January 13, 2009 Lab Errors in the News
It’s been over a year since we last talked about errors in the laboratory. Last week there was considerable attention in the media about one of the largest ever lab test recalls. USA Today reported that Quest labs had begun notifying physicians last summer about possible errant vitamin D test results (Marcus 2009). The New York Times (Pollack 2009) also reported on the issue and noted the story was first reported by The Dark Report, an industry newsletter for pathologists. The Dark Report, in fact, dedicated an entire issue to the incident.
The Times article and the Dark Report issue discuss many of the technical details about the events, which involved a proprietary test done by Quest labs using mass spectroscopy. They discuss also how some diagnostic tests fall outside the FDA’s oversight responsibilities. Importantly, they also discuss the Quest response to the errors and the Quest intentions to repeat testing at no cost.
Could a similar problem occur at your own lab? How long would it take for your hospital to recognize that its lab test results were showing an abnormal pattern? Hospital labs must undergo very rigorous certification reviews. They all have in place quality assurance programs that include frequent (usually daily) calibration of testing equipment, close monitoring of environmental conditions like ambient temperature, and cross checking results with reference laboratories. Nevertheless, problems may arise even despite such rigorous standards. That’s why it is extremely important that hospitals have methods in place to identify unusual lab result patterns.
Sometimes the abnormal patterns point to problems outside the lab. We recall a circumstance long ago where serum phenytoin levels were running higher than normal on hospital inpatients. The higher than normal levels were validated with a reference laboratory and the investigation uncovered that the pharmacy had received a formulation of generic phenytoin that differed from the previous generic formulation.
But other times the abnormal pattern means a problem with the lab itself. We recall another circumstance where abnormally high protimes were traced to faulty lab equipment.
In both cases above, the problem was noted only when multiple physician inquiries came into the lab about the abnormal results. Both incidents might have been detected sooner if automated surveillance had been operational (both incidents preceded the current sophisticated high tech environment).
With today’s sophisticated computer programs it is fairly easy to write simple applets, programs or rules to track lab value results and trigger an alert if there is a new trend in the results (eg. the values tend to be running higher or lower than normal). The logic might go something like this: “Check every 10th serum ZZZZZ level. If more than x out of y such values exceed the average value for 2008, send alert to lab director.”
You could just compare the average values for the test results this year compared to last year, but if the test is a high volume test it may take a while before there is significant change in the average value. You could also choose to sample consecutive specimens but such a trend might just reflect serial testing on a patient with a known abnormality (eg. a patient with hypokalemia may have his potassium level checked frequently during replacement therapy so the lab may see a “trend” with frequent low potassiums). The more random sampling method (eg. every 10th test) may help get around the latter problem. But that doesn’t help if the test is one that is infrequently performed.
Whatever methods your lab uses to detect unusual trends, you need to make sure you have a plan in place for what to do if you do detect such a trend. Now is a good time for your lab to review its contingency plan for communication with physicians and patients should such unusual trends be identified.
There is one last consideration worth discussion. In the hundreds of incident investigations and root cause analyses we have see over the years, there is one question that consistently fails to be asked: “Was the procedure or surgery or admission or test indicated in the first place?”. That is especially important with diagnostic tests. We try to teach our medical students and residents to ask themselves a series of key questions when they are considering ordering a diagnostic test:
In most cases, if the answer to the first 2 questions is “I’ll still do the same thing”, you have to think hard about whether the test really adds value. (There are, of course, circumstances where the test result could give important information about prognosis even if it did not alter management). And we’ve all been in the uncomfortable position of finding something unexpected (and often unrelated or incidental) when we do order a test and then having to do further tests or interventions because of that finding. Given the substantial increase in vitamin D testing in recent years, it is probably also a good time to be asking questions about the proper indications for that testing.
Marucs MB. Lab sent out number of flawed Vitamin D test results. USA Today Janary 8, 2009 http://www.usatoday.com/news/health/2009-01-08-vitaminD-testrecall_N.htm
Pollack A. Quest Acknowledges Errors in Vitamin D Tests. New York Times January 7, 209
The Dark Report. December 22, 2008. Special Issue on Quest Diagnostics Inc.