Citizen fraud detection, self-experimentation, and OOP Updates

For my American homies, we’re heading into thanksgiving and temporarily forgetting the whole food-is-medicine thing. This year, my contribution to the table is a tirzepatide turkey (aka. a chicken).

Some quick announcements and healthcare musings. 

Courses starting ASAP

>    Enrollment for How to contract with payors ends this week. If you’re trying to get in-network with payers, figure out how to negotiate better rates, what it looks like to transition from fee-for-service to value-based care, or why payers look down on you for being lesser beings…this course is the one for you.

>     The LLMs in healthcare course starts on 12/9. We limit the number of people that take this course and have about 10 slots left. We go through real-world deployments of LLMs and lessons learned from them. Look how happy that robot is! That could be you!

>     And finally the US healthcare crash course starts 1/28. Taught by yours truly, I’ll explain how US healthcare works over six sessions (which took me like 9 years and 2 forehead wrinkles to figure out). 

Fraud Waste & Abuse

Fraud has been innovating. Recently I had to tell my parents that if I ever called them asking for money, it could be a clone of my voice so we should have a password with each other. They said they wouldn’t pay even if it was me so it’s not an issue.

Even in healthcare the frauds are getting more innovative and even involve NBA players. But there hasn’t really been much innovation in fraud waste and abuse detection in healthcare. This seems to be some combo of:

1. Very few people have access to the data needed to make a fraud claim or even build the models to detect fraud (e.g Optum and Change are the largest players here I believe because they have access to claims data)
2. The business model kind of sucks, you need to either demonstrate that something didn’t happen because of you (e.g. fraud avoided) or you get paid only when you do detect fraud and take a % of collections (very lumpy and hard to build a big business around)
3. There’s very little incentive to find fraud and lots of disincentive - orgs obviously don’t want to root out fraud in their own companies because it looks bad on them. Or the dollar amounts are usually too small for massive orgs to care about or bring a case against.

Every company in fraud, waste, and abuse focuses on selling directly to businesses or the government. But I wonder if there's something more interesting by making individuals a part of spotting fraud.

It seems plausible to me that with these new gen AI tools you could make it easier to do that.

• Have patients to contribute their EHR/claims data into a pool to detect any anomalies
• Use AI tools to follow up with patients where they might be in a suspected fraud scenario
• Ask more specific questions about what happened + additional data (e.g their Explanation of Benefits, or a recorded phone call in single party consent states).
• Focus on Medicare/Medicaid, and any lawsuits/settlements that happen, 50% payouts that CMS brings goes to the pool and split between everyone who contributed their data. Or partner with a law firm that specializes in qui tam (whistleblowing) and split the proceeds of a case.

Here’s another angle. Prediction markets have become more popular lately. These are essentially places for citizens to bet on an outcome. For example, in this last presidential election the prediction markets performed surprisingly well (even better than traditional polling in a lot of dimensions).

If you're a regulator and trying to figure out where - why not create a prediction market around "is X committing fraud?" and see how people are placing their bets. The Ethereum founder has a post about this and gave one example for scientific peer review. 

“Scientific peer review - there is an ongoing "replication crisis" in science where famous results that have in some cases become part of folk wisdom end up not being reproduced at all by newer studies. We can try to identify results that need re-checking with a prediction market. Before the re-checking is done, such a market would also give readers a quick estimate of how much they should trust any specific result. Experiments of this idea have been done, and so far seem successful.”

A lot of the major paper retractions recently were essentially found by online bloggers. Imagine how much more citizen investigation might happen if there was financial incentive and market? People who believe fraud is happening have asymmetric knowledge by seeing what’s happening at the ground level, wouldn’t have to reveal who they are, but could be financially motivated by voting “yes I think fraud is taking place at X company”. 

If you're an entity that needs to root out fraud, this can build a map of where fraud might be happening. The issue is that you have to have a clear cut definition of what “fraud” is, which is tricky.

Calling Data Engineers/Data Scientists in NY

Doing something fun for data science/data engineer types in NY next month. Email me back if this sounds like you.

SELECT * 

FROM data_engineers_in_new_york 

WHERE they_work_in_healthcare 

AND they_know_happiness

Weird, no results…

The Ethics Of Using Your Own Body

I’m going down the rabbit hole of learning about the surrogacy market for a future post. 

But as I’m reading, I realize that the US doesn’t really have a consistent framework around when it’s okay to consent to selling your body in healthcare. AND there are no good black friday deals here, everything costs an arm and a leg.

1. You can be a surrogate that carries someone else’s child. Explicitly in this relationship someone is paying another person to use their body.
2. You cannot be paid to donate your kidney to someone. 
3. You cannot sell your body to science after you die, but there are lots of brokers in very legal gray areas that connect donated bodies to medical schools, military, etc.
4. You can get paid for donating liquid from your body (plasma, sperm, etc.). However, while not illegal, hospitals have policies against using blood from paid donors. So you practically cannot get paid for donating blood.
5. You can get paid to test experimental pharma products. A phase 1 clinical trial will often test drugs in healthy human volunteers and they’ll get paid to do so (depends on the trial and riskiness of the drug, but it can be lucrative! Like $3K+)

Maybe some medical ethicists or bioethicists can weigh in on what the general framework is that governs these rules because it’s hard for me to find a common theme. 

One interesting twist is that the KIND of payment also might matter. For example, here’s a group that wants to make it possible to donate a kidney in exchange for a tax credit instead of getting money. It requires a waiting period to get the credit and it is phased over years to prevent people getting cash infusions immediately. Theoretically this prevents some of the bad behavior like coercion or using the money immediately for drugs that tend to make people concerned about paid donations.

But even if we think beyond selling, what about our rights around self-experimentation?

There’s a paper that just came out about a virologist that injected herself with an oncolytic virotherapy for her stage 3 breast cancer. It seems to have worked pretty well, but that’s outside of my scope of expertise and pronunciation ability.

What’s interesting though is that many journals refused to publish this.

“Halassy felt a responsibility to publish her findings. But she received more than a dozen rejections from journals — mainly, she says, because the paper, co-authored with colleagues, involved self-experimentation….

The problem is not that Halassy used self-experimentation as such, but that publishing her results could encourage others to reject conventional treatment and try something similar, says Sherkow. People with cancer can be particularly susceptible to trying unproven treatments. Yet, he notes, it’s also important to ensure that the knowledge that comes from self-experimentation isn’t lost.”

Some of the biggest breakthroughs in the history of medicine have come from self-experimentation. There’s the infamous story of Dr. Barry Marshall experimenting on himself to prove H. Pylori cause peptic ulcers. I was recently reading how one of the dentists that made general anesthesia popular would experiment on himself and cats to figure out the right dosages. A med student discovered a cure to his own deadly Castleman’s disease via self-experimentation under his physician when he didn’t have other options.

I can see both sides - but it does feel like we’ve over rotated to discouraging any sort of self-experimentation. People who are willing to self-experiment have pretty deep conviction in their hypothesis - you can’t have any more “skin in the game” than your literal skin. Maybe that might yield actually more successful experiments or at least shorter hypothesis testing time? I know quite a few people that would rather put an experimental cancer vaccine into their body than apply for another grant.

In general I think we’re a little too paternalistic around what people can and can’t do with their own bodies. If we thought more creatively about how to structure the incentives, we could enable more individual agency and experimentation.

How To Throw A Good Healthcare Event

Me and Danielle just finished our third Knowledgefest, the small scale conference we throw every year that’s focused on healthcare operations. There’s a massive appetite for these small scale, tactical conferences. Knowledgefest is 80 people. We’re thinking about bringing this to other roles (e.g. data, product, etc.)

[Btw if you want to hear when operations Knowledgefest 2025 drops - you can sign up here and we’ll let you know]

We’ve gotten quite good at event design by hosting these. People think that if you just get people into a room, magic will happen. But there’s a lot more to it. 

A few examples of things we did at the last knowledgefest for example:

1. Curation of people - people have to apply to knowledgefest. This helps us only focus on people that are serious about coming + signals to people that the other attendees are serious too. This also gives us information about them that we use through the event (e.g. we pair people up in groups of 4 where they bring a problem they’re dealing with and get perspective. Those pairingss are informed by the application)
2. Coaching the speakers - For many conferences there’s like 1 call between the panelists beforehand and it’s basically discussing what you don’t want to talk about. Danielle coaches all 12 speakers starting MONTHS and for many hours on topic, best practices, etc. Because we know our attendees, we know how to make the breakout sessions worthwhile and we keep the breakouts interactive.
3. Interactivity - A big part of the conference is getting the attendees involved. About 20% of the people that attend speak in some capacity (breakouts, lightning talks, etc.). People are much more invested in the success of the event if they’re actually in it.
4. Experimentation - We use Knowledgefest as a launching point to experiment with someone new with the attendees. This year for example, we’re testing out an SMS “call and response” thing where we text questions, get back answers from attendees, then text back all of the responses. It’s much easier to get people excited about that when we’re in-person vs. blasting it online (we have a near 100% enrollment rate of knowledgefest attendees in this experiment and a 33% response rate on the first text we sent out).
5. Have some delightful surprises - We like to have something unexpected that’s nice for attendees. This year we had a room where you could get a professional headshot for free! I think like 50% of attendees ended up doing it.
6. Keep things weird - We always try to keep people on their toes. One thing we do at each event is make weird swag. This year we did a theme around “all good ideas come in the bathroom” - shower post-it notes, coffee, and…custom toilet paper with mad libs on it. Because WHY NOT.

Sponsorship is always weird for our events because it requires the sponsor to think about the quality of attendees vs. # of eyeballs. Big shoutout to our sponsors who not only understood the value prop but also understood the assignment and matched our weirdness and creativity.

So thank you to our sponsors - if you’re in healthcare product/ops they all do things that are very relevant to your job so check them out: Abridge, A.Team, Awell , Assured, Bland AI, Awesome CX, Vogent, Bridge. Sorry not sorry we put you on toilet paper. 

If you’re a company interested in sponsoring other Out-Of-Pocket things, reach out to us.

Headhunting

We’re taking a few more companies for our headhunting service. If you’re a healthcare company that’s hiring, we help you find people in the OOP audience that might be a good fit for your role and you only pay us if they get hired. 

Come through. 

Thinkboi out,

Nikhil aka. “Turkey sucks, sorry”

Twitter: ​@nikillinit​

IG: ​@outofpockethealth​

Other posts: ​outofpocket.health/posts​

--

{{sub-form}}

‎If you’re enjoying the newsletter, do me a solid and shoot this over to a friend or healthcare slack channel and tell them to sign up. The line between unemployment and founder of a startup is traction and whether your parents believe you have a job.