“You say I got what, doc?”

Understanding a doctor’s diagnosis can be like grasping an actor’s lines in a foreign-language film without subtitles.

But what happens if your treatment options are blurry because your doctor is confused by the language in the medical literature regarding your condition? We’re not necessarily talking nuances of Italian, French, Spanish or Russian, but basic descriptive elements of disease and treatments.

For example, a person suffering sudden paralysis on one side of the body might be diagnosed with stroke, which is also known as a cerebrovascular accident (CVA). But the underlying problem could be either blockage of a brain blood vessel (ischemic CVA) or bleeding into the brain (hemorrhagic CVA). Failure to come up with a precise diagnosis can have devastating consequences for the patient because a clot-busting drug is urgently needed to dissolve a blood clot but, if given to a patient with hemorrhagic CVA, can cause further bleeding and even death.

The often-mysterious language of medical terminology is not just a patient’s dilemma; it is one facing doctors worldwide. It’s also one that Christopher Chute, M.D., Dr. P.H., a Mayo Clinic bioinformatics specialist, is tackling. He likens the international effort he leads, under the auspices of the World Health Organization, to working with Lego building blocks.

“You can have all the data you want, but unless you’re calling things the same, unless you’re recognizing similar things where they occur in one patient’s records versus another patient’s records, and unless you have comparable and consistent data, it’s very hard to bring together past medical experience and the elaborate epidemiologic biostatistical machinery that has emerged in recent years,” Dr. Chute says.

The Chute-led effort will produce the 11th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD). The current 10th revision (ICD-10) was adopted in 1992 at the dawn of the World Wide Web.

The new one, scheduled for 2015, is in the hands of some 2,000 medical scientists, representing every discipline, from around the world. Chute is the first North American to chair the WHO’s Revision Steering Group.

Plague-Era Beginnings

French doctor Jacques Bertillon’s efforts led to the ICD in 1883, when he sought to classify diseases, symptoms, patient complaints, social circumstances and causes of injury. Chute says the concept began with the London Bills of Mortality in 1592, when parish clerks without medical training, tracked deaths amid the plague. These records expanded through the 1830s.

The WHO assumed ICD responsibility in 1948, with revisions being done internally with physicians from several countries gathering to share ideas. “These doctors traded lists and came up with a consensus,” says Bedirhan Üstün, M.D., of WHO headquarters in Geneva, Switzerland. Dr. Üstün is a revision steering committee member and WHO coordinator of classifications, terminologies and standards. “There has been no method to the consensus,” he says. “Now, in the 21st century, that approach and results are insufficient.”

“Today, no single person or organization can do this. We are going toward a more Wikipedia-like operation but one edited like a scientific journal. We needed someone who could understand all that is involved to pull this together, and no one is as well qualified as Dr. Chute.”

Simplifying and codifying linguistic variation, abbreviations, synonyms…

The vocabulary of medicine was built with word roots, often from Greek, Latin or Norman, toggled with prefixes and suffixes. The ICD seeks to sort myriad complexities and provide more simplified coding.

A problem with the ICD, Dr. Chute wrote in a journal, is that terminology is ethnocentric “since the language labels used for a medical concept by speakers of different languages may well be quite different, rendering the study of which term to use essentially a local concern.” This problem, he says, is compounded by linguistic variation, abbreviations, language variation, alternative synonyms and near synonyms.

“We can talk about lung neoplasms or we can talk about lung cancer. It turns out these are not the same thing,” Dr. Chute says. “Not all neoplasms, strictly speaking, are malignant, whereas, almost by the definition of cancer, lung cancer is malignant. When we talk about words, we often use them loosely, when, in fact, we might mean something more specific.”

Üstün, whose expertise is mental health, notes that depression is defined 39 different ways in ICD-10.

Clear language is a passion for Dr. Chute, holder of a medical degree, as well as a bachelor’s in English, from Brown University. In the 1970s, Chute and friends grasped the idea that computer science and medicine could intersect. Medicine then, Dr. Chute recalled, was more folklore and practice than a science-driven delivery system.

After completing his internal medicine internship and residency at Dartmouth College and a doctorate in epidemiology and biostatistics from Harvard University, he established Mayo Clinic’s Division of Biomedical Informatics. He served as chair for 20 years before stepping down in 2008.

“Dr. Chute is visionary,” says Harold Solbrig, a technical specialist in Mayo’s Division of Biomedical Statistics and Informatics. “He has always had a deep understanding of what has to transpire if the science of health care is to take full advantage of the tools of the information age. He has pushed, prodded and pulled the health-care community toward standardized information structures and standardized semantics, with a strong focus on semantics.”

Improving health care requires systematic ways of describing everything that occurs in the health-care process, Dr. Chute says. “There are billing standards, coding standards, information-model standards, nascent electronic medical records standards and on and on. All are good, but the rubber really hits the road when you are naming things. I call it a heart attack, you call it myocardial infarction. Are we talking about the same thing?”

Change: good for patients, conundrum for bioinformatics

Another conundrum is that our understanding of diseases and definitions can change. He notes how Swedish scientist Carl Linnaeus, known for the first organized classification structure for all living things, described rabies in 1759 in “Genera Morborum,” a small, rare book Dr. Chute proudly keeps in his office.

In 1759, rabies was classified as a disease of the mind because victims went mad and frothed at the mouth. A century later, emerging germ theory defined rabies as viral. A more modern example: “Congestive heart failure has long been considered a disease, also based on end result, but in reality it is a manifestation of some 20 diseases,” Dr. Chute says.

ICD-11 will infuse the computer-driven field of bioinformatics, of which Dr. Chute is a pioneer, with an exploding focus on genetics.

“It can be argued that too little information is available to use genetics in diagnostics, but we are pulling together discoveries to catalog them in the ICD and provide linkage to them,” says Dr. Üstün. For ICD-11, however, genetic information will be more for research purposes than diagnostics.

Apart from genes, the “poster child” of what’s missing from the ICD, Üstün says, is Australian-born Barry Marshall’s discovery that the Helicobacter pylori bacterium causes most stomach ulcers. Dr. Marshall’s find reversed long-held dogma that stress, spicy foods and too much acid were to blame — and earned he and his colleague the 2005 Nobel Prize in Physiology or Medicine.

Improving Patient Care with standardization

Gleaning clues from current medical records by researchers, Solbrig says, is like looking for a particular word on a specific page of every book in a large library. “It’s easy if you know what you are looking for when you first shelved the books, but taking them down after the fact is a daunting task.”

A doctor today maintains two sets of records, he says. One set, in plain text, describes a patient’s visit, symptoms observed, test results, a diagnosis and treatment. “Each doctor records this information in a slightly different way.”

The second is an abbreviated summary of tests, a diagnosis and treatment recorded as codes. “These records are designed for use by the billing office, insurance carriers and the general gathering of national statistics,” he says. “While these records can be read by computers, they are of limited value, because most of the details of interest to a genetic researcher are not recorded.”

“A researcher may start with an interesting variation in a gene that impacts, say, bone growth,” Solbrig says. “To understand what this gene does, you need to study a large number of medical records from patients who have this variation and compare them to similar records with patients who don’t. The only practical way to do this is a computerized search, starting with a set of records that say the same thing in the same way.”

ICD-11 will allow genetic researchers to combine genetic markers with large amounts of clinical data. In turn, Solbrig and Üstün say, new findings will be uncovered for use in diagnostics and treatment.

“The ICD today is very reminiscent of a 16th-century flat list, offering a shallow and incomplete model of disease and disease representation,” Dr. Chute says. “It was authored on a word-processing document with very little computational coherency or consistency. It is simplistically a string of words with a number next to them, such as anterior myocardial infarction. I have a clue what that is, but the ICD doesn’t tell me. How do we define it so everyone knows what it is?”

The new ICD, he says, “will be much more formal, much more usable, much more understandable and much more complete.”

It also will be in digital form, notes Üstün:

“Paper classifications are a dying breed, but we will produce paper versions for those who need them. Think of ICD-11 as the medical equivalent of E-Trade or electronic banking. We want the ICD to be a key part of the health informatics package.”

In Dr. Chute’s view, post ICD-11, a doctor will log in, type in keywords and instantly access what treatments worked the best, statistically, for other patients with similar complaints around the world. The “practice of medicine,” he says, will have been transformed to the “science of medicine.”

— Jim Barlow, May 2009