College of American Pathologists
Printable Version

  Sealing the cracks in test result followup


CAP Today




July 2011
Feature Story

Karen Titus

The core laboratory at Massachusetts General Hospital unleashes 8 million test results a year.

Then what?

Ordering physicians get the results. They contact their patients. Decisions are made, treatments are given.

Unless a result falls through the cracks. And when that result is actionable—the word favored by Anand Dighe, MD, PhD—the stakes are high.

Speaking at the CAP Foundation’s Futurescape of Pathology conference in April, Dr. Dighe, director of the core laboratory at MGH and assistant professor of pathology, Harvard Medical School, talked about the life cycle of actionable results and how his colleagues at Partners HealthCare are closing the loop to make sure every last such result—be it from pathology, radiology, or cardiology—gets where it needs to go. Closure, as it turns out, is not solely the realm of ending a bad relationship.

When test results trail off and, for all practical purposes float away, like the notes in an Arvo Pärt composition, the results are anything but dreamy. Dr. Dighe presented several nightmarish tales from recent case reports. One dropped ball, as he called them, was an elevated PSA in a 54-year-old man, with “no evidence that it was ever reviewed by anyone,” Dr. Dighe said. Two years later, the patient presented with widespread metastatic prostate cancer.

Another case involved a patient presenting with stomach pain, who was eventually diagnosed with acute cholecystitis. The radiologist reviewing the CT scan noted an incidental ovarian mass, but the finding was not followed up in outpatient care. Twelve months later came the diagnosis of ovarian cancer.

Followup failure may be even more prevalent when it comes to transitions of care. He cited the case of a 45-year-old man admitted with fever. He was discharged with pending blood cultures, which were positive for MRSA. By now, the pattern should be clear: The results weren’t followed up on, said Dr. Dighe. The patient was readmitted a month later with septic shoulder and spinal osteomyelitis.

These are no ordinary nightmares, with obvious horrors. Rather, they take a turn worthy of a David Lynch movie, where few folks notice what problems lie below a fine-seeming surface.

Everyone already knows to act on a critical result, thanks in part to the emphasis the CAP and Joint Commission have placed on reporting critical values, Dr. Dighe says. “For the most part we are all doing well on critical values, or we wouldn’t be passing our inspections.”

Rather, “We’re talking about a whole host of results that may not pose immediate risk but require action to prevent morbidity and mortality.” These include more than laboratory results, Dr. Dighe made clear. “Examples would be a highly elevated TSH, an abnormal Pap smear, a pulmonary nodule.”

What may also be hidden to some—though it shouldn’t be, according to Dr. Dighe—is the extent of the problem. Are these cases the tip of the iceberg? “Or are these just a few ice cubes in a big ocean?” he asked, one where pathologists and their colleagues generally are swimming along just fine.

Dr. Dighe sees an iceberg. A study from the Archives of Internal Medicine (2009;169: 1123–1129) looked at how 23 different groups of clinicians managed outpatient test results. Seven percent of the time—the median failure rate—there was no evidence of communication to the patient about an actionable test result, Dr. Dighe said. Furthermore, “What was interesting was that groups that heavily used their electronic medical records didn’t end up doing substantially better than places with paper-based records.”

“So the problem is there. The problem is real,” he said. And, he added, “A seven percent failure rate is completely unacceptable” for actionable test results.

Common wisdom holds that the root cause of medical error is failed communication, he said. But why do those errors occur? He maintains they take root when there’s no robust process—whether an IT process or paper one—to make sure every step of the journey is completed:

  • order the right test
  • complete the test
  • review the results
  • communicate the results to the patient
  • make a decision based on results, then make sure it’s carried out.

Every step counts. Test results, like nature, abhor a vacuum. If producing them is the only step labs worry about, they’ll miss the boat, Dr. Dighe said—or sail it toward that iceberg.

With federal mandates breathing down the necks of health care providers, that Weltanschauung is changing. Readmission rates are becoming a well-eyed metric, and payers are refusing to reimburse for errors. Dr. Dighe would like to see laboratories help pilot the boat, advocating for such changes, rather than reacting to—or possibly resisting—government edicts.

For all the standard spiels of what makes a good leader, Dr. Dighe said it’s not all that complicated. Yes, there needs to be talk of shared vision and building infrastructure around that vision, and of adding value. In practical terms, however, “You don’t have to look far to figure out what you should be doing,” he said. Look at the folks who are taking care of patients on the front lines.

In both CP and AP, he said, the first step should be to look at how orders come in. Are physicians ordering the right tests? Do the orders come in without errors? Are specimen containers already bar-coded when they reach the lab?

Take a look at computerized provider order entry, too. Its use remains small—less than 20 percent of hospitals nationwide, in a recent study Dr. Dighe mentioned. But by eliminating free-text orders and creating bar codes at the point of care, CPOE can eliminate a huge expense in hospitals: reworking flawed orders.

CPOE doesn’t offer immediate redemption because workarounds can always be cooked up. Thus, CPOE systems must be designed with flexibility and monitoring tools to permit continuous improvements to the system. When Partners began using its newly designed inpatient laboratory CPOE a few years ago, Dr. Dighe said he was dismayed by the 10 percent rate of free text usage on day one. “On the other hand, I was delighted because we knew there was a 10 percent rate on day one. We didn’t have to wait six months or hear complaints about the system, and we could set about fixing it right away.”

When he and his colleagues looked at the order details, they quickly discovered the root of the problem. Today, thanks to improvements in search algorithms and feedback, free-text orders are rare and usually occur appropriately when a physician requests an esoteric test or when a researcher is doing a particular study. “We’re now at about 4 Sigma (one in 1,000) for the use of free-text orders,” he said.

With ordering now in order, so to speak, it’s been easier to make the other steps fall into place. “Patient ID is a lot easier when you have a bar-coded label coming out of your LIS in the clinical area. Specimen collection is a lot easier when you have a bar-coded label that says, ‘Put it in formalin,’ ‘Put it in saline,’ ‘Put it in a purple top,’ ‘Put it in a red top,’ ” said Dr. Dighe, who noted that wrong tube errors have dropped almost 90 percent in units that are live on order entry.

But none of this closes the loop.

No matter how much effort the laboratory makes to provide a result, there’s no way to automatically ensure that a clinical action will follow.

Dr. Dighe urged laboratories to think of themselves in a larger context. “We lump ourselves in with the radiologists and diagnostic cardiologists in terms of how we handle test results.” They share similar concerns: order entry, decision support, patient ID, and generating and managing results. “The same steps are involved,” Dr. Dighe said.

Partners created an Actionable Test Results Task Force, chaired by Dr. Dighe and Christopher Roy, MD, a hospitalist at Brigham and Wo­men’s Hospital. The task force includes members from every hospital in the group—“We did not want to get vetoed by anyone,” Dr. Dighe said. At the table: “pathology, cardiology, and radiology, which make up the bulk of results in any patient’s medical record.”

The task force knew what it didn’t want to do—become a permanent hospital committee producing “white papers that could be summarily ignored.” It wanted a short lifespan, during which it would create concrete recommendations that in turn would be put in place by a permanent, corporate-level team at Partners.

It’s a work in progress, but Dr. Dighe clearly is proud of the progress so far.

There was nothing glamorous about one urgent task: deciding what counts as an actionable test result. “Everyone agrees that a potassium of 8 is an actionable test result, but what about a TSH? And what level of TSH?” Dr. Dighe asked. “And what about a vancomycin peak? What should be the cutoff before we alert somebody?

“All these questions had to be painfully asked,” he said. And they were, with laboratory and pathology directors gathered around the virtual table and hammering it out over six months. They decided on three levels of responses. Potassium results, for example, are considered level one and would require interruptive notification within 60 minutes. A level two result, such as MRSA-positive swabs, would require a followup within eight hours, while a level three result, such as a pulmonary nodule, would require action within 14 days.

The task force also examined their existing information systems to ensure they could support the project. “The goal here is to work with our information systems to permanently affix a criticality flag to actionable test results,” Dr. Dighe explained. “If we don’t do this, the test result risks getting lost to followup.”

Partners has an outpatient longitudinal medical record, or LMR, application for test result management known as Result Manager. “In the current iteration of Result Manager, you can acknowledge the results—but acknowledgement isn’t enough. Just because someone sees a result doesn’t mean they do anything,” Dr. Dighe said. The goal is that providers, with a single added click, will be able to add an actionable result to the patient’s list of active issues, order followup testing, send the patient for a referral, or send the patient a letter.

Partners is taking another ahead-of-the-curve step by promoting its patient health record, called Patient Gateway, anticipating future mandates to give patients full access to their medical records. For now, records include results from the clinical laboratory, radiology, and cardiology, with anatomic pathology results to follow soon. Eventually records are likely to include clinical notes. “This is the future of medicine,” Dr. Dighe predicted, “with a truly patient-centric record and the patient taking a big role in their care.”

Patient satisfaction among those enrolled in the Gateway program is high, Dr. Dighe said. He suspects this taps a hidden desire—that most patients may not realize they want their results until they get them. He himself was a skeptical patient. Now that he has access to his medical record on his home PC, however, he’s sold on the idea. “It changes the way I look at my health. I get pop-ups to remind me about my cholesterol—it changes the game, making the patient a true partner in their care.”

All of this brings Partners closer to closing the loop, Dr. Dighe said. One large, troubling issue persists, however. A recent study looking at another large health care system where alerting is widely in place found that timely followup was not significantly different for an acknowledged result than for an unacknowledged one, Dr. Dighe noted. In this 2010 report, many providers were unaware of system features, and workarounds were abundant.

At Partners the aim is to avoid similar problems through enhancements to its applications to make test results acknowledgment so well integrated into the daily workflow that providers will want to use all the features. The Actionable Test Results Task Force is developing metrics that will provide an ongoing assessment of progress on each step, including clinical action. It’s soon going to be possible to provide chief medical officers within the system, for example, with the percentage of incidental pulmonary nodules that may be lost to followup at each hospital. What’s more, they’ll soon start putting it on their systemwide Quality Dashboard. “This creates visibility for the issue within our system,” he said. It’s a way to get everyone’s attention. By blowing the problem wide open, he and his colleagues might eventually close the loop.

Karen Titus is CAP TODAY contributing editor and co-managing editor.

Related Links Related Links