Conversation with the Cancer Letter
Basch: Patient-reported outcomes data can make cancer drugs safer, more tolerable
Director of the Cancer Outcomes Research Program, professor of medicine and public health at the UNC Lineberger Comprehensive Cancer Center.
Fifteen years ago, when Ethan Basch started developing measures for aggregating patient-reported outcomes, many of his colleagues in oncology saw no promise in this enterprise.
“We conducted cooperative group studies and showed that the vast majority of patients were willing and able to self-report even when extremely ill or close to death or hospice,” said Basch, director of the Cancer Outcomes Research Program and professor of medicine and public health at the UNC Lineberger Comprehensive Cancer Center.
“These data were useful for responding to my skeptical colleagues who asserted, ‘We understand our patients well, so we don’t need them to provide their own information. We know what’s going on with them, we talk to them all the time. And patients probably aren’t willing or able to report their toxicities during trials, they have a lot going on.’ But our data showed those assertions were not true.”
Now, NCI plans to commit $3.25 million in fiscal year 2018 to fund three to five awards focused on studying patient adverse event data to better understand how patients are tolerating cancer treatments—as part of the institute’s PRO-CTCAE platform, also known as the Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events. This tool was developed by Basch’s team under contracts from the NCI.
Last year, NCI issued the first Request for Applications on PROs based on the scientific recommendations of NCI’s Blue Ribbon Panel for Vice President Joe Biden’s 2016 Cancer Moonshot.
In a wide-ranging conversation, Basch outlines the development of the PRO-CTCAE, FDA and industry’s roles, and his thoughts about the next steps for NCI.
“With the caveat that I’m a researcher, not an NCI employee, in my opinion, the day-to-day oversight of the PRO-CTCAE tool at NCI is ready to be moved into the Cancer Therapy Evaluation Program along with the CTCAE, and be managed by the same or a similar clinical research team,” Basch said.
“Enough is known about this tool for it to be integrated with the CTCAE, as was its initial intention, and be implemented in trials, through partnership with industry, the NCTN, and regulators. Research on the PRO-CTCAE like the pending Moonshot project on this topic could remain in the Division of Cancer Prevention, but this would be focused on expanding understanding of the tool and its applications.”
Basch spoke with Matthew Ong, a reporter with The Cancer Letter.
What was your role in the creation of the PRO-CTCAE? Did you design the framework? When did this work begin?
My research collaborators and I developed the PRO-CTCAE, under contracts to the NCI between 2008-2015.
The foundational work for assessing patient-reported adverse symptoms goes back further than the PRO-CTCAE itself, though. About 15 years ago, I developed an interest in this area based on an observation that, in early phase trials we were conducting, many patients experienced debilitating symptomatic side effects that impact tolerability.
Yet, when we looked at the reported CTCAE data for trials, that information was missing. It was invisible in the trial documentation because we, not the patients, were reporting on the patient experience.
It occurred to me that this was a missed opportunity—there seemed to be an excess of risk to patients from toxicities that we weren’t aware of. So, my collaborator Dr. Deb Schrag and I conducted a series of studies where we started to collect symptom information from patients about their side effects during treatment. The first thing we found was that we as clinicians miss about half of our patients’ symptoms.
We became interested in developing standardized tools, because we thought, if we can capture this information from our patients, this is an opportunity both to improve management of patient symptoms and side effects, and to improve the integrity of clinical trial data.
If you think about the downstream consequences of missing many symptomatic adverse events our patients are experiencing, when an investigator or regulatory authority is reviewing data from a trial, they are going to have a substantial underestimation of the risk of the product. I felt that by improving the tools that we use, we could improve both the quality of data and patient management.
We began developing standardized tools and testing them in various ways, both locally and nationally. We conducted cooperative group studies in the old [Cancer and Leukemia Group B] (now the Alliance), where we inserted these early tools into clinical trials and showed that the vast majority of patients were willing and able to self-report even when extremely ill or close to death or hospice.
These data were useful for responding to my skeptical colleagues who asserted, “We understand our patients well, so we don’t need them to provide their own information. We know what’s going on with them, we talk to them all the time. And patients probably aren’t willing or able to report their toxicities during trials, they have a lot going on.” But our data showed those assertions were not true.
You probably ran into people saying, “Self-reports are by definition, variable, non-discrete data and it’s not easy to turn it into something useful,” right?
Absolutely, that was another area of skepticism where some clinicians asserted, “Well, you know, my patients don’t really understand their own symptoms. I as an oncologist understand symptoms well. I see patients across the whole spectrum of illness. I can normalize or calibrate how patients feel.”
But when we studied clinician reporting of CTCAE symptoms in trials, we found it to be unreliable. In inter-rater reliability studies, we found that investigator agreement on the grading of CTCAE symptoms was low. If you and I saw the same patient on the same day, and we used the CTCAE criteria, there’s a good chance we would not agree with each other. On the other hand, in our studies of patient self-reporting, we found that patients are highly reliable in reporting symptoms over time. So we pushed forward with these empiric data.
But you’re absolutely right. There was definitely an old-school bias out there that “I know better than my patients,” and that we can’t really trust patients to be reliable. But the data show that’s not true. What we find is that, in fact, when clinicians report on their patients’ symptoms, that’s where the real subjectivity comes in.
It happens as the clinician eyeballs the patient and makes assumptions about their health status that may not be true. We know from the psycho-behavioral literature that it is very hard for a person to understand another person’s feelings or symptoms.
And so, with the standardized criteria that you and your team developed, did CTCAE itself exist way before that?
The CTCAE already existed, and provided a framework on which to build the patient-reported tools. The CTCAE is an item library designed for clinician reporting of adverse events, with a lot of technical language. It includes about 800 discrete adverse events in it. Many of these are not appropriate for patient reporting, like retinal detachment. But items like nausea, dysphagia, taste disturbance, or sleep problems are appropriate for patient self-reporting—experiences that the patient knows best.
About 78 out of 800 items, right?
Yes, about 10 percent of the CTCAE is comprised of symptoms. In our early work, we created quite literal adaptations of CTCAE criteria into patient language, called the STAR (Symptom Tracking and Reporting) system, which predated the PRO-CTCAE.
So, NCI and FDA became interested in your work?
About 10 years ago or so the NCI became interested and I was invited down to present our early work.
Several people at the NCI were particularly enthusiastic and championed the concept—there was Dr. Bryce Reeve [then at the Outcomes Research Branch in the Division of Cancer Control and Population Sciences and now director of the Center for Health Measurement and professor of population health sciences and pediatrics at the Duke University School of Medicine] and Dr. Lori Minasian [deputy director of the NCI Division of Cancer Prevention], in two different divisions at the NCI.
They issued a series of contracts, to develop a formal patient version of the CTCAE called the PRO-CTCAE. As an aside, Bryce was also the one who hatched the idea of PROMIS (Patient-Reported Outcomes Measurement Information System) and named it. Jeff Abrams in CTEP was also an early advocate for this tool, and continues to valiantly support its implementation.
Bryce subsequently left the NCI and went into academia, but he was really responsible for the NCI’s initial push to develop patient-reported tools like PROMIS and PRO-CTCAE that are coming into their own now, more than a decade later. He was important, as was Lori, who’s still at the NCI, as well as Dr. Ann O’Mara who’s overseeing the new PRO-CTCAE NCI Moonshot announcement, and Dr. Sandy Mitchell who took over Bryce’s role when he left the NCI.
NCI leadership has also been highly supportive, including Dr. Bob Croyle at DCCPS, Dr. Barry Kramer in DCP, and leadership in [the Center for Biomedical Informatics and Information Technology] and [the Division of Cancer Treatment and Diagnosis]. So overall, this effort has had broad-based and interdisciplinary support at the NCI. I’m delighted that Dr. Ned Sharpless, our new NCI director, is also very committed to PROs and the PRO-CTCAE for bringing the patient voice into cancer trials.
FDA input has also been key, and FDA stakeholders were involved from the start, particularly Lori Burke, Gini Kwitkowski, and more recently Paul Kluetz, who is a major champion for PROs in oncology at the FDA. Rick Pazdur has also been a strong supporter, and the PRO-CTCAE has been seen as a vehicle for enhancing the patient-centeredness of cancer drug development. The FDA has had several panels on the use of the PRO-CTCAE in trials as it’s been rolled out.
In terms of timeline, how long has CTCAE been around?
The CTCAE is a longstanding tool that’s been developed and maintained by the NCI, with revisions every several years. It’s been widely used in oncology both in clinical trials and in routine care. But there was no patient reporting component to it, which is the innovation of the PRO-CTCAE.
So, the NCI supported a team of researchers you led to develop the PRO-CTCAE?
With NCI contract support, we assembled a national team to build and test the PRO-CTCAE, also with terrific input from many NCI and FDA stakeholders. The team included incredible collaborators—Dr. Deb Schrag [Dana-Farber Cancer institute], Dr. Amylou Dueck [Mayo], Dr. Amy Abernethy [then at Duke, now at Flatiron Health], Dr. Deb Bruner [Emory], Dr. Charlie Cleeland [MD Anderson], Dr. Jeff Sloan [Mayo], Diane Paul and Cindy Geoghegan [patient representatives]. It was a very exciting time and project—one of the highlights of my career. We felt we were doing something new that could make a real difference for patients and improve trial data.
The first thing we did was to systematically evaluate the CTCAE and identify items that are amenable to patient self-reporting, that’s the 78 items that you referred to. Then, we developed a structure for how patients would report this information, that is, how we would ask the questions. Next, we had to generate patient-friendly language for terms like dysphasia and dyspnea in the CTCAE, which are too technical for patient questions.
Then, we did a national interviewing study, led by Drs. Jennifer Hay and Tom Atkinson at MSKCC, what we call a cognitive interviewing study, among patients from different backgrounds and cancer types. This study made sure that patients understood the terminology and that the questions truly measured the concepts of interest. This is a requirement when creating rigorous PRO tools.
Next, we conducted a large national validation study led by Dr. Amylou Dueck at Mayo where we looked at what’s called the measurement properties of the tool, to make sure that the items are valid, reliable, and sensitive to changes over time. We tested what the recall period should be, which is how far back people’s memory can reasonably go to remember these AEs (led by Dr. Tito Mendoza at MD Anderson) and the optimal recall period for the tool, which turned out ideally to be over the past 7 days, but could reasonably be stretched out to a month (led by Dr. Antonia Bennett at UNC).
We created an electronic system for administering the questions, and tested that through usability testing, led by Dr. Amy Abernethy. The whole initiative was overseen by an incredible project manager, Lauren Rogak, who I still work with closely.
And then, we implemented the tool in big national trials.
When thinking of patient safety and tolerability, could you explain how the PRO component can provide a more complete picture of what’s going on, especially in terms of how tolerability is understood?
There’s a lot of interest now in developing a more patient-centered approach to tolerability.
The current standard definition of tolerability is a bit circular: how well a patient tolerates a treatment. But, a more patient-centered approach has to do with the ability of a patient to continue treatment based on how they feel and function during that treatment. So, it’s the extent to which the impact on their symptoms or their physical functioning affects their ability to continue taking the drug.
Friends of Cancer Research is supporting a current effort to refine understanding of tolerability from the patient perspective through an international, multidisciplinary working group. This effort reflects the broader championing of PROs and the PRO-CTCAE by Friends of Cancer Research, particularly by its founder Ellen Sigal, who is an incredible advocate for cancer research, and its director Jeff Allen.
In order to operationalize that kind of approach to considering tolerability you need to have the patients’ own report of their symptomatic adverse events, in addition to essential non-PRO information such as hospitalizations and clinician-reported information from the CTCAE.
But the key point here is that it’s essential to have the patient’s self-report, for example to know if there’s an excess of rash, or myalgias, or disabling nausea and vomiting etc. that we wouldn’t otherwise appreciate. You’re going to get that best from patients to understand how well people are tolerating a drug. Once you have that information, you can make better decisions about dosing of drugs, or the frequency of giving a drug, or the balance between risk and benefits. And avoid picking a dose that’s ultimately tough on patients or will impair compliance.
There are multiple examples of drugs that were released that were really not tolerable for patients, and we didn’t figure that out until they were out on the market, because we simply didn’t collect how patients felt when they were receiving the drug in early phase development.
Which brings us to, could you explain in laymen terms, step by step, how exactly you or an investigator would use PRO-CTCAE in a clinical trial? So, sort of walk me through a case study if you would.
Sure. The PRO-CTCAE can be used in trials at any phase of development for a descriptive summary of adverse events, much like the CTCAE, or as a component of assessing tolerability—either for selecting dose in early phase work, or for comparing the tolerability of different treatments in later phase trials.
There are currently numerous NCI and industry clinical trials in which the PRO-CTCAE has been embedded. The generic way that it would be embedded would first involve selection of PRO-CTCAE items for administration in a trial. To do this, the investigators would anticipate the likely symptomatic side effects that would occur, based on what’s known about the properties of the drugs in all arms of the study, prior experience with the product, and if available some pre-trial questionnaires or interviews with patients who previously received the product(s).
Based on this information, they would create an electronic form to administer PRO-CTCAE items to patients corresponding to those pre-designated symptomatic side effects. An important point is that we don’t administer all PRO-CTCAE items in a given trial, only the salient ones, perhaps 10-15. I would also recommend including a mechanism to capture unsolicited symptoms from patients, for example through a free text box at the end of the PRO-CTCAE questionnaire. We have found this to be a rich source of supplemental information.
The PRO-CTCAE should be administered on a regular basis throughout active treatment, or at least during early cycles, either once a week, or every two weeks. Regular administration assures that we don’t miss important events during treatment. As I mentioned, this can be stretched out to as infrequently as every month, but that’s not ideal, because if a patient misses a report, they’ll have a long stretch between time points.
Reminding patients who miss self-reports is also very helpful for data completeness. For patients who don’t self-report on time, a PRO system can send an alert to an administrator, who can call the patient to capture the missing data and remind them to self-report next time. We have found that by doing this, we can boost completeness of our data by 15 percent, which is substantial.
After patients finish active treatment, you can space out PRO-CTCAE administration, for example, every six months for another one or two years depending on the context and the patient population.
In the analysis and reporting, there are a number of different ways to look at the data. We have quite a bit of active work in the area of how to analyze and visualize the data, and this is the topic of the planned NCI Moonshot project. I am particularly interested in how the data should look in drug labels.
There’s an industry PRO-CTCAE working group now with representatives from numerous companies working on this very question, which has been led by Alicyn Campbell and Sheetal Patel at Genentech, who have been important pioneers incorporating rigorous PROs in cancer trials.
The classic way to analyze adverse events in cancer trials it is to look at the proportion of patients with the worst level of each adverse event between arms, for example, the proportion of patients with any amount of nausea, and with grade 3 nausea throughout the trial.
An innovation in the PRO-CTCAE is that we can adjust for the baseline symptoms of a patient. Patients do a much better job reporting baseline symptomatology than clinicians. Baseline symptoms are common and are often misattributed to study drug. By adjusting for these in analyses, we remove noise in analyses, and improve the precision of adverse event assessment.
PRO-CTCAE information can be shown in a table form or in histograms displaying the distribution of scores overall or more granularly at each time point of measurement.
You also mentioned real-world use and how PRO-CTCAE is coming in handy in understanding the full range of symptomatic adverse events in drugs that weren’t apparent until those drugs went on the market.
So real-world use is a different application of PRO tools with some special considerations. But its value is predicated on the same early finding idea we talked about, that we as clinicians miss about half of our patients’ symptoms. This means that in clinical practice, many of our patients are quietly suffering at home without our knowledge.
If we don’t know how our patients are doing, we can’t intervene. Systematic collection of PROs can key us in and improve symptom management and clinical outcomes like quality of life and survival, as well as reduce ER visits and lengthen tolerability of chemotherapy, which we showed in a recent randomized trial presented at the last ASCO.
And, of course, a central part of our job as cancer care providers, of my job as an oncologist, is to address my patients’ symptoms. So, I need to know how my patients are doing. These kinds of tools can be inserted into clinical care and integrated into the electronic health record.
There are multiple reasons why it’s important to collect PRO information in the real world. One reason is to manage patient symptoms; the second reason is to collect population level data to understand problems that people have long-term that we simply can’t collect in very small, brief, clinical trials. Once PRO information becomes a standard part of the EHR, it can be combined with other clinical data for real world trials of drugs.
Is that done widely?
Not yet, but interest is rapidly growing. But it’s technically challenging. A major barrier is the difficulty of integrating PROs in electronic health record systems, and EHR vendors have been slow to develop PRO interfaces well, which has stalled the widespread collection of this information. There has been pressure on the big EHR vendors to improve their PRO functionality, but more is needed. That said, many health systems are implementing PRO collection, either using the limited native functionality of their EHR, building an interface to their EHR with a third-party PRO platform, or using a freestanding PRO system.
I guess it’s becoming even more relevant with immunotherapies and targeted therapies, and many of these more recent treatment modalities don’t have long-term outcomes data. Would this become a more central way of assessing those outcomes?
With newer therapies, we know that patterns of adverse events are different from cytotoxic chemotherapy. There are low-grade, chronic symptoms that people find bothersome and impair compliance. And there can be catastrophic adverse events that are preceded by milder symptoms that we could better pick up using PROs.
From what I heard at the NCAB meeting, the latest is about fully implementing ePRO, which is electronic version of this. How far have we come as a field, or rather, what is your sense of how well the cancer world understands the process, and are we as a field further along in PRO than other medical specialties?
There’s been tremendous progress in the ePRO area over the past 10 to 15 years, and much of this has been spearheaded in oncology. Stakeholders across institutions in our field are increasingly conscious of the importance of PROs, and PROs are incrementally making their way into trials and routine care as a result.
A lot of this started with oncology-specific quality of life tools developed by pioneers like Dr. Neil Aaronson and Dr. Dave Cella, which were initially administered by paper and pencil in trials. In oncology, patients have many symptoms, both from disease and treatment, so it makes sense that much of this worked stemmed from our field. More recently, it’s become more common to collect PROs electronically, and the NCI just licensed a downloadable app for collecting ePROs across cooperative group and other NCI-sponsored trials, which is an exciting development.
When I started this kind of work, very few of my clinical colleagues had heard of a patient-reported outcome, or given this much thought. I’d say there really is widespread consciousness now. The fact that we’re having this conversation reflects that. PROs and the importance of truly understanding the patient experience with disease and treatment have risen prominently in the discussion across the sector.
I attribute much of this to early advocacy by leaders in our field. I’ve done a lot of my work in the cooperative groups, particularly in the old CALGB and now in the Alliance. Strong support from leaders like Rich Schilsky who was CALGB Group Chair (now at ASCO), and now by his successor in the Alliance, Monica Bertagnolli, have been essential. And Cliff Hudis was an early collaborator and champion of this work.
I’d say the FDA and the EMA have also been central to raising the prominence of PROs in oncology and more widely. Both agencies have developed guidance documents to try to help drug developers figure out how to use PROs.
Many of the larger industry sponsors have PRO units that mostly operate on the post-marketing side, but are starting to migrate towards being involved in pre-approval trials. The inception of PCORI and PCORI’s emphasis on including the patient voice in research has also been important in raising consciousness and funding PRO research.
There are now a lot of commercial ePRO systems out there that are available. The big EHR vendors, Epic in particular, also have a rudimentary ability to collect patient reported outcomes—they could use a lot of improvement, but at least they’re out there, it’s a start.
And health systems are rapidly becoming interested in this. As accountable care organizations and bundled payment programs become more widespread, there is interest to catch symptoms early through PROs, before they lead to expensive downstream use of services. As I mentioned, some of our work showed that systematic PRO monitoring in routine care reduces ER visits.
No doubt there’s been tremendous progress in the methodology of developing questionnaires, the technology for collection of PRO data, the enthusiasm of patients for reporting it, and the consciousness of many stakeholders about the importance of this. But, we’re not quite there yet. We’re on the cusp on this coming into common use, but a number of barriers still exist.
Would you say that this movement to collect PROs did start in oncology, or was it a simultaneous effort spanning a couple of specialties?
Assessment of health status and quality of life through questionnaires goes back further and developed in a number of areas of healthcare, not just oncology. A number of generic PRO tools have been around for decades. But I would say that some of the leaps in PROs have happened in oncology over the years. And now especially, there is tremendous activity in the oncology space around PROs.
I see that FDA and CDRH have provided PRO guidances for medical devices, for instance.
Yes, as I mentioned earlier, the regulatory agencies have been key to the rise of PROs in oncology and more broadly. Both the FDA and EMA have issued guidance around methods for integrating PROs into drug development towards labeling claims. Some have criticized these as being too stringent, but they have nonetheless been very important for bringing PROs into the consciousness of drug developers. Key stakeholders at the FDA and EMA like Paul Kluetz and Dan Sherman, respectively, are pushing this field further now, particularly for thinking about how the PRO-CTCAE and other PROs fit into labeling.
Since PRO-CTCAE has been used in at least 20 NCI-sponsored trials and many industry trials, how is FDA formally engaging the framework? You said industry-sponsored trials will follow FDA guidances and that PRO-CTCAE is used to inform drug labels.
Stakeholders in the FDA and EMA have been enthusiastic about the PRO-CTCAE. They have hosted panels and published on the PRO-CTCAE specifically. There is a lot of ongoing work in the weeds between the regulators and sponsors to figure out clinical trial design and reporting standards for the PRO-CTCAE.
Some of this builds on prior guidance, and some on evaluating data from trials that already include the PRO-CTCAE. We are rapidly getting there. This FDA in particular has a mandate around patient-focused drug development, and PRO-CTCAE fits into this well.
So, beyond the challenges of the industry and getting EHR systems to incorporate this into their platforms, what else do you foresee as needing to do as a field over the next few years to really make it happen?
In my opinion, there are several things that need to happen to broadly implement the PRO-CTCAE—with the caveat that I’m simply a researcher who helped develop the tool, and an understanding that the PRO-CTCAE is a public product supported and maintained by the NCI.
At the NCI:
First, I think that the day-to-day oversight of the PRO-CTCAE tool at NCI should be moved into CTEP along with the CTCAE, and be managed by the same or a similar clinical research team. Enough is known about this tool for it to be integrated with the CTCAE and implemented in trials, through partnership with industry, the NCTN, and regulators. Research on the PRO-CTCAE like the pending Moonshot project could remain in DCP, but this would be to expand understanding of the tool and its applications.
Second, a standard operating manual for the PRO-CTCAE is greatly needed, and would ideally be developed by an interdisciplinary group of clinical investigators, PRO experts, and stakeholders from the NCI and FDA. This manual should be updated regularly as we learn more about the tool.
Third, the tool needs to be regularly updated with a standardized process, just like the CTCAE is, with new items and tweaks to existing items. We have known for several years that there are items that likely should be added, for example to better accommodate radiation oncology trials, but there has not been any visible activity to make this happen.
Fourth, the PRO-CTCAE should be fully integrated with the CTCAE in its next release.
Fifth, a mapping algorithm for converting PRO-CTCAE scores to summary grades has been developed and should be released.
Sixth, many more language translations of the PRO-CTCAE are needed. Most available languages for the PRO-CTCAE to date were actually funded by Genentech, which is wonderful, but it is unclear to me why NCI, which has already invested substantially in this tool, has not supported translations that are essential for international trials. NCI should support broader translations to make the tool optimally usable in my opinion.
Seventh, although the PRO-CTCAE can be downloaded for free from an NCI site, there is overly restrictive licensing language that appears on that site to preclude use of the PRO-CTCAE in real world settings like routine clinical care. I see no reason for this restrictiveness. This is a public tool created with public money, and should be available for improving care.
Finally, the NCI should not play a role in the interactions between the FDA and sponsors who are using this tool in drug development, unless consulted specifically.
For the FDA:
About PROs in general: Tremendous work has been done at the FDA thinking about methods and standards for integrating PROs into trials. But industry sponsors will not widely and rigorously do this until the FDA makes clear that this is an expectation of trials. In other words, that the characteristics of a product cannot be fully understood without direct reports by patients about how they feel and function during treatment.
Messaging from the FDA to sponsors currently is ‘if you decide to use PROs, here are the methods we would like to see’. But to me, the message should be ‘we need to understand the patient experience with your product, so show us your strategy for rigorously collecting that information’. I believe that until that happens, we’ll see inconsistent and sub-par PROs in cancer trials, yielding an inadequate understanding of benefit and risks.
Where is the PRO-CTCAE managed right now in NCI?
Right now, it’s managed between two population science research branches, the Division of Cancer Prevention and the Division of Cancer Control and Population Sciences.
So, it should go to CTEP.
Yes, I think so, for day-to-day maintenance, updates, SOPs, and dissemination efforts – particularly in partnership with FDA. But for research on the tool, I think the model of the Moonshot announcement makes sense, with that part living in DCP.
It is visionary that the NCI supported creation of this tool, which is truly a unique contribution for bringing the patient voice into clinical research. It is a model for other diseases. The challenge now is to move to the next step of dissemination so it can take on a life of its own and truly yield its potential value for patients.