Arun Krishnaraj, MD, MPH

With the release of the iPhone 4S and Siri, Apple has introduced speech recognition (SR) technology to the masses. Apple bills and markets Siri as a “humble personal assistant.” However, I doubt many radiologists, who have been working with SR technology for multiple years, would describe their SR software systems this way.

Radiologists who use SR software to dictate their reports often describe the experience as frustrating. Much of this frustration, I believe, stems from the introduction of SR before its maturity.

Apple is renowned for their elegant consumer gadgets. They delay release of their products until their engineers believe that a technology has improved to a point where its implementation will not cause undue frustration on the part of their users. Case in point, while the tablet computer existed long before the iPad, Jobs and Co. at Apple realized that it would not be readily embraced by the average user until the product could be very thin, very responsive, and light. Apparently the “magic” (to borrow a term Jobs liked to throw around) year when technology was robust enough to accomplish this was 2010.

Now that Apple has determined SR technology to be mature enough to incorporate it into its iconic iPhone, should radiologists stop their grumbling?

Speech recognition software, which facilitates radiology dictations, has many proven benefits. Last year in AJR I wrote about our experience at 

UNC Hospitals incorporating SR software into the radiologist’s workflow

. We noticed substantial improvements in report turn around time similar to what other observers had demonstrated, but also found that improvements in turn around time varied significantly between users due predominately to different work habits. Users who spent greater time training the system to recognize their voice and adding and revising terms in the vocabulary library, for example, benefited the most.

If you read between the lines of Apple’s marketing onslaught you will realize that they advise the same type of behavior to improve Siri (the more you use it, the better it will get). Additional benefits of SR driven dictation systems include the ability to leverage templates, mini macros, and structured reporting to improve accuracy, consistency, and billing. Moreover, advances in 

natural language processing coupled to SR software

 may allow for the ability to integrate targeted radiology decision support at the time of dictation, the ability to structure unstructured reports, and identify key content and highlight critical findings.

However, despite these benefits SR software has potential drawbacks. Most notably errors can result from incorrect capture of a radiologist’s voice and more important, the intent of their language. Most recently Sarah Basma and her group at Women’s College Hospital in Toronto, found an 

alarmingly high rate of errors using SR software vis a vis traditional transcriptionists for breast imaging reports

. After correcting for multiple variables Basma’s group found that reports generated with SR were eight times more likely to contain major errors than those generated by traditional transcription.

The conflicting reports on SR software should give us pause as we move forward. The companies producing SR software have improved their products significantly over the last few years. Despite these improvements, 

 shows that while 80 percent of users use SR software, 30 percent still find it hard to use and not as accurate as traditional transcription services.

Apple believes the technology is ready for prime time. However as in most things tech, what is good enough for the average consumer product doesn’t necessarily stand up to the rigorous standards of healthcare. If Siri misinterprets your request for the weather forecast, you may get rained on. However, if a radiologist’s breast imaging report is recorded incorrectly the potential consequences are much more substantial.

Articles

Speech Recognition Technology - Finally Ready for Prime Time?

Impression: 1) Complex nodule left lobe of thyroid.
                       2) Very small solid nodule right lobe of thyroid.

The above is an excerpt from my mother’s thyroid ultrasound report. Last week my mother noticed a lump in her neck while brushing her teeth. She was concerned about this finding and called me to inquire how to proceed. I advised her to go and see her primary care doctor. I knew her doctor would perform a physical exam and that he would likely order some imaging if he felt that the lump was suspicious.

As expected, after my mother’s visit to her primary care doctor, her physician recommended imaging: a thyroid ultrasound due to his localization of the lesion to the thyroid gland. The above report was returned to my mother’s doctor who was somewhat confused with how to proceed, and after my full review of the report I can understand why.

The report included a measurement of the complex nodule in the left lobe of the thyroid, but no explicit comment on whether the nodule had a solid component, the nature of the borders of the lesion (i.e. well defined vs. infiltrative), presence or absence of microcalcifications, or the presence of flow.

The solid nodule on the right described in the impression, in addition to excluding the aforementioned descriptors, also failed to accurately gauge the size of the lesion only noting that it was “very small.” No recommendation on whether biopsy was necessary or feasible was included.

As a radiologist who routinely reads thyroid ultrasound, I understand that these mundane studies with high likelihood of a benign result regardless of the imaging findings can become the bane of a radiologist’s existence. The sloppiness of the report, nevertheless, bothered me. Due to the size (> 2cm) of the so called “complex” nodule, my mother will proceed with a biopsy this week. However, I still do not know enough details about the lesion to justify whether the procedure is even necessary.

Clearly there is confusion in our field about how to accurately report and follow thyroid nodules detected on ultrasound. This is best illustrated by the fact that the Society of Radiologists in Ultrasound consensus conference statement on the management of thyroid nodules published in 2005 remains in the 

If mammography results were reported in the inconsistent way that thyroid ultrasound reports are currently dictated there would be outrage in the medical community. The breast imaging community recognized the need to standardize their reports long ago by creating the BI-RADS system. This simple-to-interpret seven point scale gives explicit instructions to referring providers on how to proceed with management of detected lesions. BI-RADS also roughly quantifies the suspicion of malignancy of a detected finding as determined by the radiologist.

Thyroid ultrasound imaging is becoming more common as a result of the detection of thyroid lesions on neck and chest CT exams. Moreover, thyroid ultrasound detects nodules in up to two-thirds of all patients while less than 10 percent of nodules harbor malignancy.

It is important for radiologists to be consistent with their approach and management of these lesions when imaged with ultrasound to prevent unnecessary biopsies, and just as important, undue fear on the part of patients.

One such solution has recently been proffered and evaluated by a group from Korea whose 

. In a study that involved more than 1,500 patients with over 1,600 path proven biopsied nodules, the investigators developed and validated a scoring and management system for thyroid nodules that they dubbed, TIRADS (Thyroid Imaging Reporting and Data System). My hope is that one day such a system will become embraced by the radiology community in the same manner that BI-RADS is accepted. This type of system could easily be integrated into speech recognition software with macro driven templates for consistency.

Despite the radiology report my mother received from her radiologist, I am confident that her nodule will likely be benign; and the statistics would back me up. However, if her radiologist spent more due diligence in crafting his report, we may not have even had to go down this path and I could give my mom more reassurance that she will be fine.
 

If mammography results were reported in the inconsistent way that thyroid ultrasound reports are currently dictated there would be outrage in the medical community. It is important for radiologists to be consistent with their approach and management of these lesions when imaged with ultrasound to prevent unnecessary biopsies, and just as important, undue fear on the part of patients.

Consistency in Reporting and Communication: Thyroid Ultrasound Imaging

Mrs. Anderson called her doctor and spoke with pressured speech: “I think there is a mistake. That cannot be my ultrasound. I don’t have a gallbladder.”

Mrs. Anderson was accessing our hospital’s open EHR, which allowed her to view her radiology reports, and noticed a discrepancy. An ultrasound had been performed which identified “multiple hypoechoic lesions, incompletely characterized.” Because the lesions discovered were indistinguishable from metastatic deposits by ultrasound an MRI was correctly recommended for more definitive characterization. However, the dictating radiologists overlooked a “canned” normal finding in his final report noting the presence of a normal appearing gallbladder. Mrs. Anderson’s gallbladder had been removed surgically 10 years prior.

Mrs. Anderson’s case came to my attention when I dictated her liver MRI, comparing it to the ultrasound performed earlier in the week. The liver MRI fortunately demonstrated only multiple benign lesions. My report commented on expected post cholecystectomy changes in the right upper quadrant in the findings of the dictation but not in the impression. I did not read the portion of the prior ultrasound report describing the gallbladder because I was only concerned with the positive finding for which the MRI had been ordered, the liver lesions.

I received an e-mail the next day from the nurse practitioner (NP) caring for Mrs. Anderson. She said that Mrs. Anderson had called her office frantic that there was some kind of mix up and that her patient wanted to know which of her reports were correct. I reviewed both reports and realized there was a proofreading error on the ultrasound report, paged the dictating radiologist to amend the error, and explained to the NP how such errors occur with our speech recognition software and macros. The NP understood and said she would explain to Mrs. Anderson that both her studies were accurate except for the oversight on the ultrasound dictation which would be amended.

While the NP realized how such errors may make their way on to finalized reports, I now better understand the impact an incorrect report can have on a patient. We are entering a new era of health care that is more patient centered.  Part of this new paradigm is patient access to their medical records,  easily accessible from the Internet.

While physicians possess the ability to distinguish between an occasional proofing error, patients can not, nor should they be expected to. In order to increase the efficiency of radiology reporting, most institutions have moved to speech recognition (SR) software with templated macros. The potential for typographical errors has been shown to be higher using SR software than traditional transcription services.

However, there is a paucity of literature regarding how often errors occur because of the use of pre-populated macros. I learned first hand how my words can affect a patient’s mental well being by inducing unnecessary anxiety. More and more pressure is being placed on radiologists to increase 

 However, patients will not and should not excuse us from errors in what we report. In an era of more open access to all medical records, including radiology dictations, it will be critical for radiologists to take pause and examine their finalized reports carefully.

[1] J Digit Imaging. 2008 Dec;21(4):384-9. Voice recognition dictation: radiologist as transcriptionist.

We are entering a new era of health care that is more patient centered. Part of this new paradigm is patient access to their medical records, easily accessible from the Internet.

Why Radiologists Should Review Reports Carefully

“I have reviewed your scrotal ultrasound and unfortunately I see some very suspicious findings which will need additional work up,” I said to the eager and nervous 32 year old man who I was speaking to on the telephone.

“What exactly does suspicious mean?” he replied.  “Please, doctor, tell me the truth, you are a guy, just tell it to me straight.”

“Robert, the findings on your scan most likely represent a tumor in your left testicle.”

There was silence, then Robert replied, “Is it curable?” 

Robert’s testicular ultrasound revealed diffuse microlithiasis and a circumscribed hypoechoic mass in the lest testicle, most suggestive of a seminoma or mixed germ cell tumor.  I proceeded to tell Robert everything I knew about testicular cancer in young men.  I tried to confine my conversation to the imaging work up he would undergo (my area of expertise) but also reviewed some basic epidemiology regarding testicular neoplasms.   I also spoke generally about treatment options including orchiectomy, radiation, and chemotherapy.  However, I stressed to Robert that the most important prognostic factor would be determining whether the cancer had spread, a determination that would be made by further imaging.

Unfortunately I had to speak to Robert over the telephone rather than in person.  His case came to my attention at 5:30 pm on a Friday afternoon, the typical time difficult cases seem to appear.  Robert was the last case of the day scanned by one our sonographers in an outpatient setting where the covering Radiologist had left for the day.  The sonographer phoned me immediately after the scan because of the concerning finding.  Being on call, I could not leave the inpatient tower and go to speak to Robert in person.   I regret having to convey such important news over the telephone, but I wanted Robert to know the results of his study rather than worry all weekend and not hear from anyone.  I also provided Robert with some reliable web resources he could review and my pager number if he had any questions over the weekend.

I had paged Robert’s primary care physician with the findings initially, but she was not available.  I was transferred to a covering physician who knew even less than I did about Robert.  She dutifully recorded my findings and recommendations and said she would tell Robert’s primary but did not offer to speak to Robert herself.  The sonographer had informed me when she called that Robert was anxious and wanted to speak to someone so I agreed, as a physician, to convey to him what I saw and its implications.

I wrote about a scenario regarding a patient I cared for where I knew the diagnosis was cancer but retreated from discussing this finding with the patient

.  Multiple responses to that blog post debated whether Radiologists should or are equipped to discuss this type of diagnosis with a patient.  Since caring for that patient in November I had regretted not having a discussion with the patient.   I told myself that the next time the situation came up I would speak directly with the patient.  While I would have rather spoken to Robert in person, I am still glad I spoke to him about his diagnosis based on my imaging expertise. 

“Thank you, Doctor,” Robert said at the end of our conversation.  “I appreciate you discussing the findings with me honestly.”   

One radiologist is determined to discuss results directly with patients.

Arun Krishnaraj, MD, MPH

Articles

Speech Recognition Technology - Finally Ready for Prime Time?

Consistency in Reporting and Communication: Thyroid Ultrasound Imaging

Why Radiologists Should Review Reports Carefully

Stepping Out From Behind Our Reports