College of American Pathologists
Printable Version

  Critical thinking on critical values


cap today



September 2006
Cover Story

Anne Paxton

It might seem simple on the surface. For a certain set of tests, the laboratory has established ranges in which results are considered critical, and it notifies the patient’s physician when those results occur. Theoretically, physicians want to know immediately when there is a critical laboratory result for a patient, and are grateful for the call.

But in practice, critical values reporting may mean the laboratory hears “Thanks for letting me know” far less often than it hears “Hold, please,” “Dr. X is not available,” or “Why did you wake me up for that?” And those are only a few of the real-world complications.

“Throughout my entire career working in academic medical centers and training residents, handling critical values has been a problem,” says Peter Howanitz, MD, professor and vice chair of the Department of Pathology at SUNY Downstate Medical Center, Brooklyn, NY.

In a 2002 CAP Q-Probes study of 623 institutions, Dr. Howanitz and colleagues found that critical values systems were medically important but highly variable and costly practices for the study’s participants.

Recently, however, a new focus on critical values by the Joint Commission on Accreditation of Healthcare Organizations has spurred efforts to standardize and streamline these systems.

“It’s a basic area where the failure to communicate rapidly and reliably has a great deal of potential to harm patients, so there’s a medical interest in critical values,” says Kent Lewandrowski, MD, associate chief of operations in the Department of Pathology at Massachusetts General Hospital, Boston.

But now, because the Joint Commission has mandated a critical value reporting and monitoring mechanism as part of its National Patient Safety Goals, “there’s a regulatory hammer that says whether you want to or not, you’re going to do well and deal with it.”

The Joint Commission requirements specify that hospitals “measure and assess, and if appropriate, take action to improve the timeliness of reporting, and the timeliness of receipt by the responsible licensed caregiver, of critical test results and values.”

Massachusetts General started actively working on revising its critical value reporting mechanisms about three years ago, Dr. Lewandrowski says. “This was a basic area of quality assurance and improvement that we knew we could do better on, but when it was made mandatory the hospital gave us the necessary resources and clinical contacts to make it happen.”

Just how many laboratory results are considered “critical”? Massachusetts General estimated that about one quarter of one percent of results trigger the prescribed reporting procedure. “I would say our numbers are typical for a large urban medical center that has a full line of clinical services,” Dr. Lewandrowski says, adding that lower-acuity settings or specialized hospitals would probably have different numbers.

But included among those calls are likely to be many results that really aren’t critical for a particular patient. For example, the sickle cell test for hemoglobin S often yields results outside normal range when patients are asymptomatic.

And even though critical value calls are part of good laboratory practice and hospital medical staffs adopt the policies, frequently physicians end up receiving calls for repetitive critical values that may be the same result the patient had on admission, or not critical for other reasons.

“If you’re calling a lot more items like that, it’s not a trivial thing at all,” says Anand Dighe, MD, PhD, associate director of the Massachusetts General Core Laboratory. “By diluting the critical value list with a lot of things that really aren’t critical, you run the risk of doing the whole job less well.”

So the first step Massachusetts General took in reviewing its critical value reporting was to look at the current critical values to see what adjustments could eliminate some of the “sheer bulk of calls” that were being made, Dr. Lewandrowski says. “Because we’re doing 40,000 calls a year, that’s a lot of manpower demands both on the laboratory and on physicians, nurses, and practice managers.”

Getting rid of courtesy calls that aren’t truly critical, or putting those into a separate category, and shifting the ranges to national norms are possible ways to prune and sharpen the list. “For example, the glucose critical limit here used to be less than 60,” Dr. Lewandrowski says. “It’s since changed to less than 45, and that eliminated a large number of noncritical glucoses.”

The next step was to assemble a standing interdepartmental team with representatives from nursing, the outpatient department, the laboratory, and administration to carry out the process-improvement effort. “An interdepartmental team is essential,” he says. “Without it, you aren’t able to get much done effectively.”

Two technological tools can enhance critical values reporting: the monitoring power of the LIS and automatic paging. Daily, weekly, and monthly critical value reports can be monitored to see how well the reporting meets standards. Massachusetts General has defined its time standard as 40 minutes from the time a result is available.

Says Dr. Dighe, “You can use your laboratory information system reports, so when you do have an outlier—a critical value that took a long time—you can quickly get to the root causes and start working on them.” A daily report at Massachusetts General is viewed by all laboratory directors and is one of the bases for process improvement.

Trying to eliminate phone calls is a key goal, Dr. Dighe adds. For Massachusetts General, a near-term goal is to tie the critical value reporting system to the hospital paging system, and to have a page automatically occur when one of the critical value limits is satisfied. That requires, of course, that the hospital carefully track who is the responding and responsible clinician for every moment of time for all inpatients and outpatients. Massachusetts General is engaged now in just such a project, building a system that tracks the responsible and responding clinician for every patient.

But the advantage of the phone is the person can read back the test result (another Joint Commission requirement) so the critical value call has a built-in acknowledgment and can be documented. “You don’t tend to have that with a paging system; they’re one-way,” Dr. Dighe notes.

In general, the technology that’s available in hospitals hasn’t quite caught up to the task of critical value reporting, Dr. Lewandrowski says. Massachusetts General uses the Misys Callback module which allows the technologist to document the call, whether the result was read back, and the name and title of who received it. “So it doesn’t automate the phone call, but it automates the documentation and potentially allows for centralization of critical callbacks.”

“We’re looking at several vendors out there who promise to help you tie your critical value LIS reporting to a paging system. But they all suffer from the same issues: the downtime problem and the problem of whom to page. That’s especially true for outpatients.”

If the responsible caregiver can’t be reached, “we have an escalation policy in place,” Dr. Dighe says. “Our first step is to call the practice. If it’s closed, we page the ordering clinician. If the ordering clinician is not available, we look at the last few visits the patient had and try to figure out who an appropriate person to call would be.”

Dr. Howanitz’s Q-Probes study found that in five percent of cases, the laboratory actually hit so many dead-ends in its effort to contact the physician that it gave up. But there should be no point at which the attempt to report a critical value is abandoned, Dr. Lewandrowski stresses.

“Even if you have to get out the phone book or go online and find the patient’s address, you have to keep going until you hunt them down. You may have to call the local police to go pick them up and escort them to the ER of the local hospital. But there is no ‘give-up’ point. By definition, someone is in imminent danger, and you have to do something about it.”

He has nonetheless seen cases where the process occurs too slowly—for example, a technologist at a reference lab makes a couple of calls, then turns it over to the customer service department, which may come in the next morning. “In the meantime, 12 or 24 hours have elapsed. So that’s why you need to have clear guidelines. It should never be left in a bucket or bin somewhere awaiting somebody else to take it up.”

Outside the clinical laboratory, there are other results that could be considered critical values, including many from anatomic pathology, radiology, or the cardiovascular laboratory, Dr. Lewandrowski notes. But “those are situations we view as being somewhat different from the typical situation in the laboratory.”

“In the laboratory, the medical technologist is the one who has the critical result, whereas in the other cases it’s the physician who has the result. Technologists are not permitted to make judgments about the medical consequences or appropriateness of making or not making a call. If there’s a high potassium but the specimen looks hemolyzed, there are flat rules. They still have to call.”

By contrast, if a pathologist sees something on a biopsy or resection, “in a certain sense the patient is under the care of that physician, so there is an opportunity for some medical judgment as to whether the result is critical, although there are general guidelines in place.”

On the other hand, in such interdepartmental cases, a result might not be considered critical because there is no immediate danger, but it needs to be communicated to the physician. “I think some hospitals have fallen into the trap of over-defining critical values, such that they’re creating a whole lot of calls or background noise with things that don’t need to be communicated within a very short period of time,” Dr. Lewandrowski says.

Using the LIS at SUNY Downstate Medical Center, Dr. Howanitz says, “we were able to tailor critical values specifically for neonates. But if you talk about somebody from dialysis, there is no way to separate easily that patient from somebody who might just walk in off the street and have an elevated count.”

What the laboratory does try to do is have the clinicians in dialysis provide it with lists of patients they are going to see. “So we know there will be critical values on these patients, and they’ve instructed us on which ones we need to call.”

Clinical information systems are now building in features that help ensure effective communication of critical values, Dr. Dighe points out. “If, when the physician first logs into the system, it’s highlighted that there’s a new result and the system gives the opportunity to acknowledge that, that adds a layer of safety in itself.” Since physicians have dozens of results hitting them each day, “they need an effective way to screen through, view, and acknowledge them—beyond just making everything important ‘critical.’”

At large institutions like Massachusetts General, Dr. Lewandrowski is skeptical that a system can be set up effectively for physician-specific critical value calls. “It could have certain advantages from the physician’s standpoint to not be called for things they don’t consider important—for example, artificially low glucose values.”

“But if you have thousands of doctors utilizing a laboratory and you start customizing too much, it’s also a prescription for breakdown of the system. Misunderstandings occur. Maybe an oncologist doesn’t want calls for his chemotherapy patients but he happens to be covering for another physician on the weekend who would very much want to know those results.”

Unfortunately, problems like that won’t necessarily come to light right away. “It may take three years with a system before something bad like that happens, but that rare occurrence is just what we’re trying to prevent,” he cautions. “The majority of laboratory directors have not had a situation where a failure to communicate a critical value resulted in a patient death, so the nature of these rare things is that the systems designed to protect against them have to be designed carefully and thoughtfully with that in mind.”

Critical values reporting poses many legal pitfalls for hospitals. “I would say 70 percent of the legal cases that lawyers ask me about have to do with critical values and failure to communicate them, with a resulting disaster,” Dr. Lewandrowski says.

The issues that result in claims are mainly failing to call back a critical value, and failure to confirm that a critical value was communicated to a responsible caregiver. “Those two are the biggest,” he says. “For example, the person has a critical PT-INR, and they make a call and leave it on the physician’s answering machine, the physician’s office is closed for a four-day weekend, and the patient bleeds out.”

He has also seen legal cases in which the laboratory has defined critical values very conservatively. “Say the hospital decides any glucose over 300 should be called back, then a pediatric patient has a glucose of 310. Most hospitals wouldn’t consider that a critical value so there’s no mandate to call back. But this particular hospital has the policy defined, but they fail to call back because the technologist doesn’t think it’s all that critical. It turns out the child has emerging diabetic ketoacidosis, and something bad happens.”

“Once you make your policy, that then locks you into an obligation to ensure all those values are successfully called person-to-person to a responsible caregiver.” That’s the policy at Massachusetts General.

Unifying multiple different critical value policies looked all but impossible seven years ago when Integrated Regional Laboratories started consolidating laboratory operations for 13 hospitals, says James A. Robb, MD, Integrated Regional Laboratories-Florida in Ft. Lauderdale. IRL is the core laboratory and a company within Hospital Corp. of America’s East Florida division. “All 13 hospitals had different critical call policies, and it was very confusing to the technologists on the bench.”

Luckily, the Joint Commission entered the picture, enforcing a requirement that all policies for critical value notification within consolidated systems be unified. And that gave IRL the boost it needed to form a critical value notification team under formal Six Sigma process-improvement protocols.

It was particularly important, Dr. Robb adds, because the East Florida division is the first consolidated system within HCA. “We’ve been selected to be a shared service in the HCA system, and we will help them direct and apply this consolidation throughout the country.”

How much variation is there among hospital policies? “They’re really all over the place,” Dr. Robb says. “Our core policy is we notify the hospital laboratory, then that laboratory notifies according to the specific hospital’s policy—sometimes it’s to the nurse in the hospital, sometimes to the clerk, sometimes to the physician.”

But IRL is working on using information technology solutions to accommodate variations in critical value reporting. “We want to be able to do individual values per analyte, per physician. In Canada this is very common,” he says. MDS, one of the largest national laboratories in Canada, has facilities in Toronto, Calgary, and Vancouver that are able to tailor alert values or critical values according to physician, with specification of whether they wish to be notified by phone, fax, or e-mail, and, more recently, whether to notify them on a PDA.

Dr. Robb says it is essential to log carefully all key information about calls so that the reports can be verified. “We’ve had a couple of lawsuits where the physicians said they were never called and there was a bad outcome. But we’ve had it documented in the computer that they were called and who received the notification.”

The “alert” or “urgent” values that some physicians want reported typically don’t have to be conveyed within the 30-minute time frame that IRL sets for critical values. “They may be something the nursing staff is monitoring this month or results they just want to be aware of,” says technical supervisor Gay Prillaman, MT(ASCP), who is in charge of unifying IRL’s policies. “We see that a lot with nursing homes or other care facilities.” But, Dr. Robb adds, there is no agreement on whether laboratories should have alert values and, if they do, what they should be.

In fact, Dr. Lewandrowski expresses concern about the overlapping terminology. “Panic values are traditional in some hospitals, while others say ‘We don’t panic. It’s an alert.’ But I think those terms should be dropped because they’re ambiguous. If we say ‘critical,’ that has defined properties, there are articles about it in the literature, and we all know what we’re talking about,” he says.

Dr. Dighe says that a graduated system, such as a color-coded warning system, “just causes confusion. I think you want to keep your definitions clear so you can then compare yours to other institutions’ and to national guidelines.”

At another Florida medical facility, the Joint Commission National Patient Safety Goals led the laboratory to start asking whether less might not be more in the cause of effective communication.

“We are a big health care system,” says Sherry L. Woodhouse, MD, chief of pathology at Memorial Hospital Miramar in Miramar, Fla., one of five hospitals in the Memorial Healthcare System. “And it had been a long time since we all sat down and looked at our critical values and whether they were still clinically relevant.”

In fall 2004, when the laboratory launched a review, it was doing 61,000 critical value calls a year—about one every seven minutes. Often, in talking with her technologists, Dr. Woodhouse heard stories about their being put on hold or even being told a result wasn’t critical. “We were seeing the techs spend a large amount of time calling these critical values, which interrupted their workflow and slowed the whole lab testing process down.”

So she set up a committee to review the critical values—not knowing, to a degree, what she was getting into. “At the beginning I thought this would be easy. We’ll suggest modifications as we have some particular tests we want deleted from the list and ranges we want adjusted. But one of the big realities when you’re revising your critical value list and policy is that it’s absolutely not a laboratory-centered process. It’s very high-profile, and the physicians and nurses in the institution have an enormous stake in it.” The revision process, she says, will take more time and be more complicated than expected.

Making sure to include physicians, nurses, and pharmacists in the process, they went to the literature, benchmarked against what other hospitals were reporting in Florida and nationally, and set up a spreadsheet by going through their critical values one by one.

“It was not only adding and deleting specific analytes but also adjusting the ranges,” Dr. Woodhouse says. “Some values we were calling as critical were actually results at the top of the normal range—values that weren’t really critical.”

How critical value lists evolve over the years can be fascinating, she says. “For example, some critical values had been added based on a single incident with a bad outcome that happened many years before. That value still remained on the list even though it was no longer clinically relevant.”

As a result of refining its critical value policy, Memorial Healthcare System has dropped its number of critical value calls by 40 percent, and the technologists are feeling their efforts are more worthwhile. “They feel results they’re calling are truly critical and not just convenient. They’re not spending so much time on the phone, and it’s been a very positive thing in the laboratory and for nursing,” Dr. Woodhouse says. The laboratory made a rough estimate of projected annual costs in nursing and technologist labor and reagents in repeating these tests and found it was probably saving about $120,000 a year in one of their hospitals alone.

In an e-mail after the changes were made, one of her technologists wrote, “Thank you so much for this new policy; it has improved our workflow tremendously.”

Dr. Woodhouse is quick to note, however, that critical value calls cannot be the only red flags in the system. “Our lab technologists are processing literally thousands of tests a day, and any one of those tests could be critical for a given patient. We’re trying to look at a list based on benchmarking data and the literature that indicates a result is truly critical for any patient. There is no way the lab can substitute for a physician’s constant vigilance over the results of lab tests they order for their patients.”

Despite their theoretical appeal, clinician-specific calls are not practical, in her view. “We’ve discussed them but we really can’t do it.” Nevertheless, she adds, “Our critical value list is not a policy written in stone. It can always be reviewed and modified. However, we will use the literature and benchmarking data as a standard.”

Recently the Joint Commission too has shown flexibility, Dr. Howanitz reports. “There were some major problems with the current policy on critical values reporting. It’s poorly understood and it has been controversial since it was implemented two years ago.” But the commission has backed away from some of its initial requirements.

Originally it mandated that a licensed physician personally take critical value results reports. “In our view that is really impractical. It delays care, and very frequently it’s difficult to find someone who’s available. We highly recommended that this policy be eased so that a ward clerk, for example, could take the call.”

Pathologists from the CAP’s Council on Scientific Resources met with the Joint Commission and it has now changed its policy, Dr. Howanitz says. “Although many hospitals are not yet aware of the change, the Joint Commission will allow others to receive these results.”

To further ease reporting, Dr. Howanitz’s institution calls and tells the person who answers that there is a critical value on such-and-such a patient and it is in the computer. “If you give a critical value by phone, you’re expected to give the patient name using two identifiers and have the result read back. But if you send results by computer, you can print them out without worrying about transcription errors, which is what the read-back requirement is supposed to fix.”

After learning how this process has saved time and improved care at SUNY Downstate Medical Center, Dr. Howanitz says, the Joint Commission accepted this policy, too.

The bottom line for laboratories is that critical value reporting does matter, and the literature continues to confirm its relevance. In a recent study of total serum calcium critical values conducted by Dr. Howanitz and his wife, Joan Howanitz, MD (Arch Pathol Lab Med. 2006;130:828–830), high disease severity and mortality were found in the patients with critical results. The two concluded that although broadening the critical values limits would reduce the number of required calls, the current limits were definitely warranted.

“It’s clear that medical staffs really favor having critical value reports because in our study a significant percentage of the patients go on to succumb,” Dr. Peter Howanitz points out. The Q-Probes study, too, showed that 94.9 percent of physicians found critical values lists valuable.

“The key thing to remember is that critical values are exceedingly important, they are frequently life-threatening, and clinicians need to be reached in a timely manner with critical values results.”

Anne Paxton is a writer in Seattle.