Error rates in test registration at a South African laboratory

Posted by Sten Westgard, MS

More evidence of pre-analytical error rates, this time for the Journal of Clinical Pathology. This is from a study back in 2010, my apologies for only finding it this year:

A Six Sigma approach to the rate and clinical effect of registration error in a laboratory, Naadira Vanker, Johan van Wyk, Annalise E. Zemlin, Rajiv T Erasmus, J Clin Pathol 2010:63:434-437.

In this study, they reviewed 47,543 test request forms from a 3 month period of November 2008 to February 2009. The study was conducted at the "chemical pathology laboratory at Tygerberg Hospital - an academic tertiary hospital in the Western Cape Province of South Africa. The laboratory is a division of the National Health Laboratory Services, which is a network of 265 pathology laboratories in South Africa."

Can you guess how many errors they found? And what was the impact of those errors?

As it turns out, there were only 72 errors in that batch of test requests, for an error rate of 0.151%, or 4.46 Sigma on the short term scale. Quite good! And an improvement over the first recorded rates for pre-analytical errors from Nevelainen et al back in 2000. That landmark study found that the Sigma score for order accuracy was only 3.6. So the years of focus on pre-analytical testing have paid off. Improvements have really been made.

This study did more than just count errors, however. It tried to figure out what kind of errors occurred and what their impact on patient care would be. Here's just a glimpse of their findings:

"Our study found that laboratory errors could result in five unnecessary inpatient days over a 3 month period. The potential outcome for almost half of the cases was a repetition of the tests and a delayed result. Many of these patients were seen at an outpatient clinic, and would have had to return to hospital to repeat the test or receive results."

Here is the breakdown of the potential outcomes of these errors:

• Delay in Emergency surgery (2 cases)
• Delay in Acute Drug therapy (4 cases)
• Delay in Chronic Drug therapy (9 cases)
• Delay in Elective surgery (1 case)
• Prolonged (unnecessary) treatment (2 cases)
• Delay in discharge from hospital (5 cases)
• Simple repetition of the test and delayed result (35 cases)
• Unable to review records (2 cases)
• Increased costs to the laboratory due to extra time required to change LIS test order (12 cases) 

In other words, even though the rate of errors was low, there was still potential to impact mortality, morbidity, and inflict additional costs on the hospital and patient. This is more sobering still when you consider that this is just the impact of one type of error in the total testing process.

A Six Sigma approach to the rate and clinical effect of registration error in a laboratory, Naadira Vanker, Johan van Wyk, Annalise E. Zemlin, Rajiv T Erasmus, J Clin Pathol 2010:63:434-437.