In the era of evidence-based medicine, clinical guidelines, and personalized medicine, one would think that convincing clinical trial data would influence clinical practice if disseminated in an appropriate manner. However, it has been estimated that only 50% of current medical practice is evidence-based, clearly demonstrating a compelling need to collect and analyze additional data to better inform practice. Current data are usually gathered from a variety of sources, including clinical trials, observational studies, and meta-analyses. Yet according to Jeff Forringer, CEO of IntrinsiQ, data from oncology practices provide real-world outcomes that give better insight into the efficacy of cancer therapeutics.
There is an ongoing debate in the oncology community about the clinical value of treatment pathways. Many physicians are concerned that practice guidelines limit their flexibility and their ability to “learn as they go” about the drugs and regimens they prescribe. So IntrinsiQ looked at data captured by the company’s software application IntelliDose regarding physicians’ drug selection in a certain clinical area. IntelliDose captures the treatment decisions and details of almost 20,000 cancer patients a month, creating an extensive database of accurate, detailed, and timely information about the medical oncology care process. Our aim in looking at these particular data was to see whether the physicians’ “real-world” experience helped them prescribe the most effective therapies.
|FIGURE 1 |
What Real-World Experience in One Setting Showed
We looked at the time-to-progression (TTP) associated with certain drug regimens in metastatic breast cancer in order to see how long it took for patients’ disease to progress while they were receiving selected combinations; we looked at real-world data, which are far removed from the kinds of results we see reported in randomized clinical trials. We grouped patients presenting with breast cancer stages I-III according to receptor status: patients were thus placed into a HER 2+, HER2–, or borderline group. We additionally isolated a subset of patients within the HER2– group, those with triple negative disease (Figure 1). We then grouped the women by their drug therapies and their drug therapies’ TTP performance category (Table).
The highest-performing group of regimens in the first-line metastatic setting, as judged by average TTP, varied by receptor status. Among stage I-III patients with HER2+ disease, trastuzumab(Drug information on trastuzumab) monotherapy and gemcitabine(Drug information on gemcitabine)-containing combinations yielded the longest TTP—about 350 days—while taxane-containing combinations and bevacizumab plus taxanes yielded the longest TTP in HER2– and triple negative patients, respectively.
None of the highest-performing regimens exceeded a 16% share in the metastatic treatment setting, and among stage I-III patients with HER2+ disease, the poorest-performing regimens, taxane monotherapies, had a 31% share.
For none of the lines of therapy were the peak performing drug combinations for patients with stage I-III breast cancer prescribed to more than 50% of patients. Still more striking was the fact that in the patients with metastatic stage I disease who were receiving first-line therapy, between 30% and 50% received the worst performing combinations; these data suggest that breast cancer therapy for patients with stage I-III disease can be improved significantly by using treatments available currently.
How Often Do Community Oncologists Use the Best Drug?
After analyzing the data, we drew conclusions about which regimens worked best—and we found that in the settings of first-line and adjuvant therapy, the drug therapies that showed the best TTP were never the primary therapy of choice—and in the setting of second-line therapy, the best-performing regimen was the primary choice only in patients with HER2+ disease (Figure 2). These results debunked the myth that real-world clinical experience is a pathway to better decision making. Naturally, these conclusions are based on less rigorous data, not the hard scientific analysis seen in a randomized clinical trial. The study and the findings are meant chiefly to be thought provoking, to stimulate a discussion about the real value of data in the clinical setting as it relates to decisions and outcomes.
Still, the underlying implications of these data are important, especially if we want to develop effective treatment pathways. The central question becomes: Are we going to rely on real-world data or the results of clinical trials in order to develop best practices clinical treatment pathways?
Another issue that’s germane to pathway development is analyzing the way in which doctors treat patients. Physicians tend to know how they treat patients qualitatively, but by and large, they do not really understand their treatment patterns from a quantitative point of view. And if you do not have a grasp of the quantitative aspect of your treatment patterns, whether from the provider or the payer perspective, how can you have an informed discussion about improving outcomes?
For instance, these real-world data showed that bevacizumab-containing combinations had a shorter average time to progression across all lines of treatment in patients with HER2– breast cancer, compared with combinations not containing bevacizumab. The difference was most pronounced in the adjuvant and metastatic I lines of therapy (Figure 3).
The take-away message from this exercise is that according to the data we collected, the majority of patients with breast cancer who are receiving bevacizumab do not have significantly better outcomes than their counterparts who are not receiving bevacizumab. These data were available long before the FDA decided to grant a supplemental indication for bevacizumab, for breast cancer.
From a policy standpoint, there is a data conundrum in that the new administration’s healthcare reform is largely built on a move toward research that seeks to identify the best possible therapies and treatment schemas. But to do that, you need data. So the question that needs to be answered is how we are going to use real-world data that yield different results than data collected from strictly configured clinical trials.
There is no need to wait for electronic medical records and total interoperability; quality real-world data are out there. Let’s start the discussion about how to use them to achieve better outcomes for our cancer patients.