Friday, February 24, 2017

Quality Errors - When ISO gets it WRONG!

A ways back maybe 1978 (?) when I was training to be a medical specialist I was doing a research review of the records of patients with a particular infection.  This required going through medical records going back 10 years.  
I remember this even today because after reading one particularly horrific record, I realized it had been written by ME about 6 years previous.  My more experienced self was embarrassed by my younger self’s incompleteness and clinical illogic.  It was a painful lesson.

Today, I  had a similar (sort of) experience.

I was going through, once again, the standard ISO15189:2012 (medical laboratories – requirements for quality and competence) in preparation for an upcoming presentation.  I must have read this document maybe 100 times.  But maybe (by which I mean clearly) this time I was being more attentive. 
I was looking for what we wrote about Quality Indicators, a personal topic of interest  See: [ ].  

Here is what ISO 15189 says within the definitions:
“3.19:  quality indicator -  measure of the degree to which a set of inherent characteristics fulfils requirements

NOTE 1 Measure can be expressed, for example, as % yield (% within specified requirements), % defects (% outside specified requirements), defects per million occasions (DPMO) or on the Six Sigma scale.

NOTE 2 Quality indicators can measure how well an organization meets the needs and requirements of users and the quality of all operational processes. 

EXAMPLE If the requirement is to receive all urine samples in the laboratory uncontaminated, the number of contaminated urine samples received as a % of all urine samples received (the inherent characteristic of the process) is a measure of the quality of the process. “

So what’s wrong with that you are probably saying.  Looks like reasonable stuff.  And it is, until you read the EXAMPLE.  

I was one of a very few microbiologists on the committee that wrote the standard.  I hope that I was not the person who contributed the example, because it is worse that terrible.  My more experienced self today would strenuously object to it being incorporated into the standard, because by its presence, gives license to laboratories to waste their time collecting useless information that sheds NO light on laboratories’ quality or competence.  

I feel some sense of responsibility for allowing the inclusion of a wrong and misleading example.

This is an “in the weeds” issue and one that may not be clear to folks who are not medical laboratorians, but the issue in NOT unique to medical laboratories.  There is  a general and guiding principle: there is no quality or indicator  value in collecting and monitoring information on activities that you do not and can not control.  

To address the aforementioned example, without getting too icky and biological, urine that is in your bladder is supposed to be sterile.  Having bacteria in urine in the bladder generally means the person has a urinary tract infection.

The problem is that when urine leaves the bladder on its way to the toilet or the specimen collection bottle, it can pick up bacteria inside or around the anatomic channel that goes from the bladder to the outside.  Those bacteria are considered contaminants.  Without going into detail, bacteria that are found in bladder urine and bacteria found as contaminants are often the same, so it may be difficult to look at a culture test result and sort if the bacteria came from the inside or outside.  There are washing and collecting techniques that can be used to reduce the amount of contamination that exiting urine may be exposed to.  

But  (and here is the point)  Laboratory workers do not collect urine samples; people go into a bathroom and “pee in private” and then leave the sample on a counter or in a refrigerator.  

We can explain and instruct to people about washing and how to collect the sample, but we don’t join them at the toilet (too gross!!).  The only ways we can control contamination are by explaining to people or putting explanatory pictures in the bathroom, to which people may or may not pay attention.  

So we can, as the example suggests count the “number of contaminated urine samples received as a % of all urine samples” but the number has no meaning and worse, we can do nothing to increase it or decrease it.  So while it may be an interesting number, it has nothing to do with quality or performance, and collecting the information is a waste-of-time.  There is nothing here that sheds any light on quality.  (no LAMP OIL here!!)

So here is a message to the current members of the ISO technical committee who will soon be reviewing and revising the standard:  KEEP the requirement for Quality Indicators, but KILL the example.


  1. This comment has been removed by a blog administrator.

  2. This comment has been removed by a blog administrator.


Comments, thoughts...