A Pending Test at Discharge and a Return with Sepsis

Dramatization:

My name’s Charles. I almost died last winter when I fell on some ice and went to the hospital. I was struggling to breathe so they kept me a couple of days to run some tests. They said I was fine and they sent me home. Well, guess what? They missed a test. In two days, I was headed back to the hospital in an ambulance fighting for my life.

A 68-year-old male was admitted to the hospital after falling on ice and feeling short of breath. The patient had a right-sided pneumothorax, and a chest tube was placed in his right lung. Monitoring showed good clinical improvement, and physical therapy staff assessed him for mobility and home safety.

Two days after his admission, the patient was ready to be discharged home. The patient’s discharge paperwork indicated that he had a pending result from a pleural fluid culture that had been collected on Day 2, with no indication of any recommended follow-up services. On the morning of discharge, the lab called the care unit with the result of the fluid culture, which demonstrated gram positive cocci, but they were not able to contact a member of the care team. The attending physician was paged but did not respond. The culture result was not escalated to the patient’s nurse or another member of the care team, and the patient was discharged later that day, unaware of the positive culture.

Two days after discharge, the patient arrived by ambulance at another hospital in septic shock. He required additional chest tubes and intravenous antibiotics to manage the infection. After ten days in the hospital, the patient went home with a PICC line and required several more weeks of IV antibiotics. He was able to fully recover.

The patient filed a claim against the hospital, alleging that the failure to communicate a critical lab result required admission for septic shock and several weeks of follow-up treatment. The case was settled in the low range.

To discuss the patient safety and risk management aspects of this case, we are joined by Dr. Adam Schaffer. Dr. Schaffer is a Senior Clinical Analytics Specialist at CRICO and Assistant Professor of Medicine at Harvard Medical School.

Q.) Adam, thank you for joining us:

A. Great to be here. I appreciate the opportunity to chat about this case.

Q.) What do you see as being the principal patient safety issue in this case?

A.) This case involved what is known as a test pending at discharge, which we refer to as a TPAD. Tests pending at discharge are those tests that have not come back with a final result by the time the patient is ready to be discharged from the hospital. Actually, I guess one would be considered a quasi-TPAD, since the specific facts in this case involved a test that resulted the day of discharge. As occurred in this case, many TPADs are microbiology tests, since cultures of blood or pleural fluid may grow bacteria only after several days incubating. One study by Walz and colleagues found that microbiology results were the most common category of TPAD.

When certain types of cultures come back positive for bacteria, then that is considered a critical test result, which needs to be rapidly and reliably communicated by the lab to the team caring for the patient. Such communication should involve closed-loop communication between the lab and clinical team. This case shows a breakdown in this communication.

Cultures may be especially problematic when it comes to following up TPADs, as most cultures that are still pending at discharge do not grow anything, in which case usually no action is needed. However, if the cultures DO grow something after the patient is discharged and the patient is not on an appropriate antibiotic, then this can be an urgent situation, as the patient likely has an untreated infection. 

That is what occurred in this case. Because the microbiology lab was not able to contact a clinician who could act on the positive pleural fluid culture that came back around the time the patient was discharged, the patient was not started on an appropriate antibiotic, and the patient ended up returning to the hospital with sepsis, which is a severe systemic infection.

Q.) Are tests pending at discharge a widespread problem?

A.) Specific data about this vary in different studies, but pretty much all agree that this is a substantial issue that does warrant attention from our patient safety leaders. One study, conducted by my colleague Chris Roy, which was published in the Annals of Internal Medicine—actually funded by CRICO—found that about 40 percent of patients who were discharged from the hospitalist service had TPADs. Among the roughly 2,000 TPADs in the study, 9.4 percent potentially required action to follow up on them. Based on survey data regarding a subset of the TPADs, physicians were unaware of 62 percent of them. And for this subset of 105 TPADs, 65 of them the physician was unaware of, and 24 of them required some action and eight actually required urgent action. These data speak to the need to address this issue, as they show that TPADs are common, clinicians may not be aware of them, and that these TPADs may warrant action. Sometimes this action actually is urgent.

So, although this was not a specific problem in this case, one thing we as inpatient clinicians can do (and I count myself among this group) is we need to be thoughtful about the tests that we order, when we order them, and be cognizant of how long it takes for them to come back.

Q.) That makes sense. You noted that there was a breakdown in communication between the lab and clinical team regarding the positive pleural fluid culture, which could be considered a critical test result. What are some of the approaches that have been tried to ensure that reliable communication of critical results happens?

A.) A key feature one wants to see as part of any system for notifying clinicians of critical test results is what’s known as closed-loop communication. What this means is that, for every critical test result, you provide the information about that critical test result to the designated recipient, you make sure this information is received, the person who's receiving this information acknowledges receiving it (often by reading it back to the person who’s contacting them), and the recipient agrees to take responsibility for acting on that result. In this case, the lab tried to call the primary team to let them know of the result, but they were not able to reach a member of the team, and so the result was never acted upon, and the patient ended up developing septic shock. A principle of closing the loop is that it is not an acceptable practice to send notification of a critical test result via e-mail or page without also receiving a response from the clinician being notified to confirm that they received and will act on this result. So, you can’t have this critical test result being sent off into the e-mail or pager ether without knowing that it was received, and that the person is going to act on it.

Hardeep Singh and Doris Hanna, along with colleagues, they each published really nice papers in The Joint Commission Journal on Quality and Patient Safety that came out with some really thoughtful recommendations for what policies should we think of implementing when it comes to communicating critical test results. First, there needs to be a shared understanding of what represents a critical test result. Because if I’m the clinician and I expect to be called about a certain test result because I think it’s critical, but on the other hand the lab never notifies me because they don’t consider the results critical based on their policies, then that is a setup for a breakdown in communication. There also needs to be clarity regarding who has primary responsibility for receiving information about, and acting upon, critical test results.

In this case, the lab tried to contact the attending physician to inform them of this critical test results, but they were not able to reach the attending. It is reasonable to contact the attending with this critical test result, as the attending has overall responsibility for the care of the patient. However, the attending needs to know that they are the designated contact for critical test results. Some experts have recommended that the clinician who ordered the test, so the ordering clinician, be the person who receives the notification of the critical test result, since the ordering clinician knows that the test is in process and the context (the reason it was ordered). There also should be fail-safe systems, in which there are designated back-up clinicians who can be contacted if the clinician who has principal responsibility for receiving the critical test result is unavailable or doesn’t respond. In this case, when the attending physician could not be contacted with the critical result, there was no effort to reach an alternate clinician or to escalate it who could act upon the result, and that’s problematic.

Q.) You mention that tests pending at discharge (or TPADs) are common. What are the systems that you see that have been really useful in making sure these are followed up upon?

A.) Whitehead and Darragh, along with colleagues, they each published really nice reviews of this area in 2018. To start at a very basic level, it is helpful to have a designated section of the discharge summary where tests pending at discharge, where these TPADs, can be listed. The team that ordered the tests needs to be responsible for following them up, but having the TPADs in a specific section of the discharge summary, which typically is sent to the patient’s PCP, that then allows the PCP to be aware of the TPADs, and then the PCP can serve as a backup provider when it comes to following them up.

We should also be informing patients of their TPADs, because you want patients as much as possible to be involved in their care, though of course the clinical team has the responsibility to follow up on them.

Critical lab results also need to be directly communicated to the responsible clinician; however, for non-critical tests pending at discharge (or these non-critical TPADs), technology is going to be a crucial part of the notification process. On a practical level, we need to work to get the available technology to integrate into individual clinician workflows in as reliable a way as possible. Because if that doesn’t happen any system is pretty much doomed to fail. For instance, test result notification systems may live in the electronic medical record, and if you are checking the electronic medical record every day, this may be adequate to avoid missing test results. However, if you are in a specialty in which intermittent finite periods of clinical work are typical, as with emergency medicine or hospital medicine, then you may not be checking the electronic medical record every day, which makes you vulnerable to missing a test result that appears only in the EMR notification system. Some EMR systems have the ability to notify you via regular e-mail when certain types of test results in the EMR system come back. 

This combination of refining our technology so that it can fit into multiple different workflows, along with education about what the technological options are for helping us avoid missing test results, these are some of the key practical approaches to this issue.

Thank you. Dr. Adam Schaffer is Senior Clinical Analytics Specialist at CRICO and Assistant part-time Professor of Medicine at Harvard Medical School. I’m Tom Augello.