Interview with Clinical Research Scientist Prof. Karen Raphael

Originally published on Tomorrow Edition on May 31, 2018

Karen Raphael, PhD, is professor in Oral Medicine at the New York University College of Dentistry, and in Psychiatry at the New York University School of Medicine, New York, NY, USA.

At the time of the interview below, Prof. Raphael commented: "I’ve lived as a Person with Parkinson’s for about 8 years and shared with my siblings in the care of my father who also had PD. He passed away two summers ago at age 91. At age 63, I am still working f/t as a Professor and Clinical Research Scientist, in an area unrelated to PD. I hope that my training in epidemiological and clinical research can help others to understand how the clinical research literature can be translated for personal use (or not).”

The following has been paraphrased from an interview with Dr. Stott on May 18, 2018.


It is often repeated that Parkinson’s is a progressive, irreversible disorder. Can you talk about your own experience and what you attribute to your improved condition?

I need to start by saying that I am a case study. As a research scientist, I’d say that an individual’s experience offers a tool for hypothesis generation, but itself is not proof of anything.

I was diagnosed about eight years ago. The first thing I do whenever I have a question about virtually anything is look to the primary research literature. At that time, research in neuroprotective strategies pointed to the idea of rodents running on a treadmill for a very long time having some effect on neuroplasticity, so I started with that. Then I found a movement disorder specialist who said to me that “there is no such thing as too much exercise for a person with PD.” That doctor, like myself, is very involved in clinical research, so she understood the cognitive and physical demands that my job placed on me. When I met her I had three years left on a grant from the NIH, and she told me that, given my symptoms at the time, I should expect that to be my last grant. But, here I am, eight years later, really doing quite well and still getting funded.

I had never been an athlete, but I am relatively assertive and from college until today (I’m now 63), I have always been physically active. I actually had my first clear symptom when I froze (a common mid-stage symptom of the disease) on a trek to Everest base camp. So, I was already fit. It was probably less difficult for me than it might be for some others to step it up to the next level of fitness. However, when I was diagnosed at 55, my PD was probably more advanced than most who first get diagnosed. I had to take far more than the normal amount of levodopa to get myself back to an acceptable level. Once it was clear what was happening to me, my family and I moved to a building with an elevator to prepare for the possibility that I was not going to be ambulatory for much longer. We also made sure the building had a gym so I could get up at 4:30am and exercise after walking my dogs.

After about two years of exercising every single day, which included at least 90 minutes of running on a treadmill at a pretty intense pace, I started noticing a decrease in my Off periods. With the blessing of my neurologist, I reduced the amount of levodopa I was taking. Now, another 18 months later, I don’t get Off periods anymore. Even if I skip a dose or two, I am okay, and I no longer have any tremor or rigidity.

I can’t fully explain why I have improved so much. However, I should add that I exercise after an overnight fast, which presumably puts me in a ketogenic state. It is believed that ketones are more energy efficient than the glucose we normally use to feed our brains. Parkinson’s is believed to have something to do with energy deficiency; some even call it Type 3 diabetes. I also tend to hold off eating as long as possible to feed the brain as many ketones as possible. This might explain my improved condition, as well as all the other brain benefits of exercise, like the production of neurotrophic factors and spurring neuronal growth in the hippocampus.

What do you think every person with PD should incorporate into their exercise routines?

I think the important thing is for people to work up to an increased level of fitness until you can do some kind of prolonged endurance exercise. That means a state in which you are breathing hard and you are sweating. I also think that strength training is essential, with a strong core for balance. Also, yoga or something similar for flexibility. Additionally, while exercising, it is important to be mindful of what you are doing: don’t just zone out to your favorite tune. You want your cortical arousal level to be high to promote neuroplastic changes.

What questions do you think people should ask themselves before deciding to believe the claims of a new therapy?

I consider myself a skeptic coming from a place of informed science. My training as an epidemiologist and psychologist has helped me understand why people believe certain things and how to debunk claims regarding certain treatments. The question for me always is: Where is the evidence? To start, I go to the peer-reviewed literature for at least preclinical research data to see if whatever I’m looking at has a plausible mechanism for treatment. If someone is proposing that treatment X works just based on evidence from fruit fly research, I’m a little skeptical. But if there is a convergence from basic science and other sources and there is no major known downside to the treatment, then I’d maybe recommend giving it a try. However, one needs to be very skeptical of any treatment claims, especially when there are conflicts of interest like profiting from a product purchase. Also, if somebody says they have done tons of research, and then you look at it and see that it consists of 300 case studies, that is uninterruptible. Those 300 may have come out of a sample size of 3000 people who tried the treatment.

Mannitol is a good example of this. The people promoting it are building a registry for people to report their experience, but you are probably more likely to actually fill out a report if you have had a good experience. They also said they are going to release the results when they are “statistically significant.” That is nonsensical for anybody who understands statistics, so I am skeptical of the case for mannitol.

Case studies are useful to lay the groundwork for possibilities, but any hypothesis about a treatment claim has to be falsifiable. That is what we do when we apply statistical analysis to systematically collected data. We ask: “Is it possible that this pattern of results happened by chance, or are the results so unlikely to be due to chance that the only other explanation is that the treatment is having an effect?” You can’t use your own experience to know if a treatment works, even if you think you are not prone to placebo effects. This is why, despite all their potential problems, we need controlled clinical trials.

What are some of the biggest misconceptions people have about clinical trials?

There are so many. I would say the biggest one is that people often discount clinical trials simply because the treatment supported in a clinical trial didn’t work for them. Trials are based on inferential statistical testing which is based on probability theory. If treatment A works better than treatment B, with some level of statistical significance, it just means that the average person got better from A than B, with a pattern and degree unlikely to happen by chance. As I’ve said elsewhere, the ‘average’ person is a statistical construct, not an empirical reality. Rarely is an individual going to know if a treatment is going to work for him or her. For someone wondering if a new therapy is a possibility for you, get your hands on the description of the study participants and see if you are similar to them. If not, it becomes even less likely to know whether the treatment would work for you.

PD is a syndrome. There is a lot of variability between individuals that has to be accounted for. Even in HIV, where we know the cause, people respond to certain drug regiments differently, because there are host factors that affect whether or not a given treatment works for an individual.

Can you talk about the exenatide trial and how that demonstrates some of the flaws people have in interpreting the results of clinical trials?

There have been repeated episodes in PD of people getting excited after a Phase 2 trial only for it to fail in Phase 3. It could be because there is more careful control, or higher expectancy, or very tight exclusion factors. We have seen again and again: Phase 2 studies succeed but fail to pass Phase 3.

When you have smaller Phase 2 studies, one of the problems is that random assignment can fail. When you split participants into group A or group B, we try to make sure that baseline characteristics of the people in the two groups are as similar as possible. In the exenatide trial, random assignment failed to equate groups on pretreatment characteristics. Right away, that tells me that the causal inference is more difficult. It was as if it became an observational study, and I wouldn’t take it as any more evidence than that. We do not know what the natural history of the control group vs. the treatment group was. There were clear differences between the groups at baseline. I hear a lot of people talking about getting it off-label, but I think that is really premature. We don’t have enough solid clinical research data yet, especially when this is a treatment that is not without risks.

What do you think are the most common mistakes made in the design of trials?

One of them is overly restricted inclusion and exclusion criteria, unless they are doing a precision medicine trial for people with a particular genetic polymorphism or particular subgrouping characteristic. For example, in PD, if your exclusion criteria list people with depression, a common problem in PD, then what you are left with are people who are not typical people with Parkinson’s. This is why, when you are considering starting a treatment, you have to look at the characteristics of the people in the clinical trial.

Another one that really upsets me, and which often comes from Big Pharma studies, is employing the wrong contrast group. Too often they use placebo controlled trials, but – if there is a treatment already out there that this new study drug is also supposed to target – the relevant clinical question is: “Does this treatment work better than an existing treatment for that problem?” The question is not, “Does it work better than placebo?” Yet, the FDA is approving new drugs based just upon placebo controlled trials that have not been designed to ask the clinically relevant questions.

Another common mistake is much ado about Phase 2 studies. They are really just preliminary evidence of efficacy. Only Phase 3 studies can give us a good sense of whether a treatment is going to be effective. Replication is a principle of science. I could see that multiple Phase 2 studies by independent investigative teams might be a substitute for a cumbersome Phase 3 trial, but – without replication – Phase 3 trial is hopefully powered well enough to see whether effects are only occurring at certain sites. This is not a rare disease, where a multi-site study is inherently necessary for recruitment. Now, it could be different for certain precision medicine trials, so that only a single Phase 2 trial is feasible short-term, but you would expect huge effect sizes, not marginal ones, in precision medicine trials, indicating that the treatment is really "hitting" the underlying mechanisms.

I also wish people were more sensitive to outcome measures. Some people have criticized the UPDRS (rating scale used to measure PD focused mostly on observable or clinically-measured signs and symptoms) as being very clinician based, and I agree that it is; there is more going on in PD than just movement problems. However, in patient-oriented measures, there is a tendency to underestimate how difficult it is for people to reconstruct their experience when asked: "How did you feel in the last two weeks?" This is especially true in a condition like PD in which every day can be different. From my pain research background, one of the interesting things we learned is that, when we ask people to rate their immediate pain severity regularly using random cell phone prompts over the course of a week or two and then afterwards ask them to retrospectively assess, “What was your pain like over the last two weeks?”, there is virtually no relationship between the retrospective rating and the average of the momentary ratings. It is incredibly difficult to assess retrospectively what average symptoms are like over a given period. We need to anchor ratings of any symptoms better, asking about worst, best, or perhaps when you were “on” or “off” within a relatively brief period.

What questions should patients ask before enrolling in a clinical trial?

It is very important to remember that you should not just sign a consent form. "Informed consent" is a process of asking a lot of detailed questions and getting detailed answers in language that you understand. The potential participant should understand what risks are possible, what the benefits might be, why you might be withdrawn, how you will be monitored, how you are going to be informed of results, and more. Also, if there is not a guarantee that you will see the results of the study, walk out. Of course we want to help research, but the fact that you showed up to help does not alleviate the responsibility of the clinical research team to keep you fully informed of what you are going to undergo and what they learn. You have rights as a clinical trial participant, although they vary from country to country. Know your rights.

Basically, you should ask as many "what if" questions as you want. What if I start and I decide later I want to stop? How is this going to affect my usual clinical treatment? What if I want to change my regular medication? Come in with the expectation that it is their job to convince you that they have thought through all the possibilities and potential problems. They have to win you over. If the person answering your questions doesn’t know an answer, make sure they find out.


Keep an eye out for more entries in this section soon, and go here to view all Benjamin Stecher's unique blog posts for JPD! See all entries his Tomorrow Edition blog on that website here.