husband and father
Talented, beautiful people surround me. The now page usually has the most current updates in this arena.
physician
Oncology fellow, Vanderbilt 2022-2025.
Internal medicine resident, Vanderbilt 2020-2022.
Medical school, Cleveland Clinic Lerner College of Medicine of Case
Western Reserve University 2015-2020.
I enjoy treating people with any cancer, but have a particular affinity for head-and-neck and upper GI cancers (together, 2/3 of the “aerodigestive” cancers, the other 1/3 being lung). There are many right ways to be a doctor, but the way I understand my role is this: to be with people in their suffering and their joy (suffering is everywhere, but there is a special joy known only in the oncology clinic), to see and know the soul within the body (whatever the tumor or the scalpel or the radiation beam or the drug has done to that body), to cure when I can (and deal with the fallout from that cure), and help always (to the end and beyond).
There are glimmers of hope and serious advances in recent years, but these remain horrible diseases, many of which have few treatment options once they’ve reached an advanced stage, and the treatments we do have tend to be quite difficult to tolerate. Most cancers can be disfiguring, but these especially so, whether from the tumors themselves or from the therapies. Aerodigestive cancers also disproportionately affect folks who live out in the country, or in the cities but with few resources, and these folks are my folks.
We need to help each other, and there is much to do.
educator
Harvard Macy Institute faculty, 2018-2020. Health Care Education 2.0.
https://vimbook.org,
2020-present.
The Vanderbilt Internal Medicine Handbook was started by Mike Neuss,
MD/PhD in the late 2010s when he was a resident. It’s an incredible
resource, primarily envisioned as a physical book to keep in your white
coat pocket for quick, authoritative reference. It had a website when I
came to Vandy, but the UX… left something to be desired. I rebuilt it
into its current state (website and infrastructure only - each section
has its own author(s)), and help maintain the back end.
Chase Webber is the faculty support, and has been amazing.
It’s used globally, and one of my goals is to make it easier for smaller, particularly international programs to have their own versions. It uses only free and open-source software, and fits comfortably into pretty much any free server despite being a fairly large book with multimedia.
The main problem is that updating the content, while it is just in Markdown (a very simple text format you can learn in ~7 minutes), feels a little too tech-y to be comfortable for people who haven’t spent time in a code text editor. What it really needs is funding, and a part-time developer with protected time.
data scientist
If you are interested data science consulting, contact me. Current rate is listed at that link.
Major projects are listed below in reverse chronological order (roughly), newest projects at the top.
overview
My work has morphed over the years, as everyone’s does, but the consistent thread and drive throughout has been on coaxing large, messy, complex data to tell us a story about ourselves, about all of us as societies and neighborhoods as well as each one of us individually, to empower us to speak our own sequels.
To translate that into buzzwords, for your bingo game: I’m an oncologist who uses explainable artificial intelligence among other techniques to diagnose and address healthcare disparities, including democratizing personalized medicine and pursuing synergies in the global academic-industrial complex.
I was trained as an anthropologist, where reductionism is an insult, at the same time that I was trained as a scientist, where reductionism is the central conceit. Despite the way I set up that last sentence, there is no true conflict, as the goal for most scientists and anthropologists is the same: to make things at least a little better for someone, but hopefully a lot better for everyone. Modern anthropology is inherently activist, far from the crusty sepia-toned image of a staid researcher sitting just outside the village campfire furiously scribbling in a notebook, and so is modern medical research (we all read Tuskegee and are appalled and want to do better, though precious few become Paul Farmer).
Data science was an unexpected boon, a set of tools that lets me deal with staggering complexity in a disciplined way, to a degree unifying the anthropologist’s drive to let the data be itself (messy, human) with the scientist’s drive to simplify.
I became acquainted with using code to model the world when I was doing physics research (equivalent circuit modeling and scanning laser doppler vibrometry on a Nigerian-style clay pot drum, total hoot, I presented it at the Acoustical Society of America Annual Meeting and we published it in JASA), and the research-centric medical school at Cleveland Clinic gave me the better part of five years and a supportive environment to dive in deeper to computer science, machine learning, etc.
Now that have the freedom to choose my own direction, data science projects that were informed by anthropology but had centroids in other disciplines have become squarely within the overlap of my personal Venn diagram.
Additionally, as I have moved from academia to the worlds of private practice and industry, I am increasingly inspired toward using these interests to hit cancer where it needs it the most: in the clinic of the community oncologist, where the vast majority of cancer patients receive care, by bringing timely insights from every relevant axis to bear on every visit.
Information almost wants to be free, and there is now so much of it, that a lot can be done for free or at a low cost. (I’m take special glee in byproducts, “digital waste,” some say, but there is gold in them there hills - Shigeru Ban is my hero). There are abstracts from conferences aplenty describing the cutting edge of everything, usually seen as jots and tiddles on the march of science forward but forgotten when the journal article gets published, and lab values and imaging results and clinic notes generated in the usual course of patient care, used once or a few times and then forgotten and discarded in the wake of what comes next - in all of this digital refuse I believe there are insights to be gained, efficiencies to be achieved, lives to be saved.
disparities in cancer research across space and time
This ongoing project starts with a custom metadata archive of all the abstracts from the two big annual cancer meetings (ASH and ASCO - as a starting point), made machine-readable and hence more easily accessible en masse. Here’s a rough prelim of what that looks like, with an interactive map and sorting/search/export (you can hit JSON or CSV endpoints trivially to get nicely formatted data, as opposed to the hulking web scrapers and HTML wranglers I wrote to get the data the first time). Analyses are planned on the authors and affiliations (global mapping, with a time component), and how those overlap with disease states (e.g. how much sickle cell research comes out of Sub-Saharan Africa? We know it’s bad, but nobody has quantified how bad - the idea is to have a platform for answering these questions easily), before branching out into things like conflicts of interest and various analyses using NLP, etc.
Cleveland Clinic Center for Clinical Artificial Intelligence
explainable machine learning for readmissions, with a focus on sociodemographics
While at Cleveland Clinic, I was one of the founding members of the Center for Clinical Artificial Intelligence, and its first dedicated analyst. We had the delightful opportunity to work with a rich dataset from one of the world’s largest hospitals, focused on predicting the risk of readmission (discharging from the hospital and “bouncing back” too soon) and extended length of stay. We had access to not only health data, but socioeconomic data, for millions of patients, and I further enriched this with census data.
With only an address and a date, one can learn about a person’s neighborhood in incredible detail (though not the person themselves - the US Census is wise about privacy in the data they publish). It turns out that a person’s neighborhood is a major predictor in their health outcomes, as is their insurance provider, in addition to a host of health parameters and hospital process clues.
I used interpretable machine learning techniques, with a focus on revealing actionable items that the patients and medical teams could hope to influence, as well as to call out structural issues that organizations and governments need to know. One of the problems with machine learning is that it can only ever restate (and in some cases, as it feeds back on itself, more firmly entrench) the biases that led to the historical data you fed to it. If you design it to show you these biases, show you all of its biases, explicitly, to tell you exactly how much a person’s race, ethnicity, gender, and neighborhood played a role in its predictions (alongside medical diagnoses and lab values), you have transformed an algorithm from a potentially destructive tool, a heinous thing that pushes hurting people farther down, into a tool for positive change. If these biases are shown in an easy-to-read visual, that patients, clinicians, and administrators can all understand, all the better. See what it looks like here.
Cleveland Clinic Taussig Cancer Institute
explainable machine learning for blood cancer diagnostics
Initially, when I was going to be a malignant hematologist (I still think it is a beautiful field, but was drawn away to other pastures), I wanted to level the playing field for advanced diagnostics.
A world-class hematopathologist together with a world-class clinician, preferably a whole group of these together, are required to make certain diagnoses, and these only after at least one invasive biopsy. What if we could mobilize the rich genetic and phenotypic data available in simple blood samples and from the electronic medical record to support diagnostics, and, eventually, democratize them? Further, what if the answer the machine gives could be not only accurate, but interpretable?
What if we could get the machine to explain itself and its reasoning, both as a check against biological implausibility (and more insidious problems such as systemic racism), and to reveal more areas for research? What if a high resolution bone marrow biopsy image could be read in moments by a smart phone in rural Arkansas/Kandahar/Mogadishu/Kushalnagar/etc., instead of having to be shipped to one of a handful of academic centers while the patient waits, still sick, for an answer?
We made some progress, but these problems remain largely unsolved.
internal links
If you want to know which tools I use, visit /uses.
If you are interested in the tech stack for this website, visit /colophon.
If you would like to throw money at me, for whatever reason, visit /pay.