A dozen first steps to national
  quality measures for labs

June 2005
Feature Story

Karen Lusky

Some things aren’t a matter of whether but of when. And national quality measures for laboratories that can be linked to payment incentives or inspection penalties or both are likely to be one of them.

The Institute for Quality in Laboratory Medicine—a public-private partnership with more than 65 stakeholders, including the CAP, and supported by the Centers for Disease Control and Prevention— unveiled at its April conference a set of 12 quality indicators as a potential starting point for measuring laboratory performance.

The institute’s consensus-driven effort to define laboratory quality standards and metrics is just out of the gate and aimed at encouraging labs’ voluntary participation. But many in the field expect clinical laboratories eventually to join other health care sectors that publicly report national quality measures.

Yet the ultimate message coming out of the institute’s endeavor may be that laboratory medicine is even more of a driver of overall health care quality than realized previously. Lee Hilborne, MD, MPH, co-chair of the institute’s Quality Indicator Workgroup that developed the initial set of indicators for further investigation, predicts that “over the next decade, laboratory testing will be increasingly used to assess the quality of care and the extent to which pathologists’ clinical colleagues practice evidence-based medicine.”

To identify and develop national quality measures, the CDC is working with the National Quality Forum, an independent organization with more than 260 public and private stakeholders, including the CAP, created to forge a national strategy for health care quality measurement and reporting. The CDC’s Joe Boone, PhD, says the institute’s goal is to collaborate with the forum and its own partners to bring a spectrum of stakeholders to the table to hammer out a set of national quality indicators. “The goal is to develop something that will be useful not just to labs but to the whole health care field,” Dr. Boone says.

As a first step in that direction, the institute workgroup’s 12 candidate quality indicators are a blend of preanalytical, analytical, and postanalytical measures, as well as systems measures that extend before and after the traditional beginning and end of the process. “There’s a very clear understanding that the IQLM should examine the entire testing process from the ‘pre-pre’ analytical to the ‘post-post’ analytical phases,” explains Robin Stombler, president of Auburn Health Strategies, Arlington, Va., which is helping the CDC develop the institute, or IQLM.

The institute’s Quality Indicator Workgroup culled the indicators from a much larger pool of potential measures as a “first pass” at a core set, said the workgroup’s co-chair Frederick Meier, MD, in a presentation titled “Coming Soon to Your Lab: National Quality Indicators for Laboratory Testing,” at this year’s Executive War College, sponsored by The Dark Report.

In his presentation, Dr. Meier, who is division head of system laboratories for the Henry Ford Health System in Detroit, analyzed the strengths and weaknesses of the 12 initial quality indicators and discussed some of the directions in which a quality reporting initiative might take labs and the health care system.

The candidate quality indicators are as follows:

• Diabetes monitoring (system)
• Hyperlipidemia screening (system)
• Test order accuracy (preanalytic)
• Patient identification (preanalytic)
• Blood culture contamination (preanalytic)
• Adequacy of specimen information (system/preanalytic)
• Accuracy of point-of-care testing (analytic)
• Cervical cytology/biopsy correlation (analytic)
• Critical value reporting (post analytic)
• Turnaround time (postanalytic)
• Clinician satisfaction (system/postanalytic)
• Clinician followup (system/postanalytic)

The list includes a number of process quality indicators that testify to the accuracy and efficiency of laboratory testing, Dr. Meier said. “These include doing the right test, identifying patients and specimens correctly, reporting critical values, and getting test results back rapidly.”

Blood culture contamination is a particularly good measure of specimen collection quality that is linked directly to patient outcomes. “Published studies have demonstrated that people with contaminated blood culture samples have longer hospital lengths of stay and more nosocomial infections,” Dr. Meier told CAP TODAY. The infections result from unnecessary intravenous access, antibiotic-induced super infections, and antibiotic-associated colitis.

The CDC championed the diabetes monitoring with hemoglobin A1c and hyperlipidemia screening indicators because they focus on measures that assess and monitor heart disease and diabetes, which are the two major causes of morbidity for Medicare and Medicaid populations, Dr. Meier says. Hyperlipidemia screening would include cholesterol, LDL, HDL, and triglycerides, which are part of the Health Plan Employer Data Information Set, or HEDIS, used by the National Commission on Quality Assurance to evaluate health plans.

The CDC’s Dr. Boone says diabetes monitoring and hyperlipidemia screening are health care rather than lab-service measures. “But because health care payers use the measures so broadly to evaluate quality of care—and labs play a key role in providing the basic information for that evaluation—the measures really provide a nice link between lab services and health care.”

Several of the candidate measures have been validated in the CAP’s Q-Tracks program, which offers continuous quality monitoring for clinical and anatomic pathology laboratories with longitudinal tracking of key indicators.

They are:

• Patient and specimen identification. “Measurement of patient identification is a Q-Track-validated QI that can clearly be done,” one associated with improvements related to specific institutional practices, Dr. Meier says. Labeling specimens at the bedside or point of care has also been shown to reduce patient ID errors.
• Test order accuracy, if the indicator simply involves detecting transcription errors. The Q-Tracks program has shown that redundant order-entry approaches reduce errors in test order accuracy, Dr. Meier says. There are also feasible monitors for duplicate test orders. But if the measure is defined as monitoring whether the right test for an indication is ordered, he says, most laboratorians are at a loss as to how to monitor that as a quality indicator.
• Blood culture contamination. The blood contamination Q-Tracks study found that institutions have significantly less blood culture contamination when dedicated phlebotomists under direct supervision of the laboratory collect the blood culture samples, says Dr. Meier.
• Adequacy of specimen information, if the indicator is limited to specimen labeling, which has yet to be determined. “But nonlaboratorians in the IQLM Quality Indicator Workgroup say there is often other information that should be submitted with specimens to help labs and clinicians interpret the results,” Dr. Meier says. “They want labs to monitor that information to reduce omissions.”
• Turnaround time. This indicator is also “Q-Track-able,” Dr. Meier says, and shows modest improvement just by virtue of tracking it.
• Critical value reporting. “This quality indicator would involve an institution measuring how quickly it delivered critical values and how infrequently it was unable to deliver a result designated as critical,” Dr. Meier says. He predicts that the critical value reporting indicator, if implemented, could hasten consensus among pathologists and clinical labs on which values are critical in various settings.
• Clinician satisfaction. The College has also done Q-Probes studies on clinician satisfaction. (Q-Probes studies take a statistical snapshot in time of a single primary performance indicator and possible influencing variables at participating institutions.) “CAP’s Laboratory Accreditation Program already requires a satisfaction survey at least every two years,” Dr. Meier adds, noting that satisfaction is itself a measurable outcome.

In Dr. Meier’s view, two of the candidate indicators are problematic. These are accuracy of POC testing and cervical cytology-histology correlation. “The former is a nebulous monitor,” he says. “And evidence shows that the latter does not improve process, let alone outcome quality.”

Yet the Centers for Medicare and Medicaid Services and the CDC have experienced serious problems with waived test accuracy, most of which is point of care, and with the CLIA-mandated cervical cytology/histology correlation, Dr. Meier notes. “So these federal agencies would like to use the National Quality Forum consensus process to figure out how they can discharge these two regulatory responsibilities in a useful way,” though he’s not sure such a solution will be forthcoming.

How might a quality indicator for POC accuracy be reported? Judy Yost, director of the CMS Division of Laboratory Services, says she’ll leave the design of the indicator to the experts. But she speculates that the measure may be able to be reported in the number of labs within a total population or incidents in which specific problems or indicators are identified as having an actual or potential impact on the quality of testing. “To monitor, you might do some sort of self-survey crafted to tease out these issues, or conduct focused on-site visits,” she says. She sees POC testing accuracy as an important indicator “since so much testing is moving in that direction and test quality should be assured regardless of where the test is performed.”

As a candidate indicator, clinician followup for lab test results is also still nebulous, Dr. Meier says: “A quality indicator looking at clinician followup would require evidence-based protocols for acting on specific sorts of test results. And laboratorians fear that if the clinician followup index were to require laboratories to ensure that clinicians ‘do something’ with abnormal test results, laboratorians would be placed in a policing role that would set labs up to fail” on the measure.

Addressing clinician followup is one of those “cross-cutting issues” that requires laboratorians and nonlaboratorians to work together, agrees the CDC’s Dr. Boone, who adds that one of the institute’s goals is to bring stakeholders with different perspectives to the table.

Computerized systems that link test results to evidence-based clinical protocols could play a role in helping ensure clinicians do what’s best for patients. Such systems are, in fact, already in play in some large health care systems. As an example, Vanderbilt University Medical Center’s computerized physician order-entry system includes dozens of care protocols for clinician followup of inpatient lab testing.

“The CPOE system includes protocols such as an anticoagulation adviser that suggests optimal therapy for patients with clots in their legs or lungs,” says Randolph A. Miller, MD, an internist and university professor of biomedical informatics, medicine, and nursing at Vanderbilt.

Robert Michel, editor of The Dark Report, predicts that the more that ordering physicians feel pressure to improve their own quality performance on standardized measures, the more motivation they will have to consult pathologists for help in selecting the right lab tests and using test results to improve outcomes.

Some experts predict that implementing valid national lab quality indicators could reduce the volume of ineffective laboratory testing. Dr. Meier explained the backdrop for this concept at the Executive War College by pointing to two ostensibly conflicting hypotheses about the impact of lab testing on patient care.

“The Wennberg-Fisher model presents data that lab testing, at best, has no impact on health care outcomes,” Dr. Meier said. In fact, more testing, in some cases, may actually cause harm. In that sense, less lab testing is better. According to this model, lab testing is what’s known as a “supply-sensitive service,” which means the more lab testing is available, the more it’s done, even though the services do not have any measurable impact on patient outcomes. “Supply-sensitive services account for about 30 percent of CMS’ costs, so they have the payer’s attention,” Dr. Meier said.

For example, “the amount of lab testing done in the most consumptive one-fifth of Medicare regions is almost twice the test volume done in the least consumptive quintile, so the difference is dramatic,” he added. “Yet these differences in testing are not affected by differences in diagnoses or different measures of illness severity among patients.” (The Wennberg-Fisher data are summarized in a special edition of Health Affairs published in fall 2004.)

By contrast, another economic analysis, by Frank Lichtenberg of the National Bureau of Economic Research, who was at the institute meeting in April, “shows that the amount of money spent on developing and providing certain new tests is less than the savings produced by using these tests in patient management,” Dr. Meier says. Examples are tests for hemoglobin A1c and hyperlipidemia that are used to monitor patients with diabetes mellitus and those with hyperlipidemia who are treated with effective lipid-lowering drugs.

While not cited in the candidate quality indicators, testing for microalbuminuria at levels below that detected by urine dipsticks can also be linked to measurable positive patient outcomes in diabetic or hypertensive patients when the testing triggers appropriate medical intervention, Dr. Meier says. Better diabetic control or use of ACE inhibitors in such cases, for example, has been shown to stave off progression of chronic kidney disease or at least delay the need for dialysis.

To explain how both of the hypotheses about the impact of lab testing could be valid, Dr. Meier postulates that the positive impact of effective cost-saving testing described by Lichtenberg may be diluted by the huge volume of ineffective testing that’s done routinely.

That is, “the ineffective testing can be thought of as ‘noise,’ whereas the effective testing is the ‘signal,’” Dr. Meier explains. Laboratory quality indicators can lessen the distracting noise and enhance the signal that leads to positive outcomes. “By increasing process quality in clinical lab testing that monitors and reduces defects, health care systems can dampen or reduce the noise of ineffective testing,” he says. “The effort to link testing practices to beneficial outcomes is an attempt to amplify the signal that the noise has been obscuring.”

Unnecessary tests, of course, can result in increased morbidity or mortality, says Vanderbilt’s Dr. Miller. Vanderbilt has found that using a peer-management resource-use approach to ordering of serum chemistry panels has reduced testing for all components in the panel by 30 percent without a concomitant increase in patient morbidity or mortality.

Before the new system went into effect, many Vanderbilt physicians were ordering the basic metabolic panel every morning (all seven tests daily) in an “automatic” way, says Dr. Miller. Those physicians can now order the panels only one day at a time, and even then, they must order the seven components separately. The ordering physician must view a Web page within the computerized physician order-entry system that graphs the patient’s last week of results for each component of the panel. If the graph shows unchanging or normal results for a given component test, the physician has a strong incentive to not order another test instance.

A study by Dr. Miller and his colleagues published in August 2004 in the Annals of Internal Medicine showed that when Vanderbilt physicians ordered serum chemistry panels in recurring, “autopilot” mode, they discontinued the testing, on average, one or two days later than they did after the intervention that made them decide daily whether to order any tests at all. A net decrease in testing of even 20 percent (one day out of an average hospital stay of five days) could translate into millions of tests not done each week at the national level, Dr. Miller says.

Implementing process quality measures may also lower labs’ costs. Lean labs cut their costs by adopting the single-flow manufacturing principles that catapulted Japanese automakers to world-class status. “And Lean principles used by labs address the same issues as do the Q-Tracks-validated quality measures,” says Dr. Meier. For example, Lean labs have systems to verify test orders and stop the flow of misidentified specimens.

The government is counting on quality measures to lower health care costs by improving patient outcomes and tying provider reimbursement to performance. Pay-for-performance (P4P) is the background against which the CMS sees efforts to set quality standards and indicators to monitor them, Dr. Meier said at the Executive War College. “The P4P penalties for providers in the lowest or 10th percentile performance stratum are there to keep the P4P initiative ‘cost-neutral,’” he adds.

If you listen to CMS administrator Mark McClellan, pay-for-performance is a major part of the agency’s agenda, says Kenneth W. Kizer, MD, MPH, president and CEO of the National Quality Forum, Washington, DC, who notes that CMS plans to start pay-for-performance in ambulatory settings next year. Acute-care hospitals that voluntarily report their performance on selected quality measures will receive the full Medicare rate hike again this year, or 0.4 percent more than their nonreporting counterparts.
Pay-for-performance extends beyond government payers: There are currently more than 100 P4P initiatives in the private sector. Though none involves labs as far as Dr. Kizer knows, “one has to think that private payers are considering that approach for labs.

“Labs may be less visible than hospital care, but labs are part of the health care system and are certainly important from a quality perspective.”

Even if quality indicators do not play an immediate role in laboratory pay-for-performance, they may soon become a part of laboratory inspection, Dr. Meier predicts. “Inspectors already get laboratories’ proficiency testing information before they inspect laboratories; certainly indicators about lab performance will be another item that surveyors can use as context for their on-site examination of laboratory performance.”

Some say the CMS applies a back-door form of pay-for-performance through survey fines and decertification from government payer programs triggered by poor performance on quality indicators. “In some ways, I suppose you could characterize it that way,” Dr. Kizer says. By contrast, the government’s new vision of pay-for-performance is more of a reward system. “The potential value of pay-for-performance is that it provides incentives and rewards for doing better, so it’s not just a big stick. Big sticks only go so far, since human beings respond better to positive rewards and feedback,” Dr. Kizer says.

That’s not to say the government shouldn’t levy fines and penalties on providers that do egregious things, he adds. “But using a punitive approach toward long-term care facilities, for example, over the past two decades hasn’t produced the quality improvements the government would like to see. One of the lessons we have learned and which is being stressed by the National Commission for Quality Long-Term Care, which the NQF sponsors, is that we need a new approach to promoting quality.”

Standardized national quality measures may hold some unexpected perks for laboratories competing for contracts with payers based on quality performance. CDC’s Dr. Boone notes that some laboratories have expressed concerns about payers demanding certain quality measure information which, in some cases, the labs hadn’t collected.

“And it’s hard to compete if you don’t have what payers are asking for,” Dr. Boone says. “If a common set of performance measures were developed that all payers could agree to, the laboratories would at least know what to expect.”