No rights, no wrongs in
  second opinions

Unfamiliar specimens
In-house expertise
’Put it in a micro’

July 2001
Cover Story

Karen Titus

Distilled to its essence, the matter of second opinions in surgical pathology smacks of a one-liner: How many pathologists does it take to read a slide? Like any joke, this one has the capacity to offend as well as elicit a chuckle; what lies behind it are some not-so-funny realities. No one is laughing over cancer diagnoses, or the competency of pathologists.

But if, as comedians say, humor is pain plus the passage of time, then the topic is now providing ample material for the first half of that equation. Look no further than the nightly news or the daily paper.

Not surprisingly, most pathologists express a limited fondness for the recent coverage of second opinions in the lay press. The Wall Street Journal’s health column on the subject in April? Too simple to be useful. ABC-TV’s "World News Tonight" segment in May? Sensational—and little else. Last year’s New York Times article? Slightly better, though it still left readers with the impression that no pathologist can be trusted. As Matthew Kershisnik, MD, of Logan, Utah, puts it when asked about the latest press dispatches: "You mean all those articles about how we make mistakes all the time?"

"I tell you, they’re more than welcome to come in and show me all my mistakes," he adds.

Given the often excitable nature of the lay press, their discussion of second opinions might sound daffy if the stakes weren’t so high. But if the coverage creates anger and even a bit of worry, it may also contain the seeds of opportunity. Pathologists facing concerned, even skeptical patients are likely to find those same patients willing and eager to be reassured that surgical pathologists know their stuff—and, equally important, when they don’t.

Media coverage aside, one could argue there’s nothing new to report about second opinions, given that they’re a routine part of surgical pathology and have been for decades. That doesn’t mean there isn’t plenty to say on the matter.

A likely place to begin is with second opinions done for quality assurance.

No one has a lock on the ideal QA program. "Clearly you can’t have a one-size-fits-all plan for any kind of peer review or quality improvement program," says Patrick L. Fitzgibbons, MD, chair of the CAP Surgical Pathology Committee and director of pathology at Good Samaritan Hospital, Los Angeles. "It has to be designed specifically for the institution, for the group size, for the kinds of cases that are coming through, for the kinds of concerns that the local clinicians have."

Second opinions, or peer reviews, are actually a relatively small part of overall QA programs, says Dennis G. O’Neill, MD, chairman and director of the Department of Pathology and Laboratory Services, Manchester (Conn.) Memorial Hospital. And while "there are some basic tenets to quality assurance programs, when you come to peer review, most of it is developed locally."

This means, among other things, that while practices must consider the basics when developing peer review programs—for example, looking at organ sites known to be more problematic—they also need to look at areas within their group where depth of experience may be lacking.

Dr. O’Neill is part of a five-pathologist group, what he calls "kind of the typical, private practice, hospital-based, general pathology group." What works for his group, he cautions, may not suit other groups, especially since group dynamics as well as size and locale figure into any QA program.

At the Manchester lab—in fact, at most laboratories—peer review happens prospectively and retrospectively. The former has the obvious advantage of permitting pathologists to pick up on any misinterpretations or errors in a timely matter. "So if an error does occur, there’s no harm to the patient," Dr. O’Neill notes.

Prospective peer review occurs daily. His department requires every malignancy to be seen by a second in-house pathologist prior to sign-out. In addition, departmental policy calls for difficult types of biopsies—such as all prostate and breast needle biopsies—whether malignant or benign, to be shown to a second colleague. Under these policies, Dr. O’Neill’s group prospectively reviews about 25 percent of its cases routinely.

"If a doctor or patient says to me, ’My breast cancer was diagnosed at your institution last week, and I want a second opinion,’ I can say to them, ’You already have had a second opinion within the department—another pathologist has seen those slides. If that’s still not sufficient, we’d be happy to send it outside our institution for another opinion,’" Dr. O’Neill explains.

All opinions generated from prospective peer review are documented on log sheets and retained for monthly tabulation. "We keep track of who called what what, and who was right and who was wrong, and who was a little bit off and who was a lot off in their interpretation," he says. "We spend a lot of time with that, because we think that’s important for ensuring the right diagnosis goes out."

Dr. O’Neill’s group also employs retrospective review. Classically, this means correlating frozen sections with final diagnoses—a decades-old practice, he notes. This not only ensures appropriate information makes it into the pathology reports, but it also "is a good way of determining a given pathologist’s skill level," he says.

Retrospective review also occurs when patients have their specimens interpreted at another institution after a first diagnosis at Dr. O’Neill’s laboratory. When the outside pathology report comes back, "I review that report and correlate it with my own diagnosis." His department also routinely reviews amended reports within the institution. While such reports often reflect minor clerical error—a wrong date, for example, or a name misspelling—they occasionally are triggered by a change in a pathology report. When that happens, "We need to figure out if a mistake was made—and why." Finally, if a former patient returns to Manchester for treatment, any previous related biopsies done at the institution are reviewed—and again, if any mistakes are uncovered, they are tabulated as part of the lab’s peer review program.

While retrospective opinions can be done at a pathologist’s leisure, they lose the just-in-time benefits of prospective reviews. "Depending on when you do the retrospective review, the patient may have already been harmed—a malignancy may have progressed, or a therapy may have already been started that shouldn’t have been started," Dr. O’Neill acknowledges.

Laboratory QA programs at Intermountain Health Care, based in Salt Lake City, are far-reaching, to say the least. With its three hospitals and 14 pathologists—the largest pathology practice in the state, according to its head—the burden of maintaining consistency and competency is heavy. "It requires us to be very careful about quality assurance," says M. Elizabeth Hammond, MD, who chairs the Department of Pathology for Intermountain’s Urban Central Region Hospitals.

One prong of their QA program consists of intradepartmental consultation, which is required whenever a pathologist sees a case that’s cause for uncertainty. "The slides, along with the pathologist’s impression and a little clinical history, are circulated among the members of the group who are available that day—it might be three or four different pathologists," explains Dr. Hammond, who is the former chair of the CAP Cancer Committee. After this review, the case circles back to the first pathologist, who then signs it out if there is a consensus diagnosis. "Which is usually what happens," Dr. Hammond says. Cases that do not achieve consensus are sent to an additional "referee" pathologist who practices in an outside group or another institution. Another lab policy requires every new cancer diagnosis to be reviewed by at least one other pathologist in the group prior to final diagnosis.

A third level of QA involves intermittent review of notable cases. An expert in a specific subject, such as transplant pathology or GI biopsy, will put together a group of slides representing difficult cases, accompanied by a list of potential diagnoses. After the slides and list are circulated among the group’s members, who contribute their own diagnoses, "We have a conference where that pathologist shares with us all the results from our group looking at the slides, and we discuss the potential issues around those diagnostic entities," Dr. Hammond says. "That’s an important educational activity for our group, and it helps us improve our level of skill."

The group also attempts to follow up on the cases used in these intermittent reviews. Perhaps six months after the first group review, "We try to send the slides out again and document whether there’s been consistent changes in the behavior of our pathologists, based on that learning exercise." Dr. Hammond reports being pleased with the results. "There’s been a lot of improvement in consistency by doing that." In addition, the laboratories use the CAP’s Performance Improvement Program in Surgical Pathology, or PIP.

Directed case review, which occurs randomly, is another QA component. "Directed case review occurs when a case is rereviewed at the suggestion of a clinician or because of new clinical information about a patient," she says.

If the pathologist has indeed erred, the lab’s response is nonpunitive and educational. "We assume anyone can make a random mistake. All of us have done it," Dr. Hammond says. "So we discuss the slides with the pathologist who made the mistake, and make sure they agree it is a mistake." If the pathologist remains unconvinced, then the slides are sent to an outside expert, whose opinion may serve to persuade. At that point, the followup discussion "usually causes the person to say, ’Maybe I do need to know more about that particular entity,’" Dr. Hammond says. Thanks to her hospital’s long history of SNOMED-coding all its cancer cases, it’s easy to pull samples for educational purposes. If the case in question was an endometrial biopsy, for example, and the terminology used by the first pathologist wasn’t quite right, "We can suggest that they look at 20 cases from the last year, maybe with our help. So we come up with a plan to make that person more comfortable about that particular entity."

Finally, Dr. Hammond and her colleagues meet quarterly to review any interpretive errors on frozen sections. "We keep track of the error rate by pathologist at all times for the frozen sections," she says. That information becomes part of her biannual review of her department’s pathologists and helps her assess whether they’re competent to be readmitted to the medical staff.

The group’s size means it sees a fairly steady stream of new faces, which adds a mentoring component to the QA program. "They’re encouraged to show all their cases to one of the senior people for several months after they come into the practice," Dr. Hammond says. Since the group also encourages cases to be shown during intradepartmental reviews, and because backup is required for all frozen sections, "People have no problem feeling comfortable asking for help. We make it easy for everyone to get opinions."

If Dr. Hammond oversees Utah’s largest pathology group, then Dr. Kershisnik is part of what is arguably one of the state’s smallest practices. He is one of two pathologists at Logan Regional Hospital, although the two physicians essentially work as solo practitioners within the institution.

While the Intermountain pathologists can easily turn to their own board-certified specialists for opinions, Dr. Kershisnik has no such luxury. Large-scale educational reviews are out of the question. So how does he handle QA peer review?

For starters, he and the other pathologist at Logan review 10 percent of each other’s surgical cases, which are chosen at random. Most cases are reviewed within two to four weeks of being signed out.

Another three to four percent of cases get reviewed by pathologists outside the hospital. When a patient or clinician seeks a second opinion—pathologic or clinical—at an outside institution, for example, the patient’s slides are invariably seen by the second facility. When the report generated by that review is sent to Dr. Kershisnik, "We correlate those reports with ours," he says.

Now let’s muddy the waters a bit.

Like most pathologists, Dr. Kershisnik reports that significant discrepancies with outside pathologists are rare in the aforementioned cases. More frequent are splits that occur when pathologists themselves seek an outside opinion—no surprise, since what leaves the labs in these instances are, by definition, problem cases.

While these consultations are sought for diagnostic purposes, they’re also pivotal to QA efforts. It’s not a matter of pathologist incompetence, but the reverse. "The single most important topic in surgical pathology is addressing your own level of certainty about a case," says Ronald Sirota, MD, chairman of the Department of Pathology at West Suburban Hospital Medical Center, Oak Park, Ill.

Finding outside experts to provide second opinions is relatively easy. Many of Dr. Kershisnik’s troublesome specimens find their way to LDS Hospital, which is part of the Intermountain system. Dr. Kershisnik trained there as well as at M.D. Anderson Cancer Center, Houston, where he also sends specimens for outside consults. Both groups, it goes without saying, offer an attractive selection of board-certified specialists. But just as important, says Dr. Kershisnik, is his relationship with those whose opinions he seeks.

"It’s helpful to know your consultants and to have worked with them in the past," he says. "You want to be comfortable working with them, and to know you’ll get some education in the process."

Beyond the familiar ties, pathologists choose experts based on meeting presentations, papers published, and reputation, as well as their ability to produce clear and timely reports.

Though finding outside experts to provide second opinions is relatively easy, pressing that information into service can be less so. A host of difficulties can accompany the use of outside opinions, not the least of which is they might be wrong.

"It’s important to evaluate the opinion of the consultant," says Dr. Kershisnik, who readily admits he’s disagreed with second opinions and sought additional ones. "That doesn’t happen very frequently, but I don’t worry about offending anybody when it does," he says. "Usually one second opinion is fine, but sometimes you need a third or even fourth opinion from a higher authority, a world-renowned expert at a large cancer center."

Richard Zarbo, MD, chair of the Department of Pathology at Henry Ford Hospital, Detroit, recommends that practices monitor their outside consultants by followup of patients through tumor boards and by integrating diagnoses with clinical information over time.

For their own part, he adds, referring pathologists need to make sure that consultants at treating institutions receive a copy of the original pathology report and a copy of all pathologic materials used to arrive at a diagnosis. "Recut slides are acceptable," Dr. Zarbo says. "But it’s extremely important for the first pathologist to review those recuts and make sure they accurately reflect the findings used to make his diagnosis, and to make sure that no new pathology arises in those recuts that may impact treatment." If it does, the first pathologist should dictate an addendum report modifying his diagnosis based on that review, as both an aid to the second pathologist and for legal self-protection.

When Jonathan Epstein, MD, and his pathologist colleagues at the Johns Hopkins Medical Institutions, Baltimore, arrive at a second opinion that’s seriously at odds with the opinion of a referring pathologist, they first try to explore reasons for the discrepancy. "We typically call the clinician or the pathologist at the outside institution to make sure we have all the material. Because one of the potential sources of a disagreement is we’re just not looking at the same material they are, and sometimes that’s the easiest way to resolve the discrepancy," says Dr. Epstein, a professor of pathology, urology, and oncology at Johns Hopkins.

While the outside pathologist is not obligated to track down the original pathologist in every case—"That could be very onerous," Dr. Fitzgibbons admits—common sense and common courtesy dictate that the second pathologist inform the first of any major disagreements. "I’d prefer that pathologist pick up the phone and call me, although I recognize there’s no one best way to do it," he says. "But no matter how it’s done, I’d rather hear it from the person issuing the report and try to understand their reasons, rather than hearing it from a family member"—a ticklish situation he himself has experienced.

Ronald M. Harris, MD, MBA, is in full agreement, and has plenty of opportunity to act on his beliefs. Dr. Harris, an adjunct assistant professor of pathology and assistant professor of dermatology at the University of Utah, keeps busy. Until recently he was part of the 14-member pathology practice serving Intermountain Health Care—and the group’s only dermatopathologist. In addition to handling reviews from his colleagues and from Intermountain’s three hospitals, he handles dermatopathology materials for the local VA and children’s hospitals, covers the university’s derm-path services, and oversees derm-path services for a local cancer center’s melanoma program. "A lot of my time is spent looking at other people’s slides," he says—and not only because his subspecialty has few practitioners in the region.

"I think part of the reason I get a lot of consults is because I tend to try to talk to people, to call them, about the important cases," he says. "If it’s a melanoma case particularly, I always call them. I cannot stress this enough—communication with the referring pathologist, and oftentimes the dermatologist who sent the slide, is vital." He stops short of saying calls to referring physicians are mandatory, but notes, "Boy, it will save you a lot of grief in the long run if you do make that phone call."

When he does disagree with the referring pathologists, he says, "I’ll talk to them and explain why my diagnosis is such and such." Sometimes they’ll ask Dr. Harris to show the slides to another pathologist, which he says he’s happy to do. "I cover my bases by doing that, and I think it’s good politics. Most people appreciate if you make that effort, and they’ll tend to value your opinion more."

Since this is, after all, an article about second opinions, it may be time to add another, very vocal one to the mix.

Paul Kalish, MD, admits errors occur in surgical pathology. He’s not opposed to sending out cases for expert review. He believes in working closely with clinicians. He’s a strong advocate of documenting QA efforts. He understands the pressures pathologists are under when the public and the media cry out about second opinions—he even wrote a measured response to the New York Times after it published its article on the subject last year. But he puts his foot down, firmly, on one widely accepted peer review practice.

He stomps on it, actually.

"There are many pathology departments for which internal second opinions are the rule rather than the exception. In fact, one of the departments in our own network has mandatory second opinions on all cancer cases," says Dr. Kalish, assistant director of the core laboratory at North Shore-Long Island (NY) Jewish Health System and director of pathology at North Shore University Hospital at Glen Cove.

"I’m absolutely appalled that this has gotten to be the standard of practice," he says. "Pathologists probably ought to go out of business if it takes two people to come up with the correct diagnosis."

The fact is, it doesn’t, he says. "The overwhelming majority of pathologists are perfectly capable and competent to render cancer diagnoses without having a mandatory second review. No other specialty in medicine would even think of taking that position."

Pathologists merely weaken the profession when they put such stringent policies in place, he says, and spook an already mistrustful public. "It certainly doesn’t help when you have leading academic institutions requiring a second pathologist to examine every cancer," he says. "The lay press already has people believing you need at least two people to make a real diagnosis. And now we’re contributing to that perception. It sends the message that at major university centers, where presumably the leading pathologists in the world are practicing, you need two people to diagnose a specific lesion.

"No wonder the public thinks pathologists can’t make an accurate diagnosis," he says.

A clearly wrong diagnosis, whether it’s made by the referring pathologist or an outside expert, is rarely the true problem, however. It’s the more ornery cases, the ones with no clear-cut answer in sight, where the real hand-wringing begins.

Such disagreements are healthy, in some respects, and can provide another means of evaluating consultants’ competency. "You want your outside experts to provide opinions that are roughly in concert with your own," says Dr. Epstein. At the same time, he says, "Part of having confidence in your experts is recognizing they’re going to hand back a certain number of cases where they say, ’We’re not sure either.’ If you send cases off to an expert and you’re always getting back a definitive case, that should raise some questions as well."

When his laboratory reviews material sent from other pathologists, serious disagreements—presence or absence of cancer, type of cancer—occur about one to 1.5 percent of the time. In a larger percentage of cases there may be disagreements over diagnostic aspects that could affect treatment, says Dr. Epstein, who’s quick to add, "I’m not saying that we’re right and they’re wrong. But it does flag that these may be difficult cases, that they’re not necessarily straightforward one way or the other, and that they might benefit from a third opinion or maybe additional tissue to try to resolve it."

Before seeking outside experts for help on a difficult case, Dr. Hammond and her colleagues first turn to other testing—immunocytochemistry, electron microscopy, FISH, "something that would help us evaluate the tissue differently," she says.

If the case still lacks consensus, the group turns to outside reviewers—after informing the clinician. "We do no levels of consultation without involving the clinician in that discussion," Dr. Hammond says. "If we think we’re going to send a case out, we talk to the clinician and say, ’We can’t agree on what this is, and we’re going to send it to So-and-so at the Mayo Clinic.’ This is a cooperative venture between the pathologist and the clinician."

Doesn’t that undermine clinicians’ confidence in the pathology department?

"Pathologists don’t want to look divided and inept, so there’s clearly going to be internal reasons to try to work it out without letting everyone know you can’t agree on anything," acknowledges Dr. Fitzgibbons. "On the other hand, it’s important for clinicians to know when there is substantial disagreement. It’s not unusual for pathologists to say in their report that different pathologists have reviewed the case and that there’s not a unanimous opinion." In fact, says Dr. Fitzgibbons, a group’s peer review policy might even want to reflect this practice, stating that clinicians will be told of substantial disagreements and what pathologists will do to address them. "You’d never want to hide or bury the major disagreements," he says. "The patient’s best interest has to take precedence over the group’s image."

Dr. Hammond suggests that sharing uncertainties with clinicians may enhance the relationship. That’s been the case at her institution. "Because they know we take this seriously and have a process to deal with it, I think they have a greater confidence in the diagnoses we do render," she says.

There’s another good reason to share difficulties with clinicians, says Louis P. Dehner, MD, professor of pathology at Washington University School of Medicine, St. Louis, and director of anatomic pathology, Washington University Medical Center. "When we see we are having problems, whether we eventually send that case out or not, the attending pathologist will call the surgeon or clinician and let them know we’re having trouble, and that we’re getting additional stains, or whatever. Because we know there’s a patient out there, a patient’s family, saying, ’What’s going on? What’s going on?’"

And sometimes, notes Dr. Zarbo, clinicians have to accept that occasionally diagnoses will be less than certain. While clinicians largely expect pathologists’ opinions to be black-and-white, "Gray is sometimes the honest answer," he says. "It has nothing to do with our expertise or talent, but the fact that the tissue clues just aren’t there, or that classification schemes are incomplete to account for a finding."

Are there types of cases that only experts should handle? The answer would appear to be "no." Better medicine, most agree, takes place when care—including the laboratory component—is provided locally by a group of physicians who regularly interact with one another. "That’s far better than just having an answer spit out of a fax machine by someone far away," says Dr. Fitzgibbons.

Still, the issue persists, in part a reflection of town-gown tensions. "There’s a certain arrogance out there, a concern in some of the literature that out in the community, the expertise level is not sufficient," Dr. Kalish says.

The problem is partly semantic. "I usually put the word ’error’ in quote marks," says Paul Bachner, MD, CAP president and professor and chairman of the Department of Pathology and Laboratory Medicine at the University of Kentucky Chandler Medical Center, Lexington. "When you look at any study published in the literature, you have to ask, What’s being called an error? How are the errors classified? Who’s making the decision as to whether the difference in opinion results in a change in treatment for the patient?"

"Just because an expert said the diagnosis is incorrect doesn’t mean the original pathologist was wrong," adds Dr. Fitzgibbons.

He points to a handful of reasons why an expert’s opinion may vary from that of the first pathologist. Those rendering second opinions can—and often do—benefit from access to clinical outcomes information. "Many times these second opinions take place months or even years later, although that’s not always made clear in the literature," he says. Likewise, over time new understandings of diseases evolve, as do new testing methods.

More important, long-term clinical followup sometimes shows the original diagnosis was actually correct. As Dr. Harris puts it, "One person, no matter how expert they are, is not the gold standard."

Outside consultants aren’t the only ones who disagree. Whenever pathologists review each other’s slides regularly, there’s bound to be differences of opinion. Indeed, there should be.

That doesn’t mean such disagreements are easy to swallow. "It can be very intimidating for both parties if the second pathologist says, ’I think you got this one wrong,’" says Dr. Fitzgibbons. Who wouldn’t feel defensive hearing that?

It’s tricky, admits Dr. O’Neill. "One of the reasons we spend so much time doing prospective case review is to become comfortable with each other’s opinions over time.

"If a given pathology group functions as five individuals sitting in closed offices, and they don’t share cases daily and exchange opinions and knowledge, then you can have some real disruption in a department," he continues. "And that in turn is confusing to clinicians, because they don’t know who to believe. It ultimately becomes very destructive."

The bottom line, he says, is this: When a pathology report leaves a department, it should represent the best work that department can generate. "Even though there’s one person’s name on it, and one person holds liability for it, in essence it’s a departmental product."

Dr. Sirota of West Suburban Hospital strongly encourages pathologists to blind their opinions when sending around slides for intradepartmental review. "That way you don’t have the bias introduced by reading someone’s opinion," he says. That bias can be problematic in group hierarchies, he adds, "where the junior members could be influenced by the senior members simply by reading the senior members’ opinions."

Whatever the disagreement, it eventually needs to be resolved by the pathologist who solicited additional opinions, he says. "We have no real rules about how to do that in my department, other than we trust the pathologist to make the right decision. All my pathologists are credentialed, and I trust their work.

"I feel almost sheepish saying that," he continues. "If you don’t trust the pathologists in your department, you have to look at why they’re allowed to sign out material."

Written policies can also help defuse touchy encounters. In fact, not having written policies can be a mistake. Without good paper trails it can be difficult to recreate the sequence of events when a diagnosis proves to be in error, and error-tracking and QA processes are hindered in general. "Each practice needs to document their rate of revised diagnoses, and to document that rate by organ system and by pathologist as part of their peer review and QA processes," insists Dr. Zarbo. Clear records also assure clinicians—and perhaps patients—that a second set of eyes has looked at the more problematic cases.

Those who worry that documentation will make them more vulnerable to lawsuits should relax, suggests Dr. Hammond. "Documenting your processes protects you, because it shows you have a standardized process to avoid error." In some states, those records are not discoverable.

In yet another balancing act, however, pathologists need not rush to put everything in writing—doing so, quite simply, would be useless. "You can’t really put into writing which cases are going to be sent out; it requires the judgment of that original pathologist," Dr. Fitzgibbons says. His lab’s policy states that second opinions will be obtained whenever necessary—a decision that’s left to the pathologist of record.

Another form of written reports—published studies—could help pathologists, clinicians, and patients step through the intricacies of second opinions. If pathologists are able to identify areas at higher risk for discrepant opinions, the thinking goes, no one loses-pathologists could home in on these problems in their practices, and clinicians and patients would gain a refined sense of when it might be valuable to request a second opinion.

Yet there are obvious downsides. In the ideal world, researchers would pore over materials at multiple hospitals of varying sizes throughout the country, unearthing and evaluating missed and discrepant diagnoses. "But that just doesn’t happen," Dr. Epstein says. Most research ends up being done at large academic sites, which often receive the more difficult cases. "So the studies may not be representative of the field in general."

Then there’s that niggling problem of bad PR. Internecine discussions rarely play well in the press.

In controversial areas—the ones most likely to be targeted in research studies—a slide given to 10 experts may easily, and reasonably, yield half a dozen different opinions. "It doesn’t mean anyone is right or wrong; it means the issue is controversial. But the medical community, let alone patients and the lay press, has a hard time understanding and dealing with that," Dr. O’Neill says.

When these matters reach the lay press, something altogether different can get communicated. Dr. Epstein, who has been a frequently quoted source in many of the recent stories on second opinions, is well aware of the pitfalls. "The problem with the news media is they have two minutes to show what they’re going to show. Of the two minutes, probably about a minute of it is fluff. They’re left with about one minute to present a very complicated issue."

Just what is the public being told?

Readers who pick up the July issue of Prevention magazine are greeted by a cover line that suggests they can "Beat a scary diagnosis." The story inside urges readers to obtain a second opinion to establish the accuracy of their diagnosis, and a second, third, or even fourth opinion to determine the best treatment. Harvard Medical School professor Jerome Groopman, MD, is quoted as saying, "In many cases, pathologists disagree, or test results may not be reliable." Another source, Charles Inlander, noted (in bold-faced type),"One out of every five diagnostic first opinions are not confirmed by second opinions." Up to 80 percent of second opinions don’t get confirmation, he added.

Scary stuff to a lay reader. And, frankly, pretty meaningless, given the complexity of the topic.

Contacted by CAP TODAY, Inlander, who is president of People’s Medical Society, a consumer health advocacy organization, elaborated on his remarks (after being reminded about where they had appeared—"I do eight, 10 interviews a day," he said).

"There’s two types of second opinions: There’s the diagnostic and there’s the treatment," he says. "Studies have been pretty consistent that overall, about 20 percent of second opinions don’t confirm diagnostically what the first opinion was. And studies show in some cases that eight out of 10 second opinions don’t confirm the second opinion on what to do."

Inlander credits consumers with becoming more proactive in seeking second opinions, then adds, "The docs don’t like it."

Should patients seek a second pathology opinion? "I think it’s vital," he says, noting that at least one pathologist’s research "has shown some of the terrible work that’s done in pathology, particularly oncology."

He suggests that samples be reviewed "independently from where it may have been reviewed the first time." Are QA practices and pathologists’ own judgments on whether a second opinion is needed sufficient? Nope, he says. "Remember, even if it’s within their own institution, these guys are protecting each other. It’s just a natural tendency. And the reputation of the institution is on the line."

The meaning of the word "opinion" is not lost on Inlander, who says pathologic interpretations are still more of an art than a science. Pathologists "may see the exact same thing, but two different people see it in a different way, and it means something different." Sounding more like a pathologist than a consumer advocate, he adds that when opinions collide, "You’re looking for consensus. It’s not always black and white—in most cases it’s not black and white, on any type of second opinion."

Few patients would find these amplified remarks comforting. And, since no patient is likely to have a telephone conversation with Inlander to find out just what he means anyway, they’re left with those disquieting figures: one out of five, eight out of 10. And, no doubt, with a deep suspicion that their samples probably do need a second set of eyes, if not more.

Equally disturbing numbers popped up in the Wall Street Journal’s April 17 "Health Journal" column, which noted that while serious errors in surgical pathology opinions are considered rare, mistakes that could alter treatment may happen as often as 20 percent of the time, depending on the type of cancer.

Citing research from Johns Hopkins and Dr. Epstein, the column discussed error rates for various types of cancer. And no, it didn’t discuss the number of cases in which the second opinion itself proved to be in error; the seriousness of the cases seen at Hopkins; the time that elapsed between the first and second diagnoses; QA programs; or whether "error" was indeed the mot juste.

But if pathologists are frustrated by abbreviated descriptions of their work in the lay press, so are those who pen those accounts. Tara Parker-Pope, the Wall Street Journal’s health columnist, says her column typically runs a mere 850 words. (In published studies, by comparison, researchers regularly use up several hundred words on abstracts that explain only the bare essentials of their work.) She’s familiar with complaints that consumer medical stories omit crucial information, but says, "I’m not writing a textbook."

"My goal is to give patients the highlights, the big points that will prompt them to have a conversation with a doctor on the subject," she explains. "I leave it to the medical profession to help patients decide what is best for them."

She reports an interesting mix of mail from her second opinion piece. About half of it was generated by pathologists and clinicians who applauded her tackling the subject. "Then I got another batch from [physicians] complaining the story was unfair, didn’t appreciate the ’art and science’ of pathology diagnosis, complained about the subtleties involved in prostate cancer grading and staging, and harrumphed at the notion that another set of eyes would really make any difference. That if it really were that difficult or uncertain they would already have shown it to a colleague," she says. "They really bristled at the notion that a second opinion is necessary."

"I do find the medical profession often gets frustrated at the lay press, but I think they need to start looking at it from the patient point of view," she says. "Patients want information, even if in a condensed version, because it is often more than their doctor will give them."

"Judging by my mail from readers," she adds, "doctors who otherwise provide their patients with excellent medical care regularly fail their patients when it comes to simple communication."

Words even most pathologists would have a hard time arguing with.

Karen Titus is CAP TODAY contributing editor and co-managing editor.