Tuesday, December 14, 2010

Big enough to fail?


Question 1: How many laboratories in Canada (or indeed anywhere) have been closed down because they did not get a perfect score on their proficiency testing challenges? Answer: None.  Never.  Notta. Zip. Zero.
Question 2: How many laboratories have been allowed to remain open, but lost their ability to bill their provincial medical services program for services rendered because they did not get a perfect score on their proficiency testing challenges.  Answer: see above.
Question 3: How many laboratories in Canada “game” their proficiency challenges by holding the sample back until the right technologist is available, or do repeat testing, or do extended testing, or send the sample to a referral laboratory.  More than one, disturbingly a lot more than one. 
Not to belabor the issue, but what is the point of the exercise?  Are we really that insecure about our professional competence that we have to “cheat” when there is nothing on the table, nothing to gain.  No gold stars, no extra cookies, no scholarship, nothing.

What we do lose is the opportunity to check that our standard operating procedures are getting us to the right answer.  What we do lose is the opportunity to do a competency check on our operating systems.  What we do lose is all the extra time, effort, energy and money is doing the extra testing, and scheduling. Consider that  finding a single weakness through PT can save you poor quality costs equaling your laboratory's total PT costs for a whole year, or more.

There are solutions or work-arounds that can be put into place, but they are either expensive or inconvenient, or create increased, and from my perspective unacceptable risk.  
We have been toying with the possible linking proficiency testing turnaround times to clinical sample turnaround times.  If a sample should pass through the clinical laboratory in less than 24 hours, then that becomes the upper limit for the P.T. challenge.  If it should take five days, then the PT turnaround limit is five days.  It would not be particularly difficult to do, for programs with on-line entry of results.  For paper or fax entry it may be a little more difficult.   It would result in more paper work on both ends and that would result in some inconvenience and added cost.

Another solution would be to create a link between the P.T.  provider and the laboratory information systems, so that reports would be automatically generated and sent tot he provider.  This would take some initial set-up time and save time on the laboratory side, but would, in most P.T. programs, cause increased time requirements to transfer results to the data base for analysis  with an inevitable increase in cost.  Additionally, it is stunning how unstandardized our reports are.  But that is a topic for another time.

Or we could disguise the samples completely and have them submitted as clinical samples.  But that would cause all sorts of challenges with getting site specific requisitions from each laboratory and creating patient names with appropriate identification numbers.  And then we would have to sort through the problems created with billings.  (A number of years ago when I was a resident, a microbiologist in the hospital created a  throat swab with C. diphtheriae.  It was a brilliant idea except that the result got reported to public health and much chaos ensued).

Or we could just get rid of proficiency testing, but most laboratories would lose an essential quality assessment tool and a major source of continuing education.
There is, I think, a more reasonable solution.  Laboratory management including the quality manager decide that as of this day we do things differently.  No more overwork, no more processing by the QC technologist.  Routine processing only.  It's worth a try.


The POLQM Weekend Workshop is coming together well.  For those interested in doing a poster or podium presentation, register early and send your abstract to ubcpolqm@gmail.com

1 comment:

  1. Your first two questions pose an interesting proposition. Should laboratories be closed or lose their ability to bill if they do not achieve perfection on proficiency tests? I'd argue they should not. I do think laboratories should do well-enough on proficiency testing that there is little doubt on their ability to get the right answer almost all of the time. In general though, "well-enough" is not well-defined.

    Some accreditation bodies do follow proficiency testing results and demand answers when a laboratory does the wrong thing. On occasion faulty proficiency testing does lead to further investigating by the regulatory body and may lead further to a complete on-site survey that will reveal other problems. I'm not saying the system is perfect, but PT is monitored.

    As for your third question, I think many if not all laboratories stack the deck when performing proficiency testing. Unfortunately, in this case 1+2=3. When laboratories feel threatened by poor PT results, there is even greater preferential treatment of PT samples.

    I do think repporting PT results in real-time would at least partially address this wide-spread problem. I clearly remember a discussion of accreditation standards addressing PT where a laboratory physician advised that anything prohibiting preferential treatment for PT samples should be rejected. The reasoning was that "no one obeys this; it's just not reality".

    A new approach to the way PT is handled will only come when laboratories realize the benefits they can reap if PT is handled in the same manner as patient samples.

    This comment is my opinion only and does not reflect the opinion of any current or past orgainzation I have worked for or with.


Comments, thoughts...