What can we Learn from Patients who Self-Prescribe?

It is time to support patients who choose to prescribe their own treatments and manage their own healthcare outcomes.

Clinical trials are ALS patients’ only hope towards achieving an effective treatment in their lifetime. Yet, thousands of them will never enroll. That may be due to economic or geographical constraints. However it is more likely because they are excluded from participation due to specific enrollment criteria designed to facilitate the interpretation of a treatment benefit in the clinical trial.

So where do patients turn when they are no longer eligible to participate in the regulatory trial process? They turn to “experimental” options. Options where there are often no clinical data to support the treatment. Or even data to show it is safe.
But options they can self-prescribe and take “off-label”.

Every one of us on the planet is a patient at some level, striving for a higher degree of well being through fitness, preventive medicine, healthier lifestyle, emotional happiness and stability. And ALS patients want all of these things too. There is little to debate that patients now play a more active role than ever before in their health care decisions. Patients track their exercise activity, calorie intake, calorie burn, heart rate, body temperature, pulse with a multitude of devices from the iWatch to the Fitbit. They track their wellness via web-based portals such as PatientsLikeMe that provide sophisticated tools to monitor their disease progression that is presented in a well-designed cohesive manner.

Advances in wearable devices, implantable devices, and the ability to perform routine clinical read outs associated with disease are a part of our medical future as human beings and as patients.

ALS patients want all of these things. Patients will continue to manage their own healthcare decisions until effective treatment options for ALS have been proven through the traditional clinical trial process and regulatory systems. Until then, they will take supplements. They will take off-label medications with some scientific rationale even if it is not the strongest rationale. And they will even take experimental treatments not approved for any disease indication. They will do any and all of these things if they can get access to the treatment and they can afford to do so.

And they have the right.

But what if we encouraged them to tell us about it? What if we could do clinical trials on these potential treatments? Could we improve the interpretability of the outcomes? Could a more systematic approach be utilized to benefit the whole community? To facilitate informed decisions about which treatments might be working and which ones are actually doing harm?

Randomized Clinical Trials (RCTs)

For decades the gold standard for determining the effectiveness of a drug has been a method of randomized clinical trials (RCTs), where parallel groups of patients are randomized into treatment or placebo groups. The objective of RCTs is to determine the average treatment effect in a specific cohort of patients that meet specific inclusion and exclusion criteria.

Often these studies provide limited information on whether the treatment positively impacts a limited number of participants or even individual patients. In addition, these types of trials are often long and expensive and may not meet the needs of a patient seeking treatment due to confounding variables such as other medical conditions.

N of 1 Clinical Trials

So, how can the research and clinical community work together with patients in this framework to leverage the potential positive outcomes and eliminate negative outcomes as soon as possible?

The answer: By systematically adopting the principles of N of 1 clinical trials.

N of 1 trials are randomized trials for each individual patient with the goal of pooling individual patient data retrospectively to characterize treatment benefit. They are typically “crossover” designs with multiple treatments and often a placebo with randomization of treatment over specified time intervals.

N of 1 clinical trials have been used for decades in other medical practices such as cystic fibrosis, assessment of pain medications, respiratory function, anti nicotine treatments, and gastrointestinal disease to name a few.

The major challenge with N of 1 trials in ALS or other disease indications is in their design. How can we design them to be informative and impactful and ensure that the needs of the patient are aligned with the needs of the entire community?

The successful execution and interpretation of outcomes from N of 1 trials are dependent upon:

  • The randomization of treatment for each patient into a treatment group on enrollment.
  • The commitment of every patient to be randomized into multiple treatment groups during the trial period.
  • The introduction of multiple randomized treatment intervals during the course of the study.

These processes are implemented into the design to eliminate “biases” of effective outcomes from the patients and prescribers in the trial process.

This is how it could work:

  1. Every ALS patient who wants to participate “enrolls” via an online system (similar to the system that ALS TDI has in place for its Precision Medicine Program). Once enrolled, they are assigned a patient portal where they can follow along in the data collection and data interpretation stages of the trial. They will have to comply with self-reporting of their wellness over the next 12 weeks. This period of time is called the “lead-in period” and it captures data using self-reported ALS FRS, motion tracking devices which can be provided, and in some cases, at-home blood tests for biomarkers of disease progression (I’ll get to this in a minute.)
  2. After 12 weeks, patients are randomized into any possible treatment group including a placebo if it is designed into the trial. Patients then continue weekly self-reporting and any blood tests that are required. After 12 weeks they stop taking anything for 3 weeks, otherwise known as the “wash out period. ” Patients are then re-randomized into any of the treatment or placebo arms and the process continues.
  3. The key point is that each patient who enrolls is randomly assigned a group and all the medications including the placebo look the same.

For simplicity sake, let’s say there are three treatment arms in the trial. Every patient will be randomized three times in random order into any arm of the trial. At the end of the 18 months, the data is unmasked and shared with the community.

 

No1

Figure 1: Graphical representation of N of 1 clinical trial workflow

At the end of an N of 1 trial like this one, there will be enough data to determine if a treatment actually slowed disease progression or made progression worse. Knowing either would benefit all ALS patients.

It’s not all roses

There are obviously many issues to be resolved with the execution of N of 1 trials.

  • For instance, what are the possible drugs that are available?

What are the outcome measures?

This is a particularly challenging one. Ideally the outcome measure would be easy to measure, quantifiable, very accurate, reproducible, capable of weekly measurements, not require a clinical visit, and apply to all patients. As one can guess in ALS we don’t really have the perfect outcome measure. However we may be able to leverage real time accelerometer data and compare these data to self reporting scales such as ALS FRS. We may also be able to use an N of 1 trial as described to “validate” preliminary data on promising biomarkers from blood (NF-L, creatinine). These outcomes could be utilized in addition to more traditional outcomes in ALS such as forced vital capacity, time to assisted ventilation, and survival.

  • Can we achieve compliance on the self-reported outcome measures?
  • What standard clinical chemistries could be used to assess disease progression?
  • How would the data inform future decisions for stakeholders in the ALS community?

This strategy however, still has a very significant potential upside, when it comes to the ALS drug development landscape.

Let me use a real time example:

In 2013 ALS TDI sponsored a clinical trial to test the safety of Gilenya, a treatment approved by the FDA for Multiple Sclerosis. Scientists at ALS TDI hypothesized it may provide therapeutic benefit in ALS patients. A trial was designed by a team of ALS neurologists and approved by the FDA to test the approved dose for MS patients in 30 ALS patients at 4 ALS clinics and treat them for 30 days to monitor potential acute safety issues. Since it was just a 30-day trial, there was no preconceived notion that the trial would impact any outcomes of slowing down disease or improving survival rates. It took 12 months to enroll the 30 patients at the 4 ALS clinics.

Now compare and contrast that time frame to the following scenario:

ALS TDI could design and execute an N of 1 trial for several potential “off-label” treatments that patients may be taking anyway in conjunction with our Precision Medicine Program.

We have enrolled 180 patients into the program in only 9 months. Based on patient inquiry, the program could have easily enrolled more than 400 in this time frame but we are limited by capacity per week. We have been tracking ALS FRS, speech recording, accelerometer data, biomarkers in blood, and a full genome sequence for all these patients. If we had randomized each patient into an arm of the N of 1 trial upon enrollment in the program, we would have already effectively completed a typically sized phase IIB trial in ALS for two potential treatments!

And more importantly we would now be sharing the data with the entire community to inform future decision making.

This could be a game changer.

We need to support the patient community when they do what they are going to do anyway. Seek alternative treatments that are safe until the traditional randomized clinical trial process can bring effective treatments forward. An effort has been made to accomplish this objective on patient forums as well as companies. But we need to help the community collect and aggregate the data as quickly as possible to actually see if the treatments are working..

We need to help the patients monitor their wellness during this process with web- based tools to monitor disease progression. We need to offer them technologies to more accurately capture their daily activities and quality of life. We need to develop safe, accurate and effective ways to measure biomarkers of disease progression without having to go for a clinical visit.

This is the future of Precision Medicine for patients afflicted with a disease for which there are no current treatment options. This is now a very strong option for ALS patients.

This is the solution for today.

It is imperative that we don’t miss another opportunity in the ALS patient community. I am willing to incorporate an N of 1 trial into the enrollment of our Precision Medicine Program if the ALS community is willing to give it a try.

It will take some effort to design the trial:

  • What are the treatments?
  • What are the enrollment criteria?
  • What are the exclusion criteria?
  • Where do we buy the drugs?
  • Who will re package them and ship them in a blinded fashion to patients?
  • What are the regulatory concerns?
  • What are the outcome measures?

Many of these can be addressed expeditiously. Others will take some time.

But we should start today!

15 thoughts on “What can we Learn from Patients who Self-Prescribe?

  1. “It is imperative that we don’t miss another opportunity in the ALS patient community. I am willing to incorporate an N of 1 trial into the enrollment of our Precision Medicine Program if the ALS community is willing to give it a try.”

    My PALS will be at ALS TDI for this program in September and would jump at the chance to participate in an N of 1 trial! It’s clear to us from several online support and advocacy sites that the PATIENT community is more than ready for this.

    It thrills my PALS and I to see you propose this design! My hopes are that other components of the ALS community are ready to embrace it as well.

  2. It’s a breath of fresh air to see some honesty about what goes on every day for PALS and how much information we waste! HP would be very happy.

    If only we were doing this kind of thing as each of the many “protocols” of OTC stuff has come and gone and come back.

    I do have questions —

    1. If you rotate 3 or 4 treatments into the process, is a placebo really necessary?

    2. The treatments require matching form factors, right? So if a protocol has 17 pills a day, the others would have to somehow match that? If a treatment is injected in the tush, then the others would have to be injected in the tush?

    3. I have always had trouble with the fixed lead-in periods for PALS. Can PALS not already be in your PMP tracking with far more than 12 weeks’ of data and opt in for any new n=1-style trial that you may start? It seems to me that the more you know about a person’s trajectory and the more natural history, the better.

    4. I realize that this isn’t the way scientists were raised, but I think that sometimes a PALS can identify outcomes or endpoints that are personally important that may be much more specific than the smoosh of a rating score. Mom had good grip strength but her finger dexterity got to be terrible. She would have been ecstatic with an intervention that helped that regardless of whether it moved the needle on FVC or SVC or ALSFRS or grip strength. In the big pic, the trial might be a failure, but having data publicly available would help PALS roll their personal dice better to meet their individual goals. That wasn’t a question, was it?

    5. If something were tested this way and a year later somebody figured out that substance x was a great ALS biomarker, could you go back and check that biomaker from samples of PALS who were in the test?

    6. You say that you are limited by capacity. What is needed to fix that? That’s a serious question and I trust that the answer may be more complicated than $.

    7. If the things that are tested are off-label or all legally available, are regulatory concerns all that big of a deal at this point?

    8. Do you have riluzole or no-intervention PALS in your program who could somehow provide some control data?

    9. This is more about your PMP program than this specific concept. If a PALS enrolls, I understand that learning one’s genetic information is up to the PALS. Are PALS who opt not to know their genetics able to use the portal without having those beans spilled?

    Thanks. I can’t wait for the conversations to continue.

    1. Los of good questions here.
      1. Is a placebo necessary.
      Even with multiple treatment groups a randomized placebo arm will increase ability to detect drug effects.
      2. Dosing regiment.
      Since this is not blinded but just randomized the patients would know what they are taking during each arm and follow the specified dosing regiment and route of administration for each treatment.
      3. The lead in period.
      I just gave an example of the process. However yo are correct is about trying to get each patients’ trajectory before they start the first treatment. If a patient is already in our PMP program for 3 months then they would start right on a treatment since we have trajectory data on them. For patients just enrolling it is taking about 8 weeks to get them enrolled so we would send out accelerometers now and get them to start self reporting ALS FRS and get baselines on blood chemistries. Then by the time they come to ALS TDI to provide tissue samples we would randomize them onto treatment and they would start.
      4. I just agree we need more individualized outcome measures.
      5. Yes we can always go back and do retrospective analysis.
      6. It is half economic and half wetlab infrastructure to make iPS cell lines, make DNA for genome sequencing, process blood samples for RNA and protein profiling, project management to keep all the lab workflow moving, clinical coordinators to schedule patient visits and collect data, IT support for data storage, data retrieval, help desk support, bioinformatics capabilities to analyze the complex data…..You get the picture
      7. There are some regulatory best practices that should be followed but the onus is on the community to do the best they can to establish a robust design, choose potentially high value treatments, make sure the batches of drug are quality controlled which is hard, patients comply with the protocol, and pick as many informative outcome measures as possible.
      8. There will always be some patients not on treatments but we would have the placebo for each individual patient to compare against their own data
      9.Many patients in our PMP program choose not to know their genetic results.

      I hope this helps clarify your questions

      Steve

    1. I think there are several issues that need to be resolved to get this going as you put it.
      1. What are the off the shelf treatments worth testing?
      2. Do we need to quality control them first since they are just supplements that people can but anywhere we should compare product quality and get them from a single vendor
      3. From a legal and regulatory standpoint could a single “group” buy enough drug from the chosen vendor and distribute it at the right times and doses with instructions
      4. Build a data tracking system for the “virtual” study
      5. What outcome measures can we use to observe potential treatment effects
      Address these and we have a good start

      Steve

  3. This is a step in the right direction but people like me don’t have 18 months to go through a trial and people like me don’t have the money to travel to Massachusetts to get enrolled in the precision medicine program. Can’t they establish a West Coast office?

    1. It is possible not roll out an N of 1 trial without coming to Cambridge and participating in PMP. Treatments can be shipped. Patient reported outcomes can be collected online via a web interface. Depending on the agreed outcome measures it might not even require clinical visits. The missing data going this route is that there won’t be any tissue samples collected so we won’t know the underlying genome sequence, we won’t make an iPS cell line for drug screening ect.

      These other pieces of data are very important but it is unfortunately money that limits our collections to Boston

      Steve

  4. This is a very exciting new direction for testing potential therapeutics. I also wish there was a West Coast office, or another option for enrolling those PALS who would like to participate but are unable to travel to Boston.

    Secondly, are you part of or collaborating with the AnswerALS consortium? I understand their long-term goal is to develop a system that sounds very similar, and it would be wonderfully efficient not to recreate an already successful process. Thanks.

    1. I commented below on the issue with other site collections for PMP. It is mostly financial.
      AnswerALS is trying to set up a program that looks very similar and it is unfortunate we can’t combine the efforts since we are already up and running with our program.

      Steve

  5. It has been since feb 2015( almost a year) since a company has asked about an approval for GM604 by the FDA… Not one word since? What gives!!!

Leave a Reply

Your email address will not be published.