| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, University of North Carolina, CB#7510, 503 Old
Infirmary, Chapel Hill, NC 27599-7510.
2
Lineberger Comprehensive Cancer Center 237, University of North Carolina,
CB#7295, Chapel Hill, NC 27599-7295.
3
Department of Environment and Natural Resources, North Carolina Department of
Radiation Protection, 3825 Barrett Dr., Raleigh, NC 27609-7221.
4
Department of Radiologic Sciences, University of North Carolina School of
Medical Wing E22, CB 7130, Chapel Hill, NC 27599-7130.
5
North Carolina Department of Health and Human Services, Division of Community
Health, 1330 St. Mary's St., P. O. Box 29605, Raleigh, NC 27626-0605.
Received June 1, 1999;
accepted after revision August 31, 1999.
Supported in part by the North Carolina Breast and Cervical Cancer Control
Program, as part of a grant from the Department of Environment and Natural
Resources, State of North Carolina, which is funded by the United States
Centers for Disease Control and Prevention.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. All mammography facilities were inspected during 1993-1994, and again in 1995. Both inspections evaluated mean glandular radiation dose, phantom image evaluation, darkroom fog, and developer temperature. Two mammography health specialists employed by the North Carolina Division of Radiation Protection performed all inspections and collected and codified data.
RESULTS. The percentage of facilities that met quality standards increased from the first inspection to the second inspection. Phantom scores passing rate was 31.6% versus 78.2%; darkroom fog passing rate was 74.3% versus 88.5%; and temperature difference passing rate was 62.4% versus 86.9%.
CONCLUSION. In 1995, the first year that the MQSA was in effect, there was a significant improvement in the quality of mammography in North Carolina. This improvement probably resulted from facilities' compliance with federal regulations.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The implementation of the MQSA has not been effortless. The costs of annual on-site inspections of all mammography facilities has been criticized as being excessive compared with the potential benefits of such a program [4]. The inspection cost for facilities with one mammography unit is $1178 with $152 charged for each additional unit. A study by Farria et al. [5] estimated that personnel at facilities with two mammography units spend approximately 130 hr/year on federally required activities. Furthermore, some experts have questioned the wisdom of allowing the federal government to promulgate regulations for mammography when previous ACR voluntary standards had been successful.
The FDA has made a commitment to improving the inspection and compliance program by developing an extensive program for state inspector training [6, 7]. The General Accounting Office published three reports on the impact of the MQSA, including two mandatory reviews in 1995 and 1997 [8,9,10]. The first report, published in 1995, stated that federal regulations had a positive effect on mammography quality. The positive effect was supported by ACR data showing that between October 1, 1994 and August 1, 1995 (the time between mandated compliance with the new law and the report), a total of 7525 mammography units from 5510 facilities underwent ACR accreditation review for the first time, with a 35% failure rate [8]. Approximately two thirds of the units failing accreditation on the first attempt became accredited with a minimum of two additional attempts. These data suggest that many facilities that had previously not applied for ACR mammography accreditation were in compliance with Mammography Accreditation Program standards. In addition, 97% of the facilities that closed because of noncompliance with FDA regulations were within 25 miles of a certified facility, and 62% were within 1 mile of such a facility. Therefore, the General Accounting Office reported that there was no significant impact on access to mammography services caused by closed facilities, even in states in which more than 10% of the facilities were closed.
The second General Accounting Office report, published in early 1997, provided inspection results for the first 18 months of the MQSA inspection program [9]. At the time of the report, 1503 facilities had been inspected twice. Although 26% of the facilities had significant violations detected during the first inspection, only 10% had violations during the second inspection, suggesting increased compliance with the MQSA. However, the General Accounting Office questioned whether phantom image evaluation had been adequately standardized and whether sufficient mechanisms existed to ensure that noncompliance problems were corrected.
The third General Accounting Office report, published in October 1997, compared the results of the MQSA inspections with earlier federal survey data [10]. This comparison showed an improvement in phantom scores, with 89% of facilities passing in 1992 compared with 98% passing in 1995. Also, the report revealed the correction of many earlier inspection problems. These trends are confirmed in a recent report, by the administrators of the FDA program, that summarizes the MQSA effects [11].
We compared mammography inspection data collected in North Carolina for two inspections, one in 1993-1994 (before FDA standards took effect) and one in 1995 (under MQSA, after standards were established).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
During the 1993-1994 inspection, performed before FDA regulations were implemented, inspections were unannounced, or were preceded at most by 2-days notice. During the 1995 inspection, under MQSA and FDA regulations, inspector training increased. Both mammography health specialists participated in three 2-week courses. Additionally, facilities were given at least 5-days advance notice before the 1995 inspection.
During both inspections, measurements of breast entrance exposure, half-value layer, average glandular radiation dose, process developer temperature difference, phantom scores, and darkroom fog were obtained according to a previously published methodology [12]. Because these were the only quantifiable data obtained during both inspection cycles, these were examined for this study.
For measurements of some of the technical quality tests, data on a complete census of North Carolina facilities were collected by the inspectors. With complete census data, no statistical analysis was needed to prove that the percentages of facilities that passed inspection were statistically different because the sample and population were the same for the two censuses. Simple comparison of the values was sufficient. However, because some data from North Carolina mammography facilities were never collected, statistical analysis was required to judge the applicability of data to the whole sample. For the phantom scores and darkroom fog pass rate comparisons, only a small quantity of data were not collected during the first inspection. To underestimate the value of the MQSA inspection as much as possible and to determine whether the increase in the percentage of passing units in 1995 was statistically significant, we assumed that all data from units without data recorded would be passing data. For temperature difference, a more substantial amount of data were never collected by inspectors; therefore, a more formal statistical comparison based on a finite population of units was needed. The estimated finite population variance for a proportion is [p (1 - p) / n](1 - n / N), where p, which equals n / N, is the sample proportion, n is the sample size, and N is the number of units [13].
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Table 1 summarizes the percentage of mammography units that met limits for quality control assessment measurements. During the second inspection, the pass rates for phantom scores and darkroom fog increased and were statistically significant, indicating that even if the data that were not collected came from facilities with acceptable scores, there would be an improvement over time.
|
For temperature difference, the pass rates for the first and second inspections were 62.4% (95% confidence interval, 57.0-67.8%) and 86.9% (95% confidence interval, 84.5-89.3%), respectively. We noted a significant improvement in the pass rate for processor temperature difference. The glandular dose pass rates were almost identical: 99.2% for 1993-1994 and 99.8% for 1995. However, the radiation dose level increased for 120 (73%) of 164 mammography units between the first and second inspections. Although almost all units passed both inspections, the mean glandular dose increased from 147 to 168 mrad.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The improvement in phantom scores could have been caused by the increase in advance notice before inspection. Before the implementation of the MQSA, facilities were informed a maximum of 2 days before inspection. After the implementation of the MQSA, facilities were informed at least 5 days before inspection. We believe that any quality tune-up before inspection was highly unlikely. Although it might be possible for a site to adjust one or two parameters for a short period before inspection, it is highly unlikely that a facility that was performing poor quality mammography would have the expertise to make major adjustments that would allow it to pass such a thorough inspection with only 5 days notice. The overall improvement suggests that facilities implemented new quality control practices.
Information obtained from the handwritten notations of inspectors and during educational site visits to North Carolina mammography facilities [14] suggests that many of the failing facilities were processing their images incorrectly (i.e., developer time and temperature were inappropriately elevated during the first inspection). Incorrect processing resulted in noisy images with increased image fog. The increase in mean glandular dose between the two inspections and the notations of inspectors suggest that many facilities switched to slower higher contrast film-screen systems or improved film processing before the second inspection. These changes resulted in images with less image fog and noise. The increased mean glandular dose, occurring simultaneously with an increased pass rate, strongly indicates improvement in image quality across all sites.
For the phantom score, both the total score and the score per object improved between the two inspections. Although some of this may be attributable to changes in the system of phantom scoring, the fact that the greatest change occurred in mass score suggests that most of the difference was not caused by this factor. In other words, the types of artifacts seen in mammography phantoms rarely resemble masses. Therefore, a change in scoring that deducts less per artifact would have the least effect on mass score. In addition, the mass score should be very sensitive to image contrast and poor quality processing. This suggests that the facilities adapted to federal regulations by changing to screen-film systems with improved contrast or by improving their film-processing standards.
In summary, we compared the quality of mammography facilities before and after the implementation of the MQSA. We determined that major improvements in mammography quality occurred in North Carolina facilities between the first and second inspections. Although we cannot formally attribute the mammography improvement to the MQSA, the implementation of federal standards probably improved image quality in many facilities. Improvements in mammography quality started with the ACR Mammography Accreditation Program and earlier inspection programs. Now, regulations extend to all facilities, enabling women in North Carolina to access high-quality mammography. Improved image quality will ultimately result in earlier detection and improved breast cancer outcomes for women living in North Carolina; however, further studies are required before this link can be established [10].
Of course, improved image quality alone will not improve patient care. Improved breast cancer diagnosis and reduction of breast cancer mortality will require improved image quality, improved interpretive performance by radiologists, and improved communication between physicians and patients.
Acknowledgments
We thank Ricky McGee, Anna Cleveland, Vickie Jernigan, Keith Muller,
Christopher Coffee, Bonnie Yankaskas, Joy Roberts, Kristen Harris, Jenny
Harrison, and Mary Brown for their assistance on this project.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K.-H. Ng, N. Jamal, and L. DeWerd Global quality control perspective for the physical and technical aspects of screen-film mammography--image quality and radiation dose Radiat Prot Dosimetry, December 1, 2006; 121(4): 445 - 451. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. Taplin, C. M. Rutter, and C. D. Lehman Testing the effect of computer-assisted detection on interpretive performance in screening mammography. Am. J. Roentgenol., December 1, 2006; 187(6): 1475 - 1482. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Baker and M. S. Soo Breast US: Assessment of Technical Quality and Image Interpretation Radiology, April 1, 2002; 223(1): 229 - 238. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
