Part 2: Communicating evidence of value
Roundtable moderated by Steve Halasey
The clinical laboratory community has long recognized the need to provide payers and policymakers with a clear understanding of the value of diagnostic testing. But moving that task from the ‘to do’ to the ‘done’ column has proven to be more difficult than those outside the community could ever imagine, and experts acknowledge that the need to demonstrate the value of diagnostics is as strong today as it ever was.
To find out more about how their makers and users view the value of diagnostics—and how changing views are affecting the adoption and use of such tests—CLP recently spoke with a number of experts in the field (see sidebar, Roundtable Participants). Below is the second portion of the resulting conversation, in which the participants consider what types of evidence are needed to demonstrate the value of diagnostic testing, and how that evidence can be best communicated to providers and payors.
CLP: What kind of evidence is necessary to demonstrate the value of diagnostics?
Lâle White: In the past, physician acceptance of a test was all that was necessary to support its adoption and use. But now that’s not enough, because payors don’t have confidence that tests are being used properly. As a result, health economic analysis is becoming a lot more important than it used to be.
Health economics studies are a way to demonstrate the value of a product or procedure, which ultimately depends on patient outcomes. It’s a longer scope methodology for proving clinical utility, because clinical utility really has to do with patient outcomes and the costs associated with those outcomes. That’s the evidence that payors are looking for.
FDA approval of an assay is focused on demonstrating its safety and efficacy, which speaks to the clinical value of the test, but not its utility. Health economics analysis is the clinical utility piece that payors need to support their coverage and reimbursement decisions. And it’s not just about new tests. Payors and providers have even raised questions about old tests that are considered standard of care, asking whether and under what circumstances such tests should be covered, because they still have questions about the proper use and optimization of diagnostics.
Franz Walt, MBA: In the past, we would have been communicating with laboratory specialists. But increasingly, purchasing decisions are being made from the procurement department or C-suite, and the people there are not really laboratory specialists. That’s why health economics has its own, distinct type of marketing tools.
We do have some individual studies showing clearly the positive effects of diagnostics. At the Floyd Medical Center, for example, we implemented a technical solution to expedite the evaluation of chest pain patients. That solution reduced the emergency department (ED) readmission rate by 62%, providing clear evidence of how relatively little investment in a cost-effective test can lead to significant improvements in patient care.
In another example, at North Memorial Healthcare, we improved workflow efficiencies, resulting in a 19% reduction in turnaround times for a basic metabolic panel, and a 17% reduction in turnaround times for troponin testing. Medically trained professionals understand that faster test results enable providers to do more with fewer resources. Especially for patients in critical condition, this can result in a better quality of care at a very cost-effective rate. But that benefit is very difficult to establish comprehensively across a wide range of assays. Siemens, for instance, offers approximately 900 assays.
We have also developed case studies on lab automation with NHS Tayside and with the Carlos Haya Hospital, where we have been able to demonstrate faster, higher quality results, with fewer errors. But it remains challenging to translate that innovation into a pricetag, and to figure out how to get paid for the innovation. That is also a difficult bridge to establish here in the United States.
Again, what’s needed is a holistic perspective, not individual data points. Asking ‘how much does the test cost?’ or ‘can we reduce the cost of the test?’ yields answers with a singular, very narrow perspective. A better approach would be to look at the total disease or episode-of-care costs, and then calculate the extent to which costs can be reduced—and the quality of care improved—by using the test.
Robert D. Jenison: In the realm of infectious diseases, and particularly for hospital-acquired infections, there are an increasing number of high-quality studies demonstrating that certain positive test attributes—such as accuracy and faster turnaround times—are highly correlated with improved clinical outcomes and lower treatment costs. In turn, hospitals are increasingly taking those findings into account when deciding which tests to adopt. The days of making such decisions based solely on a simple analysis of upfront test costs are fading.
Chris Bird, DPhil: What it comes down to is whether there is evidence that the diagnostic can improve outcomes—that it is going to improve guideline compliance, doctor compliance, patient compliance, and so on. Many companies, including Roche, employ medical science liaisons—field-based clinicians whose role is to communicate with physicians. Their discussions are all about evidence-based medicine—how we can help educate clinicians about outcomes, about comparative effectiveness, and about the relevant health economics.
Studies to establish a test’s turnaround time are useful for gauging optimal efficiency. Getting a test result at the right time can change care, and adds importance to the potential value of point-of-care diagnostics. The more timely a test result can be, the more likely it is that the physician can provide a quick diagnosis, which in some cases means the difference between life and death. Evidence to prove the value of diagnostics can come in a variety of different ways.
CLP: When labs are developing evidence to support tests they have developed, do they undertake the same kinds of studies as manufacturers do? Is there a difference in attitude, or in the way that labs view the question of evidence for their tests?
Brian R. Jackson, MD, MS: There are different layers of evidence. Evidence about a test’s analytic performance, for instance, must be provided at the level at which the test is developed and implemented. The hard part is taking that up to the level of clinical benefit. In the diagnostics world as a whole, we all need to figure out ways to get a much deeper and richer evidence base for the clinical impact of diagnostic tests, and in particular those areas that will distinguish between one test and another, or when to test and when not to test.
It’s possible for regulatory bodies to set the clinical evidence bar too high in certain situations and effectively quash the market for certain types of tests. By way of analogy: because of the enormous expense associated with getting a drug to market, pharmaceutical companies put a disproportional amount of effort into developing drugs expected to be blockbusters, and limit the resources they put into developing drugs that are expected to be less lucrative. Depending on how the evidence for diagnostics plays out, there’s a huge danger that the diagnostics world could end up in the same situation, where the only tests that make it to market are those that have large commercial potential. In turn, that would create a really large category of tests that either don’t get developed or don’t get improved over time.
The challenge for the diagnostics world is to figure out what kinds of evidence are most critical for particular types of tests. In its proposed regulations for laboratory-developed tests, FDA is trying to apply a framework based on patient risk, but that is very difficult to do for a diagnostic test. That paradigm fits much better for therapeutics than for diagnostics. But if one were to compare the risks of a test to the risks associated with alternatives, one might arrive at a cost-effective approach to clinical validation that would work.
Daniella Cramp: Depending on the setting and the nature of the test, the value of a diagnostic can be expressed in clinical, operational, or economic terms. Bedside blood gas testing, for example, provides both operational efficiencies and clinical benefits by streamlining processes from hours to minutes, and by enabling clinicians to determine sooner the best care pathway for patients.
In turn, the types of evidence needed to demonstrate the value of a given test may include the performance of the test in terms of sensitivity and specificity; its ability to create workflow and time-to-result efficiencies; or its ability to produce cost savings by way of reduced hospital stays, staffing needs, and process errors.
Patrick R. Murray, PhD: There are a number of ways to demonstrate the value of diagnostics. Conducting workflow and economic analyses, for instance, are good ways to build the evidence needed to demonstrate the value of diagnostics. In general, determining the value that diagnostics bring to laboratory professionals and others can be found by answering such questions as: Is the test accurate? Are the results available in a timely manner? Did the results help the patient?
CLP: How do new technologies and tests add to the value of diagnostics? For instance, Lâle, how do the automated decision support systems you mentioned earlier contribute to the clinical or economic value of diagnostics?
White: There are three major parts to a decision support system. The first part is a real-time decision support tool that enables the physician to communicate with the system and gather guidance about the selection of tests appropriate for the patient. The results provided by the system must include enough information for the physician to make a good decision.
The second part is a mechanism that enables real-time collaboration and consultation among physicians and potentially providers at the time that the test results are delivered. The technology must be sufficiently robust to permit sharing of documents, images, and so on. The purpose of this mechanism is to ensure that the therapy selected for the patient is the correct one. In the oncology arena, for instance, several studies have indicated that patients’ diagnosis and therapy can change almost 30% of the time when the oncologist, pathologist, and radiologist collaborate and speak to one another. In the case of breast cancer and prostate cancer, it’s almost 50% of the time.
And for the third part, since all of this is ultimately a reimbursement issue, the system should facilitate a health economic analysis of the whole process. In the end, companies and physicians still need to demonstrate to payors that all of their decisions ended up providing a good health economic solution that took advantage of efficiencies; produced good, effective outcomes for the patient; and reduced costs.
Costs are not necessarily reduced by controlling the overutilization of laboratory testing; that’s not where the greatest opportunity lies. The primary opportunity for cost reduction is actually in selecting the right therapy. McKenzie & Company analyzed spending on $300 billion worth of pharmaceuticals, and found that half of that amount was not effective for patients and, in some cases, it was detrimental. They concluded that half of that waste, or roughly $75 billion, could have been saved if the right genetic testing had been performed.
In this analysis, the cost of the diagnostics is just a small fraction of the potential savings they can bring about when used properly and optimized. In the end, since payors are looking at outcomes and quality, companies must perform an economic analysis for each test they are submitting for reimbursement. To get payors to approve coverage for the right circumstances and at the right reimbursement rate, the economic analysis should be attached to the financial package that a company submits to a payor. These are all parts of a good financial and economic analysis.
When delivered to a payor for a coverage and reimbursement determination, all of this information needs to have a technology solution wrapper around it that allows the submitting company to analyze and provide the right set of data for the payor.
Walt: There are of course also newer and tougher rules that the Centers for Medicare and Medicaid Services (CMS) has rolled out as part of its implementation of the Patient Protection and Affordable Care Act (PPACA). These rules penalize hospitals when patients must be readmitted as a result of conditions that they acquired in conjunction with a hospital stay—so-called hospital-acquired conditions.
As a result of these rules, and to avoid being penalized, hospitals are more likely to use IVD testing prior to admission, to determine whether a patient already has one of the specified conditions. These tests act like kind of an insurance policy, allowing hospitals to minimize or avoid penalties. But such testing may also lead to the early detection of conditions that the patient was not aware of, and can contribute to earlier treatment of otherwise undetected diseases.
There are a number of diseases that might be captured by such comprehensive preadmission testing. NHS Tayside in Scotland has a study showing that procalcitonin testing led to earlier detection of sepsis, and fewer readmissions after early treatment. Cardiac testing determines whether the physician can send a patient home or must admit them for observation or immediate treatment. Detection of diabetes, heart attack, HIV, and stroke are other good examples.
Such testing can even be of use for nonemergent conditions such as breast cancer. Siemens offers two new blood-based tests that help to identify the right patients for a very expensive pharmaceutical treatment. From the point of view of a hospital, performing these tests might be seen as cost avoidance: the hospital shouldn’t provide the treatment to those who don’t qualify. But from the point of view of a patient, the test helps to ensure that appropriate patients get the best possible treatment.
Jenison: Great Basin provides molecular diagnostic testing, which is beginning to have a very strong impact on the clinical microbiology lab. Researchers have started to perform interesting assessments on how to deal with the data generated by molecular diagnostics. One thing they’re finding is that using data to initiate broader engagement within the hospital can help to turn information into actionable results.
There had been some studies suggesting, contraintuitively, that identifying a pathogen more quickly might not correlate to better patient outcomes. But when a pharmacist was given that information sooner, and consulted with the physician about selecting the right therapy, there was a huge benefit with respect to the hospital’s economics—meaning shorter lengths of stay, lower treatment costs, and therefore also better outcomes for the patient.
Many of these infections occur as a result of a surgery or some other service that’s covered under diagnosis-related group (DRG) codes, but not reimbursed directly. There’s some cost mitigation there as well, related to lowered morbidity and shorter lengths of hospitalization, again as a result of more-rapid detection of infectious pathogens.
CLP: In the traditional model for disseminating the findings of clinical research, investigators seek out prominent peer-reviewed journals, and important findings eventually rise to the surface, to be incorporated into practice guidelines compiled by medical specialty societies. Is the current technology sufficiently robust to replace that model with a real-time decision support system?
White: Yes, I believe the technology is absolutely ready. Xifin offers a platform and product that actually performs these functions, and it is being utilized both within health centers and among separate health centers. Being able to communicate and consult with extramural partners is important, because not every health center has expertise in every area that its patients might encounter.
The major issue with such systems is where they can source the content they need to function as decision support tools. Fortunately, a number of very strong content vendors are developing—especially in the field of oncology. Just as important, the technology exists for all the data within a health center to be consolidated—essentially providing home-grown content for the decision support system. Xifin’s products permit clinicians to capture images in a variety of formats—whole-slide images, digital radiology images, DICOM images, and so on—and to consolidate them with electronic medical records (EMRs), so that the analysis needed to share the records can be performed.
The big problem for most health centers, of course, is that they have always treated diagnostics and the clinical laboratory as an afterthought, and their financial structure as it pertains to diagnostic testing has never been very well managed. As a result, they have a lot of difficulty performing economic analyses to understand the value of their operations. Private laboratories have paid attention to the health economics of their tests for many years, because they needed this information to develop their products and get them accepted for reimbursement. Health centers have not been so engaged in performing financial and health economic analyses, but they’re going to need to do so if they expect to continue receiving the coverage and reimbursement they need.
For that matter, health centers need to learn how to perform financial and health economic analyses just to determine whether their accountable care organizations (ACOs) or other cost-reduction processes are actually working. Right now, many organizations are attempting to reduce costs by reducing utilization. But that strategy only reduces a fraction of the organization’s costs, and if the right testing and therapy decisions aren’t followed, it could wind up costing the organization even more money. Organizations that can’t perform financial and health economic analyses are unlikely to achieve the cost savings they hope for.
Steve Halasey is chief editor of CLP.