| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Evidence-based Practice |
1 From the Department of Radiology (M.G.C.G., F.J.G.), the Health Services Research Unit (M.G.C.G., J.E.A., A.M.G.), and the Department of Orthopaedic Surgery (D.W.), University of Aberdeen, Foresterhill, AB25 2ZD, Scotland; the Department of Orthopaedic Surgery, Ninewells Hospital, Dundee, Scotland (N.W.V.); and the Department of Orthopaedics, Hairmyres Hospital, Glasgow, Scotland (A.C.G.). From the 2000 RSNA scientific assembly. Received October 13, 2000; revision requested November 26; revision received January 31, 2001; accepted March 2. Supported by the NHS Research & Development Health Technology Assessment Programme. The Health Services Research Unit is funded by the Chief Scientist Office, Scottish Executive Health Department. Address correspondence to M.G.C.G. (e-mail: m.g.gillan@abdn.ac.uk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: A randomized controlled before-and-after study was performed in 145 patients who had symptomatic lumbar spinal disorders and had been referred to orthopedists or neurosurgeons. Participants were a subgroup within a multicenter pragmatic randomized comparison of two imaging policies on LBP treatment: "imaging" versus "no imaging," unless a clear indication developed. Paired assessments were made of diagnosis, diagnostic confidence, proposed treatment, treatment confidence at trial entry and follow-up, and expectations of imaging. Data were analyzed according to the groups as randomized.
RESULTS: At follow-up, there were no statistically significant differences between the groups with respect to diagnosis or treatment plans. Significant increases in diagnostic and therapeutic confidence between trial entry and follow-up were observed for both groups, with a significantly greater increase in diagnostic confidence (P = .01) in the imaging group.
CONCLUSION: Imaging may increase diagnostic confidence but has minimal influence on diagnostic or therapeutic decisions for patients with LBP. The results highlight the need for evidence-based guidelines for imaging in LBP treatment.
Index terms: Efficacy study • Spine, abnormalities, 30.1211, 30.1214 • Technology assessment
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
There are demands for more rigorous scientific evaluation of both the clinical effectiveness and cost-effectiveness of new health care technologies such as MR imaging or CT prior to their widespread diffusion and use (4–7). Most health technology assessments of imaging (5,8,9) focus on technical and diagnostic performance (8,10,11), with a minority (12–20) addressing the broader influence of imaging on clinical decision making, patient treatment, and outcome.
We report a randomized controlled before-and-after study of "imaging" versus "no-imaging" policies among patients who had symptomatic lumbar spinal disorders and had been referred to orthopedists or neurosurgeons. The purpose of the study was to assess the influence of CT and MR imaging on LBP diagnosis and treatment.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The Scottish Back Trial
The Scottish Back Trial is a multicenter pragmatic (21) randomized comparison of two imaging policies on the treatment of patients with LBP. It was approved by the Scottish Multicentre Research Ethics Committee and the appropriate Local Research Ethics Committees. The goal of the trial is to establish whether early cross-sectional imaging (MR imaging or CT) influences the clinical treatment and outcome of patients with LBP and whether it is cost-effective.
Eligible patients were those with symptomatic lumbar spinal disorders for whom there was clinical uncertainty about whether or when to perform imaging. This excluded patients who required immediate referral for imaging, patients who had undergone previous imaging, patients in whom imaging was clearly not indicated, and patients who had pain of nonspinal origin. Randomization of patients giving informed consent was performed by using an automated telephone randomization service that incorporated stratification by surgeon and incorporated minimization (22) on age, sex, and clinical category. Patients were randomly assigned to either imaging (MR imaging or CT as soon as possible) or no imaging (no MR imaging or CT unless a clear clinical indication developed [eg, a decision to perform surgery] or the clinical situation deteriorated). Since the goal of the Scottish Back Trial is to assess the use of imaging in normal clinical practice, the choice of imaging modality was at the discretion of the clinician.
At each center, imaging was performed with standard commercially available MR imagers and CT scanners, by using imaging protocols at the discretion of the radiologist. The principal outcome measures are the SF-36 general health status questionnaire (23) and the Aberdeen Low Back Pain Score (24), which are completed by participants at trial entry and after 8 and 24 months. Information on health care resource use is being obtained from hospital records and patient questionnaires. The trial will be completed in 2001.
A pilot study involving approximately 30 patients performed prior to this study (25) enabled clinicians to improve the questionnaires’ design and gain experience in their completion and provided the basis for the sample size estimation.
Patient Assessment
Paired assessments were made at trial entry and follow-up appointment. Completing standardized questionnaires, the recruiting clinicians (including D.W., N.W.V., and A.C.G.) recorded a clinical category, stated their diagnostic confidence (0%–100%), the proposed treatment plan and confidence (0%–100%) in the choice of treatment, and the expectations of imaging (to establish or confirm the diagnosis, assess the extent or location of disease, exclude an abnormality, or plan treatment).
The second assessment was at the follow-up appointment after imaging, if allocated, but without reference to the first assessment. At some outpatient clinics, follow-up assessment forms had to be completed retrospectively with reference to case notes. Clinicians recorded whether imaging had been performed. For patients in the imaging group, clinicians assessed the contribution of imaging (ie, to establish or confirm the diagnosis, assess the extent or location of disease, exclude an abnormality, or plan treatment) and stated their opinion from a list of options on the scale of the contribution, from "no contribution" through "minor" and "moderate" to "considerable" contribution. The working diagnoses, or clinical categories, were (a) symptomatic lumbar disk herniation, (b) root entrapment secondary to degenerative disease, (c) neurogenic claudication, (d) chronic LBP (symptom duration > 1 year) not covered by a–c, and (e) others (not covered by a–d) including spondylosis, spondylolisthesis, spinal stenosis, sacroiliitis, pathologic fracture, and osteoporosis.
Treatment plans were ordered by increasing degrees of intervention as (a) discharge with reassurance, (b) further investigation only (eg, blood testing to exclude inflammatory or infectious disorders, bone scanning, or further lumbar spinal imaging), (c) conservative treatment (including physiotherapy, manual therapy, traction, bed rest, lumbar support, or medication), (d) injection therapy (epidural, facet, nerve root infiltration, or similar injection), and (e) surgery or chemonucleolysis.
In assessing changes from trial entry to follow-up, differences between the paired assessments were rated as "more interventionist" or "less interventionist." For example, if the proposed treatment was changed from discharge to physiotherapy, this was classified as a change to more interventionist, whereas a change from surgery to injection therapy was classified as a change to less interventionist.
Clinician Experience
Because the study was conducted in a combination of teaching hospitals and district general hospitals, specialist spinal surgeons (D.W., N.W.V.), general orthopedists (including A.C.G.) (orthopedic surgeons and one orthopedic physician), and junior orthopedic staff members were involved in patient assessment. Some patients were assessed by a different clinician at the follow-up appointment. Although limited by sample size, additional analyses were performed to assess the potential influence of some of these factors.
Sample Size
Results of our pilot study involving 27 patients suggested that diagnostic confidence was increased after imaging in 21 (78%) of 27 patients. Adopting a more conservative estimate (reflecting the small numbers on which the pilot study was based), we aimed to detect whether imaging was associated with increased diagnostic confidence in at least 70% of patients.
Accepting that a number of those randomly assigned to the no-imaging group would actually undergo imaging, and accepting that clinicians’ diagnostic confidence may have increased anyway by the time of the second assessment, we sized the trial to detect a difference in the proportion who had increased their confidence from 40% in the no-imaging group to at least 70% in the imaging group by the time of the second assessment. To detect this difference with 90% power and at the 5% significance level, approximately 130 participants required randomization. To allow for the potential loss to follow-up because of patient nonattendance at the second visit, we aimed to recruit at least a further 20% of patients.
Statistical Analysis
Univariate comparisons of categoric data were performed with the 
2 test. The Wilcoxon signed rank test was performed to test for change in diagnostic and therapeutic confidence between entry and follow-up for each group. The Mann-Whitney test was performed to test for differences in nonnormally distributed continuous data. Nonparametric data (ie, interval between trial entry and follow-up assessment forms) were presented as medians with interquartile ranges (the difference between the 25th and 75th percentiles).
All analyses and summaries were performed by using statistical computer packages (SPSS, SPSS, Chicago, Ill; and Arcus Quickstat [Biomedical], Longman Software Publishing, Cambridge, England). We recognized that analysis based on the actual use of imaging would be biased by the selective use of imaging, especially in some patients in the no-imaging group who "crossed over." To avoid such selection bias, an analysis based on the groups as randomized ("intention to treat") (26) was performed. A secondary analysis in which groups were compared according to whether patients actually underwent imaging, to assess the maximum influence of imaging, produced similar results to those of the primary analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
During the 12-month period of this study, completed trial entry and follow-up assessment forms were available for 145 of the original 190 patients recruited by participating clinicians. The progress of patients through the study is shown in the Figure. The clinical characteristics of the 190 patients recruited—the 45 lost to follow-up and the 145 for whom completed assessment forms were available—are shown in Table 1. Response analysis showed no significant difference in clinical characteristics (age, sex, diagnostic category) at trial entry between those patients who completed the study and those who were lost to follow-up. Of those patients with follow-up assessments, 81 (56%) were from the imaging arm of the trial and 64 (44%) from the no-imaging arm. Some forms were incomplete, usually because of omission of diagnostic confidence scores, and the numbers in each of the tables vary accordingly. There was no significant difference in the median time interval between trial entry and follow-up appointments for the two study groups (Table 1).
|
|
21, 1.24; P = .265) (Table 2). There was no apparent difference in the patterns of diagnostic change (data not shown).
|
22, 9.28; P = .01) (Table 2).
Therapeutic Impact
There was no significant difference between the groups in relation to changes in treatment plan (Table 3). At the follow-up assessment, treatment plans were unchanged for approximately 50% of patients in both groups. Furthermore, when treatment plans were altered, approximately one-third of patients in each group had their treatment changed to a less interventionist treatment.
|
Contribution of Imaging to Treatment
The expected contribution of imaging was most frequently confirmation of diagnosis (33% [70 of 209]), with establishment of a diagnosis (40 of 209), assessment of the extent or location of disease (41 of 209), and exclusion of abnormalities (49 of 209) all expected with similar frequencies (each approximately 20%). In contrast, imaging was expected to contribute to treatment planning for less than 5% (six of 209) of patients. Among those in the imaging group, the actual contribution to the treatment plan was rated as considerable for more than one-third (25 of 67) of patients, moderate for 25% (17 of 67), and minor for 27% (18 of 67), with no contribution for 10% (seven of 67) of patients. The actual contributions cited were confirmation of diagnosis (37% [30 of 82]), exclusion of abnormality (35% [29 of 82]), establishment of diagnosis (12%), assessment of disease location and/or extent (7%), and planning of treatment (9%).
Clinician Experience
Although there was no significant difference in diagnostic impact between the imaging and no-imaging groups overall (Table 2), a secondary analysis comparing specialist and nonspecialist spinal surgeons revealed a significant difference in changes in diagnosis for patients in the no-imaging group. The diagnostic category was altered in six (29%) of 21 cases assessed by a spinal specialist, as compared with 28 (67%) of 42 cases assessed by a nonspecialist (21, 6.72; P = .007). In the imaging group, diagnostic category was similarly altered, but not significantly (21, 1.99; P = .16). Further comparison of specialist and nonspecialist surgeons indicated no significant difference between the groups in changes in diagnostic category, diagnostic confidence, or treatment plan between trial entry and follow-up, emphasizing the importance of the control group. A similar analysis stratified according to whether the same clinician, regardless of level of experience or expertise, had performed both assessments showed no significant difference in changes in diagnostic category, treatment plan, and diagnostic or therapeutic confidence. Since the data could effectively be classified as clustered at the clinician level, further analysis was undertaken to examine whether the conclusions were sensitive to clustering; all conclusions remained unchanged.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Investigators in some studies (19,20, 29,30) who reported considerable changes in diagnoses, treatment plan, and diagnostic and therapeutic confidence after either CT or MR imaging acknowledged that patient symptoms could change in the interval between assessment and imaging. Another limitation of this type of evaluation is that clinicians may not actually follow their reported treatment plan. However, it seems likely that most previously reported before-and-after assessments of diagnostic and therapeutic impact that did not include a control group of patients have overestimated the contribution of imaging to changes in diagnoses and therapeutic intent.
The only significant effect of imaging appeared to be on diagnostic confidence. A secondary analysis showed that specialist spinal surgeons were significantly less likely to change their diagnosis over time in the no-imaging (control) group. A similar pattern was present in the imaging group, but there was no statistically significant difference between groups in this respect. The increase in diagnostic confidence may have resulted from clinician reassurance, a factor more likely to influence the less experienced surgeons.
The fact that only 13% of patients in the no-imaging group underwent imaging with no significant difference in therapeutic impact or diagnostic or therapeutic confidence, as compared with the imaging group, suggests that the 87% of patients in this group who did not undergo imaging did not require it. Therefore, it could be argued that routine imaging in this type of patient is unnecessary and a waste of costly resources. Imaging is necessary when surgical intervention is considered to enable the surgeon to decide whether there is a surgically treatable lesion and, if so, to decide the exact procedure to be undertaken.
We recognize, however, that our study has limitations that should be taken into account when interpreting our findings. This was part of a larger study conducted within normal clinical practice. Eligibility criteria for the Scottish Back Trial were relatively inclusive, since it is intended to assess the effectiveness of imaging over a wide spectrum of LBP conditions. We also recognized that there could be considerable variation between centers in the guidelines for referring patients for imaging and in local availability and waiting times for imaging examinations. Furthermore, participating clinicians were allowed to select the imaging modality used, although in practice only five (6%) of 83 patients underwent CT. The extent to which the results of the present study are applicable to CT or CT myelography is therefore debatable.
Organizational difficulties were reflected in our failure to complete some follow-up assessments; furthermore, some that were completed were performed retrospectively, which may have influenced the responses. Because hospitals had varying access to and waiting times for imaging, we could not impose fixed intervals on the follow-up appointments. Also, in retrospect, our assessment of clinicians’ expectations of imaging would have been more useful if we had asked them to prioritize the value of the various expected attributes.
Again reflecting the pragmatic design, some follow-up assessments were performed by a clinician who had not performed the first assessment. Although not optimal, this applied to only a minority of cases in both study groups, and secondary analysis suggested that this had no effect on the overall result. Our study also included clinicians with differing amounts of experience, and secondary analysis showed that this has an important influence on follow-up assessments and the effects of imaging. Less experienced spinal surgeons were more likely to change their working diagnosis, especially in the no-imaging group. This would be consistent with specialist spinal surgeons having wider experience in the assessment and treatment of patients with lumbar spinal disorders. When more junior clinicians were involved, we were not able to assess the extent to which any decisions were based on consultation with a more senior colleague or on the level of supervision of junior staff members.
We also recognize that our study may have limited generalizability. Most of the participating clinicians were orthopedists, and there was only one neurosurgeon; the latter recruited relatively few patients. Ninety-seven percent (140 of 145) of participants were recruited from orthopedic clinics, and this may not reflect the normal proportion of patients with LBP seen in secondary care facilities. One-third of participants were recruited from one teaching hospital. Although secondary analysis did not show any clear differences between these patients and those recruited elsewhere, the numbers were small, and this limited our ability to assess whether this had any effect on generalizability.
In conclusion, this study showed clear changes in diagnosis and treatment plan in both study groups, which reflected the natural history of the condition and follow-up consultation after an interval. Access to early imaging increased diagnostic confidence at follow-up, but this was not associated with any greater change in diagnosis or treatment plan. Despite these findings, it is possible that the observed changes in treatment might have been more appropriate in the imaging group and that this might have led to improved outcome. Also, it is possible that the clinicians’ greater diagnostic confidence may have been conveyed to those patients who underwent early imaging, with a subsequent positive effect on the natural history of their condition. More rapid change in diagnosis or treatment would not have occurred in this study because of the patient appointment system in the United Kingdom; nevertheless, it might lead to a more rapid change within other health care systems.
Outcome research in imaging is difficult in view of the chain of events in the assessment hierarchy between the imaging procedure and any measurable influence on patient health (6,8,9,30). Nevertheless, it is essential for assessing the contribution of potentially expensive and resource-intensive technology to patient treatment. More reliable information on the influence of imaging on actual patient treatment is likely to come from the 2-year follow-up of the participants in the larger Scottish Back Trial, which will provide information on health care resource use and patient outcomes.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviation: LBP = lower back pain
Author contributions: Guarantor of integrity of entire study, F.J.G.; study concepts, F.J.G., A.M.G., N.W.V., A.C.G.; study design, F.J.G., A.M.G., M.G.C.G., D.W.; literature research, F.J.G., M.G.C.G.; clinical studies, F.J.G., D.W., N.W.V., A.C.G.; data acquisition, M.G.C.G., J.E.A.; data analysis/interpretation and statistical analysis, J.E.A.; manuscript preparation and definition of intellectual content, M.G.C.G., J.E.A., F.J.G., A.M.G.; manuscript editing, F.J.G., A.M.G., D.W.; manuscript revision/review, F.J.G., A.M.G., D.W., M.G.C.G., J.E.A.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. K. Kuhl Current Status of Breast MR Imaging * Part 2. Clinical Applications Radiology, September 1, 2007; 244(3): 672 - 691. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M Kymes, K. Lee, J. W Fletcher, and SNAP (CSP 027) Study Group Assessing diagnostic accuracy and the clinical value of positron emission tomography imaging in patients with solitary pulmonary nodules (SNAP) Clinical Trials, February 1, 2006; 3(1): 31 - 42. [Abstract] [PDF]
|
|
|
|
 |
|
 R. H. Gottlieb Imaging for Whom: Patient or Physician? Am. J. Roentgenol., December 1, 2005; 185(6): 1399 - 1403. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. J. M. Kock, M. E. A. P. M. Adriaensen, P. M. T. Pattynama, M. R. H. M. van Sambeek, H. van Urk, T. Stijnen, and M. G. M. Hunink DSA versus Multi-Detector Row CT Angiography in Peripheral Arterial Disease: Randomized Controlled Trial Radiology, November 1, 2005; 237(2): 727 - 737. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. E. A. P. M. Adriaensen, M. C. J. M. Kock, T. Stijnen, M. R. H. M. van Sambeek, H. van Urk, P. M. T. Pattynama, and M. G. Myriam Hunink Peripheral Arterial Disease: Therapeutic Confidence of CT versus Digital Subtraction Angiography and Effects on Additional Imaging Recommendations Radiology, November 1, 2004; 233(2): 385 - 391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Gilbert, A. M. Grant, M. G. C. Gillan, L. D. Vale, M. K. Campbell, N. W. Scott, D. J. Knight, and D. Wardlaw Low Back Pain: Influence of Early MR Imaging or CT on Treatment and Outcome--Multicenter Randomized Trial Radiology, May 1, 2004; 231(2): 343 - 351. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
