Hacker News Reader

It's really pretty simple, and rather clear from the article, that Theranos has been cheating, VW-style. They use regular equipment to get through "proficiency testing", but use the Edison machine for regular patient specimens:

From the article:

> The two types of equipment gave different results when testing for vitamin D, two thyroid hormones and prostate cancer. The gap suggested to some employees that the Edison results were off, according to the internal emails and people familiar with the findings.

> Former employees say Mr. Balwani ordered lab personnel to stop using Edison machines on any of the proficiency-testing samples and report only the results from instruments bought from other companies.

> The former employees say they did what they were told but were concerned that the instructions violated federal rules, which state that a lab must handle “proficiency testing samples…in the same manner as it tests patient specimens” and by “using the laboratory’s routine methods.”

You seem to be missing the part where former employees claim that the company skirted federal regulations by using other machines for the proficiency testing, while using the Edison machines for patient testing. The accuracy of those claims have some relevance to the kind of speculation you hope to make.

exactly, the dilution of samples needs to move to the forefront of the discussion.

Theranos:

(1) takes a much smaller amount of blood, claiming they will use the Edison machine to analyze it

(2)(a) sometimes uses the Edison machine, whose accuracy is disputed

(2)(b) sometimes dilutes the sample to test using a regular machine. The dilution moves normal concentrations outside the range of detection of the machine, so when a 10x dilution has its concentration multiplied by 10, there are huge errors, often causing misdiagnosis

When Theranos receives a proficiency test, they don't receive a small amount of blood. They receive the amount of blood a normal lab receives. These are the only samples that they can run in the manner the standard machines were designed for.

Theranos competes on costing less and taking less blood. The doubts are all around whether you can get accurate results with smaller amounts of blood from a different body part.

This seems to me to be the most premise summary of what likely happened. Startup raises a ton of money for a new model that isn't ready yet.. so they use existing machines to fake it till they figure out how to get the Edison machine to work. Not terribly dissimilar in some way from jet.com just buying goods you ordered off of other sites and paying for the difference, in hopes they can test the market, build a brand and eventually make their model work. Problem is that it isn't obvious that scale helps you reach the ultimate goal here (unless you reason that more money/longer runway = higher likelihood of success). Unfortunately for Theranos, actual lives are at stake here and there is a lot of regulation in the medical space for good reason.

> 2nd theme: lab tests from Theranos differ from generally accepted standards. How do they differ if in fact they are using the same tests as everybody else? Is it just the general variability of lab results and similar variability could be find in quest diagnostics as well?

Because Edison and the dilution procedure produce low quality results. Thus for a large number of the tests Theranos is less accurate than traditional methods.

> 3rd Theme: almost all people say Theranos is dramatically cheaper than competitors. How is that possible when they are using the same equipment as everybody else for most of the tests? Is it a process innovation in operations rather than from Edison/better equipment tech? Or are they just subsidizing these costs and being cheaper and possible have bad unit economics?

The above is possible because they are not currently focused on profit, rather market share. All that money that they have raised has had to go somewhere.

It seems more like an investor story time exercise in order to attract capital in order to acquire the companies already holding the market share.

Well if they're not using their own technology for most of their tests it isn't just a marketing problem, it's an image problem that would lead investors to believe they don't have faith in their own product. Secondly, according to the article, their protocols even when using conventional machines differ from the norm in that they are forced to dilute samples to increase sample volume because they take less blood to begin with. This is generally accepted as being a less accurate practice.