Posted by Sten Westgard, MS
The news hasn't been getting any better over at Theranos. Since the Wall Street Journal pierced the facade on Theranos' promise of revolutionizing the lab test industry, there have been additional stories that Walgreens is seeking an exit from their partnership, that some of their premier clinical partners, Intermountain Healthcare and Cleveland Clinic, have not even started work on any validation studies with Theranos technologies, and that Capital BlueCross has put a stop to Theranos blood-drawing at the insurer's retail store in Pennsylvania.
But that's nothing compared the release of the full, but redacted, CMS inspection report of Theranos' Newark, California laboratory.
Earlier, we had seen the top sheet of the report, noting five major violations that meant Theranos was putting its patients in "immediate jeopardy."
The full report, more than 121 pages long, is as bad an inspection report as I have ever seen. More explanation of this, after the jump...
There are so many things that the Newark laboratory did wrong I don't have space to cover it in one post or article. Instead, I will focus this article just on the QC failures that were taking place last year. In fact, there's even a mention in the CMS report and the Theranos policies that call out the "Westgard Rules."
Missing Rule Violations
On page 65 of the CMS inspection report, it notes that
"a. CL QOP-00013 Revision D, 'Quality control in Chemistry', stated in section 220.127.116.11.2 that QC is deemed to have passed when... Westgard rules have not been violated (see following monthly QC section 6.3.2)"
"b. CL QOP-00013 Revision D also stated in section 18.104.22.168 that ten consecutive observations on the same side of the mean should be monitored."
In other words, Theranos had a policy to implement "Westgard Rules" on their chemistry tests, including a 10:x rule. Aside from debating the wisdom of that policy, once it was in place, the laboratory staff was expected to follow it. The facts appear to show that the laboratory did not. From the page 65 and 66 of the inspection report:
"iii Review of Levey-Jenning reports for January 2015 through April 2015 revealed MQ1 (Lot number 25661) had 10 consecutive results below the mean, MQ2 (Lot number of 45662) had 10 consecutive results below the mean and MultiQual Level3 (MQ3, Lot number 45663) had 10 consecutive results below the mean but within 2 SDs for all four months.
"iv. MQ1 data revealed 125 of 125 below the mean for January 2015 through April 2105
"v. MQ2 data revealed that 126 of 126 below the mean for January 2015 through April 2015 but within 2 SDs
"vi. MQ3 data revealed 125 of 125 below the mean for January 2015 through April 2015 but within 2 SDs."
In other words, the lab not only was missing the 10:x violations, they were missing at least a 350:x rule violation!!!
Bad QC Practices
The lab was running Citrol controls to monitor the Prothrombin time-Innovin PT on the Siemens BCS XP Instrument.
"c. The general supervisor stated that QC was acceptable if the values were +/- 2 SD from the mean.
"d. From April 1, 2015 through September 16, 2015, 32 of 69 days showed Citrol 3 QC values were greater than 2 SD (- 2 SD) below the mean.
"e. On 4/7/15, Citrol 3 was run six times before an acceptable QC value was obtained.
"f. On 9/7/15, Citrol 3 was run seven times without obtaining an acceptable QC value.
"g. On 9/8/15, Citrol 3 was run twelve times without obtaining an acceptable QC value.
"h. On 23 of 32 days, Citrol 3 was not rerun when the QC value was greater than - 2 SD."
in other words, the lab was repeating controls until they fell back "in", or they were outright ignoring out-of-control falgs and reporting patient results regardless.
This does not paint a pretty picture. First of all, these failures are occurring on the "typical" instruments in the laboratory, not on the proprietary Theranos technology. So this lab was staffed with people who couldn't properly run regular lab instruments, much less the advanced revolutionary miracle machines that Theranos has been promising the public.
Just to put a bit of hyperbole on this: it would have been simpler for CMS to hand the laboratory a postcard on which they wrote the few things that the Theranos lab had done right, rather than have to detail all the myriad ways the laboratory had gotten it wrong.
Now, to be fair, this was an inspection of past behavior, and we know that Theranos has fired its old laboratory director and hired new staff. From a practice perspective, all of these problems can be corrected. You can train your staff to properly run QC and interpret QC results correctly.
I'll wait for another post to break down the other failures of the Newark laboratory.