E-Medicine - On Google and Isabel as Sources of Diagnostic Ideas

Sermo.com, a doctor-based website in which doctors exchange views, recognizes doctors’ fascination with differential diagnosis. Sermo’s cases studies, stories of unusual and mysterious diagnoses, and talk of the unusual manifestations of disease, may be Sermo’s most popular single feature, which leads to today’s post.




Today, September 5, 2007, USA Today published “For Doctors, Diagnosing Gets a Technology Boost.” It describes how Dr. Stephen Borowitz, a U. of Virginia Medical Center pediatrician, used www.isabelhealthcare.com, to print a list of diagnoses to match a fevered child’s symptoms, signs, and test results. Gallbladder disease popped up in the list, an anatomic site not previously considered by Borowitz as the source of the fever, which it proved to be.



Sir William Osler observed, “To talk of diseases is a sort of Arabian nights’ entertainment.” Nothing excites doctors so much as making a difficult diagnosis based on a hunch, a flash of insight, or a brilliant stroke of logic. No section of the NEJM is read more than the Case Records of the Massachusetts General Hospital. When applied to these records, Isabel lists the right diagnosis 74% of the time.




Jerome Groopman, MD, a Harvard professor notes in How Doctors Think (Houghton Mifflin, 2006), doctors’ thinking is often flawed and results in misdiagnoses 10% to 30% of the time. Groopman says misdiagnoses result from listening to patients, relying too heavily on technologies, making snap judgments, using the wrong logic, or simply not thinking of an obscure cause – a cause is obvious in retrospect. Groopman doesn’t mention Google or Isabel as diagnostic tools.



Isabel Healthcare offers diagnostic reminders to supplement clinical judgment. Parents of Isabel Maude, a 3 year old English child, and Dr. Joseph Britto, founded Isabel Healthcare in 1999 after Isabel was misdiagnosed after developing necrotitizing fasciatis following a bout of chickenpox. Dr. Britto, a London-based pediatrician and his team, developed software to compile complete lists of diagnostic possibilities.


The Isabel System has two components: the Isabel Diagnosis Reminder System, a database containing more than 10,000 diagnoses, and 4,500 drugs; and the Isabel Knowledge Mobilizing System, which provides diagnosis-specific knowledge to help answer clinical questions and improve decision making quality.

.
A clinician enters a patient’s clinical signs and symptoms, including relevant lab and test results. Isabel checks the data against its vast database and produces a list of diagnoses and/or drugs that may cause the patient’s signs, symptoms, and test results.


Isabel may be interfaced with electronic health records, and with a single click based on patient information, will produce its diagnostic list. The list is based on pattern recognition and information garnered from text books and other authoritative sources.


The rise of Isabel and other diagnostic systems doesn’t surprise me. Back in the Internet’s early days, the late 1970s, Dr. Russell Hobbie and I developed a system called UNIPORT, which was based on patterns of abnormal laboratory results, age, and sex. We were able to list the proper diagnosis in the top five possibilities 80% of the time even without signs and symptoms of disease. With signs and symptoms and other results, hitting upon the right diagnosis 90% to 95% of the time should be attainable.


In a 2005 Letter to the Editor of the NEJM, Robert Greenwald, MD, an academic, remarked after a resident diagnosed a rare immunological syndrome by googling it. ‘Sir William Osler must be turning over in his grave. You Googled the diagnosis?’ The educations of house staff is morphing into computer search techniques. Surely this is a trend to watch.” (NEJM, 353:2089-2090, 2005).


Not to be outdone, the British Medical Journal published “Googling for Diagnosis – Use of Google as a Diagnostic Aid: An Internet Based-Study” in November of 2006.

I would like to hear Sermo readers’ opinions of diagnostic googling.