Safety nets catch at-risk patients

All kinds of designated safety nets exist in health care, from clinicians to hospitals to insurance plans. But in primary care, there's not always a safety net to catch patients with abnormal test results who slip through the cracks.

Over the past four years, the ambulatory patient safety team at Brigham and Women's Hospital in Boston has worked to change that. The team piloted the concept of ambulatory safety nets, which are systems that ensure proper follow-up of abnormal test results, said rheumatologist Sonali P. Desai, MD, MPH, director of ambulatory patient safety at the hospital, at the 2019 annual meeting of the American Public Health Association.

“We're all human. Test results can get overlooked, so we need to have these closed-loop processes to make sure that it doesn't happen,” she said. “… We make sure, from the point at which it's abnormal to the point at which you get your care, that the whole process is as seamless as possible.”

Brigham and Women's first began to pilot its ambulatory safety nets for missed and delayed diagnoses of cancer in 2016, modeling its program on SureNet, an automated safety program at Kaiser Permanente designed and implemented by Michael Kanter, MD. But one key difference is that the Brigham and Women's program is not within an integrated health delivery system.

In addition to its hospital, Brigham and Women's has about 135 ambulatory practices and 3 million ambulatory visits per year, said Dr. Desai. About two-thirds of patients admitted to the hospital have a primary care doctor who is outside of the system, and a substantial amount of primary care patients who are in the system might be getting subspecialty care elsewhere in Boston, she noted. “All of those systems are not connected, so therein lies the gap and the challenge,” said Dr. Desai.

At the American Public Health Association's meeting, held in November 2019 in Philadelphia, Dr. Desai explained how the academic tertiary care center developed ambulatory safety nets for patients at risk for colon cancer and lung cancer. “We've chosen to focus on things that have high impact, in high-risk areas, which is why we're focusing on cancer,” she said.

Colon cancer screening

Getting patients to have a screening colonoscopy is one thing. Getting patients to have timely follow-up of potentially benign findings is another.

Of course, a colon polyp doesn't mean cancer. “But it does mean that you might need something else done in the future, and that something else comes one year, three years, five years, or seven years down the line,” Dr. Desai said. However, when it comes to keeping track of these patients, she said, “Most organizations don't really have a good system for doing that.”

This is where the ambulatory safety net comes in. The team created electronic registries of all of the system's patients who had a colonoscopy with pathology and generated a report of those who needed to come back based on something in the chart, Dr. Desai said. But first, this required culture change among clinicians.

“That something in the chart needs to be updated by a human being,” she said. “Either somebody who did the procedure or their primary care doctor needs to update something in the medical record that's coded that says, ‘Hey, this person is due to come back in three years.’”

The team decided that the gastroenterologist performing the procedure should be responsible for updating the chart. It took three years for gastroenterology colleagues to adopt workflows for updating the coded data element, Dr. Desai said. “Without them doing that piece, we can't really do anything in a consistent way on the primary care side,” she said. “This took a lot of back and forth with various teams and with individual clinicians to change the culture.”

Pulling these efforts together, the ambulatory patient safety team runs the report and performs a chart review of all the patients in the report to be sure the right patients are targeted. If any data don't line up, the team reaches out to the gastroenterologist or, in some cases, the primary care physician to verify that the patient needs to return for follow-up.

Once the list of patients is confirmed, those who need follow-up receive both a mailed letter and a phone call seven to 10 days later, where a patient navigator does motivational interviewing to ensure the patient knows to return to the hospital for a colonoscopy, Dr. Desai said. “Before, there was really not a proactive process to actually outreach to all of the at-risk patients,” she said.

Once patients are notified, the team works with gastroenterology and administrative scheduling colleagues to get the order from the primary care doctor, schedule the test, make sure the patient shows up, and record the outcome of the procedure. “Obviously, there are multiple steps in that process, and we've been working with every kind of group that's involved to make sure that this goes smoothly,” Dr. Desai said.

Over the year and a half that the safety net has been in place, at-risk patients have completed 261 colonoscopies, she said. In particular, adding a patient navigator improved patient uptake, Dr. Desai said. “At first, we didn't have a patient navigator; we just sent people letters. … That's not the most effective way of doing things,” she said. Overall, 44% of patients targeted by the safety net were scheduled for or completed a colonoscopy following the team's outreach, according to a study published in the August 2019 Joint Commission Journal on Quality and Patient Safety.

Dr. Desai also offered an example of a 28-year-old patient with Crohn's disease who was identified in the report. “I think we brought her in much sooner than she probably would have come in through our existing ways of doing things,” she said. “And, in fact, she had a high-grade dysplasia … that really necessitated her getting her entire colon removed. So this really changed an outcome for a very young patient.”

Lung cancer

A second safety net aims to prevent lung cancer by focusing on incidental lung nodules.

There are many reasons why incidental lung nodules may not be communicated to primary care physicians or patients, Dr. Desai said. “It definitely happens. There are multiple studies in the literature showing that even when a radiologist makes a clear recommendation on what to do for the nodule, [patients] don't get the follow-up. Sometimes it's 50% of the time,” she said.

Follow-up is also often not done for nodules because they are very common and guideline recommendations can be confusing, Dr. Desai said. “Also, we don't want to do overtesting,” she said. “We don't want to just scan everybody just because they had a nodule, because most of them are not significant or don't lead to cancer down the road.”

The team spent a couple of years analyzing the problem and working closely with a primary care advisory council to garner feedback, Dr. Desai said. “We also found out that radiologists don't always put recommendations that are clear in terms of when the patient should come back for repeat testing,” she said. Initially, Dr. Desai and her team worked with the radiology team, using artificial intelligence tools to identify incidental lung nodules on radiology reports, conducting manual chart reviews, and communicating with primary care physicians about the need for follow-up chest CT imaging. About 57% of identified patients were scheduled for or completed a chest CT, according to the Joint Commission Journal on Quality and Patient Safety study.

To create a more robust and reliable lung cancer safety net in March 2018, the patient safety team worked with several chest radiologists to ensure they were making explicit recommendations about lung nodules that they believed required follow-up. The team launched a program called RADAR, which stands for Result Alert and Development of Automated Resolution. Through a web-based radiology result notification system, the program alerts ordering clinicians of the incidental lung nodule results and the radiologist's recommendation, as well as assists with scheduling a follow-up.

Compared to alerts for incidental lung nodules generated with a previous software tool, RADAR alerts had more explicit documentation of imaging modality and an improved time frame for follow-up (71% vs. 100%), according to a study published in the May 2019 American Journal of Roentgenology.

“We found that people are more likely to get their follow-up study done in a reliable time frame, so in a couple days versus three weeks in terms of scheduling, and that they were just more likely to complete it,” said Dr. Desai, adding that the team is about to analyze its first full year of data.

New safety nets that the team is working on include medication safety, prostate cancer, cervical cancer, breast cancer, and diagnostic errors, said Dr. Desai. Currently, the team includes one project manager, one patient navigator, and one pharmacist and takes up about 35% of Dr. Desai's time. “For an organization of our size and the number of practices, it's not a large team,” she said. “But we're still able to get quite a bit done.”