Our Daily Blood

With the growth of specialty genetic tests that can cost thousands of dollars, some large medical centers have become more aggressive in keeping tabs on physicians' orders of expensive tests. But while reining in big ticket tests may offer a quick fix to out-of-control sendout testing, another utilization problem often still lurks in plain sight: routine, low-cost tests that demand large amounts of real estate in the lab and on patients' charts.

The cost savings from reducing unnecessary tests are well documented, but a second and more important goal is patient safety. By cutting back on high-volume tests, hospitals can reduce the risk of iatrogenic anemia and transfusion for patients while also avoiding pouring human blood down the drain for no reason. Although a conservative approach to phlebotomy isn't new, rising healthcare costs and government programs that measure patient satisfaction and safety are putting a new emphasis on this area.

Even at institutions like the prestigious Cleveland Clinic in Ohio, laboratorians have discovered that busy physicians often order daily or even repeat tests on the same day, simply out of habit and hurry. Dealing with these kinds of frequent, unnecessary tests is a way for laboratorians to have a more direct impact on patients' experience of healthcare, according to Gary W. Procop, MD, who leads the organization's test utilization committee. "Decreasing unnecessary phlebotomy is a great way for laboratorians to show that they really care about patient satisfaction scores and patient experience," he said. "This is a great added value to improving test utilization of these tests beyond greater efficiency or cost savings." Procop also is chair of molecular pathology at Cleveland Clinic, and medical director of the molecular microbiology, mycology, and parasitology laboratories.

Stop and Think

Even when patients don't have out-of-pocket expenses, they pay for lab results with something else—their own blood. Although the occasional prick in the arm for outpatients seems inconsequential, observers have long expressed concern about how often inpatients have their blood drawn. In the hospital, patterns of routine or daily testing can drain patients past the point of discomfort.

In one example, Canadian researchers asked doctors at St. Paul's Hospital in Vancouver not to order complete blood counts (CBC) and metabolic panels when they didn't have a specific indication in mind, the idea being to reduce unnecessary routine CBCs and metabolic panels. Peter Dodek, MD, chair of the Critical Care Working Group at St. Paul's Center for Health Evaluation and Outcome Sciences led this quality improvement study. Dodek also is a professor of medicine at the University of British Columbia.

His team employed a handful of strategies aimed at getting physicians to consider whether the tests were necessary before ordering them. The first step was working with ICU staff physicians to develop acceptable indications for ordering CBCs and metabolic panels, including suspected anemia, infection, thrombocytopenia, and kidney injury (See Box, below).

Next the researchers introduced some educational strategies, including talking to rotating ICU residents about the indications for ordering these tests and displaying posters around the ICU that listed the indications along with the phrase, "THINK before you order!"

Dodek's team also added an item to the ICU rounds checklist reminding staff about the initiative, and provided a rubber stamp for lab orders, nursing flow sheets, and progress notes that read, "Routine blood work NOT indicated for tomorrow"—meant to catch the attention of on-call staff. Finally, in the hospital's computerized physician order entry (CPOE) system, a prompt for the tests in question read, "Please do not order without an indication."

For the next 6 months, Dodek and his colleagues monitored how many tests doctors ordered and compared these numbers to the year before the interventions. As they implemented each intervention in a step-wise fashion, they found that the number of CBC and metabolic panel orders dropped by 24% overall. However, the changes were not as dramatic over time. Across the 6-month study period, CBCs and metabolic panels per patient-day dropped by 7% and 5%, respectively. The researchers also checked to make sure that study participants didn't have worse outcomes, and found no adverse effects from their having had fewer tests.

Notably, physicians went along with the initiative and lodged few complaints. "The criteria for ordering these tests were actually not very strict," Dodek said. "We simply asked physicians to put down a reason. It seemed most of them saw this as a good thing—if we don't need to draw blood from patients, we shouldn't. On the other hand, some are quite stuck on routine testing and feel like they might miss something if they stop ordering them. They want to get a test ‘just in case.'"

The project won an award at St. Paul's during a quality improvement project competition and was credited with proving that routine blood tests could be reduced with relatively simple interventions. However, the researchers felt somewhat disappointed that the project was not nearly as successful as a previous intervention that focused on unnecessary routine chest X-rays. Using similar educational interventions, the X-ray project reduced routine chest X-rays by more than 60%, an effect that has lasted long after the intervention ended.

The most recognizable barrier for educational interventions such as these is simply the distractions in the ICU, according to Dodek. "If a little complacency sets in, something like this can quickly fall off physicians' radar," he said. "There are already so many things vying for our attention in the ICU. But I think the point to remember is that it is possible to reduce some routine tests without compromising patient care. It just requires continued attention."

Dodek did not collaborate directly with St. Paul's lab on the project, but he believes that laboratorians should embrace these utilization challenges and play a role. "I'm sure that laboratorians would rather be doing tests that make a difference. So I think they should help inform those on the ordering end that for their laboratory to be as efficient as possible, every test should be important and meaningful for the patient," he said. "That's the kind of wisdom that laboratory personnel can bring to the clinical scenario. If we're doing a lot of tests that are unnecessary, we're just cluttering up the lab's workload. No one is better suited than laboratorians to speak to the value of this."

Reducing unnecessary tests is about more than just efficiency, or even money, Dodek said. Rather, real risks like infection and anemia caused by phlebotomy worries him more. "We do know that there is evidence that the more blood we draw, the more we will transfuse patients. We also know that the more we transfuse, the higher the risk of nosocomial infections and mortality," he said. "So, given that linkage of drawing unnecessary blood, creating iatrogenic anemia, and then transfusing people, it does make sense to go right upstream and decrease the amount of blood that we draw."

Saying "No" to Same-Day Repeats

Taking on high-volume tests can mean more than just avoiding a daily order, and in fact Cleveland Clinic struggled before it was able to tamp down on same-day repeat testing. The medical center's test utilization committee tried using CPOE best practice alerts to address this issue, but found a full electronic stop was the only way to reduce overutilization on some orders.

The committee started with pop-up windows in the CPOE that would alert physicians when they tried to order a test for the second time in the same day. But physicians apparently clicked through these pop-ups without noticing and ordered the tests anyway. While same-day repeats decreased for a few molecular tests, the more common tests, such as C. difficile, continued unabated. That's when the committee took the next step: the hard-stop initiative.

Under this initiative, the CPOE blocked repeat orders of tests that the test utilization committee determined were never appropriate more than once per day. Physicians who still insisted on ordering a test had to call the lab, Procop explained. "We began with 10 tests for which it was beyond medical debate whether they were needed more than once per day, but with a ‘break-the-glass scenario,' so to speak, where if the doctors really, really wanted one of these tests, they could still have it. If someone wanted a second test, they had to call the lab, tell us who they were, who they worked for, and why they wanted it," he said. "That kind of initiative was never really tried here before. The previous culture was: if the doctor ordered it, just do it. And this gave us an opportunity to challenge the status quo. I give the leadership at the Cleveland Clinic a lot of credit for letting the test utilization team pilot this project." The initiative started with tests such as HbA1C, basic lipid panels, C-reactive protein, hypercoagulation panels, reticulocyte count, and hepatitis C. Now with more than 1,200 tests on the list, the success of the hard-stop initiative has grown, with surprisingly few complaints from physicians.

The key to building an initiative like this in which the lab could essentially tell a physician "no" was a close relationship with senior information technology (IT) staff, Procop noted. When physicians came to the test utilization committee with evidence that they did in fact need a test more than once a day, it was essential to promptly deal with it. "We had such a good relationship with the IT people that we were able to remove a test from the hard-stop list the same day if there was a valid objection," he said. "That way I could write back to the doctor right away and say, ‘we heard you; it won't be a problem anymore.' That really helped build our credibility with our clinical colleagues."

The hard-stop initiative saved Cleveland Clinic approximately $116,000 in its first year, discontinuing more than 7,000 unnecessary duplicate tests. However, the biggest payoff has been in patient satisfaction, according to Procop, because it's the high-volume tests that can be burdensome on patients.

"The motivator behind all of this was really a patient experience improvement initiative in that we felt like there was too much duplicate phlebotomy for inpatients," Procop said. "Our chief phlebotomist was pretty upset with patients being stuck twice for the same test, and rightly so."

Unhide the Price Tag

Even though many overused lab tests are not expensive—at least compared to imaging tests or drugs—simply showing the ordering physician a dollar amount also has had some success in reducing over-testing. A study at The Johns Hopkins Hospital in Baltimore examined if the number of laboratory tests ordered could be reduced by presenting clinicians with test fees at the time they were entering the orders in the CPOE (JAMA Intern Med 2013;173:903–8).

All clinicians, including physicians and non-physicians, who ordered laboratory tests through the hospital's CPOE were included in the study. Researchers randomly assigned 61 diagnostic laboratory tests to an active group with the fee displayed, or to a control group. Tests were randomly picked from two categories: the most frequently ordered tests and the most expensive. The expensive tests had to be ordered at least 50 times a year to be included.

No fee was displayed during a 6-month baseline period, November 2008–May 2009. During a 6-month intervention period 1 year later, fee data were displayed for active tests only. The fees shown to clinicians in this case were based on the Medicare allowable fee in the clinical laboratory fee schedule.

Physicians exposed to the test fees significantly trimmed their orders. Rates of test ordering declined from 3.72 tests per patient-day in the baseline period to 3.40 tests per patient-day in the intervention period, an 8.59% decrease. For the control group, ordering increased from 1.15 to 1.22 tests per patient-day from the baseline period to the intervention period, a 5.64% increase.

Daniel Brotman, MD, the senior author of the study, noted that orders of the more expensive tests did not make a significant difference in the study; instead, the frequent and relatively inexpensive tests took a significant dive. For example, there was a marked decrease in orders of comprehensive metabolic panels. "We realized in looking at the data that the total aggregate cost for the frequent tests was a heck of a lot higher than the more expensive tests. So if you were to focus on only the expensive tests, you'd only get at the tip of the iceberg." Brotman is an associate professor of medicine at the Johns Hopkins University School of Medicine and director of the hospitalist program at The Johns Hopkins Hospital.

Compared to basic tests like metabolic panels, clinicians appear more likely to think through the reasoning for ordering expensive tests, and order them less often, according to Brotman. "Expensive tests are generally ordered on a single occasion for a patient based on a perceived true need. So if you want to make sure that you understand whether someone has HIV, there is only one way to do that, and you only do it once. You wouldn't order it frequently throughout a hospitalization," he said. "I think to some extent it's reassuring that in our study the ordering of expensive tests didn't change, which to my mind means clinicians had a real curiosity about what was going on with their patient." Going forward, Brotman and his colleagues are working on a report card-style system for offering individualized feedback to clinicians on diagnostic ordering.

Brotman also believes there could be a generational gap in how physicians respond to these types of interventions. "I'm sure that in our study some were slightly annoyed to have to think about costs, but I think younger physicians are going to feel like it makes sense," he said. "They tend to see a dark cloud of healthcare expenses looming in this country, and physicians often don't feel very empowered to do much about it."

Let's Not Do That Again

When studied on a large scale, overuse of common, low-cost tests can look surprisingly costly—both in dollars and in hours of patient and physician time. A recent Veterans Affairs (VA) study examined patients with coronary heart disease (CHD) who had reached their treatment goals of low-density lipoprotein cholesterol (LDL-C) <100 mg/dL or even <70 mg/dL, and found that about a third underwent repeat lipid testing within 11 months without an identifiable aim (JAMA Intern Med 2013;doi:10.1001/jamainternmed.2013. 8198). Current performance measures for CHD, based on Adult Treatment Panel (ATP) III guidelines, recommend only one lipid test per year once a patient reaches his or her goal, unless the physician wants to intensify treatment or check for medication non-compliance.

Salim Virani, MD, PhD, of the Michael E. DeBakey VA Medical Center in Houston, and colleagues analyzed a total of 35,191 patients with CHD in a VA network of seven medical centers. Of 27,947 patients with LDL-C levels <100 mg/dL, about 9,200—32.9%—had additional lipid tests without treatment intensification during the following 11 months. All together, these patients had more than 12,000 additional lipid panels performed, at a mean cost of $16.08 per person, or about $204,000 in annual costs for one VA network.

The study demonstrates that overuse of common, inexpensive tests can still make a big dent in healthcare costs due to their high frequency, according to Virani. "The reason that we looked at this was that, while a regular lipid panel is not very costly compared to the imaging tests that other groups are looking at in terms of healthcare overuse, we believe the most important thing is that these are high-frequency tests, and it's the low-cost but high-frequency tests that go under the radar," he said. "Although the cost per individual patient might be low, overall the frequency of overuse is so high that it can become a pretty big number." Virani is an investigator for both the Health Policy, Quality and Informatics Program, and the Houston VA Health Services Research and Development Center of Excellence, as well as a staff cardiologist at the Michael E. DeBakey VA Medical Center. Virani is an assistant professor at Baylor College of Medicine, and he chairs the Strategic Planning Work Group the National Heart, Lung, and Blood Institute (NHLBI) National Program to Reduce Cardiovascular Risk (NPRCR).

"This well-conceived study on a large clinical database, which has the advantage of containing pharmacy data for use in tracking medication adherence, delivers an important message regarding a type of waste that is likely widespread in healthcare and that goes under the radar because it involves a low-cost test," wrote Joseph P. Drozda, Jr., in a commentary. "However, it is precisely these low-cost, high-volume tests and procedures that need to be addressed if significant savings from reduction of waste are to be realized" (JAMA Intern Med 2013;doi:10.1001/jamainternmed.2013.6808).

Virani believes that VA physicians ordered the additional lipid tests either out of routine or as a way to offer themselves reassurance, even when they did not plan on taking any action after seeing the results. In the study, Virani found a high correlation between primary care office visits and testing. "When a healthcare provider is going to see a patient, it may provide them some comfort in ordering labs just before or after the visit, so they feel that they are taking good care of the patient by doing something," Virani said.

In the future, additional lipid tests may become even less necessary if the ATP IV guidelines pivot and recommend a dose-based approach for statins versus the current treat-to-target approach that emphasizes lipid testing. "Although we don't know what the ATP guidelines will say, there is a debate in the literature about the fact that all of the clinical trials for statins were performed with a fixed-dose approach—the dose of the statin was the target and not the LDL-C," he said. "If ATP guidelines were to move in this direction, the need for follow-up with more lipid testing once a patient is on a stable statin dose would obviously go down."

Although Virani was able to use VA data sources to very accurately gauge lab costs for the redundant tests, he emphasized that more intangible consequences for unnecessary tests could be equally large. "The cost we estimated of course does not take into account the value of the patients' time to go to the lab and have an additional test performed, or the value of a physician's time to follow-up on these test results and notify patients," he said.