.

# Big data startup to fix healthcare

The problem:

• The cost of healthcare in US will go up for everyone, as employers will pass the buck to employees and the uninsured will have to pay.
• Even standard procedures have high price volatility. If you ask how much your hospital will charge for an above \$5,000 procedure, by explicitly asking how much they charged for the last 5 patients, they have no answer, as if they don't keep financial records. Would you buy a car if the car dealer has no idea how much they will charge you, until two months after the purchase? And then you hear your neighbor got exactly the same car from a different dealer for half the price? Any other type of businesses would quickly go bankrupt if using such business practices.

Causes:

• Despite hiring the brightest people (surgeons, doctors) hospitals lack basic analytic talent for budgeting, pricing, forecasting. They can't create a statistical financial model that includes revenues (insurance payments etc.) and expenses (clients not paying, labor, drug costs, equipment, lawsuits, etc.)
• Patients not analytically savvy, don't know that they are financially abused, want to get plenty of useless tests and expensive procedures for minor issues, and ready to sue for no reason. But things are changing. Yet no patient (except myself?) ever mentioned that his budget is say \$2,000 for XYZ procedure, and that he won't pay more and will cancel or find another provider if his budget is too low.
• Volatility is expected in prices, but must be controlled. If I provide consulting services to a client, the client (like any patient) has a finite budget, and myself (like hospitals) have constraints and issues (I need to purchase good quality software, the data might be much more messy than expected, etc.)

Solution:

An Internet start-up offering prices for 20 top medical procedures, across 5,000 hospitals and 20 patient segments. Prices would come from patients filling a web form or sharing their invoice details, and partners such as cost-efficient hospitals willing to attract new patients. Statistical inference would be made to estimate all these 20 x 20 x 500 = 200,000 prices (and their volatility) every month based on maybe as little as 8,000 data points. The statistical engine would be similar to Zillow (estimating the value of all houses based on a limited volume of sales data), but it would also have a patient review component (like Yelp), together with fake review detection.

Revenue would be coming from hospitals partners, from selling data to agencies, from advertising to pharmaceutical companies, and from membership fees (premium members having access to more granular data).

Related articles:

Views: 8201

Comment

Join Data Science Central

Comment by Devendra Sheethal Kumar on November 11, 2015 at 12:23pm

I like the idea, but as Edumund rightly pointed out the pricing depends on the nature of the insurance plan.

Comment by MedicalQuack on January 27, 2015 at 8:04pm

I'll echo Gary Miner below.  I wrote an EMR in the early days and there's not enough time to write everything here.  I have seen models similar to this and we are buried and I mean flat out buried in start ups in healthcare that don't do well with work flows.  Right now there are 38 government agencies that a practice has to report to and could shut the practice down, so there's a hint.  United Heatlhcare has code for just about everyone out there to buy since they are so huge and pretty much have taken over everything.  They do risk assessments like banks and make gobs of money.

I would suggest really, really read up before you jump in here:)

Comment by Brandon Zingsheim on December 24, 2013 at 4:46am
@Vincent - what came of this initiative....did anyone else bite? I'd be interested in exploring this with you further.
Comment by Gary D. Miner, Ph.D. on January 9, 2013 at 11:49am

"Big Data Ventures" and "Big Data Companies' may "think" they can fix healthcare, and bring down costs before USA Healthcare "bankrupts our economy", but there is much more to it than "big data" and "big companies". The "Real Solutions" that can make healthcare "cost effective", "essentially error free", and "Patient individualized" actually require, in effect, what we may think of as "simpler solutions".  Unfortunately, the "big solutions" are missing the most critical aspects, as per the analysis of me and our team of M.D.'s, J.D.'s, and Ph.D., and Nurses......  These will be discussed in our upcoming book PRACTICAL PREDICTIVE ANALYTICS & DECISIONING PLATFORMS for EFFECTIVE HEALTHCARE ADMINISTRATION, DELIVERY, & MEDICAL RESEARCH [Elsevier / Academic Press, expected 2014], and you may get insights into these from courses we are teaching at UC-IRVINE Extension PRACTICAL PREDICTIVE ANALYTICS and also through the LVIT program which will start next summer .......BOTTOM LINE:  all the "hype" about BIG DATA and all the MEDICAL ESTABLISHMENTS purchasing "Multi-Million Dollar"  solutions without "vetting" what the Big Suppliers are selling, will come crashing down when the public realizes that these "solutions" are not really doing the job demanded by ACO and Meaningful Use Directives  [and is already understood by many of those doing "data analysis / modeling"  in these medical institutions; but not understood by the medical administrators that make the decisions on "what solutions" to purchase, to be perfectly candid about this issue .....].......

Comment by Vincent Granville on January 7, 2013 at 10:26pm

@Edmund: My start-up should collect gross price (before insurance refund, that is,  what the hospital charges), net price (after refund, that is, what you actually pay), type of insurance, and time spent at the hospital (from entering the building to leaving the premises) for each procedure.

The gross price is a good indicator for most people since you have at least a vague idea of how much your insurance will pay (e.g. 80% in your case, a rather standard value). Plus, this data will tell you which hospitals have not only good prices, but also low variance in gross prices.

Comment by edmund freeman on January 7, 2013 at 2:41pm

There's a big kink in healthcare pricing that you might not know of. Insurance companies aggresively negotiate prices with providers. For instance, a while ago I got an MRI. The hospital price was \$1,000; the insurance price was \$200. So the "price" of a procedure depends a whole lot on what insurance plan you're on. Also, if you don't have insurance the price that most people get is going to be irrelevant -- it will have nothing to do with the price you're going to get charged.

Comment by Drake Pruitt on January 7, 2013 at 12:05pm

Vincent,

Like an Edmunds for car pricing or TrueCar for actual purchase referral. I like the thinking and have been mulling this over for awhile.

As I understand it, the insurance companies have all the data (they see disparities in claims) at a very detailed level. Further, since they cannot explain the massive discrepancies in price either, they are shifting the risk burden to delivery (doctors/Hospitals/ACO). For example if the spread on a gallbladder surgery is \$5k to \$125K, Insurance co's will set a fixed price of let's say, \$25K for the procedure. It would be a very small step for them to publish those rates.

Also, Medicare is also doing this by drastically altering the reimbursement rates between private practice and hospitals (hospitals getting higher reimbursement). Again, this takes on the look of 'standardized pricing' over time.

So, my question would be: does this service fulfill a need that won't be covered as the industry is driven to adopt standards of delivery, pricing and reimbursement?

Kind regards,

Drake

Comment by Henry PAN on January 7, 2013 at 11:19am

Cool idea,Vincent

Looking forward to use it once it's service ready.

Good luck

Henry

Comment by Hans-Gerlach Woudboer on January 6, 2013 at 10:39pm

Great initiative! imaging this would be combined with internal assessments of time spent for the various activities per service. Run a Monte Carlo analysis and get a perfect benchmarking model helping the whole industry to become leaner and increase performance.