|Home | About | Journals | Submit | Contact Us | Français|
To determine population-based rates of use of diagnostic imaging procedures with ionizing radiation in children, stratified by age and gender.
Retrospective cohort analysis.
All settings utilizing imaging procedures with ionizing radiation.
Individuals less than 18 years old, alive and continuously enrolled in Unitedhealthcare between January 1, 2005 and December 31, 2007 in 5 large U.S. healthcare markets.
Number and type of diagnostic imaging procedures utilizing ionizing radiation in children.
355,088 children were identified. A total of 436,711 imaging procedures using ionizing radiation were performed in 150,930 (42.5%) patients. The highest rates of use were in children greater than 10 years old, with frequent use in infants under 2 years old as well. Plain radiography accounted for nearly 85% of imaging procedures performed. Computed tomography (CT) scans – associated with substantially higher doses of radiation – were commonly used, accounting for 12% of all procedures during the study period. Overall, 7.9% of children received at least one CT and 3.5% received 2 or more, with CT of the head most frequent.
Exposure to ionizing radiation from medical diagnostic imaging procedures may occur frequently among children. Efforts to optimize and ensure appropriate use of these procedures in the pediatric population should be encouraged.
The rapid growth in use of medical diagnostic imaging procedures such as computed tomography (CT) has led to widespread concern about low-dose ionizing radiation exposure in adults.1–5 Despite widespread discussions about similar hazards for younger patients, contemporary data on the use of such imaging procedures in children are more limited. Infants and children are at higher risk for future malignancy as compared to adults because their developing tissues are more sensitive to radiation and their longer expected lifespans allow additional time for the emergence of detrimental effects.6–10
Accordingly, we set out to examine patterns of use among children of diagnostic imaging procedures that utilize ionizing radiation. For this report, we used comprehensive inpatient and outpatient claims data sources from UnitedHealthcare, a large healthcare organization that administers benefits to millions of families across the United States. Because methods are controversial in children for accurately quantifying the effective dose of ionizing radiation exposure11 (i.e., its detrimental biological effect), we focused primarily on describing the number and types of these procedures being used. As such, our goals were to: (1) determine overall population-based rates of use of imaging procedures with ionizing radiation in individuals less than 18 years old; (2) explore their use across age and gender groups; and (3) identify the most frequently used procedures.
We conducted an investigator-initiated, retrospective cohort study using exhaustive inpatient and outpatient claims data from UnitedHealthcare. These data were collected between January 1, 2005 and December 31, 2007 from 5 large regional markets: Arizona; Dallas; Orlando, Florida; South Florida; and Wisconsin. These markets were specifically selected because of their size, the stability of their enrollment population, and the similarity of their insurance products, as well as to provide a degree of geographic diversity. The study population included all individuals: 1) less than 18 years old at the beginning of the study period and 2) alive and continuously enrolled in one of the programs administered by UnitedHealthcare during the study period. To obtain a true denominator population, we included all individuals who were continuously enrolled regardless of whether or not they submitted a claim during the study period. After removing all personal identifiers, data were provided to the investigators for independent analysis and interpretation. The Institutional Review Board of the University of Michigan approved this study protocol and waived the requirement for informed consent.
All claims from hospitals, outpatient facilities and physician offices that were submitted during the study period were queried for Current Procedural Terminology (CPT) codes that identified imaging procedures utilizing radiation exposure (under Radiology Schedule – Diagnostic Imaging and Nuclear Medicine: 70010 through 76499 and 78000 through 79999; and under Medicine Schedule – Cardiovascular and Non-invasive Vascular Diagnostic Studies: 92950 through 93799 and 93875 through 94005).12 Procedures were included regardless of whether they were performed for diagnostic or therapeutic indications (e.g., interventional radiological procedures). However, claims related to the specific delivery of radiation for a therapeutic purpose and not for imaging (e.g., total body irradiation prior to stem cell transplant) were excluded. In cases where the CPT code for a procedure changed during the study period, all versions of the code were included.
We obtained information from each claim on: (1) age, (2) gender, (3) the market where the service was performed, and (4) the location of service (hospital inpatient, hospital outpatient, and physician office). We categorized these procedures into mutually exclusive categories based on the technology utilized (plain radiography, CT, fluoroscopy and/or angiography, and nuclear medicine scans) and anatomic area of focus (chest [including cardiac imaging], abdomen, pelvis, extremity, head and neck [including brain imaging], multiple areas [including whole body scans], and non-specified). To be as conservative as possible and to avoid the potential of over-estimating the number of procedures from duplicate claims, we limited individuals to 1 procedure per day for the same type of technology (e.g., CT) performed on the same anatomic area (e.g., chest).
We focused on describing the number and types of imaging procedures performed in the study population using simple descriptive statistics. Specifically, we calculated population-based rates of use where the numerators were the cumulative number of imaging procedures performed in an individual and denominators were the total number of eligible children enrolled throughout the study period. For these analyses, children were categorized based on their age at the beginning of the study period (0 to <2, 2 to <5, 5 to <10, 10 to <15, and 15 to <18 years old) and gender. Imaging procedures were then categorized by the type of technology used. All statistical analyses were carried out with the use of SAS (Version 9.2, SAS Institute, NC).
We identified 355,088 children continuously enrolled in a program administered by UnitedHealthcare during the study period. The mean age was 9.0 years (std dev, 4.9 years) and 181,795 (51.2%) were boys. The largest proportion of the study population was located in the Dallas market area (124,079 [34.9%]), while the smallest proportion came from the Orlando market area (45,466 [12.8%]). Overall, the percentage of subjects who underwent at least 1 diagnostic imaging procedure utilizing ionizing radiation ranged from 39.4% in the Orlando market area to 43.2% in the Dallas market area.
During the 3-year study period, a total of 436,711 of these imaging procedures were performed in 150,930 (42.5%) children (Table 1), resulting in an annual utilization rate of 410 procedures per 1000 children. During the study period, 89,618 (25.2%) children underwent 2 or more of these procedures while 56,754 (16.0%) underwent 3 or more. The highest annual rates of use were generally in children 10 years and older. Use of these procedures also was generally higher among boys (80,638 [44.4%] versus 70,292 [40.6%] for girls; P<0.001). Although the overall proportion of patients was smaller, similar patterns of use across age and gender groups were observed in those children who underwent at least 2 imaging procedures and those who underwent 3 or more.
Tables 2a–2d show patterns of procedure use stratified by the type of imaging procedure across age and gender groups. Rates varied substantially based on the types of procedures with 141,480 (39.8%) children receiving at least 1 plain radiography examination, 28,107 (7.9%) receiving at least 1 CT, 7492 (2.1%) receiving at least 1 fluoroscopic or angiographic procedure, and 2607 (0.7%) receiving at least 1 nuclear medicine scan. Plain radiography also was the most commonly repeated of these procedures: 79,209 (22.3%) children received 2 or more studies. Similarly, CT scans were frequently repeated, with 12,494 (3.5%) children receiving 2 or more studies. While use of CT was more frequent overall in boys, use in the 15–17 year age group was higher in girls (93 per 1000 person-years versus 85 per 1000 person-years.) Finally, overall patterns of plain radiography and nuclear medicine scans followed those described above for all imaging procedures, with use most frequent in infants under 2 years old and children 10 years and older. However, CT scans were used less frequently in young children. In contrast, the largest proportion of children undergoing fluoroscopic and/or angiographic studies was under the age of 2 years old – especially among infant girls.
Overall, plain radiography accounted for nearly 85% of the studies that were performed CT scans were the next most commonly used modality, accounting for 12% of imaging procedures, followed by fluoroscopy and/or angiography studies (2%) and then nuclear medicine scans (1%). However, this pattern of use varied across different age categories . For example, CT scans were used with a much higher frequency in older age groups, rising to nearly 18% of all imaging studies performed in children aged 15 to <18 years old.
Finally, Table 3 shows specific data on the 10 most frequently-used studies for the 4 different categories of imaging procedures. Chest radiography was the most common procedure performed overall at an annual rate of 68.2 procedures per 1000 children. This was followed by plain radiography of extremity areas, the spine and the abdomen. By far, the highest annual rate of use for CT scans involved studies of the head, followed by the abdomen and pelvis. CT scans of the abdomen and pelvis, in particular, increased dramatically with age, rising from 5.1 procedures performed per 1000 children annually in 0 to 1 year old patients to 40.9 in 15 to <18 year old patients. The use of fluoroscopic and/or angiographic studies and nuclear medicine scans were infrequent overall.
To our knowledge, we report the first large, population-based study examining the use of diagnostic imaging procedures that utilize low-dose ionizing radiation specifically in a pediatric population. Among 355,088 children across 5 large healthcare markets in the U.S., we found that use of these procedures during a 3-year study period was frequent, with at least 1 of these procedures being performed in over 40% of children. Importantly, many children underwent more than one procedure. Based on these data, the average child in this study population would have received over 7 procedures by the time they reach age 18. While plain radiography – which is associated with much lower levels of radiation exposure – was responsible for the majority of procedures, the use of other types of studies like CT scans was not rare.
The National Academies’ Biological Effects of Ionizing Radiation (BEIR) VII report13 has cautioned that there is no lower threshold of exposure to radiation that has been identified as without risk and that repeated exposure increases risk in a linear fashion. Furthermore, studies suggest that the hazards of radiation may be larger in children than adults.6–8, 14–16 For example, Brenner et al. estimated that the risk of fatal malignancy from radiation exposure with an abdominal CT was 8-fold higher in the first year of life as compared with a 50-year old adult.7 This results primarily from children having greater sensitivity to the effects of radiation due to growing tissues and having a longer life expectancy than adults, allowing more time for the latent effects of radiation to emerge.
The risks of radiation exposure in children are also not restricted to the development of cancer. For example, repeated head CT that includes imaging of the lens of the eye may increase the risk of later cataract formation. Concern has also been raised for the possibility of later developmental problems and other non-malignancy concerns. Over 3000 children in this study population- nearly 1% of the subjects, received 2 or more head CTs during the 3 year study period.
A key finding from our analysis was that the majority of imaging procedures performed in this study population were attributable to plain radiography, which typically delivers a “low” dose of radiation in comparison to other techniques. Yet while the use of CT scans, nuclear medicine scans and fluoroscopic and/or angiographic procedures was much less common, it was not rare. CT scans, in particular, were used in approximately 8% of the study population and their use increased dramatically in older age groups. As such, the overall contribution of these imaging procedures to the long-term risks of radiation should not be overlooked. This is particularly true since their use appears to be concentrated in a minority of individuals who may undergo repeat studies.
Of the imaging procedures we examined, CT scans may be the most important from the standpoint of radiation exposure. Nationally, the use of CT has rapidly grown over time, due to an increased availability of CT scanners and a lower threshold for ordering these studies in routine clinical practice.6, 17 Although there is evidence that CT volumes in pediatric hospitals have decreased since 2003 relative to the total number of cross-sectional imaging examinations, the absolute trend in use is less clear.18 If one were to extrapolate our findings to the pediatric population of the United States,19 5.8 million children less than the age of 18 would be expected to undergo at least 1 CT scan during a 3-year period. Furthermore, nearly 2.6 million would undergo 2 or more CT scans. The average child in this study population received 0.86 CT scans by the time they reached age 18. Importantly, these findings extend results recently reported by the Congressionally-chartered National Council on Radiation Protection and Measurements, which estimated that 8–10% of CT scans in the United States are performed in children but did not examine rates of use longitudinally over time within the same child or describe the specific procedures used.20
These data reinforce the importance of judicious use of imaging procedures that utilize ionizing radiation, particularly in children. Numerous regulatory groups and publications have highlighted the importance of the concept of ALARA – As Low As Reasonably Achievable – in the application of radiation for imaging procedures.6 But recent studies in the literature suggest that this practice is not being used broadly or cohesively in adult imaging.3, 21 We suspect the same is true for pediatric imaging. For many types of imaging procedures, the development of age and weight-dose protocols have been lacking, although this may be less of a problem in pediatric hospitals. In the case of CT scans, there is concern that they are not yet widely applied. The Image Gently and Step Lightly programs22 are important educational campaigns that are enlisting physicians and parents to reduce radiation exposure to children.
Appropriate use of these procedures requires balancing the long-term risks inherent in radiation exposure with the necessity for making clinical decisions at the bedside. Developing better guidelines for these procedures may help guide clinicians struggling to determine the best role for these studies. For example, CT scans have revolutionized the management of head trauma in children. Having either too high or too low a threshold for ordering a CT in this clinical situation may be problematic. Several studies have recently examined the use of head CT in this context,23–29 and have derived decision rules for its use based on identifying clinically-important traumatic brain injuries, suggesting that CT can be avoided in many children presenting to the emergency room with head trauma.
Several limitations to this study should be noted. We used administrative claims data. While this meant that we were able to exhaustively capture information on the utilization of imaging procedures in the study population, we do not know the clinical context under which these procedures were ordered. We therefore cannot comment on their appropriateness. However, the intent of our study was to examine contemporary use of these procedures – both appropriate and inappropriate. Second, we did not attempt to estimate effective doses of radiation from these imaging procedures. There is a paucity of data available on radiation dosimetry in the pediatric population. Extrapolating effective dose estimates for children from estimates available in the literature could be misleading.11,22 Instead, we focused on describing the patterns of utilization across various types of imaging procedures, which we believe adds critical new information for policy-makers and providers.
Third, data in the study population were gathered from five specific market areas. While these areas represent large and diverse communities, there may be aspects of these populations that differ from the greater United States as a whole. This could potentially limit the generalizability of these findings to the broader United States. Similarly, these data only include children who were insured and do not represent utilization in uninsured settings. Fourth, our results likely underestimated the total number of studies performed in children by restricting individuals to one procedure per day of the same technology on the same anatomic region and excluding patients who died during the study period. We did this to be as conservative as possible in reporting our findings and to focus explicitly on survivors – the group where the long-term risks of radiation are most concerning. Finally, these results represent a snapshot in time. Imaging procedures are constantly evolving and their use has been in flux during recent years.
In conclusion, our study describes patterns of use of diagnostic and therapeutic imaging procedures that utilize ionizing radiation in a large pediatric population. We found that the use of these procedures is common and that studies associated with high-doses of radiation are not infrequent, and are utilized repeatedly in a smaller group of children. These results highlight the importance of generating data-based guidelines to aid clinicians in determining the appropriateness of performing imaging procedures in children.
Dr. Dorfman had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.