Tracking and analyzing the hard data of laboratory results, once they are reported, should be a relatively straightforward process. But a murkier category exists: the tests that don’t get performed because the laboratory rejected the specimens. What happens after specimen rejection, and what impact does it have on patients? Those were the key questions of a recent Q-Probes study, “Clinical Consequences of Specimen Rejection.”
Conducted by the College’s Quality Practices Committee, the study showed that the 78 participating institutions had blood and urine specimens rejected 0.2 percent of the time, in each instance adding, on average, about an hour to the process of making the result available. But the study also revealed interesting patterns in hospitals’ and laboratories’ typical handling of rejected specimens—a few of them surprising.
“We knew intuitively that specimen rejection has clinical effects, because we see it in our own labs every day and we know it will cause a substantial delay in getting results,” says study coauthor Donald S. Karcher, MD, professor and chair of pathology and director of laboratories at George Washington University Medical Center. “With this study, we were able to quantify that a little better than in the past. We also know that when a specimen is rejected, patients will have to have a specimen re-collected, causing inconvenience and discomfort, and sometimes collecting another sample such as a urine sample can be problematic. But we did find a couple of things we didn’t really expect.”
To carry out the study, the authors used a common Q-Probes design: Participants prospectively monitored blood and urine specimens submitted to the chemistry or hematology laboratory for six weeks, or until 80 rejected specimens were identified. The study set out to measure four performance indicators: the median specimen processing delay, the rate of specimen rejection, the percent of rejected specimens that were abandoned by the laboratory, and the percent that were abandoned by the provider after the laboratory’s rejection.
Of the 2,054,702 accessions in the study, 4,783 rejected specimens were identified. For 92.4 percent, the reason for rejection was “inappropriate/inadequate specimen.” The other 7.6 percent were rejected because of improper labeling. Rejection of the specimen resulted in a median specimen processing delay of 65 minutes.
The clinical consequences of the specimen rejections were the focus of the study. “For patients, we know that getting another venipuncture is uncomfortable, and there can be complications such as hematoma or even phlebotomy-related transfusion,” Dr. Karcher notes. The latter is fairly common for patients who are very ill.
Another complication that is uncommon but does occur is cutaneous nerve damage, which is painful and can last many months, says the second study coauthor, Christopher M. Lehman, MD, co-director of Hospital Laboratories, University of Utah Health Sciences Center, and professor of pathology at the University of Utah School of Medicine. If a lot of extra blood is drawn from a patient, there is also the potential for iatrogenic anemia. For line draws, he notes, to get a good specimen most hospitals require that the first 5 mLs be discarded to get rid of any contamination, “so that contributes to blood loss in patients, and there are some studies out there that show it can actually affect patient outcomes.” Laboratories now use less blood volume than they used to, Dr. Karcher notes. “But still, any blood you take from the patient that wasn’t necessary introduces the potential that at some point they may have to be transfused.”
Similarly, re-collecting a urine sample carries risks. “One reason is that the patient may not be able to produce more urine right away, and of course for those patients who require catheterization, every time you put in a catheter you are potentially introducing bacteria, and the more the patient is catheterized, the greater the risk.”
The most obvious result of having to re-collect a specimen is that the result is delayed, and this study found that the median delay for non-stat tests was 90 minutes. “When it’s a non-stat test, one could argue it’s not that big a deal,” Dr. Karcher says. “But even non-urgent situations can be very important for the patient. Maybe the patient is waiting for discharge from the hospital and that final result is the last thing the patient is waiting for.” Delays would be far more significant for stat tests, of course. “We found a median delay of nearly an hour for stat tests, and that means a clinical decision had to wait, and conceivably things could happen to that patient. Maybe a drug was started, or the patient had to be transfused because the hematocrit was dropping.”
At least in this study, the number of rejected specimens does not appear to be a problem. “Our rejection rate was 0.2 percent, which is very low,” Dr. Karcher says, noting that previous studies have found rates of 0.3 to .75 percent. “There’s been a noticeable decrease in the incidence of rejection, particularly as a result of mislabeled specimens, probably because there are now more robust tools available in hospitals that make mislabeling more difficult. I think bar coding is definitely one of them.”
However, what did emerge as a problem—and one that took the study authors by surprise—was the number of tests abandoned. Of the rejected specimens, 21 percent of them were abandoned, either because the laboratory didn’t request a corrected or relabeled specimen or the provider or caregiver team did not comply with the laboratory request.
“We didn’t expect that,” Dr. Karcher says. “If the laboratory said your patient specimen was rejected because it was mislabeled or collected in the wrong tube, for example, we set an arbitrary time frame of six hours to see whether the specimen was re-collected and re-submitted. And we found, to our surprise, that often no repeat sample comes.”
What is going through clinicians’ minds? “We noticed that the abandonment rate is substantially higher for non-stat tests,” Dr. Karcher notes. “We sort of assume that once they were informed the specimen was rejected, they decided if it was a blood sample they might wait until the next phlebotomy rounds.” For urines, the rate of abandonment was a little higher, and he thinks clinicians may be deciding the patient may not be able to provide another urine sample, or they don’t want to re-catheterize the patient.
A standing order may be a contributing factor to abandonment, Dr. Lehman adds. “Patients come in and when they are admitted, the physician will write orders for a CBC every morning, or something else every four or eight hours; then they don’t have to think about having to order them again tomorrow. But often those tests aren’t really essential, and so the clinician might say ‘I don’t really need that right now,’ or ‘in four hours it’s going to be re-collected anyway,’ so they abandon the test.”
Perhaps even more surprising as well as troubling, the authors found, was that so many laboratories routinely allow mislabeled specimens to be relabeled. Among the participating institutions, 45 percent allow relabeling of blood specimens and 37 percent allow relabeling of urine specimens, while fully 59 percent allow correction of mislabeled blood specimens and 38 percent allow correction of unlabeled blood specimens.
For many years, before the Joint Commission released its patient safety goals, “it was not uncommon if a patient sample came into the lab and was labeled ‘Joe Smith’ as opposed to ‘Steven Smith,’ but it was assumed to be Steven Smith’s, the floor would be called and someone would come relabel the tube. This was a fairly common practice,” Dr. Karcher points out. “But it’s a very dangerous practice, because even if the clinicians are certain they know whose blood was in that tube, there are so many points at which a mix-up could occur. Maybe they were just in Joe Smith’s room and didn’t realize they had mixed up the tubes.”
Though relabeling has always been frowned upon, “this study demonstrated it’s still a common practice. And the real shocker is just how misguided a practice it is. In this study, 39 percent of the samples that were allowed to be relabeled were ultimately discovered not to be the correct sample for the patient,” Dr. Karcher says. “This was probably the biggest surprise of the study. We were expecting a very small percentage, and 39 percent is pretty earthshaking.”
Everyone in health care would accept that relabeling should be allowed in some circumstances, he points out. “Let’s say the patient has a sample that’s not easily repeatable. Maybe you’ve just drawn cerebrospinal fluid. You don’t want to do another spinal tap on that patient because those are definitely fraught with complications. Or you’ve just done a brain biopsy. You’re not going to tell the clinician to go back and re-biopsy the patient’s brain. So, oftentimes, laboratories will allow those kinds of specimens to be relabeled under very carefully regulated conditions.”
The specimen rejection rate for chemistries and urinalysis is much lower than the rate in blood banking, says Dr. Lehman, who is board certified in transfusion medicine. “Usually the criteria for acceptance of specimens in blood banking are more stringent. Almost everybody rejects blood bank specimens that aren’t perfect. You can see that more labs are willing to allow clinicians to correct non-blood-bank specimens.”
In fact, there is a downside to forbidding relabeling entirely, he notes. “If you just say we’re going to force you to re-collect, in theory you would minimize the danger of having the wrong blood labeled improperly. However, the other side of that is you’re then requiring people to go back, spend time re-collecting, and expose the patient to potential harm.” So some laboratories consider the risk versus benefit. “Do I allow them to correct parts of a label, such as date, time, and initials, and not allow them to correct the name or medical record number? Or is a name all right to correct if it’s one letter off? Do I really want the nurse to go back and re-collect if everything is correct except for one letter? So I think institutions vary in terms of being strict versus weighing the risks and benefits of a policy of ‘no corrections.’”
Intriguingly, the Q-Probes study showed that relabeling doesn’t even save time, on the average. Those institutions that allow specimen relabeling reported no significant shortening of the specimen processing delay resulting from specimen rejection. That finding is a little hard to interpret, Dr. Lehman says. But, he adds: “If you are a nurse on the floor and get a call from the lab saying ‘I have a specimen and there’s a problem with it’—assuming you’re not talking about a wrong tube, not enough specimen, or a hemolyzed sample—it’s often easier for the nurse to re-collect than to come all the way down to the lab and spend their time relabeling.”
Larger institutions in this study—more than 450 beds—were associated with an even longer median specimen processing delay of 91 minutes, and the authors consider it likely this is a function of the greater complexity of large hospitals. In general, patients who come into a university medical center have multiple conditions and are sicker than the patients seen in a community hospital. “As a result, they often have more lab draws, and there may be more physicians, more nurses, and more medical students involved in their care,” Dr. Lehman says. “In addition, most of the ordering of labs is done by interns, and of course they’re inexperienced trainees. They’re still learning how to take care of patients.”
Moreover, he says, interns and residents may have three or four different attending physicians while they’re in a hospital, and as those rotations occur, the attendings often have different opinions about which tests to order. “So residents learn very quickly: To cover themselves, order everything. That way, if an attending asks for a result, they have it.”
One reason Dr. Lehman was interested in conducting this study was to give laboratories an opportunity to discuss preanalytical issues in a context important to clinicians: patient safety and delays in reporting of results. “Clinicians are usually not that concerned if something is a lot of work for the lab, but they probably don’t realize how much work specimen rejection is for nursing and also the implications for patients. This study gives labs some ammunition to talk to clinicians and say, ‘It’s not just that it’s work for us to call you, but there are also clear implications for patient care.’”
At his institution, monthly summary reports are sent on specimen issues by ward and by nurse manager, and the focus on preanalytical processes has helped engage nurses in following how their units are doing in collecting and labeling specimens. “[CAP’s program] Q-Tracks has shown that when institutions concentrate on specimen quality, they frequently improve.”
Hospitals are also taking the Joint Commission national patient safety goals seriously, Dr. Karcher believes. “We have work committees and task forces that are working on those goals all the time, more hospitals are using bar-code scanning for process control, and there are tools now that didn’t exist a few years ago. So I think the goals have really had an impact.”
This Q-Probes led to two important recommendations beyond those the authors originally contemplated. “Our original purpose was just to remind everyone that specimen rejection has consequences: delay in making lab results available and inconvenience and discomfort for the patient,” Dr. Karcher says. “We ended up discovering that specimen rejection leads to abandonment of samples, and that that problem is worse when the lab doesn’t instruct the provider to re-collect the samples.
“The study showed it’s important for labs to be proactive and say, ‘Your specimen has been rejected. Please re-collect and re-submit a specimen to us.’ When that didn’t happen, in our study, there was a larger percentage of provider abandonment of the test.”
Second, this study convincingly reinforces the recommendation that laboratories not allow significant relabeling of samples, Dr. Karcher says.
“There may be extenuating circumstances with some specimens. But for blood and urine samples, if allowed at all, relabeling should be restricted to correction of only the simplest labeling defects, such as a near-perfect misspelling, minor date error, et cetera. Relabeling is a practice that clearly introduces a degree of risk to patient safety, unless done thoughtfully and with great care.”
Anne Paxton is a writer in Seattle.