Interested in Covid-19 – read the Westgard blog

May 19, 2020


Everyone should read the articles on the Westgard blog about Covid-19. Start with this one.

The bad and the good in lab medicine

May 5, 2020

First, the bad. Well bad is maybe too strong a word.


If you have ever read an ISO standard, you will notice that something is missing. There is no list of authors or committee members. The people who write the standard should be listed!

ISO 9001, ISO 15189

In the 90s, companies would display banners stating there were ISO 9001 certified. From the ISO website “Using ISO 9001 helps ensure that customers get consistent, good-quality products and services, which in turn brings many business benefits.” But accreditation success is judged by the documentation that the organization has, to show that it is following the processes that it has developed. A company could have poor quality but if they can prove through documentation that they follow their processes, they will be accredited. The same applies to ISO 15189, the clinical laboratory version.

Krouwer JS ISO 9001 has had no effect on quality in the in-vitro medical diagnostics industry Accred. Qual. Assur., 9: 39-43 (2004)

ISO 15197

This standard describes the required accuracy for patients with diabetes who self-monitor their glucose with glucose meters. It came out in 2003 and was updated in 2013. The 2003 version allowed 5% of the results to have any difference from reference (hence 5% unspecified). The 2013 version reduced the unspecified amount to 1%. People with diabetes do a lot of testing. To have 5% unspecified meant that once a week you could get a result that could kill you from an ISO acceptable glucose meter 2003 version). For the 2013 version this was once a month. I was invited to attend an early meeting of the 15197 committee. It was not run by endocrinologists, but rather by regulatory affairs people from industry.

Krouwer JS Wrong thinking about glucose standards. Clin Chem, 2010;56:874-875.


I spent many years contributing to CLSI in the area of evaluations. This group is dominated by regulatory affairs people and it was always difficult to finish any evaluation standard. For example, when I became chairholder of the committee, I finally published a bunch of standards which had been sitting around for 14 years!

I thought Jim Westgard’s original idea about total error made a lot of sense, so I established and chaired a standard about total error (EP21). When it was time to revise the standard, it occurred to me that the original document was about “total analytical error.” I suggested that in the revision, we include pre- and post-analytical error. There was strong opposition to this and not just by the regulatory affairs people but also by hospital clinical chemists. After a while, the CLSI management threw me out of CLSI.

Clinical Chemistry (Journal)

There have been a lot of good things in the journal Clinical Chemistry but here is one that is not so good. The journal will not accept for review a Letter to the Editor except if the Letter is about an original article. That means that about half of the content in the journal (case studies, opinions, editorials, and so on) are off limits. I asked the editor of the journal during a local AACC meeting and his response went along the lines – we vet our articles very carefully and don’t wish to burden the journal with useless blather. My Letter to Clinical Chemistry published in 2010 about the ISO glucose standard would not be considered today.

The good

I like all sections of the AACC Artery. I look at it every day.

Interesting things happen on an instrument approach

May 4, 2020

In flying, to remain instrument “current”, you have to fly six instrument approaches every six months. If you don’t, you can’t legally file (and fly) under instrument conditions. There are three ways to satisfy the requirement:

  1. Fly 4 approaches in a simulator and 2 in the plane
  2. Fly 6 approaches in non-instrument conditions with a safety pilot
  3. Fly 6 approaches in instrument conditions

Due to covid-19, 1 and 2 are out as simulator locations are closed and having a safety pilot violates self-distancing (unless a family member is a safety pilot). Flying approaches in actual instrument conditions in winter is difficult due to icing. So April 30th, I had 1 instrument approach to go and the weather was good for instrument flying.

So I flew the ILS 11 approach at Worcester (KORH) where the ceilings were low IFR. The unusual thing was the wind. As the picture shows, the wind was blowing 52 knots at 3,000 feet. I was flying toward the final approach fix (RILOC) with a true airspeed of 122 knots, but my ground speed was a whopping 170 knots!

And to cap off the day, upon returning to Norwood (KOWD), a family of deer ran across the runway after I landed.

Commutability – again

May 3, 2020

I have previously commented on a series of articles in Clinical Chemistry about commutability.

These posts are:

The point of my posts was when a calibrator is non-commutable, the sky is not falling. You can make the calibrator commutable by a correction factor. This is how we assigned calibrators values at Technicon 40 years ago.

But now there is a new article in Clinical Chemistry about removing bias from a non-commutable calibrator by using a correction factor. What is the purpose of these Clinical Chemistry articles?

Flying the Boston city tour

April 27, 2020





Returning from a flight to Martha’s Vineyard, as I crossed the south coast at Buzzard’s Bay, I asked approach control if I could fly the Boston city tour. The answer was “probably” and to fly direct Boston at 3,000 feet.

(Peases’s Point Mattapoisett, MA)

Boston is surrounded by the “Bravo” airspace and since I was flying VFR (visual flight rules), a clearance is required to enter this airspace. As I neared the airspace I asked for and was granted this clearance. Approach control then switched me over to Boston tower who told me to remain clear of the Bravo airspace and fly towards the Needham towers. That’s shown on the VFR chart below. The blue line is my flight path and the tower’s instructions start at the bottom right of the picture in red. The squiggly lines with the blue dot is Norwood airport my starting and ending point.

So, to stay out of the Bravo airspace, I descended to 2,600 feet and flew around the Bravo ring marked 70/20. This figure means if you are lower than 2,000 or higher than 7,000, you are not in the Bravo. I also had to be 2,600 or higher to stay out of Norwood’s airspace.

The Needham towers are tv antennas, familiar to people on the ground as well. When I was close to the Needham towers, I called Boston tower and was told to start the tour but remain west of the buildings.
My flight path is shown below on a street map. As I approached downtown, I turned away from the buildings as instructed – tower then told me I could start the tour and fly to the Zakim bridge but stay in the inner harbor. I wasn’t sure which direction to fly to remain in the inner harbor but the main point was not to fly over Logan airport.


So, after the Zakim bridge, I turned right and asked to fly back to Norwood – the tour was over. So photos from the tour below.


The specs dilemma

February 18, 2020

In my previous post, I mentioned that there are cases where glucose values are in the E zone of a glucose meter error grid. The evidence for this comes from the adverse event database (commonly called MAUDE). This occurs for both glucose meters (BGMs) and continuous glucose monitors (CGMs) and can be considered real world data because it comes from the population of actual users.

The Parkes glucose meter error grid is used in the glucose meter specification ISO 15197. Both are used for continuous glucose monitors even though this is off-label use.

The problem is that the ISO 15197 spec is used for “acceptability” even though both BGMs and CGMs have unacceptable results. To further complicate things, no unacceptable results are observed during method comparisons used for FDA clearance – the unacceptable results only occur in the MAUDE database.

Hence one has the situation that FDA cleared BGMs and CGMs meet the spec yet unacceptable results occur. The problem is there is no incentive to reduce the rate of unacceptable results when products have met regulatory requirements. Actually, no manufacturer advertises that their product meets specs – they always say the product has exceeded specs.

What we need is real world specs to match real world data.

Two different meanings of an acceptable assay

January 29, 2020

For glucose meters or continuous glucose monitors, an error grid defines potential patient outcomes for results compared to reference. Results in zone A cause no harm, even though results at the outer edge of the A zone have noticeable error. Results in the B zone have the potential to cause slight harm whereas results in the E zone are potentially life threatening.

Results in the E zone are thus unacceptable – we would never want them to occur. But in the real world, they do occur.

So is this assay unacceptable by the FDA, meaning that the assay should be unavailable to users? The answer is no, the assay is acceptable because more harm would occur if the assay were removed.

Hence, the challenge is to focus on reducing the number of “unacceptable” results for an “acceptable” assay.