What do cardiologists think about the unread echocardiograms at Harlem Hospital?

A few days ago the New York Times reported that Harlem Hospital Center had finally completed its investigation into thousands of echocardiograms that after receiving an initial reading by a technician had never been read by a physician. It turns out that out of 7,000 tests that went unread, only 14 patients “might have been misdiagnosed because their tests had not been handled properly,” according to the hospital spokeswoman quoted in the story. Of the 12 patients who have been contacted, none “were found to have suffered adverse effects from the failure to properly read their tests.”

I’ve received two thoughtful and provocative responses from cardiologists about the latest development in this fascinating and disturbing story. Here they are:

Cardiologist #1 (John Osborne):

The story of the unread echoes in Harlem continues to ripen – and possibly to instruct.

As we all know (thanks to the New York Times reporting and the efforts of Larry Husten to get this story out to the cardiology community), that over a five year period of time, some 7000 echocardiograms that were ordered at the Harlem Hospital Center were only reviewed by the echo technicians and never officially reviewed and signed off by physician readers. After this story was “outed”, the appropriate damage control was performed, and – what the hey, let’s actually read those echos – we were just informed that of the 7000-ish echos, only 14 patients were “mis-diagnosed’. Note that 12 of the 14 patients were subsequently found and that none of them were “found to have suffered adverse effects from the failure to properly read their tests”, according to a hospital spokesperson. While these events are truly scandalous, there may be some very interesting hypotheses that could come of this “echo disaster”.

Before we go any further, I acknowledge the very broad caveats of any conclusions drawn from this lapse of care and professional responsibility. We don’t know anything about the population, the indications for performing the echos are unknown, and the details of how they found that there where only 14 misreads of 7,000 studies are also undefined.

Having stated these caveats, there are a few intriguing hypotheses, Firstly, does this suggest that there is a scandlous level of overuse of echo (incompetence, greed, defensive medicine?). Clearly, as the results of thousands of echo were not demanded (or even noticed to be absent!), the data apparently was not needed for clinical decision making.

Furthermore, I was particularly struck by the claim in this story that of the 7,000 echos done over five years that were never read by docs, that in only 14 patients did the information make a difference in their diagnosis. Given that the role of any testing procedure is to potentially allow one to make a new diagnosis and/or change therapy, that they only found that 0.2% of these patients, once their echo data was “unblinded”, resulted in a misdiagnosis is amazing! Does it say that the vast majority of the echos (99.8%) that were ordered were useless – in that they did not change the diagnosis or therapy, or that the tech did a great job of interpreting them (as a shudder goes up the spine of all the ASE-certified echo readers), or perhaps that the real number of patients affected by this is being dramatically low-balled.

Lastly, it is interesting that the number of unread echos continues to increase dramatically, much like the estimates of oil in the gulf. Just as in the gulf, incompetence and lack of oversight conspire to spell a disaster.

It will be interesting to see how this story evolves.

–John A Osborne MD, PhD, FACC, State of the Heart Cardiology, Grapevine, Texas

Cardiologist #2 (anonymous):

1. Of the 7000 echos reviewed, there were only 17 significant misdiagnoses? That’s a 99.8% accuracy rate for Harlem’s echo techs’ interpretation, which flies in the face of established literature for non-expert diagnosis. I doubt that Harlem Hospital has been able to recruit the world’s best cardiac sonographers.

2. Doctors complained that they were understaffed to read 2500 echos per year — that’s <10 per working day for one FTE. What are their staffing expectations if one FTE can’t handle 10 echos?

Finally, here’s an observation from a medical student who will rotate through Harlem Hospital next year:

… many of the patients (largely Black and Latino) in the neighborhood don’t trust the hospital or their doctors.  Even before this happened.  Health care disparities are being highlighted here… and sadly there are echoes to Tuskegee.

Here are links to previous CardioBrief coverage of this story:

Speak Your Mind