When an information architect goes searching for a health answer
It sounds like the setup for a bad joke, right? But I put on my information architect hat — a well-honed ability to navigate information based on structure, patterns, and cognition — to solve a health issue doctors couldn’t figure out.
So, what was going on?
It started on August 18, 2016 after sweating buckets on a high heat/humidity day, followed by days of continued sweating (even after the weather calmed down and getting into air conditioning) and months of night sweats. My gynecologist believed it was probably menopause. Constipation kicked in, and while it was mostly just a little concerning and something that I kept trying (unsuccessfully) to get & keep moving, nothing worked. I eventually wound up with RUQ pain and SIBO.
The RUQ pain is like an icepick, on par or worse than when a bit of bone came loose in my knee. That bone fragment would get stuck between bones in the joint, then get unstuck, then get stuck again on random impact, around and around in a lovely merry-go-round of excruciating pain. The RUQ pain was more random, and would happen multiple times a day for months.
The SIBO helped me figure out a dietary path that got my transit started (sporadic, slow, but going). It took time, but I got to the point where I wouldn’t have torso-curling RUQ pain randomly, and when it did show up I had known instigation points (I ate something wrong for me) and a timeline (a week or two). I quickly got to the point where I could just hitch my step a little instead of actively curling around the pain, unless I was hyper-focused on something.
That continued for a while until I became tired, then fatigued. When the exhaustion hit, so did tremors. Then the tremors were around even when I didn’t feel especially tired, and I started having sporadic memory issues.
Now, the doctors gave me lots of good news. There were no mechanical issues, neither in my gut nor neurologically (although by November 2021 my transverse colon was sitting in my pelvis — not considered an outright problem, not unique, just probably the source of slow transit). My memory was “nothing I should be worried about” and tested high against population. They could find nothing to explain my exhaustion — no inflammation markers or primary-suspect vitamin deficiencies.
As far as the doctors were concerned, they’d done their job. There was nothing physically wrong with me.
As I kept on keeping on through the years, I was getting worse. With no guidance available other than trying to live within my energy, I still got worse. Spoiler alert! Because the core issue was going untreated.
What I did, the sky-high view
So, I’m an information architect. One aspect of my version of information architecture means I periodically spelunk in vast data seas to find patterns to bubble up with the intent of making sense of them. I know I haven’t dug deep enough yet in a new data sea if I don’t hit a point of overwhelm. It’s still as overwhelming as anyone else experiences, but mine rarely has panic and has an underlying, whispered, yes, I almost have enough data now, cue dance music to hit play soon…
Approaching our western medicine health data set was daunting, but not beyond reason. Western medicine is a hierarchical information architecture — it’s right there in that siloing of information between different specialists. But that’s not how the body works. Try to move your toe without anything else moving, not even the bones on top of your foot. Didn’t work, right? Since you were prompted by what you read, your toe didn’t move based on reflex. The signals were initiated in the brain by your conscious mind, and transmitted through the entire body with no experiential lag time. The body is a network information structure that involves multiple systems (for a toe wiggle: bone, ligaments, muscles, nerves, ions, neurotransmitters) to do an experientially simple thing. It’s reasonably safe to work on the premise that western medicine information architecture is a hierarchical structure imposed on a network structure, even without a detailed view into the information of the human body.
Connectome — those links between data that gives it meaning and context — has to be put aside in order to make the data fit in that transition. There are gaps as a matter of course. My assumption walking in was that I had to find a way to navigate the gaps.
What I have listed below didn’t happen in a particular order, and I often backtracked and looked again with new ideas/facets/lenses in mind. There were days when I was absolutely certain that I had butted up agains the edge of my universe, didn’t have any more ideas, and had exhausted the ones at hand. But my energy would bounce and another idea or six would come bouncing along behind it. It’s the nature of my flavor of persistence.
The goal wasn’t about fitting my issue to the right diagnosis, but to understand what was happening to find leverage points. If it wasn’t ultimately named, I didn’t care. I just wanted to feel better.
I built a model of my experience.
Agnostic of what I was told by health professionals (they were often interpreting through their own models and, as shown later, very much leveraging confirmation biases), I started faceting the issue. I looked at it by timeline, what it impacted, and symptoms from multiple viewpoints. I looked at my history of tweaky systems and reactions declared ‘unusual’ by health care. I even noted some of the oddities of the diet I naturally chose as a child. Finally — after wrapping my head around all those facets — I came back to how I experienced an acute episode. Then I refined it until it was reasonably simple.
Doctors said upon glancing at it, “it doesn’t work that way.”
Started sussing out models of health care methodologies.
Western medicine, traditional Chinese medicine, Ayurvedic medicine, chiropractic medicine, functional medicine. I looked primarily at overall diagnostic pathway(s), as well as where they shined. FYI, I was not good about pulling in a review from multiple practitioners — or even practitioners for each of the practice methodologies. It wasn’t robust, just a way to springboard into a different lens.
Made my own model of health categories, leveraging the idea of body-wide systems.
Most of the methodologies covered systems from a process-oriented information architecture, and/or leveraging points of immediate physical contact — e.g., liver is connected to colon, so western medicine often lumps care of either under gastroenterology. On closer review, often methodologies had stronger bets on particular interpretive mediums. In other words, Western medicine focused on mechanics; chiropractic medicine focused on nerve impingement; functional medicine focused on nutrition. Stepping back with these few guideposts, acknowledging that my issues were systemic and not nodal (e.g., not confined to an organ), I devised a different viewport. Then, I looked at what my six years of experience and trying to find help had already ruled out.
Once I had my model, I backed into what I’d already explored through the years. As I crossed more potentials off my list and added research, I went back and kept my list updated, until what I was left with was mostly an information issue.
I had many — albeit tiny — recognitions as I read about the far-out-of-whack manifestations of receptors and ion channels in the peripheral nervous system. In my model, I considered that information. Ions code events on a positive/negative binary sequence through time and space. Neurotransmitters are barely understood, but the bet is solid that they are involved in emotion, memory, autonomic and semi-autonomic processes like breathing and heart rate, and pain — to get started. And while inflammation wasn’t showing up on the accredited markers, I couldn’t rule out the immune system in a few of its less understood manifestations.
The immune system is constantly rewriting its own genetic code. It is literally updating the information substrate of its own system on an ad hoc basis to deal with every pathogen and invader; manage (start, stop, moderate) inflammation to protect healthy cells for an assumed truncated period of time; and set macrophages (self-eaters) on systems gone amok and dead detritus on a cellular level. We have multiple information systems running in our bodies and sharing information, materials, and functions.
Corralled all my data into 2 finite, searchable sources.
Instead of medical data spread through 7 different systems as owned by each practice (and still everything before 2003 lost in time), I compiled all metrics and tests into one spreadsheet file. In that, I could build metadata to connect the bits and pieces along multiple facets and ranges. As I dove into formal research papers, I downloaded PDF’s of those with interesting/potentially useful bits. Those I collated in another spreadsheet, expanding the file set in that software. The intent was to be able to reasonably easily share research papers; it hasn’t been needed to date.
I managed all my models and non-research study in one other file set. In all, I could review notes and think through various viewpoints with only 2 programs open. This cut down unnecessary noise — cognitive load — while maintaining availability and fungibility.
Logs I still maintained on paper. I find it personally easier to jot down what I want to capture rather than opening a program.
Stopped sharing medical records.
Why? Doctors looked at the labels, that a bunch of tests were done, who did them — and confirmed, over and over again, that the other guys did their job. They weren’t looking for what had been looked into to see if they could come up with a different idea. They weren’t interested in trying to understand what was going on now that a course was definitely not working. They were looking at whether the other doctor(s) followed protocol: best practice, as indicated by acknowledged factors, and accepted by insurance. The best they would do was run the same tests that had already been run to see if anything had shifted enough to provide a springboard. Correlation and movement over time was not of interest, just the updated metrics.
The experiences that every doctor was asked about and never got in my medical record? Still ignored. Push back about my collapsing energy? Not my lane. Ask pointedly about cognitive biases? Looks like you need a psych eval!
To be fair, this is not laziness or disinterest on the doctors’ part. This is time constraint. This is approaching every case under speed, efficiency, and mandated guidelines. This is reimpressing who owns what cognitive territory. Spoiler alert! The core issue was never going to be found with these constraints.
Researched.
Targeted by my models, relensed through the silos of western medicine as the easiest modality for me to gain insight through published English-language papers, I dug into information until I hit overwhelm and then dug deeper.
I avoided diagnostic labels as entry into my research; they had not helped in my I’ll-leave-it-to-the-doctors years, not in first person and not in third person. To troubleshoot, I needed to understand current and moving understanding of the function to be able to back into my particular malfunction. This is where my skills as an information architect really helped.
After my research, going back into diagnosis lists to see if there were any potentially known/widely-accepted remediations for what I was positing, the structure of diagnosis is…really fucked up. They are built based on catastrophe, which in a way makes sense since many ah-ha moments have come from seeing a body in extreme dysfunction and using that to compare/contrast to function. The methodology itself supports an on/off view of health, so that anyone who isn’t fine can only find instances that will lead to death. It is not intended to provide remediation before a situation hits the failure state, but to provide cotton batting and bandaids once a failure state has been reached. Seriously: minimize your time in them.
Correlated to my models.
correlation is not causation, it’s a reason to dig
To make sure I followed leads, to shut down wrong paths, to pull in supporting symptoms along several lines of inquiry, and so I would have at least an indicator of why/how I had moved away from lines of thought, I continued to go in and review and annotate my models. Over time, the core model started shifting and growing as I started to understand what was going on.
Tested.
Because tests are easier to tolerate in a failing system than experiments — and I say this knowing that a blood draw would put me down for two days — I tested. I targeted even the tests. Not everything can be tested as a first step, but anything I could find that was relevant and I could find tests for, was. Some tests weren’t direct, but were intended to rule out easier facets of what I thought might be contributing.
Made bets to experiment.
My initial focus was to not repeat what had been tried before and didn’t work. Clear out everything possible to refine the signal (e.g., stop taking vitamin C unless stopping made things worse; while validating, I found I was already refined, probably due to my hypersensitivity), and experiment in focused, singular roles. With the timespan running up to 4 weeks when a mistake was made, this was not quick. By testing, I could make better bets and avoid vast ranges of potential experiments. The best bets were evident in hours.
Find the right help.
This was actually the hardest thing. Many health care practitioners don’t want to work without access to digital medical history. The tests that were the most out of whack were not something used in every office, required a provider ID to talk to the lab for patterns it might have sussed out, and practitioners were willing to go outside their comfort zone. By the time I picked up the reins myself, I felt like I was failing physically. It had to be fixed, and I had to do it smartly — no ‘best practices’, no throwing spaghetti at the wall to see what stuck. It had to be precise and as close to target as possible.
I knew these constraints were triggers and outright untenable by the vast majority of health care, so I stated them baldly when I reached out. I was frequently ghosted.
And now?
I’m in the process of healing, finally.
The tests got me started in the right direction, and my energy was very slowly getting more stable.
The person I found who was willing to call the lab on the most unusual of the tests put the last piece of the puzzle in place.
The kickoff event? Lead poisoning; other heavy metals, too, but lead is the big one and some of its more unusual properties (like how it usurps calcium’s role) correlate to my experience with more strength. By treating with lead in mind, all my issues are rolling back. Some are faster, some are slower.
I’m already expanding my diet — with glee, and when missteps happen I’m able to remediate much more quickly because I know what’s being skewed in my system. I still get icepick stabs, but the pain intensity is decreasing. My energy envelope is already increasing.
I am of the opinion that the highly restricted diet that I found by SIBO-informed trial and error in 2017 (low FODMAP, no starches, only honey for added sugars) kept the genetic potential for primary biliary cirrhosis (96th percentile) or primary biliary cholangitis (84th percentile) from rolling into clinical expression, but contributed to the RUQ pain.
One of the ideas we’re getting skewed in the zeitgeist: genes aren’t your fate, not unless they are mutations that affect systems so much a balance is near or at impossibility. Genes are the cards you were dealt, and you have to figure out which game you’re playing and how to make the hand work for you in that game. If you go in looking at your cards like you’re playing gin rummy when you’re actually playing bridge, you are going to play them wrong. Western medical practice is playing Go Fish, and no other set of rules is on radar until you are already in a fail state. It will also very strongly voice that you should adhere to their Go Fish rules more closely as you feel worse.
I have 5 autoimmune disorders above the 90th percentile for susceptibility according to genetics and today’s understanding, and I didn’t kick into clinical in any of them — it was one of the paths revisited both clinically and in my own research. I think I’ve been skirting around them all, but I listen to my body well enough that I managed to stay as much in balance as the lead would let me. I say this even while remembering how many people tried to tell me that I was problematic, too focused on a balance that didn’t make sense to them.
The progression of how the dysfunction built is in line with my original model, with one cross. That said, the fatigue was something I powered through for years — I had good months and bad months, but could mostly still work in line with my workaholic soul, and I kept trying to refine what seemed to be working. I didn’t really concede to fatigue until I started hitting energy walls. I didn’t push with the doctors until I started having energy collapses: times where I would fall asleep, either sitting where I was if I didn’t read the oncoming collapse, or an unavoidable nap if I took a break.
I’m lucky and had a certain set of skills
I’m lucky in that, while the ultimate issue is one that can kill me if mishandled or continued to go untreated, it does have guideposts and can be managed. I’m lucky in that I found the right someone to help me get that last, super important piece in place. Without her and that piece, I would be continuing to just manage downstream issues.
The skillset was information architecture. I was able to leverage my understanding of how information structures interact for gross navigation and spelunking. The signs of memory, cognition, and social architectures skewing the structures were noted and used to modulate how I approached the information. I was able to mostly figure out what was going on, with no medical training, and finally had metrics that were so outside of the box that it helped me locate a person who would listen, call the lab, and recognized the constellation of symptoms well enough to suggest another test that found the kickoff issue.
We know our medicine is broken. However broken you think it is with maintenance, reasonably easily found but expensive remediations, and the exhaustion of everyone involved in care during COVID, it’s exponentially more broken if your illness doesn’t fit neatly in the existing information architectures.
Lead poisoning is not unknown; I simply didn’t fit the expected demographic.
The brokenness
These are spark points. Sparks are only the beginnings. People are multivariate, and deserve multivariate potential solutions.
Some of the bigger patterns that skewed my particular experience:
- Mismatch, not only between diagnostic expectations and me, but in the information itself. Hierarchical information architecture impressed on a network architecture is common; but the more complex the information gets, the more it needs alternative models to navigate more of the underlying potential connectome and data.
- Western medical practice uses history to define the present and future. It’s not wrong, but it’s also not right. We do not know everything, change is constant, and living according to history emphasizes core precepts and cognitive biases.
- We have a tendency, as a culture, to prefer solitary, universal answers with supreme confidence. I approached the problem looking for potentially multivariate answers and a search for truthiness.
- We are getting reductive too far down in the details, following a mean of a population set in which there is no person who actually embodies the mean. Pilot cockpits and Norma shows just how unrealistic this approach can get.
- Doctors are human. They slip into the cognitive support systems of our ouruborus of perception just like everyone else. When these systems are emphasized, like with an urgency for speed and efficiency, it takes more time and intense effort to break out of it.
- Patients have had their agency stripped away. This comes through in behavior, how data is managed, how much access we have to our own data, and in the repetitive data confirmations.
- I don’t really have enough information (yet?), but I suspect physicians also have a lack of agency.
These are not easy things to navigate, let alone shift. They are patterns that are involved in our large-scale culture systems, and some of them are generations old.
The brokenness of our health care complex goes back to how we manage information and data, the agreements we come to as we form our social architectures, our agreements surrounding money and sharing the efforts of our specializations, and how the systems interconnect. When the health care complex starts working sustainably for all patients, it will be based on many solutions.
We can’t ignore that the reason it’s functioning this way now is that people somewhere along the many input and use lines think the positives outweigh the negatives. They will fight for the status quo, as will anyone who isn’t comfortable with change.
Even with that backdrop, the leverage points are numinous and the factoring of buy-in or rejection of the status quo is on a spectrum: all, none, and anything in between. Some things we can do with fewer people and see if it takes root, some things we need a cohort with us before we do anything more than talk.
