Sen. Arlen Specter’s field hearing on the Philadelphia VA mess is set to begin today. It will be interesting to see if any new details will come out. Apparently, Dr. Kao will be answering questions at the hearing, so at least we will hear his side of the story. Before the spinning starts, though, we can look at the NRC special inspection report to get a better idea of the facts of the case. The NRC report can be found by going the NRC ADAMS search site and searching for document accession number ML090900382.
The report states that the investigators from the NRC looking into the prostate brachytherapy program at the Philadelphia VA found 92 medical events that occurred between February 2002 and June 5, 2008. The NRC defines a medical event as a treatment that failed to deliver dose according to the physician’s prescription. A medical event can be triggered for various reasons. One example is if the dose administered to a patient differs from the prescribed dose by at least 20 percent, either too high or too low. Another is if the patient receives a dose to a part of the body other than the intended treatment site that exceeds by 50 percent or more the dose expected by proper administration of the prescription. A medical event does not necessarily mean that a patient was harmed; only that there was a technical problem with the delivery. Still, a medical event is a big deal, and it must be reported to the hospital’s regulatory agency, either federal or state, within 15 days.
Fortunately, medical events are rare. There were only 100 brachytherapy related medical events in the entire US in 2008, out of 50,000 procedures. Ninety two in a six year span is a staggeringly high number; even more so since there were only 114 procedures performed. How could such a high percentage of treatments be of low quality? The NRC found six violations of federal regulations and several more areas of concern. The violations centered around one main issue: the failure to ensure the ultimate quality of the treatment.
The first two violations listed were “failure to develop adequate written procedures to provide high confidence that each prostate brachytherapy treatment administration is in accordance with the [physician's] written directive” and “[failure] to develop procedures that address methods for verifying that administration is in accordance with the treatment plan and written directive.” For prostate brachytherapy procedures, standard practice is to obtain at least one CT scan of a patient some time after the operation to determine the final placement of seeds. These images are sent to a treatment planning computer that can reconstruct the dose delivered to the patient from the seeds’ position. This can verify that the proper dose was delivered to the prostate and that the dose to the surrounding organs was sufficiently low. The investigators found that post implant CTs were performed, but due to a computer interface problem, the images were never sent to the treatment planning computer and no dose reconstruction was performed. Damningly, even after the interface issue was fixed, post treatment dose reconstruction was not performed on seven patients. Apparently, the department’s written procedures did not specify that this task was to be performed.
When investigators were alerted to the problem, patients were brought back to the center, given a CT scan, and their dose reconstructed. This is how the NRC determined that their procedure was performed incorrectly. They found that 35 patients had a dose to an organ that was more than 50% higher than intended, and that 57 recieved a dose to the prostate that was more than 20% low. There are other violations relating to improper training and failure to report the medical events, but report blames the failure to perform post implant dosimetry as the main culprit. If the dose had been reconstructed, the low quality of the implants would have been discovered much sooner.
The report goes on to blame the lack of peer review of the procedures and the lack of a “safety culture” for the errors. Two of the physicists involved tried to raise concerns, one to the physician involved and the other to a physician at an affiliated institution. Nothing came of it, and the concerns went no farther. Each facility is also required to perform a quarterly audit of conformance of treatments to the physician’s written directive. These audits failed to uncover the problem. Instead they stated that the treatments were ok, even while post implant CTs were unable to be performed and the dose distribution could not be checked. The report states that these issues pointed to a lack of a “safety culture for reporting radiation concerns to the appropriate individuals”.
To sum up, the two problems that are most blamed for the errors are the lack of post implant dosimetry and the lack of training and oversight by the radiation safety committee. I am guessing that the blame for those issues will be laid at the feet of the medical physics staff. They were the ones who are responsible for calculating the dose using the post implant CTs and for auditing the treatments. While I certainly agree that not performing these tasks correctly was a serious error, I am not convinced that this is where the ultimate responsibility lies.
If you recall, the NRC investigated this center twice before it was ultimately shut down. Each visit was triggered by a procedure where half of the seeds ended up in the bladder instead of the prostate. In each case, the investigation was closed when the physician stated that he changed his written directive to reflect the lack of available seeds. As I discussed before, the NRC does not have the authority to second guess the doctor’s decision. They were forced to defer to his decision. However, placing 40 seeds in a patient’s bladder should have been a clear signal to Dr. Kao that he was doing something terribly wrong. The technical staff involved should definitely have spoken up in a more effective manner, but his refusal to use fluoroscopy was the ultimate root of all of the problems.
Doctors are given a lot of deference in the clinic, and rightly so. The responsibility for the patient’s outcome is theirs. Therefore, they have the final word on how a patient will be treated. Of course, that deference should only go so far, and a physician that is acting in an unsafe manner needs to be stopped. Unfortunately, crossing a doctor is an easy way to end your career. Dr. Kao placed the staff of the Philadelphia VA in a very difficult situation. They should have a share of the blame, but the real problem was Dr. Kao’s poor technique. I hope that he is appearing at the hearing today to accept some responsibility and not to throw the physics staff under the bus.
