form of consent via professional committees- IRBs Consent may cause relevant data to be filtered.

I discussed these issues above in response to Question #3.

5. What challenges and successes have you had to date collecting and utilizing data from EHRs and other health IT systems? Would a JASON like architecture help address the challenges you encountered? 

For more than 25 years, I and my colleagues have worked with data from the Indiana Network for Patient Care (INPC), one of the country’s oldest, largest, and most comprehensive health information exchanges.^1,2 The INPC uses a standard data transmission format (HL7) to download data in real time from health information subsystems (lab, pharmacy, registration) in 103 of Indiana’s approximately 120 hospitals into a single cloud-based data repository. The INPC also captures and merges identified patient data from health insurers, Indiana Medicaid, and the Indiana State Department of Health. Through a document delivery system that sends most of the laboratory and imaging reports in Indiana to more than 20,000 subscribing physicians,^2,[21] the INPC captures most laboratory test results in Indiana as numeric or coded results, when available, or text reports. As a result, since it was established in 1987, the INPC has captured more than 5 billion observations for more than 14 million individual patients, including more than 11 million in 2012 alone.

The INPC data use standard codes such as ICD9-CM, CPT-4, and LOINC for diagnoses, procedures, and diagnostic tests and clinical observations, respectively. A core of code mappers maintains the integrity of the coding scheme and deals with exceptions that don’t automatically match the INPC’s concept dictionary. The INPC also provides data to its 103 hospitals and affiliated outpatient practices via a standard messaging format, the continuity of care document (CCD).

The architecture JASON recommends would not help the INPC, as it has been able to maintain analyzable data across 103 institutions using the current interoperability and coding standards. In fact, moving to new an entirely new architecture as recommended by the JASON report would place a substantial burden on the INPC and all other existing HIEs and data sharing within multi-hospital health systems. The INPC’s developers have shown that current technologies and coding schemes are sufficient to deal with the vast majority of health information systems (and subsystems) and data types. Scrapping these for a different set of standards would create a lot of work with minimal benefit. Although it has been difficult to establish and maintain health information exchanges in the U.S., this has not been due to technical barriers but rather sociological, political, and economic barriers to the sharing of information between health care provider organizations. Certainly the current technologies can be improved: for example, Fast Healthcare Interoperability Resources (FHIR), an improved health care messaging standard.[22]

All of that being said, I have also had experience developing and implementing electronic health records in developing countries. I led a group of Kenyan and American informaticists that deployed sub-Saharan Africa’s first outpatient electronic health record system in 2001[23] which became the platform for a comprehensive HIV/AIDS information system[24] that, in turn, was transformed into OpenMRS.[25] Closely following the architecture outlined by JASON, OpenMRS is an open-source platform for EHR development that has been implemented in more than 40 countries, most in sub-Saharan Africa and Asia. We have shown the system to be interoperable transnationally in sub-Saharan Africa[26] where analyses have included data for tens of thousands of patients from dozens of HIV/AIDS clinics.[27] Because these countries had centrally controlled health care delivery systems and no pre-existing EHRs, they were able to establish a single open-source standard that could be implemented. In fact, five countries are currently in the process of implementing OpenMRS in all national health clinics: Rwanda,[28] Kenya, Tanzania, Mozambique, Bangladesh, and the Philippines.

So based on this experience, I agree with JASON that in countries other than the U.S. that have a centrally controlled national health system, especially where there are no pre-existing EHRs, an open-source platform with common interoperability APIs and open-source development can more easily yield interoperable electronic health information. I just don’t believe it is possible to force such an architecture at one time in the U.S. where we have no national health system and the majority of hospitals and physicians’ practices already have EHRs or have contracted to install them.^5,6 An incremental approach is warranted.

Responses to Additional Questions

• Given currently implemented information technology (IT) architectures and enterprises, what challenges will the industry face with respect to transitioning to a JASON like architecture? What challenges will your organization face? Do you see an evolutionary path for the industry to move from currently implemented approaches to a JASON like architecture?

Implementing the architecture recommended by JASON is not possible in the U.S. at this time or the near future, given the current status of vendor-driven EHR implementation. However, the path to realizing a JASON-like architecture will not be evolutionary—the country will not gravitate in that direction spontaneously, since the main financial commitment to EHRs is through commercial vendors. Moving in that direction will take political action by ONC, CMS, Congress, or some other governmental agency with the power to engage the health care and biomedical informatics communities and wrestle control of the EHR infrastructure from the vendors. Such political actors don’t (and won’t) make such fundamental changes quickly, especially in a sector of our economy (health care) that generates 18% of GDP.

My organizations (the Regenstrief Institute, the Indiana University School of Medicine, Eskenazi Health, and IU Health) don’t face the challenges of implementing a JASON-like architecture because we already have a highly functional HIE, the Indiana Network for Patient Care, that includes most of the citizens of Indiana, effectively uses established messaging and coding standards, has been serving our clinical, managerial, and research needs for decades.^1,2  Moreover, Regenstrief’s Global Health Informatics Program has already implemented a JASON-like architecture in the developing world and is using it to roll out national EHR and HIE systems in Rwanda, Kenya, Tanzania, Mozambique, Bangladesh, and the Philippines.

• What policy and technology developments would be necessary to assure the privacy and security of information in a JASON like architecture?

I don’t have the expertise to comment usefully on technologies to assure privacy and security of health information.

• What existing efforts (standards, initiatives, pilots etc.) in the marketplace are advancing a JASON like infrastructure?

OpenMRS²⁵ and OpenHIE[29] follow an organizational structure and function quite similar to what JASON recommends, with a standard open-source platform and APIs supporting EHR development by a community of developers. As shown in the map to the right, OpenMRS has been implemented in more than 40 countries and is being implemented in all Ministry of Health venues in five countries: Rwanda, Kenya, Tanzania, Mozambique, Bangladesh, and the Philippines. As mentioned above, this is possible because (a) these countries have a national health system that is centrally funded and governed, and (b) there are very few existing EHRs.

• A key recommendation of the JASON Report is that EHR vendors should be required to develop and publish APIs for medical records data, searching and indexing, semantic harmonization and vocabulary translation, and user interface applications. What existing  efforts are underway in health care that could inform the implementation of this recommendation?

There are several serious efforts to develop a standard platform and APIs for data sharing in the U.S. Most HIEs and RHIOs have done so (or are doing so). Mini-Sentinel[30] and PCORNet[31] are two similar large-scale, multi-institutional efforts. As mentioned above, OpenMRS²⁵ is becoming quite successful with this approach in the developing world.

• What standards, implementation specifications, certification criteria, and certification processes for electronic health record (EHR) technology and other HIT would be required to implement the JASON reports’ recommendation that ONC require open published APIs through Stage 3 of Meaningful Use?

I don’t have the expertise to make such recommendations.

• What processes and approaches would facilitate the rapid development and use of these standards, implementation specifications, certification criteria and certification processes?

To be frank, I don’t think these standards, specifications, and certification policies and processes will be developed rapidly. There are too many large EHR vendors who have their products in place in too many health care institutions to fundamentally change their model of software development and interfacing overnight. I think this requires a national dialogue, led by ONC and perhaps AMIA—the governmental agency in charge of EHR development and rollout and the world’s oldest academic society without significant conflicts of interest—that would include health care providers, EHR developers, HIEs, the research community, and patients. The latter should be represented by real patients and not advocates for a particular approach to health care information storage and management. This development process, guided by ONC and CMS, will take 5-10 years to complete.

• How might ONC and other Federal agencies best integrate the changes envisioned by the JASON report into their future work?

I believe that health information exchange is required for the interchange of health care information. I don’t believe an “N by N matrix” of every EHR being connected to, or available for searching to, every other EHR is a realistic model, at least at this point. If one believes that statewide or regional HIEs are a near-term solution, they will require financial support from ONC which, in turn, can then drive the conversation on evolving the U.S. EHR community to a standard platform, APIs, and interface technologies.

• What actions would you recommend ONC take to help the industry advance towards a JASON like architecture that supports interoperability for primary and secondary uses of health information?

1. Commit to health information exchange as the near-term solution for wide access to health care information. Generate a plan for financially supporting statewide or regional HIEs, and then implement it. Support should be based, to a significant extent, on the degree to which health systems and providers participate in an HIE and use HIE data for clinical care, care management, improving quality and safety, and research.
2. Support the use of existing standards and interoperability frameworks and work with their governing bodies to improve them to better serve the needs of EHR data analysts aiming to improve care and generate new knowledge.
3. Support developmental and implementation research on natural language processing (NLP) and other technologies for identifying and tagging EHR concepts and using the meta-data tags to enhance the usefulness of EHR data to clinicians, clinical managers, quality improvement and safety enhancement coordinators, and researchers.
4. Engage health systems, clinicians, biomedical informaticists, researchers, EHR vendors, and especially patients in a national dialogue on how best to balance Fair Information Practice principles and patient control over EHR access with providers’ need for EHR data to deliver safe, high-quality, cost-effective care.
5. Convene health system leaders, the vendor industry, academic biomedical informaticists, and patients to develop a set of high-level policies for HIEs and their capture, management, transmission, sharing, and analysis of health care information.
6. Convene technical experts to design interface technologies and detailed implementation schemes based on the policies developed in E above.
7. Once the groups in bullets 5 and 6 complete their work on HIEs, have them begin working on the next generation architecture of health information infrastructure for the U.S. At that point, the current JASON report, though even more out of date, will provide a decent starting point for the conversation.

 

REFERENCES

[1] Biondich PG, Grannis SJ. The Indiana network for patient care: an integrated clinical information system informed by over thirty years of experience. J Public Health Manag Pract 2004;Suppl:S81-6.

[2] McDonald CJ, Overhage JM, Barnes M, et al. The Indiana network for patient care: a working local health information infrastructure. An example of a working infrastructure collaboration that links data from five health systems and hundreds of millions of entries. Health Aff (Millwood) 2005, 24:1214–1220.

[3] Rosenman M, He J, Martin J, Nutakki K, Eckert G, Lane K, Gradus-Pizlo I, Hui SL. Database queries for hospitalizations for acute congestive heart failure: flexible methods and validation based on set theory. J Am Med Inform Assoc. 2014 Mar-Apr;21(2):345-52.

[4] D’Amore JD, Mandel JC, Kreda DA, Swain A, Koromia GA, Sundareswaran S, Alschuler L, Dolin RH, Mandl KD, Kohane IS, Ramoni RB. Are Meaningful Use Stage 2 certified EHRs ready for interoperability? Findings from the SMART C-CDA Collaborative. J Am Med Inform Assoc 2014 (in press).

[5] National Coordinator for Health Information Technology. Adoption of electronic health record systems among U.S. non-federal acute care hospitals: 2008-2012 (ONC Data Brief No. 9). Available at: http://www.healthit.gov/sites/default/files/oncdatabrief9final.pdf, accessed April 17, 2014.

[6] National Coordinator for Health Information Technology. Physician adoption of electronic health record technology to meet meaningful use objectives: 2009-2012. Available at: http://www.healthit.gov/sites/default/files/onc-data-brief-7-december-2012.pdf, accessed April 17, 2014.

[7] Adler-Milstein J, Bates DW, Jha AK. Operational health information exchanges show substantial growth, but long-term funding remains a concern. Health affairs (Project Hope). 2013;32(8):1486-92.

[8] Wu L, Abbey R, Daniel J, et al. ONC Issue Brief: Health IT for Public Health Reporting and Information Systems. Washington, DC: Office of the National Coordinator for Health Information Technology; 2014 [cited 2014 May 5]; Available from: http://www.healthit.gov/sites/default/files/phissuebrief04-24-14.pdf. Accessed August 1, 2014.

[9] Phillips AB, Wilson RV, Kaushal R, Merrill JA; HITEC investigators. Implementing health information exchange for public health reporting: a comparison of decision and risk management of three regional health information organizations in New York state. J Am Med Inform Assoc. 2014 Feb;21(e1):e173-7.

[10] Central Indiana Beacon Community Fact Sheet. Available at: http://mpcms.blob.core.windows.net/bd985247-f489-435f-a7b4-49df92ec868e/docs/1224405a-2518-4a36-bf9b-4fc16ffce10e/default_57_1135785195.pdf. Accessed July 30, 2014.

[11] Wilbur L, Huffman G, Lofton S, Finnell JT. The use of a computer reminder system in an emergency department universal HIV screening program. Ann Emerg Med. 2011 Jul;58(1 Suppl 1):S71-3.e1.

[12] Overhage JM, Dexter PR, Perkins SM, Cordell WH, McGoff J, McGrath R, McDonald CJ. A randomized, controlled trial of clinical information shared from another institution. Ann Emerg Med. 2002 Jan;39(1):14-23.

[13] Overhage JM, Grannis S, McDonald CJ. A comparison of the completeness and timeliness of automated electronic laboratory reporting and spontaneous reporting of notifiable conditions. Am J Public Health. 2008 Feb;98(2):344-50.

[14] National Institutes of Health. Big knowledge to data (BD2K). Available at: http://http://bd2k.nih.gov, accessed February 24, 2014.

[15] Carr JM. National Committee on Vital and Health Statistics Recommendations Regarding Sensitive Health Information. Available at: http://www.ncvhs.hhs.gov/101110lt.pdf. Accessed March 28, 2014.

[16] Leventhal JC, Cummins JA, Schwartz PH, Martin DK, Tierney WM. Patient control of provider access to their electronic health records: Technical and organizational challenges. J Gen Intern Med 2014 (in press).

[17] Were MC, Gorbachev S, Cadwallader J, Kesterson J, Li X, Overhage JM, Friedlin J. Natural language processing to extract follow-up provider information from hospital discharge summaries. AMIA Annu Symp Proc. 2010 Nov 13;2010:872-6.

[18] Caine K, Hanania R. Patients want granular privacy control over health information in electronic health records. J Am Inform Assoc 2013; 20:7-15.

[19] Caine K, Kohn S, Lawrence C, Hanania R, Tierney WM. Access, understanding, control and notification: Implications for the design of a patient-centered tool to allow individual choice in the disclosure of EHR data. J Gen Intern Med 2014 (in press).

[20] Tierney WM, Alpert SA, Byrket A, Caine K, Leventhal JC, Meslin EM, Schwartz PH. Provider responses to patients controlling access to their electronic health records: A prospective cohort study in primary care. J Gen Intern Med 2014 (in press).

[21] Barnes M. Lessons learned from the implementation of clinical messaging systems. AMIA Annu Symp Proc. 2007 Oct 11:36-40.

[22] FHIR. Available at http://www.hl7.org/implement/standards/fhir/. Accessed July 28, 2014.

[23] Rotich JK, Hannan TJ, Smith FE, Bii J, Odero WW, Vu N, Mamlin BW, Mamlin JJ, Einterz RM, Tierney WM. Installing and implementing a computer-based patient record system in sub-Saharan Africa: The Mosoriot Medical Record System. J Am Med Informat Assoc 2003; 10:293-303.

[24] Siika AM, Rotich JK, Simiyu CJ, Kigotho EM, Smith FE, Sidle JE, Wools-Kaloustian K, Kimaiyo SN, Nyandiko WN, Hannan TJ, Tierney WM. An electronic medical record system for ambulatory care of HIV-infected patients in Kenya. Int J Med Informat 2005; 74:345-355.

[25] OpenMRS. Available at http://www.openmrs.org/. Accessed July 28, 2014.

[26] Tierney WM, Achieng M, Baker E, Bell A, Biondich P, Braitstein P, Kayiwa D, Kimaiyo S, Mamlin B, McKown B, Musinguzi N, Nyandiko W, Rotich J, Sidle J, Siika A, Were M, Wolfe B, Wools-Kaloustian K, Yeung A, Yiannoutsos C, the Tanzania-Uganda OpenMRS Consortium. Experience implementing electronic health records in three East African countries. Stud Health Technol Inform 2010; 160:371-375).

[27] Ciaranello A, Lu Z, Ayaya S, Losina E, Musick B, Vreeman R, Freedberg KA, Abrams EJ, Dillabaugh L, Doherty K, Ssali J, Yiannoutsos CT, Wools-Kaloustian K. Incidence of World Health Organization stage 3 and 4 events, tuberculosis and mortality in untreated, HIV-infected children enrolling in care before 1 year of age: an IeDEA (International Epidemiologic Databases To Evaluate AIDS) east Africa regional analysis. Pediatr Infect Dis J. 2014 Jun;33(6):623-9.

[28] Allen C, Jazayeri D, Miranda J, Biondich PG, Mamlin BW, Wolfe BA, Seebregts C, Lesh N, Tierney WM, Fraser HS. Experience in implementing the OpenMRS medical record system to support HIV treatment in Rwanda. Stud Health Technol Inform. 2007;129(Pt 1):382-6.

[29] OpenHIE. Available at: www.openhie.org. Accessed July 31, 2014.

[30] McClure DL, Raebel MA, Yih WK, Shoaibi A, Mullersman JE, Anderson-Smits C, Glanz JM. Mini-Sentinel methods: framework for assessment of positive results from signal refinement. Pharmacoepidemiol Drug Saf. 2014 Jan;23(1):3-8.

[31] Fleurence RL, Curtis LH, Califf RM, Platt R, Selby JV, Brown JS. Launching PCORnet, a national patient-centered clinical research network. J Am Med Inform Assoc. 2014 Jul-Aug;21(4):578-82.

Problems with the Problem-Oriented Medical Record

The problem-oriented medical record got us started breaking down the documentation of complex clinical encounters into subjective and objective observations, assessment (processing of the data), and then plans (acting on what the data told us). This was focused on individual problems. Unfortunately, as our population has aged and we’ve become very good at preventing death from things like myocardial infarctions, strokes, cancer, and infections, patients live long enough to accumulate multiple chronic problems and acute episodes that often have conflated and conflicting data and care guidelines. Breaking this down into individual problems (which I still do in my own notes and in my own mind when caring for patients) is artificial and perhaps contributing to errors by enhancing confusion and over-simplification.

What we need, IMO, is a patient-oriented medical record that has input from each of the team members, including the patient and the patient’s family and personal caregivers (what we called “subjective” data in the past, denigrating such information as not as reliable as “objective” data that we collect with our senses and tests) and all clinical and para-clinical team members. We still need an assessment by someone in charge of the patient’s overall care, which will usually be a physician. But the assessment has to be on the holistic patient and not just on his or her problems in isolation. And the EMR (especially documentation) needs to support this approach. The electronic information in the patient-oriented EHR can be reformatted to fit decision-making schemes for different problems, care venues, and needs (direct clinical care, communication between providers, billing, quality improvement, disability assessment, etc.). That is, we need to separate the medical record and documentation from the various notes, letters, and reports into which the data are collated from the various report formats through which those data are communicated.

In the paper-based record, most information was in the visit note, and the data entered therein could generally be used only be used once—for direct clinical care. (Thank God for copiers and fax machines.) The EMR of the future will be a dynamic record with inputs from patients, home caregivers, and each contact with health systems and potentially any other data source that might be useful (think of data from public health entities, the EPA, or the criminal justice system).

We simply gotta stop trying to recreate the paper-based, problem-oriented medical record in electronic format, IMHO.