| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Mallinckrodt Institute of Radiology, Washington University School of Medicine,
510 S Kingshighway Blvd., St. Louis, MO 63110.
2 Department of Orthopedic Surgery, Washington University School of Medicine,
St. Louis, MO 63130.
Received January 13, 2004;
accepted after revision March 9, 2004.
Address correspondence to W. D. Middleton.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. One hundred eighteen patients with shoulder pain and a clinically suspected rotator cuff tear underwent both MRI and sonography and filled out satisfaction surveys after both tests. Patients were asked the following questions: Did the test cause pain? If it did, they were asked to grade the pain on a scale of 1–10 (1, minimal pain; 10, severe pain). Did the test take too long? Would they be willing to undergo the test again? How would they grade their overall satisfaction with the test (1, poor; 2, fair; 3, good; 4, very good; and 5, excellent)? Which test did they prefer if both were equally accurate?
RESULTS. Sonography caused pain above the baseline in 39 patients, with an average pain score (mean ± SD) of 5.4 ± 2.3, and MRI caused pain above the baseline in 40 patients, with an average pain score of 6.1 ± 2.7 (p = 0.36). Two patients thought the sonography examination took too long, and 28 patients though the MRI examination was too long (p < 0.001). The average satisfaction level for sonography was 4.3 ± 0.7 and for MRI, 3.6 ± 1.2 (p < 0.001). The satisfaction score was higher for sonography in 54 patients, higher for MRI in 13 patients, and the same for both in 50 patients (p < 0.001). All patients were willing to repeat the sonography, but 10 patients were unwilling to repeat the MRI (p = 0.002). Ninety-three patients preferred sonography, eight patients preferred MRI, and 17 patients had no preference (p < 0.001).
CONCLUSION. Most patients with shoulder pain prefer sonography to MRI.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Many factors influence the choice of imaging tests used to evaluate specific clinical problems. In addition to accuracy and cost, patient tolerance and preference are also important considerations. The purpose of this study was to compare patient satisfaction with shoulder MRI and sonography.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The order of the tests was randomized. Standard sonography included evaluation of the entire rotator cuff and the tendon of the long head of the biceps muscle of the symptomatic shoulder. In addition, the contralateral shoulder was scanned in all patients. In 61 patients, a second scanning of the symptomatic shoulder was performed by a second radiologist. These additions to the standard unilateral shoulder sonogram were performed because the study was designed to analyze additional aspects of shoulder sonography other than patient satisfaction. All examinations were performed on an Elegra unit (Siemens Medical Solutions) using a 7.5-9.0–MHz multifrequency linear array transducer. Patients were scanned while seated on a cushioned stool. Their arms were moved into various positions to visualize the different components of the rotator cuff and biceps tendon as previously described [2]. The time required to perform sonography was not recorded, but we estimated that the average time, including the time required to obtain the additional nonstandard views, was 15–30 min.
All MRI examinations were unilateral and were supervised by a single radiologist. MRI was performed on the same day as sonography in all except three patients for whom the time between the two studies was 6 days for two patients and 1 day for one patient. Scanning was performed on either a Horizon LX scanner (GE Healthcare), equipped with a two-piece shoulder-array coil, or an Impact, Vision, or Symphony MRI scanner (Siemens Medical Solutions), equipped with flexible local coils. The MRI technique has been previously described [3]. Oblique sagittal and oblique coronal fat-suppressed fast spin-echo images were obtained in all patients. Additional proton density–weighted spin-echo or fast spin-echo and transverse T2-weighted fast spin-echo images with or without fat suppression were also acquired. Two to three signals were averaged for each pulse sequence. All MRI examinations were performed without intraarticular injections. The time required to perform the MRI examinations was not recorded, but we estimated that the average was 45–60 min.
The satisfaction survey was divided into five sections. The first section on the survey dealt with pain. The patients were asked if the test increased the baseline level of pain in their shoulders. If they answered yes, they were asked to quantify the degree of pain on a 10-point scale, with 10 being severe pain and 1 being minimal pain. They were also asked if the test caused pain or discomfort other than in the shoulder and to describe this pain or discomfort.
The second question concerned the length of the examination. Patients were asked to respond yes or no if they felt that the test took too long.
The third question dealt with overall satisfaction. Patients were asked to grade their overall satisfaction with the examination on a 5-point scale (1, poor; 2, fair; 3, good; 4, very good; and 5, excellent). If the satisfaction was rated as poor, they were asked to give the reason.
The fourth section dealt with direct comparison of the examinations. Patients were asked, "If both tests are equally accurate, which test would you prefer?" The choices were sonography, MRI, or no preference. They were also asked to indicate if they would be willing to undergo both tests in the future.
Because data for continuous variables tended to have nonnormal
distributions (Shapiro-Wilk W tests, p 
0.05), both parametric
and nonparametric statistical tests were used. These included analysis of
variance, stepwise multiple regression analysis, the paired t tests,
repeated measures analysis of variance, Wilcoxon's signed rank tests, the
McNemar tests, the Fisher-Freeman-Halton tests, and marginal homogeneity
tests. Statistical testing was performed with JMP statistical software (SAS
Institute) and StatXact 5 statistical software for exact nonparametric
inference (Cytel).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
0.31). However, two significant associations were seen between the other variables of the study and examination type. For MRI, the pain scores were inversely correlated with overall satisfaction (the higher the pain score, the lower the overall satisfaction; p = 0.008, r2 = 0.17). In addition, four of the 39 patients who experienced shoulder pain during sonography also experienced discomfort or pain other than that in the scanned shoulder. For these four patients, the average pain score was 9.0 ± 1.4. For the 35 patients who did not experience discomfort or pain elsewhere, the average pain score was 4.9 ± 2.0 (p = 0.005).
Thirty-two patients reported discomfort or pain other than in the shoulder during the MRI examination. The most common pain was experienced elsewhere in the upper extremity (elbow, n = 7; hand, n = 2), back (n = 3), and heels (n = 1). In 16 patients, the requirement to stay motionless led to cramping, stiffness, numbness, and so forth. Three patients reported claustrophobia but completed the examination. Two patients were too cold and one was bored. Fifteen patients reported discomfort or pain other than in the shoulder during sonography. The most common response was "pain elsewhere," including in the back (n = 2), neck (n = 2), hand (n = 1), bone (n = 1), arm (n = 1), and buttocks (n = 1). Two patients reported being too cold. One patient each reported fatigue, a prickly feeling from the transducer, and heat from the transducer. Two patients reported unspecified pain or tenderness. These differences between the MRI and sonography were significant (p = 0.005).
Examination Length
Two (1.7%) of the 117 patients who responded to the question regarding
length of time on the sonography survey thought that the examination took too
long. Twenty-eight (24.6%) of the 114 patients who responded to this question
on the MRI survey thought that the examination took too long (p <
0.001).
Overall Satisfaction
Figure 1 shows the results
of the patients' overall satisfaction with the two tests. One patient did not
answer this question on the sonography form. The average score for sonography
satisfaction was 4.28 ± 0.7 and the average score for MRI was 3.57
± 1.2. Sonography received a higher score than MRI in 54 surveys. MRI
received a higher score than sonography in 13 surveys. The scores were the
same for sonography and MRI in 50 surveys. A very good or excellent rating was
given 99 times for sonography and 65 times for MRI. A poor or fair rating was
given zero times for sonography and 24 times for MRI. Of the seven patients
who rated MRI as poor, the reasons given were pain (n = 3),
claustrophobia (n = 3), and a sense of abandonment (n = 1).
No matter how these ratings were analyzed, the sonography ratings were higher
than the MRI ratings (p < 0.001).
|
For sonography, none of the other variables of the study was significantly associated with overall satisfaction except that those patients who preferred sonography (n = 93) had significantly higher satisfaction ratings (4.3 ± 0.7) than did the patients who preferred MRI (n = 8, 3.6 ± 0.8, p = 0.01).
As previously noted, for MRI, overall satisfaction was inversely related to pain in the shoulder being examined (p = 0.008, r2 = 0.17). For the 32 patients who reported discomfort or pain other than in the shoulder during MRI, overall satisfaction was significantly lower than that in the 86 patients who did not experience remote discomfort or pain (2.9 ± 1.2 vs 3.8 ± 1.1, p < 0.001). The 28 patients who thought that their MRI examinations were too long had significantly lower overall satisfaction than did the 86 patients who thought the length was acceptable (2.7 ± 1.0 vs 3.9 ± 1.0, p < 0.001). When shoulder pain, remote pain, and the length of the scanning were entered into a backward stepwise analysis (p to enter = 0.25, p to leave = 0.10), only shoulder pain (p = 0.007) and the length of scanning (p = 0.03) were retained.
Examination Preference
When asked if they would be willing to repeat the test, 118 (100%) of 118
patients responded yes for sonography and 105 (91%) of 115 patients responded
yes for MRI (p = 0.002). When asked if they had a preference for one
examination, 93 patients preferred sonography, eight preferred MRI, and 17 had
no preference (p < 0.001). A repeated measures analysis of
variance with overall satisfaction (scores for sonography and MRI) as the
within-subject (repeated-measures) factor and examination preference as the
between-group factor indicated that the interaction of the two factors was
significant (p < 0.001). Overall satisfaction scores were equal
(4.6 ± 0.7 vs 4.6 ± 0.6) when there was no preference between
examination types. MRI scores were higher when MRI examinations were preferred
(4.4 ± 0.9 vs 3.6 ± 0.8), and sonography scores were higher when
sonography was preferred (4.3 ± 0.7 vs 3.4 ± 1.2).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
One of the most important factors in a patient's satisfaction with a diagnostic test is the degree of pain or discomfort caused by the test. In this group of patients, shoulder pain was equally common during both sonography and MRI. Thirty-three percent of patients (39/118) had an increase in the baseline level of shoulder pain during their sonography examinations. This increase was expected for sonography because the patient is required to move his or her arm into positions that may be uncomfortable [2]. Discomfort with positioning can be a particular problem in patients with rotator cuff tears. Fortunately, uncomfortable positions need not be maintained for prolonged periods because the patient can change position at any time without compromising the examination. Compression of the shoulder with the transducer, an important part of a sonography examination, can also produce pain.
We were also not surprised to find a substantial increase in the baseline level of shoulder pain during MRI. The surface coil used to image the shoulder can result in localized pain due to its proximity to the shoulder, and this pain is occasionally aggravated by the requirement of shoulder immobilization for relatively long periods of time.
Pain and discomfort also occurred in locations other than the shoulder. These were reported in 13% (15/118) of the patients who underwent sonography examinations. As with the shoulder pain, the explanation for this pain is most likely related to patient positioning. Patients are seated on a stool for the entire examination, and despite heavy cushioning, this position can cause pain in the back and buttocks. In addition, the examined extremity is moved into various positions to expose the rotator cuff, and this movement can cause pain in the hand, arm, and neck.
Pain remote from the shoulder occurred approximately twice as often during MRI examinations (n = 32) than during sonography examinations (n = 15). This pain was particularly common in the elbow. Seven patients had elbow pain, and on the basis of written comments, the pain was most often due to positioning of the elbow against the hard internal surface of the magnet. The requirement for the patients to remain motionless also led to a variety of complaints in 16 patients. Most often the discomfort from lack of motion was described as cramping, numbness, or stiffness. Although patients with overt claustrophobia were excluded from the study, three patients complained of claustrophobia but nevertheless completed the examination.
Another factor influencing patient satisfaction is the length of the examination. Under normal circumstances, shoulder sonography can be completed in less time than shoulder MRI. In our experience, a routine unilateral shoulder examination takes 15 min with sonography and 45 min with MRI. We did not attempt to compare the actual time required to perform the two tests. Instead, we asked patients if they thought that the test took too long. With respect to patient satisfaction, we believe that this is more important than the actual length of the examination. Twenty-five percent of our patients (28/114) thought that the MRI was too lengthy, whereas less than 2% of patients (2/117) thought that the sonography examination was too lengthy. This response occurred despite the fact that sonography included scanning of both shoulders, and in more than half of the patients, a second radiologist scanned the symptomatic shoulder. Under normal circumstances a single radiologist scans the patient, and only the symptomatic shoulder is examined. Scanning of the contralateral shoulder is required only when comparison is necessary. The patients in this study underwent a more extensive sonography examination because the study was designed to investigate additional aspects of shoulder sonography besides patient satisfaction.
We wanted to gauge the patients' overall level of satisfaction with each test. We incorporated this aspect into the survey so that patients could integrate additional subjective impressions about the test other than the amount of pain involved or the length of the examination. In general, the level of satisfaction was high with both tests. The most common level of satisfaction for both sonography and MRI was "excellent." However, 20% of the patients (24/118) graded MRI as either "poor" or "fair," whereas none of the patients graded sonography similarly. When the overall satisfaction scores were compared on a patient-by-patient basis, similar scores were given to both tests by 43% of patients (50/117). When the scores for the two tests differed, sonography was four times more likely to receive a higher score (54:13).
Given the results described, it is not surprising that most patients with a preference for one test or the other, preferred sonography. However, the fact that patients preferred sonography 10 times (93:8) as often as they preferred MRI was unexpected. This means that many patients who gave sonography and MRI similar satisfaction scores still preferred sonography. The exact reason for this preference is not clear from our study. MRI can be an anxiety-producing examination [4], and this perception undoubtedly contributed to the patients' preference for sonography. In fact, in a study that compared patient preference for MRI versus arthrography, Blanchard et al. [5] found that a significantly higher percentage of patients thought that MRI was unpleasant or extremely unpleasant when compared to arthrography. This finding was despite the invasive nature of arthrography and the noninvasive nature of MRI. From patients' comments, Blanchard et al. speculated that the close personal contact with the radiologist reassured patients during arthrography, whereas the diminished personal contact caused a sense of isolation and additional anxiety during the MRI examinations. Like arthrography, sonography is an interactive examination. Not only can the patient and the examiner converse, but the patient can observe the real-time examination and ask questions as it occurs. In many situations, the results of sonography are immediately available. This type of patient interaction is rarely possible with MRI.
In summary, the overall satisfaction of 118 patients who underwent both shoulder MRI and sonography on the same day and at the same institution was significantly higher for sonography. Most patients (93/101) with a preference for one examination preferred sonography. These results should be considered when deciding how best to image patients with shoulder pain.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. N. Nazarian The Top 10 Reasons Musculoskeletal Sonography Is an Important Complementary or Alternative Technique to MRI Am. J. Roentgenol., June 1, 2008; 190(6): 1621 - 1626. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B. Weiss, J. A. Jacobson, and M. A. Karunakar The Use of Ultrasound in Evaluating Orthopaedic Trauma Patients J. Am. Acad. Ortho. Surg., December 1, 2005; 13(8): 525 - 533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Teefey, W. D. Middleton, W. T. Payne, and K. Yamaguchi Detection and Measurement of Rotator Cuff Tears with Sonography: Analysis of Diagnostic Errors Am. J. Roentgenol., June 1, 2005; 184(6): 1768 - 1773. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
