Measuring Quality of Anticoagulation Therapy
Study Tests Risk-Adjusted Percent Time in Therapeutic Range
By Bill Malone
An estimated 2 million Americans begin anticoagulation therapy with warfarin every year, yet this common and potentially dangerous drug lacks a uniform method of monitoring quality of care. This issue of Strategies explores a recent study that used risk-adjusted time in therapeutic range to profile the performance of 100 anticoagulation clinics in the Veterans Health Administration.
Numerous studies have shown that oral anticoagulation therapy is difficult to get right. Excessive anticoagulation can lead to serious or even fatal bleeding, and inadequate anticoagulation makes patients vulnerable to stroke and other events. However, without a robust means of measuring quality of anticoagulation therapy, broad quality improvement has been challenging. This situation prompted researchers from Boston University School of Medicine and the Bedford VA Medical Center in Bedford, Mass. to investigate using risk-adjusted percent time in therapeutic range (TTR) as a quality indicator (Circ Cardiovasc Qual Outcomes 2011;4:22–29).
“Anticoagulants account for more emergency room visits and more complications than all other drugs combined. And yet nobody measures how well we deliver them,” said lead author Adam Rose, MD, MSc, assistant professor of medicine at Boston University School of Medicine and investigator at the Bedford VA Medical Center. “If a surgical procedure had a lifetime complications rate similar to that of lifelong warfarin, you bet we would have a huge industry devoted to measuring it. But for some reason no one has done that with warfarin. I think that’s ridiculous. I think we should measure quality for everything that’s important and can be measured.”
The researchers profiled 100 sites of care within the Veterans Health Administration and calculated TTR for 124,551 patients and mean TTR for each site. They also calculated expected TTR for each patient and each site based on a custom risk adjustment model that took into account comorbid conditions, medications, and hospitalizations, among other variables. The researchers found a mean TTR of 58% for the entire sample, with performance varying widely among the VA anticoagulation clinics—from 38% to 69%. They found that risk adjustment made an important difference for some sites in performance rankings. For example, the clinic that was originally ranked 27th out of 100 had a risk-adjusted rank of 7th. The study also found that site performance on risk-adjusted TTR was consistent from year-to-year, suggesting that risk-adjusted TTR measures quality of care, which is stable over time, rather than merely measuring statistical variations from year to year.
Showing that a risk adjustment model works for anticoagulation is an important step on the road to having well-accepted, reliable quality profiling for warfarin therapy, Rose said. “Without risk adjustment, sites that compare unfavorably to others will dismiss their low ranking as due to their having sicker patients, so risk adjustment adds credibility to a quality indicator,” he said. “This study is not a breakthrough in the sense of discovering something completely new: time in therapeutic range is not new, risk adjustment is not new, the idea of profiling sites of care on performance is not new. Rather, the message of this paper is that everyone needs to wake up and profile performance in oral anticoagulation because it’s important. It’s actually not that hard to do, and we have a good example of how it can be done.”
Hospitals routinely use risk-adjusted mortality rates, Rose noted, after coronary artery bypass and other surgeries. Risk adjustment is also used to predict costs of care and as a basis for compensating physicians. Although other off-the-shelf risk adjustment models are available, none seemed to fit well for measuring quality of anticoagulation care, Rose said. “Most models look at general outcomes like mortality, cost, or length of stay. However, for this particular problem, the variables you need are very different, and with a situation like this that’s unique, it makes sense to create your own.” Their model took into account traditional comorbid conditions like cancer, chronic kidney disease, diabetes, hyperlipidemia, and hypertension. They also added socioeconomic variables like poverty in zip code of residence.
The researchers’ risk model is a good start, but could still be improved, noted Gwen McMillin, PhD, medical director of ARUP Laboratories’ toxicology lab and associate professor in the department of pathology at the University of Utah. “Clearly the VA is a good system to test this in, and some other major healthcare provider systems would probably benefit from a quality index like this,” she said. “I would, however, have liked to see pharmacogenetics incorporated into this, as well as diet. We know there is increased risk associated with some genotypes, and certainly a huge food-drug interaction exists relative to warfarin and vitamin K that would impact risk.”
While Rose and his team intend to use this model to push a performance improvement program within the VA, other institutions would do well to at least measure the proportion of values in range every month within their own clinic, Rose suggested. “Risk adjustment does matter for some sites, and is probably the best way to go if you want to tie reimbursement to performance and it has to be really credible,” he said. “On the other hand, if an institution just wants to track how they’re doing month-to-month, they don’t need to risk adjust for that, and they don’t need to compute a time in range for that either. But tracking proportion of values in range is definitely doable, and if there is an anticoagulation clinic out there that isn’t at least measuring this, they should be embarrassed, because it’s easy.”
The researchers’ wider goal is to improve the mean TTR in the VA to at least 70%, up from the current 58%. “Achieving this goal will require a focus on system-level approaches to improve processes of care. For example, our group has shown that more judicious decisions regarding when to change the dose of warfarin can improve TTR considerably. Addressing this and other processes of care in a systematic way could greatly improve TTR in the VA; however, continually measuring risk-adjusted TTR will be a precondition to any program of quality improvement,” the authors wrote.
Labs can help providers improve quality of care for anticoagulation therapy by carefully monitoring point-of-care and other onsite devices that are being used more and more frequently, McMillin emphasized. “The lab is in the best position to make sure that people really know how to use those devices and that each device is calibrated properly,” she said. “The success of warfarin therapy really depends on a very skilled clinic that takes a comprehensive view of quality in every area.”