“Why Not New Delhi? Is the Local Pathologist Dispensable?” That’s what Michael Laposata, MD, PhD, asked when he spoke last year at the Advancing Practice, Instruction, and Innovation Through Informatics Conference, or APIII. Here, in the form of an edited transcript, is Dr. Laposata’s answer.
Dr. Laposata is pathologist-in-chief, Vanderbilt University Hospital, and professor of pathology and medicine, Vanderbilt University School of Medicine. Before that, he spent 18 years at Massachusetts General Hospital, 17 of them as director of the clinical laboratory.
I was invited in late 2007 to give the keynote address at the G-2 Reports symposium on the business of pathology. I attended the presentation before my own talk with no particular reason to be there except to be early. It was delivered by a pathologist from India. As I listened to the first few slides I thought this had to be the 10th or 11th iteration of a story of someone who wants to back a truck up to my hospital, take everything out of it, and make all of us unemployed. So I listened.
Then I realized that the story is different now, and the biggest reason is all the work you have been doing in information technology. Other barriers that had been around for a long time seemed to be disappearing too.
So at that time I went back to Massachusetts General—this was in December 2007—and I gave this talk in a journal club format. The presentation was to tell our pathologists and residents about this fellow who has a lab in India and who wants all of our specimens. Of course, after I gave the presentation everybody said this could never happen.
Finally, John Gilbertson [MD, associate chief, Department of Pathology, director of pathology informatics, MGH] said, ‘Would you all stop being so defensive? This is real and it’s coming like a big truck.’
There is this disconnect between the phenomenal technology presented at this conference in Pittsburgh and seven or eight pathologists in a group somewhere who say: ‘No way. We are still going to be reading out this glass and none of that stuff is going to happen to me.’
What’s been going on? First, we began long ago to learn how to transmit data electronically. The second is improved images. Pathologists can see the best image for making a diagnosis, but they’re still going to be convinced that the image isn’t as good as what they see through the microscope. And third is the expansion of health care networks that connect providers in remote areas with experts. The fact is we all need to be connected to transfer our expertise.
While all this technology has been developing, what has gotten worse? The cost of health care. Who would have thought we would have a lab test that’s $5,000? The Athena ataxia panel—$5,000 for one test. So now if you’re off balance you can be evaluated clinically, and you can find out which of the many different kinds of ataxia you have. And you can also now get a genetic test to at least rule out five or six inherited forms of ataxia. Oncotype DX for breast cancer costs $3,300 per test. And radiology doesn’t do much of anything for less than $1,000.
And what else has happened? Data appeared that said if you’re awake for 36 hours as an intern or a resident, you deliver substandard care. The data made it obvious even to those running the residency programs that if you kept everybody awake that long, they’d start saying ‘40 milligrams, 20 milligrams, just pick one.’ The long-held argument for the concept of continuity of care—‘I’m following my diabetic for 36 straight hours; I know exactly what’s going on’—was now less important than having someone fresh come in, to ask, ‘What’s been going on with this person?’ Fewer mistakes.
People have realized over time that practicing primary care medicine means you never get off the phone with your patients, and that in some specialties as well, you don’t control your life. And there are other specialties where control of your life is possible. Remember when anesthesiology couldn’t fill?
For those specialties where you don’t have as much acute care, you can actually get away. Your weekend is yours. A lot of people are going into these specialties, and we don’t have as many physicians at the front.
So what’s the answer? Well, the person who came to the G-2 Reports event and told us about the lab he had in India said, ‘I’ll tell you what the answer is: A pathology service that is available 24/7. And we’re going to address all the barriers that you previously told us were problems.’
No. 1: trained in U.S.-accredited programs. I remember looking at a list of pathologists who were signing out anatomic pathology for a commercial laboratory in the U.S. When I looked at that list, now 10 years ago, there was almost no one on it who was trained in the United States. It was a problem. The G-2 speaker said, ‘Here’s my list.’ The U.S. was training many of their residents. They were of Indian extraction; they came, they learned, and they went home. They’re working for him.
No. 2: English as a first language. He said, ‘We speak better English in India than you do in America.’ Probably true. No. 3: ‘Have a willing attitude for consultation on test results.’ There’s an important one. Physicians call us up and we need to say, ‘Sure, I’d be happy to tell you why your patient is hypercoagulable.’ Or ‘Sure, let’s discuss the biopsy in the special stain’ as opposed to ‘Didn’t you read my report?’
Recently I had a conversation with a friend at the University of Alabama at Birmingham—a physician in a solo endocrinology practice. So how does he get his patient dictations typed? For years he hired someone in Birmingham, gave the tapes to that person, and three days later he would receive the typed reports.
One day this woman told him, ‘I’m taking two weeks off, so I don’t know what you’re going to do about your reports.’ He just Googled ‘transcription’ and something popped up that said ‘We’ll do your transcription.’ He records on a digital tape now and the person who types it is in India and is awake when he is sleeping.
The reports arrive the next morning, and they arrive every day. The spelling is perfect. There are no problems, and guess what? It’s less than half of what he was paying the person in Birmingham. An improvement in service for less money. Those two things usually don’t go together.
Is it possible that could be done for pathology? If the people in India are going to process and test your specimen, or sign it out, yes, it’s going to cost 40 to 60 percent less. And the G-2 speaker’s argument is that they can do it just as well. Not only that, we’re flying CAP inspectors to India. So anything about their lab not being as good, less quality control, all those types of arguments—gone. The CAP has accredited many labs in India [26 as of February 2009] that have undergone a full CAP inspection. He said, ‘We take one more step: We bring in the European Union inspectors too.’
What’s the downside? There must be some reason why it would be a bad idea to send non-acute specimens to India. All that skin? All those GI biopsies? He has this slide of a global map that shows the direct flights to India. We’re not going through Germany from the U.S. now, so we’re not going to lose specimens in Germany as they travel to India. Direct flights between the U.S. and India—more than 15 each day.
What’s the difference in time? It takes about 10 extra hours to fly there compared with flying samples across the U.S. If I were in Boston and using a lab in California, that’s six hours. I send CP and AP samples to California and don’t think anything of it. Why not add 10 more hours in the airplane? It’s only going to make one stop at the end, and as he shows, there’s a truck pulling up to the airport in New Delhi. It’s got the name of his lab on it. They pull this container with samples off the airplane and they’re taking it to their laboratory.
Maybe there is another down-side—no consultation. David Hicks, MD, who is the head of anatomic pathology at the University of Rochester, gave a case example at CAP ’08 on what it means to be the indispensable pathologist. It’s the story of a woman who has breast cancer and what his role was in the case. He goes to the operating room and talks to the surgeon. The surgeon says, ‘How about this and this, Dave?’ Dr. Hicks is totally involved in decision-making. They’re talking about special studies that should be done, the merits of linking molecular and anatomic pathology. How about that for a consultation? The surgeon says, ‘Wow, that was phenomenal. It was completely indispensable.’ The question for the audience at CAP ’08 was, how many anatomic pathologists are doing this? Very few—less than 10 percent. How many anatomic pathologists in the U.S. are going to spend this amount of time?
What are the drivers to not interacting as effectively in the U.S.? One driver is you make more money by signing out more cases. The other is you may be stepping out of your comfort zone. Dr. Hicks knows a lot about Oncotype DX and all the other molecular pathology that goes with breast cancer, and he can confidently have this discussion. If you haven’t gotten there yet as an anatomic pathologist, you’re feeling a little uncomfortable, so you only want to talk about what you know. As a result, more and more pathologists are saying, ‘Well, it’s in my report. Just read my report.’ If that’s the case, then the doctor has more accessibility to somebody in New Delhi, which is possible with all of this information technology.
Most pathologists would like to think they’re indispensable. Are we? For clinical pathology tests, we are still throwing test results over the wall: To every clinician out there we say, ‘I hope you know what 40 percent protein S means for your patient.’ Since I’ve been at Vanderbilt we have a case conference every Monday morning, two cases by the residents. I’d say 80 percent of them involve a molecular test. Acute myelocytic leukemia—if this mutation is present or that one isn’t, or maybe they’re both present, it’s affecting the chemotherapy choices.
Personalized medicine is looking at the molecular findings. We have the anatomic pathology findings, too, but often whether or not we’re going to treat with this agent or that agent depends on a molecular finding—and it makes a big difference for the patient. You don’t want to spend $20,000 on a chemotherapeutic agent if you can do a genetic test to find out that that person is not suited for that chemotherapy.
At Vanderbilt Ingram Cancer Center here, we’re having discussions about taking all the tumors and understanding the different mutations that might be associated with those tumors. We’re going to try to figure out which ones affect therapy, which ones affect diagnosis and prognosis, and we’re going to try to figure out how to make it one-stop diagnostic shopping for the oncologist.
We can’t do the Oncotype DX in our lab because it’s a proprietary test. We have to send it out, but we hope to have a panel of genetic tests for each tumor for either in-house or send-out testing. So the first and the simple answer is, ‘Yes, that’s prostate cancer,’ and then after that, there are many additional diagnostic pieces of information you have to factor into this, much of it molecular.
Is the doctor confused on the other end? Absolutely. Are we the ones who have the information? Yes! It’s great that we pathologists are the focus, but we also have to know what we’re talking about, because if what comes from us doesn’t make sense and does not solve the problem, we’re not useful.
Doctors want more than a short report. They need their answer. It’s costing money. Their patient is in the hospital for an extra day. They need to figure out what’s going on now. If all they get from the pathologist is a short written report, they could get it much less expensively outside their institutions.
When it comes to managing the laboratory, if that’s all the clinical pathologists are supposed to do, there are just way too many non-pathologists who can do that for you. There is no reason to pay a clinical pathologist $200,000 to run a clinical lab, simply to hire the people, buy the reagents, and pick the instruments. There is a $100,000 administrator who would be delighted to do that job, and there are several programs already active and probably 10 more in the offing to create a new doctoral degree called the doctorate in clinical laboratory sciences, or DCLS. These graduates are going to be like the doctors of pharmacy, or PharmDs. They are going to have the operations of the lab down because they’re going to get their bachelor of science degree in medical technology and continue through the system to the doctoral level, with all of their training involving the clinical laboratory. Who would you hire to run your laboratory operation if not someone like that?
Let’s look at history. The first time I saw clinical pathology was where I grew up—Johnstown, Pa.—in a lab at a 200-bed community hospital where they had 1.5 pathologists. This principal pathologist had me in as an observer while I was a college student. At that time they had one instrument in the clinical lab—the SMA analyzer. It took a huge sample, 2 mLs of serum, and put it through tubing. The way you separated the sample from the next one was a bubble. Talk about carryover!
So here is a 2-mL sample flowing through the tube, and after a few reagents are added in the line it goes through a spectrophotometer that reads it. The page for the report is 8½ by 11 inches, and there’s a red line that goes from left to right across this page, and there’s a gray zone marked on the page that is the reference range. If the red line goes through that gray zone for a given lab test, it’s normal. The first time I looked at it I said, ‘What are the gray zones?’ He said, ‘Those represent normal values.’ Then I did the interpretations. I was a college student. I said, ‘Oh, look at this: This guy has a low calcium. This guy has a high bilirubin.’ The pathologist signed them all, and I said, ‘What are you doing?’ He said, ‘I’m interpreting the results.’
So when clinical pathology was the cash cow, we were getting paid for everything. We got that consultative fee for ‘interpreting’ the iron results. We didn’t quite ‘interpret’ the iron result, but we did set the test up. By the 1980s the government said, ‘Come on, you’re just signing your names. Forget it.’
If you were a resident after that time you said, ‘Well, let’s see, I could do clinical pathology and make $1,000 a year, or I could do anatomic pathology and earn a real living.’ Justifiably, residents started learning anatomic pathology almost exclusively; it really tipped the balance for what people knew. So we lost a whole area of expertise within our discipline in clinical laboratory testing.
Then by the 1990s, the government and other payers said, ‘All right, it’s kind of complicated. We’ll pay you for interpreting about 18 different things and we’ll assign codes for them.’ What about anatomic pathology by this time? Between the ’70s and ’90s more test options appeared. This became a pretty good living.
Then in the 1990s off-site pathologists became available to read cases at lower costs. The gastroenterologists had an idea: ‘If we’re doing all the scoping,’ they asked, ‘why don’t we hire our own pathologists? They could be part of our practice. We’ll make it lucrative for him or her.’ Thus, you could avoid sending your samples to the hospital pathology department. Then the question came: If we could send it to the pod lab down the street, maybe we could send it anywhere. That really has brought this issue to a head, and now we realize we have highly developed communications, which you at this meeting made possible with impressive information technology.
We have to face the facts about who could do what we do. In clinical pathology, one is the lab administrator who can tell you more about the budget of the laboratory than most pathologists know. The PhD chemists and microbiologists know all the people in industry and can buy all those instruments. They know what they’re doing. The doctors of clinical laboratory sciences are going to come in a wave, and the first group to graduate is all of about four years away. On the AP side, now we’re talking about pathologists in pod labs and foreign countries. So is there a real need for the consultation? Are we pushing it too much? Is it enough to be able to say ‘40 percent protein S,’ or say something about the breast cancer in only a sentence or two, and say that’s my report?
Clearly we provide more accurate diagnoses when we have a conversation. We did a survey at Mass General where we had for the past 10 years interpreted the special coagulation tests, those beyond the PT and PTT. I did that six months each year. We generated a patient-specific narrative—not a cookie-cutter narrative. We read the reports, the discharge summary, all of the radiology, and we generated a narrative that said this is the story about why your patient had a blood clot or is bleeding.
With those data we went back to the clinicians and asked: If we had not given you the narrative, in how many of these might you have made a misdiagnosis? The answer was about three-quarters. It was stunning. Three-quarters of the cases would have had a misdiagnosis if we hadn’t provided the narrative interpretation [Arch Pathol Lab Med. 2004;128:1424–1427].
What about length of stay? Every time you enter a discussion with a hospital administrator about how to save money, length of stay is always an issue. At Mass General we used to get a monthly report about our length of stay. When it got to 6.3 days there was panic. It was like the stocks. What is going on—6.3 days for length of stay? If we don’t get this down to under six, they said, we’re going to lose millions this year.
When 5.8 was reached—think about what that means: half a day, about eight hours—it was a savings of millions of dollars. So increasing the length of stay is expensive. If you’re going to stay for one day in the Mass General Hospital or the Vanderbilt University Hospital, you’re spending about $1,000 to $1,500, and you’re not getting another patient in the door.
When we first started doing our narrative reports, we started with one of the most complicated things—coagulation. Many doctors see it as a big black box. Factor VIII, factor IX, what are they?
Recently, there have been a number of papers in the literature from across the globe—one from Canada, another from the UK. They started surveying physicians to see what lab tests they were comfortable interpreting, what they really know. Turns out we’re down to routine tests like liver function tests. ALP, AST, bilirubin. You want a pathologist to interpret those? You want a pathologist to tell you when you need a hepatitis test? The answer is yes. That was from the recent study in the UK [Ann Clin Biochem. 2008;45:33–38].
Another study was summarized in the “Dark Daily” [from the publisher of the Dark Report]. It was from Canada and it says ‘Canadian laboratory system in shambles.’ The reason? They are ordering the wrong tests and they’re more expensive now. It’s a socialized system, so there are only so many dollars. If you order the wrong test, the system is out of money and cannot do the right tests.
The clinical and financial importance of picking the right tests and being able to interpret the test results correctly is huge. As pathologists, we have grossly overestimated, especially in clinical pathology, what our colleagues know on the other end. We’ve always assumed they cannot look through the microscope at a slide and determine if something is benign or malignant. But for the past 30 or 40 years we have deluded ourselves into thinking they know a lot about laboratory tests, and only in the past few years have we learned many want help even with liver function tests.
I had an interesting discussion with a good friend at Mass General who has been a pathologist there for some time. As I was leaving MGH in the early part of 2008, she said, ‘There are a lot of molecular tests out there, but really it’s far away. I’ve talked a few times about this and nobody is thinking about it.’ I think the opposite—that it’s all out there now, and it’s moving very fast. The indispensable pathologist knows that digging deeper is important, and he or she is going to have all the necessary conversations to make the right things happen.
We have to take a lesson from our friends in radiology. There’s a story written in the New England Journal about a radiologist from Altoona, Pa., at a 200-bed hospital. He’s the only radiologist, and he did not want to be awakened at 2 AM for a patient who had a car accident, to look at an MRI to see if the patient had a subdural hematoma.
So they said, ‘Well, what are we going to do?’ The radiologist said, ‘I don’t know.’ Guess what they found out? They can digitize the image and send it to India. So somebody in India read it and said, ‘Yeah, there’s a big subdural hematoma.’ The radiologist in Altoona comes to work the next day and says, ‘I’m glad you got that MRI read.’ He starts reading the images and the hospital administrator is there and says, ‘Are you doing any better than that person in India?’ He said, ‘Of course I am.’ She said, ‘Why is radiology better read by you at 2 PM and not better by you at 2 AM?’ Ah, now we have a problem. Saying that I’m the better radiologist at 2 in the afternoon, but at 2 in the morning anyone could do it just as well is a hard argument to make. And we have to ask ourselves the same question. This requires a bit of honesty.
Let’s take an AP case example. A biopsy of skin is performed. It can be sent to a U.S. pathologist the way it’s always done. Maybe that’s to save time, to permit in-house dialogue about the case. ‘I’ll see you in the doctor’s lounge and we’ll talk about the sample I sent you.’ Maybe.
Or the biopsy could be processed in the U.S. and the image could be digitized and sent to a pathologist in India for interpretation. Or the whole piece of tissue could be put on a plane and sent to India. Someone [at this conference] talked about tissue processors on airplanes. What a cool idea! By the time the tissue gets there, it’s all fixed.
Let’s take clinical pathology. A 200-bed community hospital has two pathologists who spend 90 percent of their effort in anatomic pathology and interact infrequently with their clinical colleagues. A 45-year-old patient on warfarin has a massive pulmonary embolism and almost dies. The clinician caring for the patient says: ‘Wow, does this guy have a hypercoagulable state? What do we do? Do we modify our treatment regimen for anticoagulation? How risky is this? Do we need to find out about a similar risk for his kids?’
You could do the test in the main hospital laboratory or you could send it to a commercial laboratory in the U.S., and for the latter you’re paying top dollar. Why not put it on an airplane to India? You’re going to package it up anyway, right? So the only difference is the address you write on the box.
This is what this is about—are you going to send this to California or are you going to send it to India? Either way it’s going to get onto a plane. Maybe the person from India is going to talk to the doctor who sent the sample from the U.S. about what that factor V Leiden means. Oh, and by the way, you’re going to save 40 percent on your lab tests. Let’s say that we’re spending $25 million on lab tests. We’re now going to save 40 percent of that. That’s a $10 million savings with no loss in productivity or clinically meaningful turnaround time.
Say you’re the hospital administrator. How does a hospital administrator get to be a vice president or chief operating officer? The administrator saves money, and the trick is to save money without incurring a new cost on the patient care end. Well, here’s an opportunity to save money and have all the clinicians say, ‘Yeah, I didn’t notice any difference. In fact, I actually found there were advantages. Let me tell you what they are.’
What would be the downside of sending something that has an acceptable turnaround time of two to four days to India, because after all that’s how long it might take in the teaching hospitals with processing in the U.S., or maybe in transit? Especially if you’re going to save millions and the consultative support is the same? It’s pretty hard for the administrator to find the downside.
Let’s think about administrators again for a moment. Who do they report to? Who fills out their evaluations? Who decides if they’re going to get an extra $25,000? It’s usually not the pathologist. It’s usually the chief operating officer or the hospital president. If that’s the case, then it’s in the best interest of the lab administrator to realize the savings and figure out the consequences for his or her hospital physicians later.
Today the pathologist in my estimation must be highly consultative. You read out the anatomic pathology, you see a set of test results, and your job is to talk to people and make sure the patient gets a quick and accurate diagnosis.
Does the lab administrator have to know what we do and value it? Yes, that’s part of the overlap. The lab administrator has to figure out the budgets and hire people and all those other things, and we have to understand how that works.
Large overlap is what happens instead in clinical pathology. That means you don’t have time for the consultative activities. There’s too much activity by the pathologist in the lab administrative piece and not enough clinical consultation.
For the AP piece, the two spheres are completely separated. The pathologists in anatomic pathology often don’t want to have anything to do with administrative roles. Two totally separate circles.
Let’s say you are the president of the hospital. You have a lab director who costs you $200,000 in salary. You have an administrator who costs $100,000. The administrator reports to you so he or she is at least particularly attentive to what you need to know. The net effect: The hospital administrator may have more influence than the laboratory director.
What about our colleagues? Our surgical and internist friends? At least 25 hospitals closed in the Boston area in the 18 years I was there. Closed—couldn’t make it. There were other larger hospitals.
Sixty-five percent of our pathologists are working in that 200 or so bed community hospital. Do you think that if there was an issue to keep the hospital alive that our surgeon and internist friends wouldn’t cut out pathology and let it go? If they figure they could send out all specimens and still keep the hospital going so they could work there, they’re going to do it.
So we must do a good job. That’s our bottom line. We have to make it worth the higher costs to use local pathology services. We have to make it valuable. We also have to broaden our knowledge and our capabilities.
We have to improve patient outcome, and we have to document it in some way. We have to show that the mortality is different, the length of stay is different, documented before and after a new version of our current pathology consult service is initiated. All of this has to come in an objective way from us and say this is why you’re not supposed to back a truck up to the hospital and take it all away.
What do we have to do as a field? We have to understand that the molecular age is here. People are coming to me now in my practice and saying, ‘I want my gene sequenced.’ You’re kidding? ‘Yes, I’m a carrier for factor VIII and I want my entire gene sequenced.’ That’s what’s happening. That’s today. If you don’t know well the interface between molecular and anatomic or clinical pathology, it’s going to be a problem. You have to get there.