Jonathan Halvorson

Complexity and a small number of cases are certainly problems for any predictive statistical model. Now that machine learning can assess millions-to-billions of correlations across many thousands of variables, I think we will see substantial improvements in predictive power anyway. These will still be statistical models, and their extension to new cases will be vulnerable to changes in social, institutional and technical conditions. I'm not talking about technocratic perfection here. Not a discovery of new causal laws. Just a substantial improvement from the current state.

A good algorithm cannot exist for the 1% of highest complexity tasks which are undertaken by modern humanity. Most of those tasks are medical. Too many inputs + poor correlation to outputs (wide standard deviations). That's the problem. It's not feasible to assume that an accurate algorithm can even exist for such complexities until they become so routine that the data sets are enormous! A great example of this is optometry. 200+ years ago very few people could afford glasses and the skills needed to assess patients and manufacture to lens were extremely limited. Starting in the 1990s, you could just walk into a Walmart and get a pair made for you without an appointment. When kidney transplants become as routine as contact lenses, the Free Market might start working for kidney transplants. Might... Yet still, I pray to God that we never reach the point where ⅓ of Americans need a new kidney every 2 years. Scenarios that can't exist... don't matter very much. Don't cling to the poison pill of Economists where everything is based on deviation from established assumptions. Be a scientist/engineer who acknowledges the constraints of reality.

Minimum conditions for a Free Market to exist: 1) Many suppliers 2) Many consumers 3) Most consumers understand alternative suppliers 4) Minimal duress 5) Low barriers to entry for new suppliers Medicine often fails 3, 4 and 5. Medicine fails Free Market applicability test.

@mtm14 @realdocspeaks It's National Health Expenditures data. Extremely valuable if you want to understand macro trends. CMS collects it. cms.gov/data-research/…

@son_of_halvor @realdocspeaks Source?

We were told Obamacare was going to save healthcare and that has been an expensive failure. But now you want to double down on another government solution, that is a terrible idea.

@clearSNR @__dwr__ My last comment about not understanding the algorithm referred to a system in which the consumer simply reads an output and has to trust it. A good algorithm will earn its trust. The average consumer already doesn't understand the science behind any modern medicine.

@son_of_halvor @__dwr__ If and only if the potential consumer can understand corelated data. Joe Q. Public can't understand how inputs map to outputs for complex processes (kidney transplant). Joe Q. Public can see it for simple processes (public school, new cell phone, new car, plane ticket, etc.)

I think it’s notable that the price increases are primarily inelastic goods — we need housing, we need healthcare, we need food — while price decreases are almost entirely optional luxury goods This feels worse than the opposite because we can’t tighten our belt out of it

@clearSNR @__dwr__ What you said is consistent with what I said.

@son_of_halvor @__dwr__ There are too many variables during a kidney transplant for a median consumer to form a statiatically significanf distinction between the quality Dr. A and the quality of Dr. B. This is very different from deciding whether to buy Apple or Samsung for the next smartphone.

You could look at a statistic of the # of prior amputations or transplants. Other things equal, the higher the number the more likely the procedure will go well in your case. But that's a crude measure, and you don't know if other things are equal. My point is that the data exists to get more fine-grained to help you judge who is most likely to treat your particular case the best, but the data is not assessed and shared. You as consumer wouldn't have to *understand* the algorithm, just read off the result.

@son_of_halvor @__dwr__ I view #3 as a consequence of success rather than an education problem. In 1890 when all we knew about diabetes was stop eating carbs or die, the general public understood "state of the art" diabetes care. Amputations & kidney transplants? Incomprehensible to Joe Q. Public.

@realdocspeaks The ACA was created because health insurance wasn't affordable in the individual market. We've gone from 16.7% uninsured to 8% uninsured. Totally fair to argue the ACA didn't solve the underlying cost problem. But it didn't exacerbate it, or we would see that in the NHE.

The reality is that over 90% of people on the ACA require subsidies. If it was affordable why are subsidies required? Many people on the ACA exchange can’t afford the deductibles and other costs sharing. Then if you can afford the premiums and deductibles, 20% if claims are denied. It isn’t all bad news though, the insurance companies have made billions of profits and their stocks have rocketed in value. It has also hypercharged vertical integration and gas killed off independent physicians.

@__dwr__ @clearSNR #3 is partly an education problem, but is much more than that. We don't have great measures of how likely a physician is to solve your particular problem vs other physicians. The data simply isn't assessed and available.

@clearSNR #3 is a citizenry education problem. If the assumption is "consumers are too stupid to breathe properly" then you call them "patients" and charge them 10x.

@clearSNR Great post. You can make your point stronger by noting that medicine often fails #1 as well. Especially for things like hospital care and on-patent drugs.

@SurgeryCenterOK Why don't we see the skyrocketing prices in the National Health Expenditures data? The ACA was passed in 2010.

The ACA caused the prices to skyrocket. Until that is acknowledged/repealed your next move (whatever that is) will bring more DC slop and worsen the situation.

@Hstlinghosptlis @mcuban Also illegal, if the labs and tests were in-network.

@mcuban Oddly looking at itemized bills, I had to pay more during deductible for the same labs and tests than the insurance company did after deductible was met…. That was weird

Explainer: If you need care and you can't afford your deductible, you will have to borrow money to pay for it. Typically the healthcare provider will loan you the money. Why ? They want to get your insurance company's money. Now, by definition, the hospital is a subprime lender. Not only is the patient in a huge financial hole, across their similarly situation patients, so is the hospital. They know they won't collect however much your deductible is. And that's just for your in-network deductible, nor does it account for your family max out of pocket I don't know the % of bankruptcies this causes What I do know is that the loss of hundreds of billions of dollars for patients and hospitals, starts with the plans offered by insurance companies. They know damn well when someone picks a plan and they can't afford the deductible. And to them, it's not a bad thing. If you can't afford your deductible, the chances they pay anything from your premiums, go way way down So patients go broke or can't afford care. Hospitals have huge uncollected debt, so they make it up elsewhere And then they aggregate those amounts, among others, and use them to get payments from state and federal programs. See how all that works together ?

@nocorporatemed Would you also deregulate the supply of doctors, nurses and drugs? If you deregulate everything except the supply of doctors, that would give doctors enormous power as the limiting factor in the supply-demand equation, which would then get to dictate reimbursement.

Absolutely! I can picture micro hospitals and centers of excellence, run by highly qualified American DRs. Rheumatolgy specialty center, specialty hospitals for rare diseases, robotic surgery centers, cardiac centers of excellence. deregulate and watch American innovation blossom

@sergeanttheodor @pmarca You're talking about managed care networks. That only arose in the 90s (on a national level). They were created because price inflation was out of control under the old system, which just used "usual and customary" prices and did no negotiation.

@pmarca Isn’t what actually happened that insurance companies passed on the entire cost of service by artificially inflating the costs in order to make it appear like they provided value by “negotiating” lower payments? The way I see it, insurance companies are a major problem.

Compete based on cost and quality.

@pmarca Health care, especially life-saving health care, has the most price-insensitive demand curve possible. No price is too high for the "best" doctor when it could mean the difference between life and death. This is why health insurance was created. We're going in circles.

@ssdhall This is like showing a graph of the percent of the ticket price that goes to pilots and flight attendants, and being upset that it's so small. Meanwhile, the cost of the airplane, fuel, maintenance, etc., gets treated as "administration."

Not convinced this is true. But remember it’s not healthcare , it’s healthcare administration

I've been through hundreds of small towns. My parents grew up in them and I've spent probably a year of my life in them. I've been interested in urbanism for my entire adult life and moved to an older rail suburb in the NE to minimize driving. I know everything you know about small towns and urbanism. I'm telling you that after looking at the same information you have, you're underestimating what will be required to do away with cars. Having a fraction of the old bones isn't enough. We need all of the bones back, and some meat added on, and changes in our way of life. I'm in favor of densification and rebuilding main streets. I think the paranoia around "15 minute cities" is stupid. But you need to take the magnitude of change far more seriously and take a long-term incrementalist approach. And even after all is said and done, the ideal state is a world where we have an average of 1 car per household, not 2.5 cars. A majority of trips by foot or bike...or horse. That's still a huge change from the isolated car culture world we have now, and a good mix of community closeness, efficiency, practicality, and freedom to roam. Never forget that if half of car trips are under 6 miles, half are over. You'll be taking small kids on these trips, groceries, and so on.

@son_of_halvor @19th_c_Houston @jasonc_nc 1. you'd be shocked by the number of small US towns that developed pre 1910, and still have the bones of walkability. 2. statistically, half of all car trips are under 6 miles, a bikeable distance. If the infrastructure supported alternative travel methods more would use them

Random conversation with our son made me look at the stats of the town I’m from. -1,500 people, ~same as when I grew up there. -a poverty rate of 40%. Also ~same -avg inflation-adjusted income same as 1990 Truth is a lot of rural places haven’t been left behind as much as they’re exactly what they’ve always been. And for those in any position to change them there is no desire to do so. Which is why a lot of the conversations about “how do we fix rural areas” are largely pointless. There is nothing to fix. It’s what they are.

Your employer is paying 80% of the bill in the scenario described in the article. They save money on the NP. You could argue they should give you a share of that and cut your copay. But the employer's costs are increasing by 8% a year. Any amount they cut from your copay is getting taken out of your paycheck.

If I have to see an NP instead of a real doctor, I want a 50% discount on my bill This is quite the racket healthcare providers have devised. Most of the time you go to the “doctor“ you’ll never see a doctor at all. You’ll see a nurse practitioner. But they’ll still bill you as if you had seen a doctor. Even though the nurse practitioners are far less educated and trained, and paid much less.

Approximately zero small towns and cities in the US today are laid out as compactly as the pre-industrial revolution towns of the old world. You can't just get them to walk everywhere without redesigning the built environment. Not everyone had a horse, but most families did outside the 10% living in urban cores. Public transport works in dense cities, not really small towns or sprawled cities. And definitely not in farming communities and deeply rural areas. I agree bikes are a big improvement that would go some distance to making the current built environment work. We would still need to change a lot, both in the built environment and in our way of life. You cannot expect most people to bike 12+ miles each way to work. Millions would need to change jobs or move. Highly disruptive.

@son_of_halvor @19th_c_Houston @jasonc_nc Wrong. Cars were not widespread until maybe 115 years ago. How do you think humanity lived for millennia without them? You think everyone had a horse? No, they had walkability, public transport, and now we have bike as an option as well.

@ColinEHunter @19th_c_Houston @jasonc_nc Yes, if I want a stroll. No, if I have four bags of groceries. No, if I need to bring myself or my kid to a doctor. You will need to do a more radical restructuring of daily life to make things work without cars. And you will need horses.

@son_of_halvor @19th_c_Houston @jasonc_nc I think you underestimate how compact a lot of small towns are. For example, in a town of 10,000 people, the majority of people would live within 1 mile of the downtown. This would be an easily walkable distance.