Abstract
E-patients are everywhere. They are sifting through health sites on Google, posting their experiences on YouTube and Twitter, and sharing with others on sites such as PatientsLikeMe.com and
In clinical research, we have a long history of trying to centralize our data. The scope of clinical trials often includes many site locations producing data from patients who could be across the country or around the world. Early on in clinical trials, we started to centralize diagnostic lab work where we would ship tubes into central labs. As we started to incorporate other kinds of end points and diagnostic data into our trials, particularly data that may be digitally acquired, we would look to follow a similar model of centralization.
One of my earlier jobs in the industry was creating a medical imaging core lab business and doing what those in the telemedicine community have been doing for years—to transmit digital data wherever possible, centrally collect it, and centrally review it. In our environment, it wasn't only to introduce efficiency, but to reduce “noise” and improve the power of our studies. So in the clinical trial space, as we've started to collect other types of diagnostic data from patients, we've also tried to centralize it, be it ambulatory blood pressure monitoring, electrocardiograms, medical images, and so on. There is an interesting history of clinical trials trying to draw off the work that the telemedicine industry has been doing for years.
Thinking about what telemedicine and e-health bring to clinical trials, I believe there are three different categories in which these types of technologies can impact how we collect data. First is the data that we collect today, which requires the patient to travel and meet with the physician investigator, such as blood pressure readings. In those cases, perhaps some of these tools can enable us to collect that data more efficiently. Could we spare the patient and the investigator the burden of an office visit if it was only scheduled to capture some basic diagnostic data?
Second, there are opportunities for us to collect existing data, but do it with greater frequency—with higher rates of sampling. In some cases the thesis may be that with higher rates of sampling we can further reduce noise and variability. Some of these data elements, which were previously individual time points, can be turned into a nearly continuous data stream. This might mean sensors that a patient wears on an ongoing basis or other diagnostic assessments that a patient does daily in their home as opposed to weekly or monthly with their investigator.
A third category may be new types of data, which historically we haven't been able to collect, but may be able to with telemedicine and e-health tools. It might be a new type of wearable accelerometer or an ingestible electronic, for instance, that wasn't previously available.
I think that there are a number of different telemedicine opportunities that impact both the data we are already collecting and the new data that we have not been able to see in the past, all of which can either improve the efficiency of our studies or provide answers to questions about our drug safety and efficacy.
What's exciting to me is that for clinical trials we have this great opportunity to repurpose tools and technologies that are coming out of the telemedicine and ambulatory monitoring communities. A lot of the companies that have developed new tools such as sensors and monitors may struggle to find near-term revenue models, because the payers are looking for data on (monetary) return and clinical benefit. But for us in the clinical trial space, that's not an issue—we will pay to get the data we need for our studies.
Many innovative companies in this space enjoy partnering with us early in their development because they can look to companies such as pharma as an early revenue source. Maybe it won't make up their larger revenue goals that may include the greater healthcare space, but for many it is faster revenue as we are not looking for the clinical benefit data needed by a payer. What we do need to know is that data transfers and handoffs are validated, and ultimately, we look for validation around the clinical data to show that it's correlated to a meaningful endpoint.
Reduce the Noise
A good example of reducing the “noise” in a clinical trial is with radiology and imaging. Picture 100 radiology investigator sites, each with very competent and board-certified radiologists, reviewing the images of the participating oncology patients. But, the challenge is, by having 100 or more different radiologists looking at these images, even with a published and standardized set of criteria, there is a lot of room for subjectivity. However, if we can bring those images together, we have the opportunity to reduce some of that variability or noise. It is especially important when we have very subtle signals of efficacy or safety that we are trying to look for, which can easily get lost with the many sites needed for our studies.
Also, a lot of current excitement is around convergence and integration. I like the idea to have an “app store,” if you will, or an ecosystem of different tools that I can draw on as I have different clinical trial needs. For instance, I may have one study where I'll need a wearable accelerometer to understand sleep patterns of a patient or an ambulatory blood pressure monitor, etc. Having a toolkit of different sensors that I can pull together individually or in different combinations brings interesting possibilities.
We have access to some. There are some companies in the clinical trial space that have tried to aggregate that data for us. There are some companies with useful tools that are not in the clinical trial space that we work with to try to bring those technologies in wherever it makes sense for both of us—to those companies as potential partners and for us where that data could be very important for answering research questions.
I think there are a number of wearable sensors developed for the commercial space, for direct-to-consumer, for example. The providers of these sensors tend to come up with their own algorithms to use the accelerometer data and make it meaningful to patients and consumers so they can interpret it.
For us, we've actually found that we like the raw data from these devices. We have different sorts of questions as opposed to a patient, so in many cases we like the device and we're not as concerned about the software that's been developed around that device. By having access to the raw data, we may be able to answer questions in different ways—say an accelerometer to understand sleep and, in a few instances, avoid the effort and expense of sleep studies. Some of these sensors are pretty unobtrusive and allow you to stay home and measure different information, which, in some cases can be an acceptable proxy.
That is always welcome. We're always looking for ways to be more efficient today, given the cost and complexity of drug development.
From a research perspective, we have a number of avenues that we're pursuing. In individual clinical trials, there are many instances where we are doing remote patient monitoring and that might take the form of ambulatory blood pressure monitoring or some other discreet technology being used in specific trials.
Another example, which is somewhat related, is our use of electronic diaries and electronic patient-reported outcomes, where we are capturing the voice of the patient, either through validated instruments or questionnaires from home using mobile devices. At the same time, we're exploring opportunities to allow us to empower a patient to be much more actively engaged. This may include being participants in our trials from home and we believe that it's good for patients as participants and it's good for Pfizer as we will be able to capture more data of use to us in a more efficient manner.
What is interesting today is how consumers in healthcare are using these tools for self-tracking, self-monitoring, and “quantified self.” There is the opportunity for convergence between those trends and the types of data that we routinely collect from patients in clinical trials. For example, in drug development, we use different types of ambulatory sensors with patients, and at the same time, you see similar technology being marketed directly to consumers to self-track their physical activity using sensors such as a Fitbit or a Phillips DirectLife. Likewise, you see no shortage of apps for the iPhone enabling patients and consumers to self-track, it could be their diet and exercise or it might be the symptoms associated with their migraines. These apps can start to look very similar to the electronic diary and patient-reported outcome tools we use in trials.
Ultimately, there is an opportunity for convergence between these different e-health tools in healthcare and the tools we are using in the clinical trial space. But we can't just collect data in clinical trials, it must be validated data. Diary instruments for a patient in a trial have been through validation studies to test the questions being asked. Apps in an app store don't follow the same path. But can we take the types of instruments that we've validated in clinical research and make them available to consumers? If consumers were using validated instruments, could that provide more real-world research data? If patients were self-tracking using the same instruments that we used to test the safety and efficacy of their medicine in clinical trials, could that improve the dialog between patients and their healthcare providers? It is an interesting convergence opportunity.
The reason people use these apps and tools is because they get something back. They're inputting information so they can get some greater value in the way that it's presented or the way it's shared. On the research side, we can do more to enable patients to have access to the information they're giving us. We can share it back with them so it benefits them as well. On the consumer side, in the spirit of PatientsLikeMe.com and crowdsourcing, we need to enable those same patients who are using those apps the opportunity to aggregate it for research purposes.
One of the first opportunities for us with consumers and patients interested in research is information and engagement. Many consumers struggle to find trusted accessible information around clinical trials. One of the most trusted resources available to patients and providers is
Some of the nonprofits and foundations are also trying to introduce trusted information for their patients. For instance, there is the TrialMatch from the Alzheimer's Association. This is an interesting example of a trusted entity repackaging trial content in a way that's more accessible for their patient community. Information and engagement are important aspects of using the Internet with patients as participants in trials.
Increasingly, as patients have access to their personal health information through personal health records or other tools and technologies, their ability to use that to match into trials or authorize sharing data during their participation in trials will be increasingly valuable. For instance, if you're a patient participating in one of our studies, we may have an investigator sit down with you and ask about your health history or prescription history, but if you have that aggregated already in your personal health record, that would be much more efficient.
Making a Match: Clinical Trials Accessible Online
The Alzheimer's Association has launched TrialMatch, a confidential and free interactive tool that provides comprehensive clinical trial information and an individualized trial matching service for people with Alzheime'rs disease and related dementias.
For more information, visit:
We have many regulatory responsibilities that relate to the information that we share online. Any of the information about our marketed products, for instance, is very tightly regulated. Our ability to participate in social media is a bit ambiguous for many of us in pharma right now. One area that is particularly challenging is with information presented on sites that are outside of our control; we are very tightly regulated on sites we do control but those are a small minority of sites you might see if you entered a brand-name drug into Google, for instance.
There is an increasing expectation around transparency, and companies such as Pfizer have been very active in living up to the expectations. That might include transparency around the results of our studies or transparency and disclosure around financial relationships with prescribers.
Over time, it will be interesting to see if there may be even more opportunities for transparency around data. You see that with open government initiatives, such as
In the research space, there are interesting opportunities for patient recruitment as it is related to social media. I think most people are optimistic that there will be guidance coming from the U.S. Food and Drug Administration that will provide clarity for other aspects of our business around how to engage in social media. That hasn't stopped Pfizer from participating wherever they can, from YouTube channels for patient recruitment or externally facing blogs with comments enabled. I participate as a blogger for Pfizer at ThinkScienceNow (science.pfizer.com), where I am given tremendous freedom for what I would like to blog about at a Pfizer-sponsored site, which is great.
I am excited about the potential that exists for us in this space; the tools available from telemedicine and e-health are low-hanging fruit solutions for some of the cost and efficiency challenges that really call into question the sustainability of drug development today. These tools exist so we don't have to start from scratch, there are clever one-off uses and integration opportunities that can really bring some necessary disruption and innovation into the clinical trial process.