back to XHP | CNR | radiation | rat haus | Index | Search | tree

plain HTML | ASCII text formats

How X-Ray Doses Vary
From One X-Ray Practitioner to Another:
Nationwide Surveys

By John W. Gofman, M.D., Ph.D., Professor Emeritus,
Molecular & Cell Biology, Univ. of Calif. Berkeley
and Egan O'Connor, Exec. Director, CNR and XaHP.

XaHP Document 103, September 2000

  • Part 1 -- Measured Doses Reveal Key Facts: Surveys
  • Part 2 -- Do 75% of Places Give Doses Higher Than Needed?
  • Part 3 -- What Are "The Lowest Doses Technically Possible"
  • Part 4 -- What Are the Main Sources of Evidence?
  • References

XaHP: The X-rays and Health Project.
      An educational project of the
      Committee for Nuclear Responsibility.

The plan:  Lower x-ray doses per x-ray procedure.
The result:  Fewer cases of x-ray-induced cancer
and coronary heart disease.


o  Part 1.  Measured Doses Reveal Key Facts: Surveys

        o  An assumption, widely held among physicians and patients alike, is that x-ray practitioners in general already give x-ray doses which are as low as technically possible, consistent with obtaining good images.

        o  That assumption is demonstrably false, as illustrated by the table below, which is based on nationwide surveys of about 300 facilities (CRCPD 1989 + CRCPD 1994).

        o  Each row in the table shows how dose varies from place to place for the same exam, on a "patient" (really a dummy) of the same standard size. Size matters (thicker patients require more x-ray dosage than thinner patients). Earlier and more comprehensive nationwide surveys reveal dose-variations of even greater magnitude (Wochos 1977 + 1979). Part 2, below, explains more about the table.

        o  The two major sources of dose in x-ray imaging are CT exams (computed tomography) and fluoroscopy (the use of a continuous x-ray beam to view the motion of organs, arteries, catheters, needles, surgical instruments, etc. on a screen in "real-time").

        o  Some fluoroscopists use more minutes of viewing than do others, for a comparable case. Skillful technique by a fluoroscopist can reduce patient exposure from fluoroscopy by a factor of 2 to 10 (NCRP 1989, p.37). In the table, the row "GI: Rate/Min" reveals the variation in dose-rate per minute of fluoroscopy in the Gastro-Intestinal examination; it does not reveal the variation in the number of minutes used.

|Entrance doses    |      (A) |      (B)* | (C) Median |       (D) |      (E) |     (F) |     (G) |     (H) |
|all are in        |   Lowest |      Dose |       Dose |      Dose |  Highest |   Dose- |   Dose- |   Dose- |
|milliRoentgens    | Observed |  Dividing |   Dividing |  Dividing | Observed |  Ratio: |  Ratio: |  Ratio: |
|   --->           |     Dose | 1st & 2nd |  2nd & 3rd | 3rd & 4th |     Dose | Col.C / | Col.D / | Col.E / |
|                  |          | Quartiles |  Quartiles | Quartiles |          |  Col.B* |  Col.B* |  Col.B* |
|  * Doses equal to or lower than Col. B are "the lowest doses technically possible"                        |
|    for getting good images (text, Part 2).                                                                |
|Chest (PA)        |      2.4 |       9.2 |       11.6 |      16.0 |     81.0 |     1.3 |     1.7 |     8.8 |
|Lumbar Spine (AP) |      6.2 |     252.0 |      333.0 |     487.0 |   2154.0 |     1.3 |     1.9 |     8.5 |
|GI: Rate/Minute   |    700.0 |    3500.0 |     5000.0 |    6800.0 |  16200.0 |     1.4 |     1.9 |     4.6 |
|GI: Spot Films    |     38.0 |     175.0 |      268.0 |     398.0 |   4815.0 |     1.5 |     2.3 |    27.5 |
|CT Scan: Head     |   1600.0 |    4500.0 |     5100.0 |    6600.0 |  14000.0 |     1.1 |     1.5 |     3.1 |

o  Part 2.  Do 75% of Places Give Doses Higher Than Needed?

        o  Suppose that dose received during a specific x-ray procedure is measured at 300 different facilities, and then analysts arrange the doses in order, from the lowest to highest one observed. The 75 lowest observed doses constitute the First Quartile of observations. The adjacent set of 75 higher doses constitutes the Second Quartile. The next set of 75 even higher doses constitutes the Third Quartile. And the last set of 75 highest doses constitutes the Fourth Quartile. The three dividing points between the four quartiles have names too: First Quartile, Median, and Third Quartile.

        o  It is self-evident that, when observations are divided into quartiles, twenty-five percent of facilities (in the First Quartile) give doses lower than do the other 75% of facilities.

        o  Does this justify a conclusion that at least 75% of x-ray practitioners give higher doses than the doses required to achieve good images? Probably, but not necessarily. We have no proof that any x-ray practitioners in the First Quartile obtain good images. But we have a hint that they do.

        o  That hint is provided in an article by Joel E. Gray, Ph.D., who was Prof. of Medical Physics at the Mayo Clinic until his recent retirement. He reports that, at the Mayo Clinic, the measured x-ray dose for the AP Lumbar Spine exam is 250 milliRoentgens (Gray 1998b, p.70). That dose is just inside the First Quartile in our table. While Dr. Gray does not explicity say that image-quality on spine films is good at the Mayo Clinic, it seems safe to assume that quality is very good, because of the Mayo Clinic's reputation for excellence.

        o  What we can say, logically, is that if any x-ray practitioners in the First Quartile can obtain very good images, then all practitioners could do it.

        o  The bottom line is that probably, but not certainly, about 75% of x-ray practitioners give higher doses than the doses required to achieve good images.

        o  For x-ray imaging, the guideline on dosage is a very permissive one, called "as low as reasonably achievable" or ALARA --- which may sometimes mean whatever is convenient.


o  Part 3.  What Are "The Lowest Doses Technically Possible"?

        o  A point to remember: The minimum dose, observed in surveys, does not reflect the ideal or "best practice" because images, made at the lowest extreme of the dose-range, are generally useless or poor. Poor quality causes re-takes and additional exposure. Therefore, the minimum observed doses in Col.A of the table are not the "lowest doses technically possible."

        o  An appropriate definition of "a dose as low as technically possible" is probably any dose which is equal to or lower than the dose which divides the First Quartile of current practitioners from the Second Quartile.


o  Part 4.  What Are the Main Sources of Evidence?

        o  National surveys of a few x-ray procedures have been conducted (one procedure per year, generally) at about 300 facilities, as a cooperative effort between (a) the U.S. Food and Drug Administration's Center for Devices and Radiological Health and (b) the Conference of Radiation Control Program Directors (CRCPD), whose voting members are directors of state and local radiation control programs (website: The measurements are made at each facility on a dummy of standard size.

        o  The CRCPD publishes a summary of the survey results. The reports do not identify where the measurements were made.

        o  The most recent observations published by the CRCPD are the 1994 Chest Radiography Survey, 1993 Dental Survey, 1992 Mammography Survey, 1990 CT Survey (head exam only) and 1991 Fluoroscopy Survey (upper gastro-intestinal exam only), 1989 Abdomen and Lumbar Sacral Spine Survey in Private Practice and Chiropractor's Offices.

        o  If the American College of Radiology, the Radiological Society of North America, or any other group has done more comprehensive and recent dose-surveys than the CRCPD, we hope that such groups will open the full results to non-members.

        o  In 1989, the National Council on Radiation Protection and Measurements (website: published an annotated table entitled "Procedures to Reduce Collective Dose Equivalent in Diagnostic X-Ray Examinations" (NCRP 1989, p.37). From the table, it is clear that x-ray practitioners who use all the procedures can give doses which are 10 or 20 times lower than practitioners who use none of them. The table is reproduced in Gofman 1999, Chap. 1 (available online at

        o  Joel E. Gray, Ph.D. --- using the CRCPD data plus his own vast experience as a professor and medical physicist at the Mayo Clinic (and currently as a private consultant) --- has published two articles which further demolish the assumption that doses received during x-ray imaging are already as low as technically possible (Gray 1998a+b).

        o  Helpfully, Dr. Gray's articles also describe "relatively simple, inexpensive, and easily applied" techniques which, combined, can reduce dose in various circumstances to one-fifth the amount given if such measures are not used. Unfortunately, the two articles are very hard to obtain, and they probably have been seen by almost no one in medicine and certainly not by the public.

# # # # #

The References

This document is also available in plain HTML and ASCII text formats

and can always be found at

back to XHP | CNR | radiation | rat haus | Index | Search | tree