If, on average, in group A the outcome is higher then in group B, we select treatment A. Let us examine in detail how the RCT addresses the three problems highlighted above:. The RCT tackles the problem of high-dimensional learning from noisy data in two ways;. First, the RCT heavily limits the problem space by pre-selecting a very small number of actions and contexts. The RCT compares only two treatments, and, only when the focus is on personalized healthcare, includes a very small number of descriptions of the context.
When there is no focus on personalization the context is fully ignored. Exactly which treatments and which contexts to focus on is determined by our theoretical understanding of the process involved. The RCT tackles the problem of learning the causal effect of the actions by virtue of its use of randomization. To appreciate how the RCT solves the last challenge, we have to view the RCT not just on its own, but we have to include the treatments that are administered after the RCT has been carried out. Approached in this way we can see that the RCT balances learning and earning by first spending a pre-determined number of interactions on learning the trial itself , and subsequently moving to earning: after the trial, the results are accepted with full certainty, and future patients will receive the treatment that performed best during the trial.
At this point I have to note that the RCT is not inherently a method for personalization; rather, it is a method for selecting one out of two competing treatments.
Now that we understand how the RCT addresses our three challenges, we can evaluate the quality of this approach. Let me start by discussing the strengths of the RCT. The RCTs approach to high dimensional learning is appealing since by severely restricting the space of actions and context the outcomes of the trial become transparent and human-understandable.
Next, the RCTs approach to the problem of learning causal relationships is extremely solid Rubin, ; Imbens and Rubin, There is no better method to assess causal effects than randomization, which is exactly what the RCT excels at. Our analysis also allows us to identify drawbacks of the RCT.
We effectively assume that only very small parts of the context and treatment are important and we ignore all others. Already in our simple weight-dose example introduced earlier, the RCT would only examine a small number of specific points in the 3d space, as opposed to examining or modeling the whole plane of outcomes. Furthermore, perhaps implicitly, we assume that the relationship between context and actions is only as complex as our theories allow us to understand.
Another disadvantage of the RCT originates from our insistence on a hard cut-off between learning and earning. However, these certain decisions are made based on noisy data, and hence full certainty is too much to ask. Given limited and noisy data there is always a non-zero probability of making the wrong choice. And, the more we try to personalize treatments, the more severe this problem becomes since at the level of small groups of patients we have very limited data at our disposal.
If we truly believe in treatment heterogeneity, then we have to accept that each patient is unique and hence we will never have a large homogenous sample available to make deterministic decisions. Regretfully, this not the last disadvantage of the RCT as a method of solving Equation 1; because of our determinism, the data that we collect after a trial also turn out to be very hard to re-use: once the probability of receiving chemo-therapy for breast cancer patients with a severe tumor is 1, and for those with a mild tumor is 0, we cannot use the future data to evaluate alternatives simply because no such data is collected.
Our deterministic decisions prohibit our future learning. I would like to sketch a possible alternative method to the RCT. Note that I will only provide an intuition for this alternative method; some technical details are provided in the footnotes of the transcript of this talk. I propose to do the following: First of all, I propose to use a modern and flexible machine learning model to learn the relationships between the actions, context, and rewards. In recent years we have seen a revolution in our abilities to learn flexible, extremely high-dimensional functions Hastie et al.
Second, we can utilize novel breakthroughs in our understanding of causality; as it turns out, it is strictly not necessary to resort to uniform random allocation as is done in the clinical trial. Rather, as long as we can compute and store the probability of receiving a treatment conditional on the patient characteristics, we can use the collected data to estimate causal effects Bang and Robins, ; Funk et al.
Finally, we can use novel methods of balancing earning and learning: as opposed to going instantly from pure learning to a deterministic choice as in the RCT, we can gradually balance the two. An allocation scheme called Thompson sampling allows us to, over time, gradually change the probabilities of receiving different treatments. Thompson sampling selects treatments with a probability that is proportional to its effectiveness. Thus, as we gain more evidence that an action is effective, we will increase the probability of selecting it. This way we can optimally balance exploration and exploitation Ortega and Braun, ; Osband, ; Eckles and Kaptein, Next, this central server estimates a model that relates the context, the actions, and the rewards.
This model is our estimate of the illustrious function f in our definition of personalization. Finally, the central server selects an action based on this model while balancing learning and earning.
Note that as a result of this method we never make a definite choice between different treatments. However, we do make the best choice we can given all the information available. Admittedly, this computational method to personalization might look a bit distant from reality, but the models I propose, and the methods by which earning and learning can be balanced, are already, at least conceptually, developed.
Also, we can already transmit large amounts of data around the world in a split second; large web companies like Facebook and Google do this constantly. Hence I believe that, in the near future, my suggestion is technically feasible. Contrary to the RCT, I will start by discussing the disadvantages of my computational approach to personalization. Actually, my proposed approach allows for much greater flexibility than the RCT: we can include a larger number of contextual variables and we can potentially collect data regarding multiple outcomes.
Thus, if anything, my proposal makes answering these questions easier as opposed to harder. However, there are more serious concerns: First of all, my proposed approach looses, at least superficially, all notions of transparency. It is not at all clear anymore why a specific patient, at some specific point in time, receives a specific treatment. While the underlying logic can theoretically still be distilled from the model parameters, such distilling is not easy. Next, the proposed method, at least in theory, never leads to a definite, deterministic, choice.
Hence, there will always be a non-zero probability of receiving a specific treatment. This might be fine for things like eHealth coaching and health education, but we will be presented with a logistic nightmare if we intend to keep all possible pills available at all pharmacies all around the world for the unlikely event that we should administer one of them. By abandoning the RCT assessing causality becomes more challenging.
How can we still be sure that the model we learn is actually learning the effects of our treatments, and not learning some spurious, non-causal, relationship? In recent decades this problem has however largely been solved Bang and Robins, ; Funk et al. Finally, implementing computational personalization at the scale that I am suggesting will not be easy; the underlying models and methods are still being developed, and many details are not yet finished.
For example, we need to be able to deal large volumes of dependent data that are collected continuously; a technical topic my recently graduated PhD student Lianne Ippel has made a large contribution to Ippel et al. Furthermore, we need the infrastructure to make all of this technically work; recently Jules Kruijswijk and Robin van Emden have gone through great lengths to build an open source platform that allows us to do exactly this, but it needs further development Kaptein and Kruijswijk, ; Kaptein et al.
Next, we need to develop methods to fit these models faster, on large datasets; work that is currently being done by my colleagues and collaborators Matthew Pratola and Reza Mohammadi Pratola et al. We also need to understand much better how we can combine multiple outcomes measures into a single reward; a problem Xynthia Kavelaars will be contributing to in her PhD project.
This is a promising project that I am honored to supervise together with dr. Joris Mulder. By now you might wonder why I have bothered to propose this new method. My proposal seems plagued with challenges and needs lots of work; probablyenough work to keep me and my PhD students busy for the next few years. Cynically, you could imagine that I propose this method precisely because I want to keep myself and my PhD students busy, but this is not the core motivation.
My actual motivation comes from the advantages of the method. Or, to be more precise, its single advantage: with this method we will have a better outcome. The number of future interactions, the number of possible actions, and the number of meaningful contextual factors is simply too large to say anything precise. However, at smaller scales, for simple versions of the personalization problem, we can quantify the benefits. The performance of a personalization method can be measured in terms of its regret: the realized outcome of a method compared to the outcome we could have achieved with full information.
Suppose we compare the RCT to my proposal in a simple case in which we choose one of two possible treatments for homogenous patients, and where the true probabilities of success are. In the worst case we would obtain an expected successes, while in the best case we expect to obtain successes. Thus, a strategy that always selects the poorest treatment obtains a regret of , while randomly picking treatments results in an expected regret of In this setting, the RCT has an expected regret of about 36, while my proposal weighs in at about 12; a difference of 25 successes as shown in Figure 2a.
This difference results from a better balancing of earning and learning. Furthermore, the difference is magnified when we include a context and focus on smaller and smaller groups of patients; this is exactly what we want to do when personalizing our treatments. Scaling the problem to This latter difference is caused by stepping away from simple binary tests to learn a complex relationship, as is the case with the RCT, towards examining and comparing multiple treatments in one go.
Also this difference is magnified when we consider personalized treatments since the more we expand the context-action space, thus, the more characteristics of the patient or the treatment we consider, the poorer the performance of the RCT will be. Finally, as long as we store the probabilities of receiving a specific treatment conditional on the context, we can effectively re-use the data that we collect; something that is almost impossible when using RCTs. A recent theoretical analysis by Agarwal et al. Figure 2c shows the estimated standard errors as a function of the number of datapoints collected using the different methods.
Simply put, using a computational approach to personalization allows us to learn more efficiently than using repeated RCTs. These simple computations show that the RCT is grossly outperformed by my suggested alternative. Furthermore, it is reasonable to expect that the RCT will comparatively suffer more from making the problem more realistic than the method I propose. Thus, if anything, the presented differences in expected out comes are underestimates of the actual outcomes rather than overestimates.
I was told there nothing more to worry about. Fast forward to March of , a routine ultrasound revealed a suspicious node in my right axilla. The Biopsy Removing a sample of tissue so that it can be evaluated under a microscope for purposes of diagnosis. In August of , I started a double blind clinical trial. In December, I was unblinded from the study. The drug was deemed ineffective. I will be starting Nivolumab sometime soon. I believe if you sit at home and wait to be sick, you will be sick.
I refuse to give in and I absolutely will never give up. This disease is tough but I am tougher. A mole I had had on my thigh since childhood began to change. It was a MM. Wide excision with reported clean margins. This was before Sentinel Node Biopsy was done. Biopsy Positive Describing a tissue sample in which cancer cells are found.
Entered clinical trial at MD Anderson. Took interferon injections for one year. A few scares along the way, but still NED now 21 years after first diagnosis. I went in for a random physical because I had gained some weight and could not loose it. That day he removed the mole and another one on my back. Two weeks later when I went back for my results, the one on my breast was Benign Noncancerous. Found so randomly I thank God every day! I had a wide excision surgery that removed the mole which was only at stage T1a, so no other testing was needed. I was Cancer A general term for more than different diseases that involve the uncontrolled growth and division of abnormal cells.
These cells form collections called tumors that can destroy surrounding normal tissue and spread throughout the body. Or so we thought. My results came back that I had an enlarged lymph-node. I had surgery on August 2, to remove the lymph nodes in my axillary and now I have no evidence of disease. Used in reference to immunization. I was going to have a different treatment, but since the melanoma research is advancing so rapidly, a new treatment was found and I have no side effects at all!
Thank goodness for all these new revelations with immunotherapy. I just completed one year of interferon—30 days of IV and 11 months of injections. I have always been a beachgoer and have experienced many bad sunburns. Living in FL for the last 22 years has not helped. I strictly follow any advice or instruction from my Dr. He advised me to cut down on my sugar intake and walk every day. I have wonderful support from my husband, kids and grandson. Planning a long awaited trip to Alaska next month and more to come! My world was turned upside down in April Gray tinged capillaries around the base of the CA were the tip off that something was going on.
Indeed it was! Biopsy and melanoma of 1. Adil Daud.
Editorial Reviews. About the Author. Biography: I was born in the Bensonhurst section of A Simple Cure: Melanoma, Our Worst Nightmare - Kindle edition by. A Simple Cure: Melanoma, Our Worst Nightmare eBook: Lawrence Gold, Dawné Dominique, Donna Meares: dynipalo.tk: Kindle Store.
That has been the greatest gift on my journey. Daud ordered a PET scan to see if indeed there was any metastasis. A sentinel node under my left arm lit up on the PET. I had a subsequent fine needle aspiration of the node and it was positive for melanoma.
Scary — actually, beyond scary! Daud then had me see a surgeon, Dr. Michael Alvarado, who, together with Dr. Daud, suggested I have all nodes removed. So, June 16, I had 22 nodes removed. Biopsy revealed just the one node was positive for melanoma. A blessing. At this point I had a choice of observation for five years, including two PET scans a year or a clinical trial. The trial was testing the preventative effects if any of Keytruda v.
The trial, a double blind study, would involve 18 infusions over the course of one year, plus quarterly CT scans. I chose the trial as it provided the highest level of surveillance along with regular contact with Dr. Daud and the melanoma team at UCSF. To date, I have completed 15 infusions and all scans have been clear. I have no side effects, but that does not automatically translate into my being on the placebo, as many people have no side effects with Keytruda.
But, with regular scans and a pantry full of new and ever-more effective therapies, I need not live in fear. I refuse to do that. In fact, I live joyously — in love with my wife, my children, my grandchildren, great-grandchild, and dear friends. I exercise like a demon, eat right ok, too many oatmeal cookies , I stopped drinking cold turkey, pray daily, receive weekly therapy from a psychologist, and meditate still learning.
First and foremost, I thank the Lord and all those angels who have helped me grow and learn this past year-plus. I thought, upon diagnosis, that the days and months ahead would be the worst of my life. I have learned to genuinely appreciate life. We are alive today — no one…absolutely no one — is guaranteed tomorrow! I wish all my fellow journeymen and women nothing but the very best. We are here for each other, always. I had a mole removed from my right calf in I was told it was not indicative of melanoma. I went on with my life and in February of I felt a lump in my right leg above my knee.
Alex Haynes removed the mass which was the size of a golf ball, along with a sentinel node biopsy. The margins were clear and no detection of melanoma in my lymph nodes per pathology. I spoke with my oncologist Dr. Ryan Sullivan Available treatments offered were immunotherapy or Interferon, both with significant side affects and we agreed not to treat any further since there was no other signs of melanoma.
I saw a dermatologist every six months along with CT scans and visits to my surgeon and oncologist. Five months after the surgery, I found another lump in my groin. It turned out to be melanoma in my lymph nodes. Jedd Wolchuk.
He concurred that as long as the recurrence was operable that was the best choice of treatment. Again five months later, my dermatologist found another lump near the last surgical area and I just had surgery to remove that mass. My melanoma is aggressive, however, not spreading YET. There is no way to detect single cells of melanoma.
I am still considered stage IIIc since it has stayed localized to my leg. Interferon can give year long flu symptoms and immunotherapy can create autoimmune disease which becomes life long treatment. In between surgeries I feel good, although I fatigue easily. I am cutting back my work hours so I have more time for me. The sun still makes me happy but I now protect myself from harmful rays.