The New Jersey Radiographic Quality Assurance Program at 5 Years

Transcription

1 The New Jersey Radiographic Quality Assurance Program at 5 Years Julie Timins, MD a, Paul Orlando, BS b, Jill Lipoti, PhD b Purpose: Five years ago, the New Jersey Bureau of Radiological Health decided to develop and implement a proactive, outcome-based quality assurance and quality control program. The specific goals were reducing unnecessary patient radiation exposure and improving image quality (IQ). Methods: From inception, input was obtained from the regulated community, which included all medical facilities providing diagnostic radiographic services. The program monitored entrance skin exposure (ESE) and quantitative IQ for 3 radiographic examinations: posteroanterior chest, anteroposterior lumbar spine, and anteroposterior foot. Results: Within 1 year, the mean ESE for postanterior chest x-rays had dropped by 39%, the ESE for anteroposterior lumbar spine films had decreased by 18%, and the ESE for anteroposterior foot x-rays had declined by 48%. After the second year, the decreases in ESE were 44%, 29%, and 61%, respectively. There was a corresponding improvement in quantitative IQ. Conclusion: Reductions in mean patient radiation exposure can be accomplished for specific radiographic examinations while achieving significant improvement in quantitative IQ. This requires the education of and cooperation with the regulated community. Details of the program and the response of the regulated community are presented. Key Words: Regulation of radiography, radiation exposure, radiation dose, radiographic quality assurance, radiographic quality control, radiographic image quality J Am Coll Radiol 2007;4: Copyright 2007 American College of Radiology INTRODUCTION In the late 1990s, third-party payers were looking for means to limit reimbursement for radiographic procedures performed in physicians offices. One method was to require quality assurance (QA) programs applicable to all providers of imaging services, radiologists and nonradiologists alike. Each insurance company had different QA criteria, making compliance cumbersome. Medical practitioners in New Jersey turned to the state for a uniform QA program for imaging facilities. Concurrently, the New Jersey Bureau of Radiological Health was reviewing its inspection program for radiologic equipment and came to the conclusion that the program was not having a sufficient impact on reducing radiation exposure to the public. The bureau s inspections focused on x-ray equipment functionality, such as peak kilovoltage, timer accuracy, and collimators, rather than on outcome goals such as image quality (IQ) and entrance skin exposure (ESE). If equipment passed inspection, no additional effort was made to cut radiation exposure or improve IQ. The Mammography Quality Standards Act of 1992 [1] had demonstrated that a constructive regulatory program could uniformly improve the quality of mammography while decreasing radiation dose. The bureau decided to develop a proactive, outcome-based QA and quality control (QC) program, with the specific goals of reducing unnecessary patient radiation exposure and improving IQ. The details of the program were carefully crafted and included an initial phase to obtain baseline information on patient radiation exposure and IQ, followed by an implementation phase. A database format was devised to record the findings of QA inspections. The state regulation, titled Quality Assurance Programs for Medical Diagnostic X-Ray Installations [2], was promulgated on January 12, a Christ Hospital, Jersey City, New Jersey. b New Jersey Department of Environmental Protection, Trenton, New Jersey. Corresponding author and reprints: Julie Timins, MD, 20 Foote s Lane, Morristown, NJ 07960; jtmns@verizon.net American College of Radiology /07/$32.00 DOI /j.jacr METHODS All facilities with diagnostic radiographic equipment were subject to the QA and QC program, both hospitals 691

2 692 Journal of the American College of Radiology/ Vol. 4 No. 10 October 2007 Table 1. Entrance skin exposures in mr for PA chest, AP lumbar (LS) spine, and AP foot radiographs Study/Action Level Low Average High Extremely High Chest Less than 5 5 to to 30 Greater than 30 LS spine Less than to to 600 Greater than 600 Foot Less than 5 5 to to 40 Greater than 40 and outpatient facilities, including the offices of physicians, chiropractors, and podiatrists. X-ray modalities surveyed included radiography, fluoroscopy, x-ray bone densitometry, and computed tomography. Each facility was required to have a QA program, which was audited during inspections. Parameters to be followed included patient radiation exposure, periodic QC examinations, processor functionality, and IQ. An annual review and analysis by a qualified medical physicist was also required. The initial radiographic examinations chosen for evaluation were posteroanterior (PA) chest, anteroposterior (AP) lumbar spine, and AP foot. Specific measurable goals were established: to decrease radiation dose in terms of ESE by and to improve IQ by 15% within 2 years. Metrics were developed to quantify these goals. This included developing protocols on how to measure and report ESE and developing a system for quantifying IQ. Radiation exposure was classified as low, average, high, or extremely high for chest, lumbar spine, and foot radiographs (Table 1). With the assistance of a private contractor, a customized image phantom was developed to evaluate and quantitatively score IQ in the following 7 areas: background density (25%), low contrast resolution (), low contrast detail (), high contrast resolution (20%), noise or artifacts (20%), film contrast (5%), and density uniformity () (Figure 1). Quantitative IQ scores were broken down into categories of poor, fair, good, and excellent (Table 2). The minimum acceptable quality was fair, requiring that the submitted phantom images meet the following criteria: background density: 0.8 to 1.8 optical density; high contrast resolution 2.0 line pairs/mm; noise or artifacts: moderate or less; density uniformity: moderate or better; low contrast resolution: at least 3 of 8 holes; low contrast detail: at least 2 pairs of holes; and film contrast (step wedge): 0 steps. The step wedge contributed 5% to the IQ score but was not a cause of performance failure, because it was satisfactory in almost all cases. The automatic processing of film was not required. Hand processing was acceptable if the appropriate QC was performed. Quality control tests for manual and automatic film processing were detailed. Daily sensitometry and densitometry measurements were required to ensure appropriate film density levels. If the established limits were not met, corrective action was required before processing patient radiographs. A schedule of QC tests at various time intervals was specified, including the documentation of acceptable processor operating temperature, proper chemical solution levels and freshness, fixer retention levels, evaluation of darkroom cleanliness, and checks for darkroom light leaks. Facilities inspected under the QA program were sent postinspection letters with their measured ESEs depicted on a graph of ESEs of all facilities performing the same radiographic studies (Appendix A). Image quality scores were also reported, including individual scores in 6 components that contribute to the overall score (Appendix B). Facilities with high ESEs or poor IQ scores were instructed to consult with their medical physicists. They were given 30 days to find and correct their problems and to send reports of their findings and corrective actions to the Bureau of Radiological Health. During the development and implementation of the QA and QC regulations, meetings were held with stakeholders, including medical physicists, physicians and Fig. 1. New Jersey Bureau of Radiological Health image quality phantom. Table 2. Image quality scale Poor Fair Good Excellent

3 Timins et al/new Jersey Radiographic Quality Assurance Program Image Quality Average Image Quality Average ESE (mr) N.E.X.T. Avg ESE= 14.2 mr ESE Average IQ Score ESE (mr) N.E.X.T. Avg ESE= 388 mr ESE Average IQ Score Baseline yr 1 yr 2 yr 3 yr 4 yr Baseline yr 1 yr 2 yr 3 yr 4 yr 5 0 Fig. 2. Chest enhance skin exposure and image quality. NEXT refers to the average ESE determined in the Nationwide Evaluation of X-Ray Trends 2001 survey, which was on chest radiography. their staffs, radiologic technologists, chiropractors, and podiatrists. Twenty-one workshops were held throughout the state of New Jersey to assist medical facilities in developing their own QA and QC programs. Detailed guidance documents, titled Quality Assurance Manual [3] and Radiographic Quality Control [4], were composed and distributed by mail and online, explaining the components of the QA program, with detailed descriptions of how to perform the QC tests. Subsequent guidance documents were developed for fluoroscopy and computed tomography: Fluoroscopic Quality Control [5] and CT Quality Control [6]. Baseline data were obtained from June 1, 2000, through February 15, The first year of the program was February 16, 2001 through September 11, 2002; the bureau did not close year 1 until all facilities had their initial QA inspections. For data analysis, subsequent years were calendar years, starting on September 12. Half of the regulated community was reinspected in year 2 and the other half in year 3. A smaller number of facilities were inspected in year 4 than in preceding and Fig. 4. Lumbar spine enhance skin exposure and image quality. NEXT refers to the average ESE determined in the Nationwide Evaluation of X-Ray Trends 2002 survey, which was on lumbar spine radiography. subsequent years. Year 5 ended on September 11, Five QA years are presented here. RESULTS The radiation exposure for PA chest, AP foot, and AP lumbar spine radiographs dropped significantly during the study period. At baseline, the mean ESE for PA chest examinations was 22.2 mr. This dropped by 39% to 13.5 mr in the first year and by 44% to 12.5 mr in the second year (Figure 2). The mean radiation exposure from an AP foot x-ray at baseline was 31.1 mr, which dropped to 17.2 mr (a 48% decrease) and 12.8 mr (a 61% decrease), respectively, in years 1 and 2 of the program (Figure 3). Mean AP lumbar spine ESE was mr at baseline, decreasing to mr (down 18%) in year 1 and mr (a drop of 29%) in year 2 (Figure 4). Further reductions were observed in years 3 and 4 in all imaging modalities. Slight increases in radiation exposure were observed in all modalities in year 5. In the first year of the QA program, dose reductions exceeded the set Image Quality Average ESE (mr) ESE Average IQ Score Baseline yr 1 yr 2 yr 3 yr 4 yr 5 0 Fig. 3. Foot enhance skin exposure and image quality. Fig. 5. Image quality distribution by year.

4 694 Journal of the American College of Radiology/ Vol. 4 No. 10 October 2007 Table 3. Distribution of image quality scores Number of Data Points All Poor Studies Good or Exc Chest Poor Studies Good or Exc Foot Poor Studies Good or Exc Lumbar Spine Poor Studies Good or Exc Year 1 1,762 58% 5% 74% 7% 62% 14% 49% Year 2 2,328 2% 84% 2% 85% 1% 89% 2% 83% Year 3 2,555 2% 87% 2% 84% 1% 90% 2% 86% Year % 89% 1% 93% 0% 93% 1% 86% Year 5 1,672 1% 88% 0% 92% 0% 93% 2% 84% Note: Exc Excellent; # of Data Points # of Facilities. goal of a reduction in ESE for all examination types. After 5 years, decreases of 36% for PA chest, 65% for AP foot, and 33% for AP lumbar spine radiographic ESE have been observed. Because of the time involved in developing and refining the phantom and IQ scoring systems, the baseline IQ scores were obtained in year 1 of the QA and QC program. Ninety percent of facilities had acceptable IQ, and had failing IQ scores at baseline. The following year, only 2% of facilities had poor IQ. The average IQ score in year 1 was 51.3 and improved by increasing to 60.6 in year 2 (Figure 5). This 18% increase in average IQ scores after 1 year exceeded the original goal of a 15% improvement in IQ in 2 years. In year 1, 58% of facilities produced images rated as good or excellent; a year later, this increased to 84%. Years 3 and 4 saw continued improvement. In year 5, the number of facilities with poor IQ was less than 1% of those evaluated (Table 3). DISCUSSION The rapid positive response of the regulated community was striking. Within 1 year, the goals of decreasing patient radiation exposure and improving IQ by 15% were more than met, with further improvements in years 2 to 4. The findings were so impressive that a preliminary report at the November 2003 meeting of the Radiological Society of North America promoted the Table 4. New Jersey facilities subject to quality assurance regulations Facility Type Nov 2003 March 2005 Dec 2006 Hospitals Physician Medical Offices 1,494 1,295 1,343 Chiropractors 1, Podiatrists Industrial, Schools, Gov t Total 3,542 2,683 2,664 New Jersey QA and QC program as a model for other states to follow [7]. Compliance with the QA regulations was perceived as burdensome by some of the regulated community. The bureau observed a 24% decline in the number of facilities with medical radiographic equipment from November 2003 to March 2005 (Table 4). The percentage decrease in facilities performing radiography varied by type of medical practice: physician offices were down 13%, chiropractic offices down 34%, and podiatric offices down 33%. Since March 2005, the number of physician and podiatric facilities providing radiographic services has stabilized because of adaptation to the record keeping and periodic QC tests required by the QA program. However, the number of chiropractic facilities with x-ray equipment has dropped an additional 8%. Interviews with chiropractors indicated that many felt that the income from their relatively small volume of x-rays did not adequately compensate for the added burden of compliance with the QA program; instead, many chiropractors opted to refer their patients to outside imaging facilities. Meanwhile, physician-run imaging facilities increased by 4% over the same time period. Although individual facilities had varying opinions on the necessity and benefits of the QA regulations, the professional organizations representing the regulated community (the Medical Society of New Jersey, the Radiological Society of New Jersey, the Chiropractic Forum, the Podiatric Society, the Health Physics Society, the American Association of Physicists in Medicine, the New Jersey Hospital Association, etc) were supportive of the QA program and participated in both the rule development process and the implementation outreach efforts conducted throughout the state. The New Jersey medical physicist and health physicist communities were particularly helpful in identifying pertinent QC tests to be performed, establishing meaningful standards for QC tests, and assisting in the evaluation and development of the bureau s IQ phantom. The US Food and Drug Administration and the Conference of Radiation Control Program Directors cooper-

5 Timins et al/new Jersey Radiographic Quality Assurance Program % 50% Percent of Facilities 90% 80% 70% 60% 50% 40% 30% 20% 0% 58% 84% 2% 87% Facilities with Good or Excellent Image Quality Facilities with Poor Image Quality 89% 88% N Values Yr 1= 1762 Yr 2= 2328 Yr 3= 2555 Yr 4= 773 Yr 5= % 1% 1% Year 1 Year 2 Year 3 Year 4 Year 5 45% 40% 35% 30% 25% 20% 15% 5% Mean IQ Yr Yr Yr Yr Yr Yr 1 Yr 2 Yr 3 Yr 4 Yr 5 Fig. 6. Image quality distribution, all studies combined. 0% atively conduct periodic national surveys of radiation dose and IQ for several commonly performed radiologic examinations. These surveys are published as the Nationwide Evaluation of X-Ray Trends (NEXT). Data are solicited from hospital and outpatient facilities in all states and are analyzed to determine regional trends. Previously, exposure levels in New Jersey were substantially higher than the NEXT mean; now they are lower. The 1995 NEXT survey for AP lumbar spine radiography found a national mean ESE of 368 mr [8].Atthe beginning of the QA and QC study, the New Jersey baseline ESE for a lumbar spine radiograph was 525 mr. The 2002 NEXT lumbar spine survey found that mean national exposure had increased to 388 mr [9]. In contrast, after 5 years of the QA and QC program, the mean in New Jersey has decreased to 350 mr (Figure 4). In 2001, the NEXT survey on chest radiography found a national mean ESE of 14.2 mr for a PA chest radiograph, which was down from 16.0 mr in the 1994 NEXT survey [10,11]. At baseline, the New Jersey QA and QC program reported a mean ESE of 22.2 mr. By year 4 of the program, the mean PA chest exposure had decreased to 11.9 mr, but it increased somewhat in year Fig. 8. Foot image quality score distribution. 5 to 14.1 mr, just below the 2001 NEXT average of 14.2 mr (Figure 2). This is still a substantial improvement from the baseline at the start of the QA program. The New Jersey QA program measures IQ in addition to radiation exposure, thus determining whether reductions in ESE are clinically valid and result in high-quality diagnostic images. If the QA program documented reductions in radiation exposure but revealed a corresponding reduction in IQ, the radiation reduction could be interpreted as counterproductive to quality medical care. Average IQ for all 3 radiographic procedures combined improved in years 2 through 4, compared with year 1, and average IQ scores for year 5 were comparable with those for year 3. During all comparison years, the distribution of IQ scores was in a tight range of good to excellent, with a small percentage (2% or less) having poor IQ (Table 3, Figure 5). Comparing IQ scores by type of radiograph resulted in similar IQ improvements, with the majority of facilities obtaining good to excellent IQ (Figures 6 to 9). 50% 45% 45% Mean IQ Yr 1 40% Mean IQ 40% 35% 30% 25% 20% Yr Yr Yr Yr Yr Yr 2 Yr 3 Yr 4 Y4 5 35% 30% 25% 20% Yr Yr Yr Yr Yr Yr 1 Yr 2 Yr 3 Yr 4 Y4 5 15% 15% 5% 5% 0% % Fig. 7. Chest image quality score distribution. Fig. 9. Lumbar spine image quality score distribution.

6 696 Journal of the American College of Radiology/ Vol. 4 No. 10 October 2007 SUMMARY The New Jersey QA and QC program has been a resounding success, reducing mean patient radiation exposure by 36%, 65%, and 33%, respectively, for PA chest, AP foot, and AP lumbar spine radiographs. Image quality has improved quantitatively by 21% for these examinations. The regulated community continues to be monitored for compliance. The program has been expanded to include fluoroscopy and computed tomography. Solutions to standardizing methods for calculating and reporting computed tomographic radiation dose had to be worked out. Monitoring computed tomographic dose is especially important because, although computed tomography accounts for a relatively small percentage of radiologic procedures (recently reported at 16%, but increasing progressively), it contributes at least 65% of patient radiation dose. Also, there is the potential for increased radiation dose with multidetector CT equipment. Radiation exposure from dental radiography is now being tracked, with dentists receiving similar recommendations from the Bureau of Radiological Health on how to reduce patient radiation exposure. Regulations to protect people performing veterinary radiography are also being developed. REFERENCES 1. Mammography Quality Standards Act of 1992, 42 USC 263b (1992). 2. Quality assurance programs for medical diagnostic x-ray installations, New Jersey Administrative Code 7:28-22 (2001). 3. New Jersey Bureau of Radiological Health. Quality assurance manual. Trenton: New Jersey Bureau of Radiological Health; New Jersey Bureau of Radiological Health. Radiographic quality control. Trenton: New Jersey Bureau of Radiological Health; New Jersey Bureau of Radiological Health. Fluoroscopic quality control. Trenton: New Jersey Bureau of Radiological Health; New Jersey Bureau of Radiological Health. CT quality control. Trenton: New Jersey Bureau of Radiological Health; Orrison W, Lipoti J. Benefits of quality assurance for diagnostic x-ray. Presented at: Annual meeting of the Radiological Society of North America; November Nationwide Evaluation of X-Ray Trends (NEXT) 1995 abdomen and LS spine x-ray data. CRCPD Publication 00-2 (revised 2001). Frankfort, Ken: Conference of Radiation Control Program Directors; Nationwide Evaluation of X-Ray Trends (NEXT) tabulation and graphical summary of 2002 abdomen/lumbosacral spine. CRCPD Publication E-06-2b. Frankfort, Ken: Conference of Radiation Control Program Directors; Nationwide Evaluation of X-Ray Trends (NEXT) tabulation and graphical summary of 2001 survey of adult chest radiography. CRCPD Publication E Frankfort, Ken: Conference of Radiation Control Program Directors; Nationwide Evaluation of X-Ray Trends (NEXT) 1994 P/A chest x-ray data. CRCPD Publication Frankfort, Ken: Conference of Radiation Control Program Directors; 1998.

7 Timins et al/new Jersey Radiographic Quality Assurance Program 697 APPENDIX A Note: AP anteroposterior; BRH New Jersey Bureau of Radiological Health.

8 698 Journal of the American College of Radiology/ Vol. 4 No. 10 October 2007 APPENDIX B