British Journal of Radiology (2005) 78, 34-38
© 2005 British Institute of Radiology
doi: 10.1259/bjr/79694026
Radiation risk and cost-benefit analysis of a paediatric radiology procedure: results from a national study
H B L Pettersson, PhD1,
K Fälth-Magnusson, MD, PhD2,
J Persliden, PhD1,3 and
M Scott, PhD4
1 Department of Radiation Physics, IMV and 2 Department of Paediatrics, IMK, Faculty of Health Science, University of Linköping, S-581 85 Linköping, Sweden, 3 Department of Medical Physics, Örebro University Hospital, S-701 85 Örebro, Sweden and 4 Department of Statistics, University of Glasgow, Glasgow G12 8QW, UK
 |
Abstract
|
|---|
A national study was performed to investigate radiation doses and associated risks to patients during X-ray fluoroscopy-guided small intestinal biopsies in the investigation of coeliac disease. Thermoluminescent dosemeters (TLD) and questionnaires were sent to 42 of the 43 paediatric departments in Sweden performing these biopsies. During the study period (2 x 3 weeks) 257 biopsies were recorded, representing about 10% of annually performed paediatric investigations. The results show that the absorbed dose during biopsy ranged from 0.04 mGy to 23.8 mGy (mean 1.87 mGy). The fluoroscopy time ranged from 2 s to 663 s (mean 60 s). The collective dose from the procedure amounts to 4.7 manSv year–1. Thus, the annual excess cancer mortality, including severe hereditary effects, can be estimated at 0.6–0.7 cases per year. However, significant dose saving can be obtained by proper choice of sedation and biopsy equipment.
|
Introduction
|
|---|
Several studies have been performed on dose and related risk to the patient from different diagnostic radiological procedures. The vast majority of these studies have concerned adult patients. However, the risk of lethal cancer from radiation exposure of children is expected to be 2–4 times higher than for adults per dose unit [1, 2]. The reason for this difference is not fully clear, but greater cell proliferation rate and longer life expectancy for children both result in a higher risk of developing late effects. Therefore, initiatives have been taken among radiation protection authorities [3–5] to give priority to investigations of dose levels and frequencies of X-ray examinations among children. The main objective is to establish recommendations of upper dose limits for various diagnostic procedures and to implement minimum requirements for equipment standards [6].
In a previous study [7] we measured the radiation dose and assessed the associated long-term risk of paediatric small intestinal biopsies performed at our hospital during a 10-year period. In a follow-up study, performed in co-operation with a neighbouring hospital [8], we found considerably lower radiation doses associated with their biopsy investigations. This urged us to extend our investigations to a national study to obtain extensive information on this diagnostic procedure, including information about equipment and the resulting radiation dose to assess the potential for dose reduction. A full report of patient absorbed doses and fluoroscopy times was published elsewhere [9].
In this study we examined parameters such as (i) X–ray equipment; (ii) sedation techniques; (iii) capsule device; (iv) examination frequency, with the purpose of identifying their importance for the dose and associated risk to the patient, and from a cost-benefit point of view to suggest optimal examination conditions.
|
Material and methods
|
|---|
Dose measurements
All Swedish paediatric departments that performed small intestinal biopsies were invited to participate in a co-operative study, and 42 out of 43 units accepted. The study was approved by the Paediatric Section of Gastroenterology and Nutrition and performed under the consent of the Swedish Paediatric Association. The dose measurements were carried out during two periods of 3 weeks each in 1992. Dosemeters were sent by mail and returned after each irradiation, and the measurements were performed by Li2B4O7 thermoluminescence dosimetry. Two Li2B4O7 chips were placed together in thin plastic bags for each dosemeter. Four dosemeters were each placed on the anterior and posterior side of the trunk in the area of irradiation. Reproducible positioning of the dosemeters was ensured by providing dosemeters fixed in plastic bags, with paper backing and pre-cut circles for the umbilicus, and a drawing marked out for the spine. The thickness of each child was measured at the umbilical level, with the child lying supine. This measure is entered in the calculations of the absorbed dose.
The absorbed doses, D [Gy], were calculated as the mean absorbed dose in the irradiated volume from the anterior and posterior dosemeter readings according to methods previously reported [7]. Based on the reported number of biopsies and the calculated average of all absorbed doses, the annual collective effective dose, Ec [Sv] for this procedure in Sweden could be evaluated.
In addition, all participating departments were asked to report the conditions of their examinations, i.e. method of sedation, type of biopsy capsule, type of X-ray equipment and exposure conditions (fluoroscopic time, image system sensitivity and age, image memory, grid ratio, distance focus spot–image intensifier surface).
We chose to study the conditions at each paediatric unit, and not ask for the identity of each paediatrician performing the biopsy procedure, since we feared that this would have significantly decreased the participation in the study.
Cancer risk assessment
The cancer risk due to radiation exposure, e.g. expressed as the cancer excess lifetime mortality risk (CELMR), is well documented. BEIR-V [2] give estimates of CELMR for exposure of different age groups; the unisex CELMR of radiation exposure at high dose rate at 5 years of age being 0.14 per Sv whole body dose (effective dose).
Fluoroscopy during coeliac biopsy does not correspond to a whole-body exposure situation, i.e. it affects only a limited part of the body (lungs, liver, spleen, kidneys, bladder, digestive system). However, if we examine the relative sensitivity of the different organs (using tissue weighting factors, wt) [1] we find that most of the "high-risk" organs are exposed, which means that the risk factor is not significantly overestimated. One could of course argue that the chosen values of wt do not explicitly reflect the relative sensitivity of organs in children. However, recent work [10] indicates that significant differences in these factors between children and adults are not to be expected.
Loss of life expectancy
The reduction of life expectancy is the mathematical expectation of the loss of lifetime due to accidents, diseases, exposure of pollutants etc. In this context we use this concept of loss of life expectancy (LLE) to evaluate the effect of radiation exposure [1]. It is calculated from the relationship: LLE=R x Y, where R is the attributal lifetime probability of death, i.e. here equal to CELMR, and Y is the mean loss of lifetime if the radiation exposure is the cause of death. The LLE concept is only applicable to exposure of a cohort to estimate the collective loss of lifetime, i.e. for the individual the LLE is either zero or Y. In our assessment, we assume a mean loss of lifetime of 40–70 years if the fluoroscopic radiation causes cancer death, to account for both cancers with relatively short latency period, i.e. leukaemia, thyroid cancer and those with longer latency period. This gives the following equation: 
Generally speaking, when diagnostic radiological procedures are used, the practice should produce a net benefit for the patient, i.e. the risk attributed to radiation should be much smaller than the diagnostic benefit for the patient, a concept known as "justification of the practice" [11]. In most situations the net benefit is easily obtained. In addition, all exposures shall be kept as low as reasonably achievable, a radiation protection concept known as the ALARA principle [11]. Thus, for a given procedure, a cost-benefit analysis, which establishes optimum levels of radiation protection for the patient, should be performed. The action taken depends on the exposure situation. In diagnostic radiology it could involve introduction of different radiation shields, proper choice of diagnostic procedure and equipment and adequate training of personnel. The cost to society of a lost life is taken into consideration and could be expressed as a monetary value, 
. The actual value of is not fixed since it depends on decisions made at a national level. For the Nordic countries, the Nordic radiation protection authorities advocate that up to US $100 000 would be a reasonable amount to spend for reducing the collective dose by 1 manSv [12].
|
Results and discussion
|
|---|
Number of biopsies and age of the children
The total number of biopsies performed during the 6 week study period was 257 (130 in the spring period and 127 in the autumn). This frequency (an average of 43 biopsies per week) corresponds rather well with the information previously received in a national questionnaire, reporting a total of 2500 small intestinal biopsies per year (48 biopsies/week). As both central, regional and university hospitals perform these examinations, the number of biopsies performed per clinic showed great variation; from 0 to 18. This introduces a risk of bias when results from different clinics are compared.
The age of the children ranged from 0.2 years to 18.7 years, showing a skewed distribution. Normality was obtained by log transforming the data, resulting in a mean age of 2.7 years (95% CI: 2.4–3.1 years), compared with a median age of 2.5 years.
Dose results
Figure 1 shows the distribution of the absorbed doses. Large variations in dose, both between different departments and also within the same department, were observed. The departmental median absorbed doses ranged between 0.11 mGy and 14.4 mGy and the individual doses ranged between 0.04 mGy and 23.8 mGy (mean 1.87 mGy, median 0.69 mGy).
The corresponding range of fluoroscopy time was 2–663 s (mean 60 s, median 30 s), i.e. more than two orders of magnitude variability.
Cancer risk assessment
Based on the annual number of X-ray examinations for the investigation of paediatric coeliac disease in Sweden (2500), the annual collective dose derived from the individual doses is 2500 x 1.87 mSv=4.7 manSv. Accordingly, the annual number of excess lethal cancers per year in Sweden due to this procedure is 0.6–0.7 (4.7 Sv x 0.14 Sv–1) and the collective annual LLE is 20–50 years (4.7 Sv x 5–10 years Sv–1).
Exposure conditions and cost-benefit analysis
All the results on fluoroscopy time (t) and absorbed dose (D) obtained for the 257 biopsy examinations were tested statistically (standard parametric 1 and 2 sample t-tests, analysis of variance and regression methods) against the different examination conditions, i.e. X-ray equipment parameters (image system sensitivity and age, image memory, grid ratio, distance focus spot–image intensifier surface), sedation technique and the departmental experience (number of biopsies performed annually). Both the fluoroscopy time and absorbed dose data displayed clearly skewed and non-normal distributions. However normal distributions were obtained by log-transforming the data, for which parametric statistics were applied. Multi-regression analysis was used initially to identify any linear relationships between t and D and the different variables. Only the relation to different sedation techniques was identified as strongly significant. However, due to the large number of variables and the rather limited number of departments (42) the result cannot exclude other significant variables.
Sedation
The paediatric units reported that they used one of three different levels of sedation: full anaesthesia, parenteral sedation or oral/rectal sedation. In full anaesthesia, intubation is used and ventilation is taken over (no spontaneous breathing). In parenteral sedation, the child is under deep sedation but has maintained spontaneous breathing. In oral/rectal sedation, the child is more superficially sedated.
24 of the 257 (9.3%) biopsies were performed with oral/rectal sedation, 188 of 257 (73.2%) by parenteral sedation and 35 of 257 (13.6%) by full anaesthesia. The dose distributions for the three sedation techniques used (Figure 2) were studied by two sample t-test. A significant difference was obtained both for (a) oral vs parenteral sedation (p=0.02, D(mean)=3.40 and 1.66 mSv, respectively) and (b) oral sedation vs full anaesthesia (p<0.00001, D(mean)=3.40 and 0.81 mSv, respectively), but parenteral sedation vs full anaesthesia showed a statistically non significant difference (p=0.15). Using a 95% confidence interval we can conclude that oral sedation results in 0.5–2.9 mSv higher dose than for parenteral sedation and 1.5–3.7 mSv higher dose than for full anaesthesia.
Thus, on a national basis, the collective doses can be reduced by changing sedation techniques. If instead of the oral/rectal sedation technique parenteral sedation or full anaesthesia were to be used, 0.12–0.67 manSv and 0.35–0.86 manSv, respectively, could be saved annually. If full anaesthesia were adopted at all clinics, the net gain would be 1–3 manSv annually. From a cost-benefit point of view, up to $100 000–300 000 US gross could thereby be spent annually to cover costs for changing sedation technique. The full anaesthesia procedure brings additional cost of personnel for surveillance of the patient. Information from the hospitals showed that the calculated additional cost of change of practice from oral sedation to full anaesthesia amounts to at least US $200–250 per patient. Thus, if full anaesthesia is adopted at all clinics, the total additional annual medical cost amounts to about US $500 000.
Hence, from a cost-benefit point of view, a change of sedation practice is not justified. However, according to the opinion of most paediatricians performing this procedure, full anaesthesia or parenteral sedation is considered far less traumatic for the child, which could justify a change of sedation practice. But, in daily life, the availability of staff at the anaesthetic unit may be a limiting factor for such change.
Image memory
Overall, 48.5% of the biopsy examinations were carried out with X-ray units equipped with image memory. In order to facilitate comparison in dose distribution between the two options, only the cases where parenteral sedation technique was used were considered. Of these investigations, 50% were performed with image memory and 50% without. No significant difference in the two dose distributions (Figure 3) was observed; p=0.90, D(mean)=1.38 and 1.32 mSv, respectively, i.e. from a cost-benefit point of view it does not seem worthwhile to change the current practice. However, if for example the investigation is performed in a teaching hospital, our experience is a 30% reduction in fluoroscopy time using image memory in the teaching situation.
View larger version (18K):
[in this window]
[in a new window]
|
Figure 3. Frequency distribution of absorbed doses for X-ray equipment with and without image memory (only parenteral sedation cases shown).
|
|
Choice of biopsy capsule
Three types of biopsy capsule were used. The Crosby and Watson capsules are both one-port devices, giving only one chance to obtain a mucosal specimen, while the Storz capsule enables repeated firing, thereby decreasing the risk of a failure in the sampling. 76% of the investigations were performed with Watson capsule, 16% with Storz capsule and 8% with Crosby capsule. Figure 4 shows the frequency distributions of fluoroscopy times for the three options, clearly displaying skewed distributions. The mean fluoroscopy times, calculated on log-transformed data, and the median fluoroscopy times were: Watson capsule; t(mean)=37 s (95% CI: 32–44 s), and t(median)=36 s, Storz capsule; t(mean)=11 s (95% CI: 6–19 s) and t(median)=12 s, Crosby capsule; t(mean)=12 s (95% CI: 5–29 s) and t(median)=5 s. One-way analysis of variance (ANOVA) showed statistically significant differences (p<0.001) between the distributions of fluoroscopy times for Watson vs Storz capsule and for Watson vs Crosby capsule. The data show that at the 95% confidence level the use of the Crosby capsule results in 0.16 to 0.6 times lower fluoroscopy times compared with using the Watson capsule, and using the Storz capsule results in 0.17 to 0.5 times lower fluoroscopy times compared with using the Watson capsule. These results would suggest significant dose savings by changing from the use of the Watson capsule to the use of either the Crosby or Storz capsule. However, owing to the limited number of cases, 16 cases using the Crosby capsule and 34 cases using the Storz capsule, the risk of bias is clear, which makes an estimate of the cost-benefit uncertain. In recent years there has been a discussion on whether the endoscopic technique to obtain a biopsy is better than the peroral capsule technique [13–16]. The endoscopic technique is becoming more commonly used [15], but the capsule technique is still often used, making it important to optimize the conditions for its use.
Age of patient
The age of the child may influence the degree of difficulty experienced in performing the biopsy procedure and thus the time needed for the procedure and the fluoroscopy time. However, analysis of the fluoroscopy and dose data for the different sedation techniques used, did not show any significant correlation with the age of children (R2<0.13). In addition, correlation analysis of fluoroscopy time and child age data from individual hospitals, where the X-ray equipment, sedation technique and biopsy device is fixed, confirmed the lack of correlation (R2<0.03).
Staff experience
The staff experience of the biopsy procedure is likely to influence the fluoroscopy time. The trend outlined in Figure 5 is however rather weak. Even so, one could argue that it could be beneficial to reduce the number of clinics performing this kind of procedure, thereby increasing the number of patients per clinic, i.e. maintaining staff experience. However, such changes may increase the cost of transportation and inconvenience for the patient. It was not possible to study the influence of experience at the individual level, i.e. how many biopsies per year were performed by each investigator.
View larger version (18K):
[in this window]
[in a new window]
|
Figure 5. The departmental mean fluoroscopy time (±1 standard deviation) as a function of the number of annually performed biopsies (only parenteral sedation cases shown).
|
|
X-ray equipment
It is often not possible to choose X-ray equipment at a radiology department. The best equipment should be chosen (low dose, good image quality, easy to operate), especially for paediatric radiology. It has been shown in colon examinations [17] that the KAP-value (kerma–area product=DAP-value (dose–area product)) to patients can be reduced by as much as 60% by a change from analogue fluoroscopy, including screen–film overview images, to digital equipment with pulsed fluoroscopy and exclusion of the large overview images. Pulsed fluoroscopy alone reduced the KAP-value by 30%. Unfortunately, in our study, we were not able to obtain information about the use of pulsed or non-pulsed fluoroscopy by individual radiologists/paediatricians.
|
Conclusions
|
|---|
Several factors influence the fluoroscopy time in paediatric small intestinal biopsies. X-ray units equipped with image memory represent an obvious advantage in teaching and training situations. The sedation technique was found to be an important aspect of this procedure as well as the choice of capsule. Since the biopsy device is not a part of the X-ray equipment, its importance may be overlooked and not taken into consideration in investigations of patient absorbed doses. The training of staff, reflected in the mean fluoroscopy time versus number of annual biopsies performed, did not have a great influence on patient exposures. However, the identities of the paediatricians were not revealed, which makes it impossible to evaluate the impact of personal experience. It could be hypothesized that it is more beneficial to study and learn the procedure at centres with optimized procedures than to perform many biopsies in a non-optimized manner. Taking the different steps in this X-ray investigation into consideration, it is possible to reduce the absorbed dose to the patient considerably. The general results shown here are likely to be important also in other types of paediatric X-ray examinations.
|
Acknowledgments
|
|---|
We gratefully acknowledge the cooperation of the staff at paediatric and radiation physics units at the participating hospitals.
|
Footnotes
|
|---|
This work was supported by grants (#SSI 815/94) from the Swedish Radiation Protection Authority. 
Received for publication May 6, 2003.
Revision received August 4, 2004.
Accepted for publication August 18, 2004.
|
References
|
|---|
- ICRP 60. International Commission on Radiological Protection. 1990 Recommendations of the International Commission on Radiological Protection. Oxford: Pergamon Press, 1990.
- BEIR-V. Committee on the Biological Effects of Ionizing Radiations. Health effects of exposure to low levels of ionizing radiation. BEIR V Report. Washington: National Academy Press, 1990.
- Fendel H, Schneider K, Kohn M, Bakowski C. Specific principles for optimization of image quality and patient exposure in paediatric diagnostic imaging. BIR Report 20: Optimization of image quality and patient exposure in diagnostic radiology. London: BIR Press, 1989:91–101.
- Schneider K, Fendel H, Bakawski C, Stein E, Kohn M, Kellner M, et al. Results of a dosimetry study in the European Community on frequent X-ray examinations in infants. Radiat Prot Dosimetry 1992;43:31–6.[Abstract]
- Schneider K, Kohn M, Bakowski C, Stein E, Freidhof C, Horwitz AE, et al. Impact of radiographic imaging criteria on dose and image quality in infants in an EC-wide survey. Radiat Prot Dosimetry 1993;49:73–6.[Abstract]
- European Guidelines on Quality Criteria for Diagnostic Radiographic Images in Paediatrics, EUR 16261. Luxembourg: Office of Official Publications of the European Communities, 1996.
- Persliden J, Pettersson HBL, Fälth-Magnusson K. Small intestinal biopsy in children with coeliac disease: measurement of radiation dose and analysis of risk. Acta Paediatr 1993;82:296–9.[Medline]
- Persliden J, Pettersson HBL, Stenhammar L, Fälth-Magnusson K. Small intestinal biopsy of children with coeliac disease: the influence of X-ray equipment on radiation dose. Eur Radiol 1994;4:458–61.[CrossRef]
- Persliden J, Pettersson HBL, Fälth-Magnusson K. Intestinal biopsy in children with coeliac disease; a Swedish national study of radiation dose and risk. Acta Paediatr 1996;85:1042–6.[Medline]
- Almén A. Radiation dose to children in diagnostic radiology. Thesis. Malmö University Hospital, Lund University, 1995.
- ICRP 42. International Commission on Radiological Protection. A compilation of the major concepts and quantities in use by the ICRP. Oxford: Pergamon Press, 1984.
- News from SSI. Stockholm: National Radiation Protection Institute (SSI), 1991; l4. 0280-0357.
- Branski D, Faber J, Freier S, Gottschalk-Sabag S, Shiner M. Histologic evaluation of endoscopic versus suction biopsies of small intestinal mucosae in children with and without coeliac disease. J Pediatr Gastroenterol Nutr 1998;27:6–11.[CrossRef][Medline]
- Magliocca FM, Bonamico M, Petrozza V, Danesi H, Liuzzi M, Velucci O, et al. Usefulness of endoscopic small intestinal biopsies in children with coeliac disease. Ital J Anat Embryol 2001;106 (2 Suppl 1):329–35.
- Stenhammar L, Ascher H, Daniesson L, Dannaeus A, Hernell O, Ivarsson A, et al. Small bowel biopsy in Swedish paediatric clinics. Acta Paediatr 2002;91:1126–9.[CrossRef][Medline]
- Thomson M, Kitching P, Jones A, Walker-Smith JA, Phillips A. Are endoscopic biopsies of small bowel as good as suction biopsies for diagnosis of enteropathy? J Pediatr Gastroenterol Nutr 1999;29:438–41.[CrossRef][Medline]
- Sjöholm B, Andersson T, Jonsson B, Persliden P. Does digital radiology increase or decrease the dose to the patient? Svenska Läkarsällskapet Hygiea 1999;108/3:239. (In Swedish.)
Top
Abstract
Introduction
