The Neuroinflammation, Pain, and Fatigue Laboratory calls it a chronic disease survey. I call it “The GREAT Survey”! The definition of great is “of an extent, amount, or intensity considerably above the normal or average” and it qualifies on all accounts. It’s longer, more comprehensive, and ultimately dives more intensely into the chronic illnesses than any survey I’ve ever taken – and I’ve taken lots of them.
(This survey was introduced in the recent Jarred Younger blog and is fleshed out more here.)
One clue that it’s a level above most surveys is that it’s using machine learning to uncover the gems it will produce. Machine learning is just like it sounds: in machine learning, the computer uses the data it gets to create its own algorithms – which it then uses to reanalyze – which provides the opportunity to produce an even better algorithm, and so on.
Throwing machine learning at a standard survey form would be like using calculus to answer a simple math problem. You only throw machine learning at complex data sets and this data set is intentionally complex.
Instead of focusing only on people with chronic fatigue syndrome (ME/CFS), or fibromyalgia (FM), or long COVID, this survey is embedding people with all these diseases (as well as Ehlers-Danlos Syndrome (EDS), Post Treatment Lyme Disease Syndrome (PTLDS), irritable bowel syndrome (IBS), multiple chemical sensitivity (MCS), multiple sclerosis, migraine, lupus, etc.) in the same data set and then is going to have “the machine” explore how they are related.
Essentially, everyone is getting thrown into the same basket – then the basket is going to get shaken up – and then the “machine” is going to put things together again. The specific aim is to better understand the big suite of pain and fatigue disorders (ME/CFS, FM, IBS, etc.) all of which (with the exception of long COVID) are damatically underfunded. We may see connections pop up that we’ve never suspected before.
It’s a very long survey, but you can stop and resume it at any time (using the same device). Some people finish in under an hour, but it will probably take between one and two hours in total. I did part of it about a month ago – clicked on the link yesterday – and was right back to where I had stopped. As always in these surveys, some of the questions are phrased in a kind of weird and off-putting way – that’s because these surveys have to use validated questionnaires that were phrased in sometimes weird and off-putting ways.
The survey can be completed on a smartphone, but a tablet or computer with a larger screen is probably best. Anyone, anywhere can take the survey.
Chloe Jones – the researcher running the study – answered some questions about herself and the survey.
Chloe Jones Interview
“My mother became severely ill with fibromyalgia and ME/CFS when I was 4 years old.” Chloe Jones
I asked Chloe Jones, the researcher behind the study about it and herself. It turns out she was inspired to take on these conditions through her mother – who has ME/CFS/FM.
What is your academic background?
I am currently a graduate student within the Neuroinflammation, Pain, & Fatigue Lab with Jarred Younger at the University of Alabama at Birmingham (UAB). I am enrolled in the Medical/Clinical Psychology doctoral program at UAB. Prior to UAB, I conducted cognitive neuroscience research at the University of Connecticut. I have been conducting neuroimaging research for the past five years and conducting neuroinflammation research with Dr. Younger for the last three.
How did you get interested in ME/CFS?
My mother became severely ill with fibromyalgia and ME/CFS when I was 4 years old. As many of your readers did, my parents became their own investigative team, as they received little help from clinicians at the time. From that age, I always tried to understand the illness, constantly making my own hypotheses, and always wishing that answers were near. I never stopped asking questions or searching for answers.
How is the survey different from others of its kind?
Thank you for your interest in the project! This survey aims to investigate different facets of illness that may exist between and across chronic illnesses, particularly pain and fatigue disorders that are poorly understood. Typically, researchers control as many variables as possible to isolate a factor of interest, but this is not always reflective of reality.
The complexity of patients’ medical histories and illnesses may be central to uncovering critical underlying features. Rather than trying to isolate a singular, distinctive, or unifying feature among all ME/CFS patients, (such as a particular genetic mutation or virus that could explain all patients), it may be more fruitful to investigate the various dysfunctions that might exist.
Furthermore, we can investigate if any of those identified facets overlap with other medical conditions. This is especially true in that our ‘label’ for ME/CFS is based on a clinically useful definition, but this label is not based on any clear biomarkers or etiology, and therefore might be applied to a variety of underlying causes.
Although there have been research findings identifying certain biomarkers of interest, these are not used clinically to diagnose patients. Some patients may receive different diagnoses but actually have a shared etiology, while other patients may be given the same diagnoses but have different underlying dysfunction. The complexity and diversity within the ME/CFS patient population has traditionally been a challenge in research, while this project embraces that complexity.
Why are you using machine learning? What is it and what does it bring to this particular project?
There are two main benefits to applying this type of analysis. One, this type of analysis can be applied to varied and complex datasets to uncover patterns that may not be readily identified with more traditional methods. It is particularly powerful with large datasets, and in this project we aim for a large sampling of a variety of illnesses.
Secondly, one goal of this project is to minimize the bias from typical research or clinical definitions that are often useful, but flawed. Rather than apply the traditional definitions of ME/CFS to patient groups, we can assess more nuanced connections with emerging patterns in the data.
What do you hope to get out of the project?
There are many research questions that we can investigate with this type of dataset, however, the main goal for this project is to identify facets of illness that exist between and across pain and fatigue disorders that may more accurately represent underlying mechanisms.
How many participants are you aiming for (and why), and how many you do have right now?
We currently have collected around 2,000 responses for this survey and hope to capture 10,000. Because we want to capture the true reality and complexity of humans’ health and functioning, we are interested in those with all different health conditions and health statuses to respond. That includes those with pain and fatigue illnesses like ME/CFS and FM, related illnesses like EDS, IBS, MCS, MS, SLE, as well as seemingly unrelated illnesses that are still informative to assess the relationships between and across disorders. Those without any medical conditions or health concerns are also encouraged to participate if interested.
Lastly – This is a VERY long survey – the longest I’ve ever taken. Why is it so long? How does its length support what you want to get out of the survey?
We understand that this is much longer than your typical online survey. I truly wish that we could capture this degree of detail in a fraction of the time, but we are aiming to collect information that is typically not asked, and may point us in important directions. This is a very comprehensive survey, and we are grateful to anyone who is willing to provide their data to this dataset. Each respondent is providing uniquely valuable data, and the more data we collect the more we may be able to uncover.
For those who are unable to finish, the amount of data they did provide is still very useful, even if they can only spend 15-20 minutes. There is a core question set that takes most users around 45-60 minutes and an extended set of questions afterward for those who would like to provide even more detail. The survey also varies in length depending on one’s responses. Thankfully, responses are automatically saved to the server and participants can take as many pauses as needed. 20 respondents will also be randomly selected for $50 e-gift cards of their choosing.
Plus, anything you’d like to add? 🙂
The survey focuses on ME/CFS, FM, and related disorders, but it is open to all participants who are adults and are proficient in English language, including those without any medical conditions at all. Some of the survey questions may seem strangely worded or confusing; unfortunately, this is usually because it is taken from a standardized instrument and cannot be changed. However, any feedback about any aspect of the survey is always appreciated. I would also like to say thank you to you, Cort! Thank you very much for your interest, your curiosity, advocacy, and outreach.
Use the link below to start the survey