Clinicians order at least one laboratory test in about a third of all clinical encounters. However, a new study published in the Journal of Applied Laboratory Medicine finds wide variation in the use of such tests to make clinical decisions, ranging from 98% in inpatient settings to 56% and 29% in emergency department (ED) and outpatient populations, respectively. 

The investigators, from Virginia Commonwealth University in Richmond, used electronic medical records from the University’s medical system to assess 72,197 typical patient encounters throughout the year. 

They found that laboratory tests were the most frequently ordered diagnostic test, followed by radiology, vital signs, respiratory, cardiology, and others. Half the laboratory tests ordered came from the inpatient setting even though there were about 20-fold fewer inpatients than outpatients; 42% from the outpatient setting; and 8% from the ED. 

The most common diagnosis orders for inpatients were vital signs (53%), laboratory (26%), and respiratory (18%); for outpatients, the most common orders were for laboratory (70%), vital signs (18%), and radiology (7%); and for the ED, the most common orders were for laboratory (57%), radiology (19%), and cardiology (13%). The authors speculated that vital signs were ordered more often in the inpatient setting because of the need for multiple vital signs compared with outpatient encounters, which typically required only a single vital sign, but which may need multiple laboratory tests. 

The lower number of tests ordered in the ED compared with the inpatient setting, they suggested, may be related to the wide range of conditions seen in that setting, from minor injuries and infections to serious, life-threatening conditions. Finally, the low number ordered in the outpatient setting likely reflects the chronic and elective nature of conditions seen in that setting, which typically require fewer diagnostic tests. 

The study appears to debunk the oft-used claim that laboratory tests inform 70% of medical decisions. Instead, the researchers wrote, “Our observations support that it is not possible to use a single number to categorize the frequency with which laboratory tests occur in patient encounters.”