Welcome back. Thank you for visiting this series on Risks and the Clinical Laboratory. This is our last installment. In it, we will discuss the third area in the laboratory where mistakes occur: the post-analytical (PA) portion. We will begin by defining this area and share some data on the number of errors that occur in the PA phase. We follow this with suggestions on detecting PA errors, and then end this first part of the installment with suggestions on avoiding PA errors.
The last installment will conclude with a review of parts of the earlier installments, and another list of sources of errors/risk and the frequency of them.
Once a test has completed the analytical phase and has passed QC, the result is most often reported in a timely manner using an electronic process, or by accurately transcribing the results to a paper or written medical record – in some cases, by a telephone call to the physician who ordered the test or other person if the laboratory protocol states that critical values be called.
The three highlighted words in the previous paragraph are places where there are risks for possible errors. Some of these can significantly affect the patient. For example, should a troponin (I or T) be elevated on a patient in the ED, it is imperative that the result be given to the ED physician so that treatment can begin (“Time is muscle”). “Accurately” and “written” are well known pitfalls in the PA phase. Approximately 50 percent of proficiency survey ‘failures’ are due to transcription errors. As many as 25 percent of errors given to the clinician were PA.
How can these types of errors be detected? Agarwal et al. performed an in-depth study of all three general areas of risks and errors. They first quantified the number of PA errors, and then, after discussions with the staff, they noted substantial improvement. In this phase, critical value reporting, urgent sample reporting and TATs were monitored. On scrutiny of urgent sample and critical value sample reporting, poor awareness among nursing and lab staff was observed. As reluctant lab staff was to communicate these reports, the nursing staff was also not aware of their role in this regard.
Critical value reporting is an important aspect of the post-analytical phase of the clinical laboratory testing process. It is defined as values that represent situations that could be life threatening without treatment. Ineffectiveness of critical values notification or the failure to provide notification within the target time might prove to be life threatening in certain cases. The literature reports frequency of critical value reporting from 1 in 2,000 to 14 per 1,000, whereas other studies reported 2 in 1000 which improved to 4 in 10 000 after intervention. Hence, the sensitization of staff improved the reporting of such cases by a factor of 5. It has been suggested that regular ‘pep talks’ play an active role in improving the reporting system in the lab.
Since dividing the possible sources of risks and errors into three parts, others have added two categories and quantified all five:
1. Pre-pre-analytical (46-68 percent)
Inappropriate test request, order entry, patient/specimen misidentification, sample collected from infusion route, sample collection (hemolysis, lipemic, icteric, clotting, insufficient volume, etc.), incorrect collection tube type, inappropriate container, handling, storage and transportation.
2. Pre-analytical (3-5 percent)
Sorting and routing, pour-off, aliquoting, centrifugation (time and/or speed), pipetting and labeling.
3. Analytical (7-13 percent)
Equipment malfunction, sample mix-ups, interference (endogenous or exogenous) and undetected failure in quality control.
4. Post-analytical (13-20 percent)
Erroneous validation of analytical data, failure in reporting/addressing the report, excessive turn-around-time, improper data entry and manual transcription error and failure/delay in reporting critical values.
5. Post-post-analytical (25-46 percent)
Delayed/missed reaction to laboratory reporting, incorrect interpretation, inappropriate/inadequate follow-up plan and failure to order appropriate consultation. Sometimes, in the P-PA, not enough attention is spent regarding the method’s reporting units.
Examples may be CRP/hsCRP and, in the near future: when the latest version of the very sensitive troponin assays come on line, the reporting units will change from ng/ml to ng/L (actually pg/mL). Also, from time to time, laboratories will change methods or on a grander scale, such as with the instrumentation (i.e., change from Beckman Coulter to Siemens). These are significant changes. Communication to those in the PA arena need to receive proper instructions of these changes through meetings or written lab communications signed by the lab director/pathologist about such changes regarding effective date of changes and, if applicable, the interpretative comment.
In this scheme, one part in the pre-pre-analytical and all of post-post-analytical errors cannot be ascribed to the laboratory (highlighted). However, the data certainly point out that there is still room for improvement – in all areas. Some of the improvements can be achieved by those who are “at the bench,” but only some. Other changes need to be made by those high up the line. In some cases, changes may be needed by the director of the laboratory and even higher.
Years ago, I read in a book about biomarkers and errors the question, “How many babies is it OK to drop on their head in the nursery?” Of course, the answer is “None.” The same answer applies to laboratory results to the physician -“No errors is the goal.” Granted, that is not possible. Given, it can be improved is the approach that needs to be taken.
I wish to thank Tammy Taylor, Nathalie Lepage and William McLellan for their comments.