Sustainable Hospitals

Selecting Non-Mercury Thermometers
Alternatives to glass/mercury thermometers are quite appealing as they are easier and faster to use and avoid the shortcomings of glass/mercury thermometers. The risks of broken glass and exposure to mercury are eliminated, as well as the cost of a clean-up and disposal of mercury from a broken thermometer. With the variety of alternatives available, it is essential that one make an educated choice, to ensure that the tool satisfies the task. Here are some points worth thinking about when you consider thermometers:
1. Acceptable standards of accuracy
Thermometers for medical use are typically tested to voluntary standards set by the American Society of Testing and Materials (ASTM)(See reference 1). The following table shows the maximum error allowed. One sees that glass/mercury and electronic thermometers have the same requirements over the range of 96.4 - 106 F.
Maximum Error over Temperature Range Shown
Thermometer Type ASTM Procedure
(see reference 1)
<96.4 F 96.4 < to 98 F 98.0 to 102 F >102 to 106 F >106 F
Mercury in Glass E667-86 (reapproved 1991)1 ±0.4 ±0.3 ±0.2 ±0.3 ±0.4
Electronic Thermometers E1112-86 (reapproved 1991)1 ±0.5 ±0.3 ±0.2 ±0.3 ±0.5
It is important to note that many thermometers read out to a smaller division than the accuracy of the thermometer itself. For example, digital thermometers which read to 0.1 degrees F may be accurate only to +/- 0.2 F or less. If the accuracy is +/- 0.2 degrees F, the true temperature of a thermometer reading 98.9 F is in the range of 98.7 - 99.1 degrees Fahrenheit. Therefore when selecting a thermometer, one must look closely at the accuracy, rather than the smallest increment reported.
2. Accuracy of glass/mercury thermometers
Inherent in any discussion of alternatives is the assumption that glass/mercury thermometers are accurate. Data suggests that our faith in glass/mercury thermometers may be misplaced.
Leick-Rude and Bloom (See reference 2) describe a study in which axillary temperature in neonates was taken with non-mercury thermometers and compared with a "standard" of glass/mercury thermometers. For the purpose of the study, the accuracy of each glass/mercury thermometer was tested as a condition of accepting it for the study. 25% of the glass/mercury thermometers tested differed from the reference thermometer by >0.2 degrees Centigrade and were deemed unacceptable for use in the study. The authors cite another study in which 28% of glass/mercury thermometers were discarded because they differed by more than 0.1 degree Centigrade from the reference thermometer. The authors raise concern as to the accuracy of glass/mercury thermometers for general use, when one out of four of those tested was not deemed accurate enough. (In fact, the ASTM standard for glass/mercury medical thermometers specifies a maximum allowable error of + 0.1 C in the cited range).
3. Favoring the old standard
Chamberlain and Terndrup (See reference 3) remind us that "Whenever a new clinical test is introduced, investigators measure its accuracy by comparing it to an accepted standard, termed the ‘gold standard’. Because of this comparison to the old standard, initial testing will, by definition, favor the old method, even if the new clinical test is a better test."
4. Use of rectal, oral, or axillary readings as a reference for tympanic temperature
The publication The Clinical Utility of Ear Thermometers (See reference 4) describes different methods and their limitations for measuring body temperature. It cites that the medically accepted "gold standard" for core temperature is pulmonary artery blood temperature. However this is an invasive technique, so rectal, oral, or axillary readings are often used as a crude estimate of body core temperature. Each site is reflective of a different blood supply, with separate rates of change with a rising or falling body temperature. Additionally, each site has variables unique to that site that influence the body temperature measured. The publication concludes that since each site provides its own characteristic temperature properties, comparing a tympanic temperature directly with oral, axillary, or rectal temperatures is inherently flawed.
The lesson here is that with an understanding of how tympanic thermometers work, they offer a safe, convenient alternative to oral, axillary, or rectal temperature measurement. Education is critical to satisfactory performance, and manufacturers are well prepared to advise and coach clinicians on the use of their products.
5. Customer Satisfaction
Numerous interviews with users of non-mercury thermometers provide convincing evidence that alternatives are viable and well-received in health care facilities. For more information on mercury and on product alternatives, check our Sustainable Hospitals web site at or feel free to contact us at or (978) 934-3386.
1) 1997 Annual Book of ASTM Standards, Roberta A. Storer, Editorial Services Director, American Society of Testing and Materials (ASTM), West Conshohocken, PA
2) MK Leick-Rude and Bloom LF, "A comparison of temperature-taking methods in neonates", Neonatal Network; August, 1998, Volume 17 No. 5, pp. 21-37
3) James M. Chamberlain, MD, and Thomas E. Terndrup, MD, "New light on ear thermometer readings" Contemporary Pediatrics; March, 1994.
4) The Clinical Utility of Ear Thermometers, Published by Braun Thermoscan, Pub. No. 0996-267P-R1097
Last Revised: March 31, 1999

back to factsheet listing page


Mercury Reduction
R E G I S T E R     G L O S S A R Y     F E E D B A C K     S I T E  M A P     H O M E

Copyright © 2003 Sustainable Hospitals / Lowell Center for Sustainable Production
All rights reserved.
Images copyright © 1998 PhotoDisc, Inc.