Charna Albert
November 2023—At the point of care, there are testing wins, some losses, and plenty of pitfalls. “Point-of-care coordinators all have the same problems,” says Meaghan Gladstone, applications consultant at Werfen.
Until recently, Gladstone was point-of-care supervisor at the University of Pittsburgh Medical Center. In a session at this year’s Association for Diagnostics and Laboratory Medicine meeting, she told the story of UPMC’s point-of-care successes as well as what didn’t work, and how the clinical committee that now governs point-of-care requests is in place to prevent failures.
That committee has approved more than 50 requests for point-of-care testing since 2017, Gladstone said. “Our failures have been tests that were implemented outside that process.”
The first success she shared comes from the UPMC Magee-Womens Pregnancy and Women’s Recovery Center, which provides care to women with opioid use disorder and where a request was made for urine drug screening at the point of care. Staff travel between the center’s five community locations. Nurses collected samples at the community clinics and drove them to the main lab for testing. Sometimes couriers were used and samples could be lost or delayed. Patients wouldn’t return for follow-ups or repeat testing, and patients who returned for treatment after initial drug testing could in the interim have used additional substances.
“Urine drug screen is not a common test at the point of care in our institution,” Gladstone said. “We prefer the main lab testing.” The clinical committee denied the center’s initial request because there was an insufficient understanding of why the test was needed at the point of care. “However, after the point-of-care coordinator was able to create a dialogue with the center to understand the need and present it to the clinical committee, it was then approved.”
The first step was to validate the new test, a dip card immunochromatographic assay, “because this was a new method for us,” she said. “We took two brands of dip card and compared 15 clinic patients to the main lab.” Nine of the 15 patients underdetected on at least one drug on both cards, “so not very sensitive,” and four patients overdetected but on only one of the cards. They selected the other brand, “and the clinical committee made the decision to require confirmatory testing for both positive and negative results, so it’s truly a screening test.”
Average collect-to-result time for a stat urine drug screen in the main lab is about four hours, she said, compared with five minutes at the point of care. The time from receipt in the main lab to result is only about an hour, “but you have a three-hour time difference between collection and delivery to the lab. They are losing these patients.”
“The reason this is a success story seems to be about turnaround times,” she continued, “and it is to some degree, but it also had an effect on patient treatment and communication.” With results available in five minutes, “the nurses say the patient does not hide if they have used. It opens up that line of conversation with them that they never had before because of that wait for lab results. So patients are more open about what they’ve used and will tell their care provider, sometimes even before the test is complete, because they know they’re going to have to anyway.”
In another of UPMC’s successes, point-of-care whole blood β-hCG testing was implemented in the emergency department at level one trauma centers, where it can be difficult, amid imaging and emergency surgery, to obtain a urine sample.
Average collect-to-result time for the point-of-care β-hCG test is 10 minutes, compared with 80.3 minutes for main lab stat testing at one trauma center and 105 minutes at another.
Gladstone shared a case in which a 31-year-old woman presented to the emergency department for evaluation after a motor vehicle collision. “The trauma team noted open fractures, and after initial stabilization they knew right away the patient needed imaging, despite her age and potential for pregnancy.”

The patient arrived at the emergency department at 11:50 PM. “That point-of-care β-hCG results 13 minutes later as positive, at 12:03 AM.” Surgery began at 1:55 AM. Had the care team waited for central lab results, which would have been available at 1:41 AM, there would have been little time to make medical decisions, she said. “But in this case, the surgical team had enough time to develop a plan and weigh the risks and benefits.” Ultimately, the team decided to use a lead apron to shield the patient’s abdomen during surgery.
“If you look at a case like this,” she said, “you have to call this test a success, regardless of turnaround times or cost.”
Point-of-care creatinine testing in UPMC’s radiology areas also has worked well, Gladstone said. Before the test was implemented, patients who did not comply with preprocedural bloodwork before imaging would have to wait for their bloodwork from the lab, delaying their treatment time and disrupting the flow of patients. Average collect-to-result time for the point-of-care test is four minutes; for laboratory-based testing it was 35.7 minutes in 2021 and 32.7 minutes in 2022. “That faster four-minute result would include improved patient throughput and wait times. This is now the system standard in all our radiology departments, and it’s generally a quick approval in our clinical committee,” she said.
Another win: home-based point-of-care testing for an in-home urgent care program, which “meets the patient where they are” while reducing the number of ED encounters, inpatient hospitalizations, and ambulance transports, Gladstone said. At the program’s inception in September 2021, five waived point-of-care tests were selected: glucose, creatinine, hemoglobin, urine dipstick, and SARS-CoV-2 antigen. As of last spring, urine hCG had been added, fecal occult blood and white blood cell testing were being added, and strep and point-of-care chemistry testing were under consideration. More long term are blood gas testing and prothrombin time/INR. “Our state regulations do not allow nonwaived testing in this type of mobile screening environment, so we’re trying to work through how we can get gases and PT/INR in there,” she said.
Of 564 patients seen in the program’s first year, 498 were able to be treated at home. Median response time for the paramedics and EMTs who travel to homes to provide care is one hour and 24 minutes, she said. “These patients don’t have to decide where they’re going, drive there, check in, wait in a waiting room, wait for somebody to see them. Most of that takes longer than an hour and 24 minutes, so this is a beneficial program for a certain patient population.” The program has since expanded to cancer care.
“They have now visited over 1,000 patients, and they have expanded the program by providing next-day care as an alternative to emergency department 24-hour observation, so they’re freeing up beds,” she said. The program also has expanded from daytime-only to 24/7 service. Fifty-seven percent of patients seen in March had samples collected and dropped off for laboratory-based testing, and 54 percent had point-of-care testing. “And as we add more [point-of-care] tests that number is only going to increase.”
Less successful in the EDs of two community hospitals was the use of a basic metabolic panel, undertaken without guidance from the laboratory.
Lab turnaround times were the problem, “and in this case it was for dehydrated patients,” she said. “But they didn’t consult the lab,” and because there was no dedicated point-of-care coordinator at the hospitals, there was no one to manage the program.
In the end, the physicians preferred the lab test, and orders for the point-of-care test declined within the first few months. In December 2017, the month the test was implemented at one of the two hospitals, 17 point-of-care basic metabolic panels were ordered. One month later the number was nine, and in February 2018 only one was ordered. “They were no longer interested,” she said. But quality checks were still being run on the two instruments purchased for the testing. “That’s 44 QCs,” she said, in addition to the training, procedure development, competency testing, and cost of the instruments. A similar trajectory played out at the other hospital, and the point-of-care test was discontinued.
“It was this failure that drove the decision in our institution to develop a clinical committee to vet these types of tests before we get to this point,” Gladstone said.
At another UPMC hospital, the anesthesiology department and hospital administration purchased an instrument for point-of-care viscoelastic testing in the cardiovascular surgical operating room, without laboratory consult. They were under the impression “they could just order the instrument and plug it in and run it. Once we found out, we did explain that was not the case,” she said.
“The test hit roadblocks right from the beginning,” from the pandemic, to security not approving the software for a year, to the manufacturer quality control not being released during validation, among other things.
On other point-of-care testing the jury is still out, largely due to low volumes, Gladstone said. One is the urine drug and pregnancy testing in the Center for Interventional Psychiatry, which specializes in electroconvulsive therapy, transcranial magnetic stimulation, and esketamine intranasal therapy. “For patient safety, they wanted to be able to do a urine drug screen and pregnancy test prior to treatment. Before they implemented point of care, they were relying on verbal confirmation from patients.”
Pregnancy testing is now run on every woman of childbearing age receiving ECT, “which has only been 20 to 40 tests in the first half of 2023.” Urine drug screens are performed as needed, which amounted to 10 tests in the first half of this year. “The volumes are low, but the staff like the flexibility of having the tests to make more rapid treatment decisions.”
In one case, a patient woke up agitated after electroconvulsive therapy and the providers were concerned the patient was reacting to the anesthesia because of recreational drug use. “So the provider was then able to order that urine drug screen prior to the next treatment,” she said. In another situation, a patient’s insurer requested documentation of no drug use before treatment. “They were able right then and there to run the urine drug screen and provide that insurance company with the results to prevent delay in the insurance authorization.”
“Now the question is whether the low volume of tests is worth these benefits,” she added.
Test volumes have been low, too, for point-of-care HIV testing at a women’s hospital, where it is needed for rapid HIV screening in the pregnant patient at risk for imminent delivery but no prenatal care or at least no documentation of HIV status, among other reasons. There was no off-hours laboratory coverage, and “we were not meeting that American College of Obstetricians and Gynecologists’ recommendation of [HIV] testing being available 24 hours with a turnaround time of less than one hour,” Gladstone said. “The physicians also needed a faster initiation of antiretroviral therapy when they weren’t getting these results overnight for mothers and infants who are HIV positive.”
The clinical committee approved the point-of-care test, with the caveat that all results—reactive and nonreactive—would require confirmatory lab testing. “But treatment and newborn management may be considered prior to laboratory testing, so they can run point of care and if it’s positive it’s up to them whether they want to start treatment before they get that confirmatory result.” The test is run only during the lab’s off-hours.
About 22 tests were run over a two-year period, she said. “Is it a success because 22 at-risk patients without prenatal care and their babies were appropriately treated? Or is it a failure because they are running less than one test per month? If the goal is equitable health care, one could see how those 22 patients are a success story.” The clinical committee will determine the future of the testing.
Charna Albert is CAP TODAY associate contributing editor.