Standardized Imaging Protocols - The Next Step For Providers Seeking Consistency, Added Value

Over the last several years, there has been an increased focus throughout the imaging industry on imaging protocol standardization, or more often, the lack thereof.

At the University of Texas Southwestern (UTSW) Medical Center in Dallas, for instance, problems related to protocol standardization

stemmed from the fact that it belongs to a massive health system made up of four large hospitals, five outpatient centers and several remote locations. The system also maintains three PACS, 21 CT scanners and numerous other scanners, which are all different ages and were designed by a variety of manufacturers.

Prabhakar Rajiah, MD, associate professor of radiology and associate director of cardiac CT and MR in UTSW’s department of radiology, says things were simply getting out of hand. Patients were receiving different radiation doses for the same protocols at different sites, for instance, and image quality was inconsistent.

“We have a very complex health system and we basically found that our protocols are all over the place,” Rajiah says.

With this in mind, UTSW created a CT radiation task force with the purpose of:

  • Optimizing patient radiation dose across scanners
  • Standardizing and homogenizing scan protocols and
  • their names
  • Maintaining/improving image quality
  • Establishing mechanisms to continually track dose
  • Establishing reliable training and dissemination processes
  • Making protocols readily available
  • Ensuring adherence to imaging protocols

According to Rajiah, the CT radiation task force “homogenized” protocols by reviewing existing protocols, eliminating redundant protocols across all sites and optimizing radiation dose within each one. Studies were selected for each protocol and each scanner and then evaluated by a radiologist, who focused on image quality. Next, a physicist would quantify the radiation dose and image quality using anthropomorphic phantoms.

“So a combination of radiologists and physicists were able to optimize the protocols, selecting the least possible dose for a particular study,” Rajiah says. That optimized protocol, he added, was then programmed into the scanner.

Optimizing all of the protocols took about nine months, and the team succeeded in reducing the total number of CT protocol types from 222 to 136. In addition, the initiative resulted in reductions in radiation dose ranging from 23 percent to 58 percent, depending on the protocol.

“It is a model that can be adapted in other systems,” Rajiah says. “So [radiologists and technologists] can maintain a homogenous protocol and also ensure that they provide a high quality service without any adverse effects.”

While much attention has been paid to the need to standardize CT protocols in order to achieve the lowest possible radiation dose while achieving images of diagnostic quality, standardizing imaging protocols can provide other benefits as well.

[[{"fid":"22563","view_mode":"default","type":"media","attributes":{"height":355,"width":460,"style":"height: 139px; width: 180px; margin: 5px 10px; float: left;","alt":"p-rajiah.jpg - p-rajiah","class":"media-element file-default"}}]]

Maintaining the brand

After more than two decades as a standalone academic site, the University of Texas MD Anderson Cancer Center’s breast radiology section has recently expanded its reach to become the exclusive provider of professional breast radiology services to 15 community breast care centers around the Houston region.

“When the opportunity for partnership came along, we anticipated there would be a lot of growth in terms of faculty recruitment, as well as the number of patients that are seen, and ultimately the number of technologists we interact with,” says Megan Kalambo, MD, department of breast imaging at the University of Texas MD Anderson Cancer Center.

It also became clear that steps would have to be taken to standardize clinical care among these community breast care centers. “The greatest benefit to [standardization] is reproducibility,” says Kalambo. “It ensures that every patient who comes into our system and sees an MD Anderson breast imaging radiologist will receive the same care, regardless of which facility they go to.”

It always helps the facility maintain its brand, she adds. “As an imaging facility, MD Anderson has a particular brand—a typical way that we actually work patients up—and we wanted to make sure our patients have the same experience regardless of what facility they step into.”

At MD Anderson, Kalambo says that the feeling among radiologists in the breast radiology section was that the best way to move forward was to generate some kind of an imaging protocol document.

To accomplish that goal, Kalambo and her colleagues set up a standardization committee of seven radiologists to generate the document. Initial steps were taken to come up with an outline identifying the most commonly encountered breast imaging care line services and each committee member was tasked to tackle various modalities such as screening mammography, diagnostic mammography, breast MRI, and breast ultrasound.

Each section of the protocol document was drafted with current American College of Radiology and National Cancer Center Network Clinical Practice guidelines as a framework. The sections were then formatted to include a list of approved indications, a procedural protocol with detailed information on the images that are required for interpretation, and information about how the technologist should implement the exam. Once reviewed and approved by the entire faculty, the document was distributed to all community practice sites.

Kalambo also co-authored an article in the Journal of the American College of Radiology about the entire process, giving specialists who encounter this issue in the future an additional resource to help them get started (J Am Coll Radiol. 2016 Nov 3. pii: S1546-1440(16)30630-5).

According to Kalambo, the document she and her colleagues have put together is not final, and updates can be made as necessary. “The protocol document is a living, breathing document,” she says.

For example, after implementing the guidelines, it was determined that some of the existing protocols driven by the academic main campus needed some tailoring when applied in the community setting.

One change her team made was related to whole-breast sonography. “Our approach to working up patients was much more comprehensive when working at a single facility,” Kalambo says. “In terms of technique with ultrasound, we do whole breast ultrasound on all patients who come to the main campus. But, we realized very quickly that when we went out into the community that this particular protocol wasn’t really needed or time efficient, because the patient population was much healthier. So we don’t have to do as extensive a workup on these particular patients.”

[[{"fid":"22564","view_mode":"default","type":"media","attributes":{"height":355,"width":460,"style":"height: 139px; width: 180px; margin: 5px 10px; float: left;","alt":"m-kalambo.jpg - m-kalambo","class":"media-element file-default"}}]]

Protocol creep

Like UT Southwestern and MD Anderson, University of Colorado Health—a five-hospital provider network—also is dealing with the necessity of managing imaging protocols in a large and growing system.

“I think large or small, anytime you have more than one radiologist, you’re going to see protocols being fussed with,” says Peter Sachs, MD, department of radiology at the University of Colorado School of Medicine, “And I think in large health systems, everyone has this problem and is thinking about how to deal with it.”

Beware of “protocol creep,” Sachs warns. It tends to happen when departments don’t have an organized system of protocols and radiologists and technologists end up changing protocols on a case-by-case basis. “A radiologist might decide he would like to do an extra sequence, or one with thinner slices than what is standard,” Sachs says. “And when that happens, there are some unpleasant downstream effects.”

One example of the confusion this can create, Sachs explains, is that a technologist can end up with multiple protocols on the same scanner for the same study type and indication. “So the next time a technologist goes to do a study there could be two, or four, or 16, or 64 different versions of this CT of the abdomen for liver cancer, and he or she may have a hard time figuring out which one to do,” he says. “Then if they do ‘the Dr. Sachs protocol,’ and it turns out I’m out of town and Dr. Smith is reading the exam, all of a sudden he’s going to see that the images look different.”

Using too many protocols also can cause issues for referring physicians, because two different studies performed on the same patient for the same indication can end up looking so different that a side-by-side comparison is almost impossible. Even patients can be impacted by protocol creep, Sachs says, since patients “are much more educated and involved in their own healthcare” in today’s healthcare environment than in the past. 

Protocol creep can also lead to direct issues with image quality. “There’s pretty good evidence of what an imaging protocol should look like to answer a question or evaluate the body part you are looking at,” Sachs says. “And when you deviate from that, the image quality suffers, and your diagnostic accuracy decreases.”

A group effort

At the University of Colorado, the goal was to come up with a set of standardized CT and MR protocols. That process was led by the radiology department’s full-time process improvement consultant and involved two sets of meetings—one for MR and one for CT. Each meeting —called a rapid improvement event or RIE—was attended by a section specific radiologist from each region of the hospital network, as well as one or two lead technologists from each region.

To allow for any changes in electronic storage and management of protocols, standardized protocols for MR and CT were formatted in Excel. Similar protocols for the same indication were matched and reviewed and a protocol review list distributed to each of the CT and MR team members.

“We got a consensus on what standard imaging protocol should like for CT, then did the same for MRI” says Sachs, who co-authored an article for the Journal of Digital Imaging about his team’s experience (J Digit Imaging. 2016 Jul 22). “And then we had to get all that information into an electronic format.”

Sachs says they also had to take into account that the system had multiple scanners from multiple vendors, “which meant these [standardized] protocols had to be sent back to the technologists so they could enter any machine-specific parameters.” Finally, the protocols had to be submitted to a physicist for review.

The protocol standardization process succeeded in reducing the number of CT protocols from 248 to 97, while 168 MR protocols were combined into 66 protocols.

What was done at the University of Colorado Health is something “that can be done anywhere,” says Sachs. “The real issue is once you do this, how do you prevent protocol creep from occurring again. Other people have working to standardize protocols and the honest answer is we don’t have a guarantee that [protocol creep] won’t reoccur.”

One way the University of Colorado Health is addressing this issue is the establishment of a protocol governance committee, which consists of one radiologist, one CT technologist and on MRT technologists from each of the system’s three regions, as well as two ad hoc members.

“In the past, technologists have been reluctant to say no when a radiologist wants to change something,” Sachs says. “We as a system are saying you can’t do that anymore, and we have empowered our technologists to tell radiologists that there is now a process in place for requesting that change.”

Sachs noted that standardizing protocols within the electronic health record (EHR) could result in other benefits as well, including “some nice scheduling and patient communication tricks.”

For example, he explains, if one patient comes in to have his abdomen scanned for cancer as another patient in for a scan because of fever and abdominal pain, those exams are going to follow different protocols and won’t take the same amount of time.

“By creating unique imaging protocols, we can electronically handle those scheduling events so that the system schedules it in the appropriate room, for the appropriate time slot,” Sachs says.  In addition, he pointed out it would be possible to communicate through the patient portal in the EHR what the actual preparation for a scan might be, considering that the preparation for a CT scan of the abdomen for cancer may be different from that of another CT scan of the abdomen.

[[{"fid":"22565","view_mode":"default","type":"media","attributes":{"height":355,"width":460,"style":"height: 139px; width: 180px; margin: 5px 10px; float: left;","alt":"p-sachs.jpg - p-sachs","class":"media-element file-default"}}]]

Reducing dose, one protocol at a time

Yet another reason more providers are working to standardize imaging protocols is that it helps them complete their goal of reducing radiation dose for all patients.

Rebecca Smith-Bindman, MD, a professor in the departments of radiology, epidemiology and biostatistics, obstetrics, gynecology and reproductive medicine at the University of California, San Francisco, has focused much of her research assessing the amount of radiation patients receive when they undergo CT scans. It turns out, she says, that “when we look at doses that we use for CT across institutions, we see that they are profoundly variable and higher than they should be for diagnosis.”

In 2009, Smith-Bindman and her colleagues published a landmark study in the Archives of Internal Medicine that evaluated the radiation dose associated with the 11 most common types of CT exams on more than 1,000 patients in four institutions in the San Francisco Bay area (Arch Intern Med. 2009 Dec 14;169(22):2078-86). They found that not only did radiation doses vary significantly between the different types of CT studies, but that they varied within CT study types as well.

Why do these profound differences exist? While there are many factors that go into scanning individual patients, Smith-Bindman says a key reason we see this variability is the large number of existing CT protocols.  And when there are a lot of protocols for specialists to juggle, it makes reducing patient dose much more of a challenge.

“The number of protocols we have is enormous—far greater than what we need for the clinical/patient problems we are trying to scan for,” she says. “And because there are so many, and because facilities have such different ones, it makes the whole process of standardizing protocols really tricky. And in general, I don’t think we are doing a very good job of it all.”

In Smith-Bindman’s own experience, the facilities that are successful in reducing the amount of dose in their CT exams are those with fewer protocols.

As leading radiology departments and health systems build and refine protocols, those at the front of the pack are already seeing benefits in both dose and department efficiency and image quality. Relationships with referring physicians and scanner utilization are improving as well. It may be time for all imaging leaders to take a closer look at their own protocols.