NEW YORK CITY - Who really "owns" medical images? And why isn't image compression used more often? These are a few of the questions addressed by Dr. Eliot Siegel in a talk on Monday at the New York Medical Imaging Informatics Symposium (NYMIIS), which we're covering in a two-part series on medicolegal concerns in PACS.
Every facility using PACS must navigate image retention requirements and grasp the rights and responsibilities associated with "ownership" of images. An understanding of the benefits of lossy compression is also helpful, said Siegel, who holds appointments at the University of Maryland and the Baltimore Veterans Affairs (VA) Medical Center.
Who owns the images?
Patients are often confused about who has ownership of their images and the extent to which they may exert rights over their records, Siegel said. He collaborated on his NYMIIS 2015 presentation with Dr. Jonathan Mezrich, JD, of the University of Maryland.
Dr. Eliot Siegel from the University of Maryland.While access to health information is regulated by HIPAA and the Health Information Technology for Economic and Clinical Health (HITECH) Act of 2009, this protection for patients does not constitute "ownership." In fact, patient rights to the underlying records are limited, he said.
In general, the facility that performs the imaging maintains "ownership" rights to the images. Individuals have the right to inspect their images and obtain copies, but they may not have their medical records modified or stricken, Siegel said.
"An individual may request that a medical record be amended to correct inaccuracies, but if a provider does not agree, the patient's recourse is limited to submitting a statement of disagreement into the record," he said. "Individuals in a hospital's care are not purchasing the medical records or images, but rather the professional services of individuals who generate opinions on the basis of those records."
Facilities have the obligation to maintain copies of all images for the purpose of comparison studies, as evidence in malpractice suits, and to comply with statutory requirements for record retention. Medicare and most states require that all medical records be maintained for at least five years, but longer retention periods may be necessary for specific types of images such as mammography, in cases related to toxic exposure, or for patients who are minors, Siegel said. State statutes can also impose fines for failure to retain imaging.
Images associated with transferred patients can pose a management dilemma. American College of Radiology (ACR) guidelines indicate that facilities receiving studies generated at another site are not required to store images that are already being archived at the transmitting site, Siegel noted. However, a facility that relies on those images in making a patient decision probably needs to maintain a copy for malpractice defense reasons.
The legal concept of "spoliation" provides that when evidence -- including imaging -- is lost, destroyed, or suppressed, a jury is entitled to presume that the missing evidence would be unfavorable to the party responsible for its unavailability, Siegel said.
For the purpose of teaching, research, and quality, HIPAA allows facilities to broadly use or disclose "deidentified health information" even when the information is not entirely deidentified, he said. A facility may still use limited dataset information without a patient's consent for purposes of research, public health, and healthcare operations.
Bankruptcy issues
When an imaging center goes out of business, U.S. bankruptcy law applies. If a healthcare provider files for bankruptcy and lacks the funds to store its images, the bankruptcy court trustee will give notice to patients that that they have one year to claim their records before they are destroyed. An ombudsman may also be appointed to monitor the quality of patient care and represent patient interests, Siegel said.
Some states also provide for the appointment of a custodian of records if a facility closes outside of federal bankruptcy, while other states may provide for the transfer of records to another facility in the same vicinity, he said. In some jurisdictions, the department of health itself may become involved in the storage of medical records.
In a recent online survey conducted among members of the general public, the majority of laypeople actually understood that the facility that generated their imaging study retained ownership rights to those images, Siegel said. However, more than half believed -- incorrectly -- that their images could not be used for any purpose without their consent.
"It's really important that clinicians educate patients regarding the limitations and extent of their control over their own imaging," he said.
A number of interesting image storage questions are emerging, according to Siegel.
"As the ability to store large images and large quantities of digital imaging increases, one question is: Is the five-year retention [requirement] something that we should revisit?" he said.
Other issues to be addressed include the following:
- Should lossy compression be allowed on images older than 5 years?
- What are the requirements for notifying patients that images are being deleted?
- Who is responsible for maintaining images when they are uploaded to a cloud platform and are accessible by multiple PACS databases?
Image compression
The legality of image compression in medical imaging often generates a lot of controversy and has held back its utilization.
"There's an incredible mismatch between the technology that's available to compress and the fact that practically nobody in the U.S. is doing image compression," Siegel said.
Lossless compression methods can achieve compression ratios of 2:1 to 4:1 in medical images without losing any data, but lossy -- or irreversible -- techniques can yield much higher ratios. While these lossy techniques may not perfectly reproduce the original image, it may be that no image degradation is perceptible and diagnostic value is not compromised.
At the same time, there's also a significant mismatch between clinical image production and expectations for storage and network capacity and connectivity bandwidth, he said.
Current enterprise demands often meet or exceed the ability of traditional lossless compression methods or models that require delivery of the complete study to begin image interpretation, Siegel said. For example, the volume of data created at the University of Maryland and the Baltimore VA Medical Center has increased per month by as much as 50-fold over the past 20 years.
Out of the hands of radiologists
Siegel also noted that decisions about image compression are increasingly being taken out of the hands of radiologists and radiologic technologists -- to be made instead by nonclinicians.
"I personally think a lot of those decisions should probably come back [to radiologists] because a lot of those are clinical decisions," he said.
ACR's stance on compression is that data compression may be performed to facilitate transmission and storage. Several methods -- including reversible and nonreversible techniques -- may be used under the direction of a qualified physician with no reduction in clinically diagnostic image quality, Siegel said. Furthermore, the types and ratios of compression used for different imaging studies should be selected and periodically reviewed by the responsible physician to ensure appropriate clinical image quality.
The U.S. Food and Drug Administration (FDA) is also OK with the use of irreversible compression as long as it's noted when the images are displayed. The technique and ratio are left to the radiologist's discretion. Per the Mammography Quality Standards Act (MQSA), however, lossy compression cannot be used with mammography.
Siegel pointed out that many factors contribute to the final image, such as scratches and physical imperfections on a computed radiography (CR) plate, radiologist visual acuity limitations, dirt or dust on a monitor, and vendor image processing algorithms.
"Image compression that's typically used clinically and may be 10:1 or 20:1, might be a 1% change in comparison with all of those other factors," he said. "However, image compression from a medicolegal perspective is the one thing that everybody really pays attention to."
Compression pros and cons
Those who don't advocate the use of lossy compression point to the ever-decreasing cost of storage space and constantly improving network speeds. However, the number and size of images are increasing as fast as or faster than storage costs are dropping. Furthermore, the network bandwidth at most facilities is reaching a plateau, even while they are facing big demands such as multislice CT, Siegel said.
"What's happening is that we're starting to see a bottleneck now and again of images waiting at scanners to be transferred to a PACS, waiting to be transferred to a network," he said. "Image compression would allow us to be able to deal with that."
More than 1,000 articles in the imaging literature have concluded that the loss of image data is unlikely to be clinically relevant at lossy compression ratios up to 20:1 for wavelet compression for computed radiography (CR) images and up to 8:1 for CT, ultrasound, and MRI.
"Most of the research was performed 15 years or more ago, and by comparison there has been very limited research in a lot of other areas of image processing and image modification," Siegel said.
Some critics believe the liability risk is too high to use lossy compression; they fear having to go to court after a missed nodule or pneumothorax and having the compression algorithm blamed for the miss, according to Siegel
"That argument has been made for years, and there has never been a single case, to my knowledge, that has hinged on image compression," he said. "Typically, lawyers and courts have not interfered with technical image production and display."
It's an interesting irony that for medicolegal reasons, endoscopists, bronchoscopists, and cardiologists typically don't keep any images at all or only a small subset, and there is much less worry about clinically lossless compression, he noted. Scans using mild lossy compression have also been found in some studies to be rated higher by radiologists than uncompressed images.
"I strongly believe that the responsibility for the decision, as the ACR recommends, should be 'under the direction of a qualified physician' with radiologists unquestionably the most qualified to determine how it should or should not be used," Siegel said. "The benefits of compression are not being realized fully in the PACS community."
Part 2 of AuntMinnie.com's coverage of Dr. Siegel's talk at NYMIIS 2015 will highlight medicolegal issues related to computer-aided detection (CAD) software.
