- Email: [email protected]
Magnetic Resonance Imaging in Evaluating Workers' Compensation Patients
SCIENTIFIC ARTICLE
Magnetic Resonance Imaging in Evaluating Workers’ Compensation Patients Daniel Babbel, MD, Ghazi Rayan, MD
Purpose We studied the utility of magnetic resonance imaging (MRI) studies for workers’ compensation patients with hand conditions in which the referring doctor obtained the images. We compared the MRI findings with the eventual clinical findings. We also investigated the approximate cost of these MRI studies. Methods We retrospectively reviewed the charts of all workers’ compensation patients seen in a hand and upper extremity practice over the course of 3 years. We selected patients who had MRI studies of the affected upper extremities before referral to the senior author (G.R.). We reviewed the charts for information regarding demographics, referral diagnoses, MRI diagnoses made by the radiologist, the area of the upper extremity studied, and eventual clinical diagnoses by the senior author. We made a determination as to whether a hand surgeon could have adequately diagnosed and treated the patients’ conditions without the imaging studies. We also investigated the cost associated with these MRIs. Results We included 62 patients with a total of 67 MRI scans in this study. The MRI studies did not contribute to clinically diagnosing the patients’ conditions in any of the cases we reviewed. The hand surgeon’s clinical diagnosis disagreed with the radiologist’s MRI diagnosis in 63% of patients. The MRI was unnecessary to arrive at the clinical diagnosis and did not influence the treatment offered for any of the 62 patients. The total cost for the 67 non-contrast MRI studies was approximately $53,000. Conclusions Costly imaging studies are frequently done to determine the validity of a patient’s reported problems; unfortunately, these tests are frequently unnecessary and waste resources. Magnetic resonance imaging scans may not be the standard for accurate diagnosis and can misdirect care. (J Hand Surg 2012;37A:811–815. Copyright © 2012 by the American Society for Surgery of the Hand. All rights reserved.) Type of study/level of evidence Therapeutic III. Key words MRI, upper extremity, workers’ compensation. frequently present to a hand surgeon for treatment of a variety of upper extremity disorders. According to the Survey of Occupational Injuries and Illnesses conducted by the U.S. Department of Labor, the upper extremity is the most common site of work-
W
ORKERS’ COMPENSATION PATIENTS
From the Hand Surgery Section, INTEGRIS Baptist Medical Center; and the Orthopedic Surgery Department, University of Oklahoma, Oklahoma City, OK; and Private Practice, Wichita Falls, TX.
related injuries.1 Many workers’ compensation patients are referred to hand surgeons after being evaluated by primary care physicians. The initial treating physician may order a magnetic resonance imaging (MRI) study as part of the patient’s evaluation. Magnetic resonance imaging can be a useful diagnostic tool in the evaluation No benefits in any form have been received or will be received related directly or indirectly to the subject of this article.
Received for publication August 25, 2011; accepted in revised form December 6, 2011.
Correspondingauthor:GhaziRayan,MD,3366N.W.Expressway,Suite700,PhysiciansBuildingD, Oklahoma City, OK 73132; e-mail: [email protected].
The authors thank Dr. David M. Thompson for statistical help and Annette Kezbers for administrative assistance.
0363-5023/12/37A04-0029$36.00/0 doi:10.1016/j.jhsa.2011.12.008
© ASSH 䉬 Published by Elsevier, Inc. All rights reserved. 䉬 811
812
MRI IN EVALUATING WORKERS’ COMPENSATION
of certain upper extremity conditions such as rotator cuff tears and soft tissue tumors. Upper extremity MRI findings can support a clinical diagnosis or yield new information that may not be detected by the clinical examination alone. The utility of the MRI in these settings may guide the hand surgeon toward providing appropriate treatment. Despite its benefits, MRI of the upper extremity has limitations, and its diagnostic accuracy has been questioned. For instance, a review article2 reported the diagnostic accuracy of MRI to be inconsistent in evaluating ulnar-sided wrist pain. In a study3 examining the wrists of asymptomatic volunteers, ganglions were identified in 51% of subjects. Another study of healthy asymptomatic subjects4 found 50% to have abnormally high triangular fibrocartilage complex signals. In a comparative study of MRI and arthroscopy for evaluating wrist pain without a specific diagnosis,5 the MRI was found to have the following diagnostic values for tears of the scapholunate ligament: sensitivity, 19%; specificity, 95%; positive predictive value, 71%; negative predictive value, 66%; and accuracy, 66%. A recent study assessed the reliability and validity of MRI in diagnosing lateral epicondylitis and found that the degree of tendinosis manifest on MRI has no correlation to the clinical symptoms.6 The lack of validity with clinical symptoms suggests that the results of MRI should be interpreted with caution in assessing lateral epicondylitis and should be used only as part of a comprehensive clinical assessment. Other studies also have questioned the utility of MRI as both a screening and diagnostic tool for shoulder symptoms compared with arthroscopic findings.7,8 These studies underscore the potential for overuse of pre-referral MRI and for treating hand surgeons to encounter discrepancies in MRI and clinical diagnoses of hand conditions. These issues may adversely alter patient management including those in workers’ compensation settings. Overreliance on MRI diagnoses may misdirect care and affect treatment outcomes. The purposes of this study were to determine the utility of pre-referral upper extremity MRI in helping the treating surgeon arrive at a clinical diagnosis and to decide on a course of treatment among a group of workers’ compensation patients who presented with upper extremity conditions. We determined the utility by comparing the MRI and eventual clinical diagnoses. In addition, we investigated the approximate costs of the MRI studies and the areas of the upper extremity for which they were most commonly obtained.
MATERIALS AND METHODS We retrospectively reviewed the charts of all workers’ compensation patients seen in a single hand and upper extremity practice over the course of 3 years (2008 – 2010). We examined each chart to determine whether the patients had an MRI of the affected extremity before referral. The inclusion criterion was patients who had no surgical intervention on the affected part of the upper extremity between the date of the MRI study and the referral to the treating hand surgeon and the availability of the official MRI report. We did not review actual images for this study. Patients who underwent at least 1 MRI evaluation of the upper extremity before referral were the subjects of this study. Patients were excluded if the MRI did not image the affected area. Clinical information We reviewed patients’ charts for information regarding demographics and diagnoses including the age, gender, mechanism of injury if any, chief problem at initial visit and referred preliminary diagnosis, definitive clinical diagnosis by the senior author (G.R.), treatment offered to the patient after clinical examination, and specialty of the physician ordering the MRI. MRI information We obtained information about the MRI diagnoses from radiologists’ reports and specific related information, if applicable, from each chart, including stated reason for ordering the MRI, anatomical area studied by the MRI, and MRI diagnosis or diagnoses. MRI cost To estimate the cost of the MRIs obtained, we contacted 4 local MRI providers and queried them about the total cost of obtaining a non-contrast upper extremity MRI study and of generating the radiologist’s report. We then averaged those figures to arrive at an estimated cost for all MRI studies obtained. MRI utility After extracting the needed information from charts, we evaluated the utility of each patient’s MRI by answering the following questions: (1) Was the MRI available for review at the initial visit? (2) Did the MRI and clinical diagnoses agree? (3) Did the MRI contribute to the hand surgeon’s arriving at a clinical diagnosis? (4) Did the MRI and referral diagnoses agree? (5) Was the MRI necessary for the hand surgeon to make the clinical diagnosis? (6) Did the MRI alter or influence the treatment prescribed?
JHS 䉬 Vol A, April
MRI IN EVALUATING WORKERS’ COMPENSATION
Only the senior author determined each patient’s clinical diagnosis after collecting a careful history and thorough upper extremity examination. In addition, the surgeon reviewed available radiographic studies, including each patient’s MRI report after clinical evaluation. Statistical analysis We used chi-square tests to compare subject groups on the proportion of patients for whom diagnoses were in concordance. We calculated exact confidence intervals on estimates of certain binomial proportions. In some cases, a yes/yes-but/no variable was merged into a yes/no variable so that the appropriate tests could be carried out. RESULTS Clinical information A total of 62 patients met the inclusion criteria and were the subject of this study. Of these, 34 were women, which was statistically consistent with equal gender groups. Mean age was 44 years (range, 23– 67 y). Twelve patients reported no accident or trauma, but rather repetitive use as the cause of the symptoms. The remaining 50 patients reported a specific action or accident that caused the symptoms. Within this group, a fall was the most common cause of injury, occurring in 14 patients. The remaining patients reported injuries from different mechanisms including twisting, motor vehicle accidents, and strikes by heavy objects. The most often reported problem at the time of evaluation by the hand surgeon was pain in the affected extremity. Other symptoms included stiffness, swelling, paresthesias, and weakness. Although the clinical diagnoses of the hand surgeon varied within this group of 62 patients, the most common diagnoses were triangular fibrocartilage complex injury, lateral epicondylitis, median nerve compression, scapholunate ligament injury, and trapeziometacarpal joint injury. The hand surgeon offered 8 patients surgery after evaluation. MRI information There were 67 MRI examinations used in evaluating this study group. The MRI studies were ordered by physician assistants and physicians including 23 orthopedic surgeons, 15 occupational medicine physicians, 12 hand surgeons other than the senior author, 4 family practice physicians, 2 physical medicine or rehabilitation physicians, and 1 internist. There were a total of 5 physician assis-
813
tants from different disciplines, including family practice, general medicine, and occupational medicine. Although a few of these studies were ordered by surgeons who treat hand and upper extremity conditions, all reports were generated by radiologists. For 48 patients, MRI studies were ordered primarily to evaluate pain. In the remaining 14 patients, other reasons for ordering an MRI were to rule out fractures, ensure fracture healing, and evaluate wrist sprains, thumb sprains, and lateral epicondylitis. Of the MRI studies, radiologists read 12 as normal. In 27 patients, radiologists gave a single MRI diagnosis. In the remaining 28 MRI reports, 2 to 6 different diagnoses were given. A few examples of multiple diagnoses were tendinosis, increased signal suggesting soft tissue inflammation, scapholunate ligament tears, abnormally high triangular fibrocartilage complex signals, and triangular fibrocartilage complex tears. The MRI studies included wrist (n ⫽ 37), elbow (n ⫽ 13), shoulder (n ⫽ 4), hand (n ⫽ 4), finger or thumb (n ⫽ 4), forearm (n ⫽ 3), cervical spine (n ⫽ 1), and brachial plexus (n ⫽ 1). MRI cost The average cost for each non-contrast MRI study of 1 anatomic area of the upper extremity including a radiologist report was approximately $787. We did not determine the cost of an MRI study excluding the radiologist fee. One patient had 3 MRI studies, 3 patients had 2 MRI studies, and the remaining patients each had 1 MRI study. The total cost for all these MRIs was estimated to be $53,000. MRI utility Only 16 patients brought their MRI images for review at the time of evaluation by the senior author. All patients brought the official MRI report without the actual or digital images. The senior author determined the diagnosis in all cases before consulting images or reports and did not modify the diagnosis after reviewing MRI reports in any case. He did not find that the MRI contributed to the diagnosis in any case. The MRI diagnosis disagreed with the senior author’s diagnosis in 39 patients (95% confidence interval, 51% to 75%). There were 23 cases of agreement between the MRI and the clinical diagnosis. In only 11 of the 23 cases of concordance was there complete agreement between the MRI and clinical diagnoses. In the remaining 12 cases there was only partial agreement. In 8 of these 12 cases, the MRI report listed several diagnoses, not every one of which was clinically
JHS 䉬 Vol A, April
814
MRI IN EVALUATING WORKERS’ COMPENSATION
concordant. There were also 3 cases of partial agreement in which the clinical examination established an additional diagnosis that was not part of the MRI report. As an example of only partial agreement, an MRI report stated that the patient had a tear of common flexor tendon and a partial tear of ulnar collateral ligament. The clinical diagnosis agreed with the finding of a partial tendon tear but did not confirm ulnar collateral ligament tear. The surgeon additionally diagnosed cubital tunnel syndrome. The proportion of cases in which the MRI and clinical diagnoses agreed did not differ when the referring physician was a hand surgeon (agreement in 4 of 12 cases) or not a hand surgeon (agreement in 19 of 50 cases) (P ⬎ .9999, Fisher exact test). We examined the referring diagnosis and its agreement with the MRI diagnosis. Because the referring physician’s notes were unavailable for 1 patient, we were able to compare agreement between the referring and the MRI diagnoses for 61 patients. There was agreement between the MRI and referring diagnoses in 32 cases (95% confidence interval, 40% to 65%). The referring diagnosis often lacked a specific pathology but described the symptoms. Examples were wrist pain, shoulder pain, hand stiffness, finger injury, wrist sprain, and soft tissue injury. When we separated the referral diagnoses for MRIs ordered by the 12 hand surgeons from those ordered by the 50 other referring professionals, we found a disparate concordance rate. For MRIs ordered by hand surgeons, the referring and MRI diagnosis agreed in 9 of 11 cases. For other referring professionals, the agreement rate of 23 of 50 cases was substantially lower (P ⫽ .046, Fisher exact test). When we compared acute versus chronic injury, the concordance rate of MRI diagnosis versus referring or clinical diagnosis did not change dramatically (P ⫽ .520, Fisher exact test). The MRI diagnosis was in agreement with the referring diagnosis in 50% of acute and 64% of chronic cases. There was agreement between the MRI and definitive clinical diagnosis in 40% of the acute cases and in 27% of the chronic cases. These differences did not meet statistical significance (P ⫽ .510, Fisher exact test). DISCUSSION This study determined the utility of obtaining upper extremity MRI studies among a group of workers’ compensation patients before referral to the treating hand surgeon. In addition, it assessed the agreement of the MRI diagnosis by the MRI reports in comparison with a definitive clinical diagnosis. We found that the
clinical diagnosis agreed with the MRIs diagnosis in only 37% of this group of workers’ compensation patients. This degree of agreement, although low, is misleading. This is because even when the MRI and clinical diagnoses agreed, in no patient was the MRI necessary to establish the patient’s definitive diagnosis. The MRI studies in this group of patients neither contributed to the diagnosis of the problem nor influenced the treatment prescribed. We found disparate concordance between the diagnoses for MRIs ordered by the 12 hand surgeons (82% agreement) and by the 50 other referring professionals (46% agreement). This suggests that the hand specialist has a tendency to order MRI studies for more specific reasons. Our results are in agreement with previous studies that questioned the utility of obtaining MRI study before to referral to a specialist for shoulder symptoms and for foot and ankle problems. Bradley et al8 evaluated 101 consecutive patients (104 shoulders) with chronic, atraumatic shoulder pain and found that routine pre-evaluation with MRI did not have a significant effect on the treatment or outcome. They concluded that MRI should not be used as a screening tool for atraumatic shoulder pain before completing a comprehensive clinical evaluation of the shoulder. Tocci et al9 reviewed 221 consecutive patients referred for treatment of lower extremity conditions. They found that 27 of 31 prereferral MRI scans were unnecessary and 15 of 31 yielded radiographic interpretations that were immaterial to the patient’s clinical diagnosis or inconsistent with the specialist’s interpretations. In contrast to our study, a recent article by Henn et al10 looked at 396 consecutive patients referred to a hand surgery practice. A total of 14 patients presented with an MRI. The authors found that of those MRIs, 10 were useful. They did state, however, that they did not exclude subsets of patients, and that excluding patients with acute problems may effectively decrease the reported usefulness of MRI. A recent article evaluating the use of MRI for low back pain in a Medicare population found that the acquisition of MRI equipment was strongly correlated with patients receiving MRI scans.11 We were not able to evaluate whether financial interest in an MRI scanner was a factor in obtaining a scan among the patients we studied. This factor should be considered when attempting to evaluate the utility and usefulness of MRI scan for all regions. One drawback of our study is its retrospective nature. In our patient population, however, we found that obtaining MRI studies before referral to a hand surgeon is not likely to be useful and may be a misuse of re-
JHS 䉬 Vol A, April
MRI IN EVALUATING WORKERS’ COMPENSATION
sources. It is also disconcerting that the total cost of these studies was approximately $53,000. In an era of seeking to reduce the cost of health care, this is a monetary resource that can be better used. In some instances, the workers’ compensation insurance carrier may require an objective test before authorizing treatment, which becomes a factor in increasing the cost of patient care. Our study suggests that in the population of workers’ compensation patients referred to a hand surgery practice, MRI studies ordered before referral are not likely to be useful for definitive diagnosis or treatment and may in fact be a waste of resources. Hence, this practice trend should be reconsidered. REFERENCES 1. Survey of Occupational Injuries and Illnesses. Counts, rates, and characteristics, 2007. Bureau of Labor Statistics, US Department of Labor; 2008. 2. Faber KJ, Iordache S, Grewal R. Magnetic resonance imaging for ulnar wrist pain. J Hand Surg 2010;35A:303–307. 3. Lowden CM, Attiah M, Garvin G, Macdermid JC, Osman S, Faber KJ. The prevalence of wrist ganglia in an asymptomatic population: magnetic resonance evaluation. J Hand Surg 2005;30B:302–306.
815
4. Sugimoto H, Shinozaki T, Ohsawa T. Triangular fibrocartilage in asymptomatic subjects: investigation of abnormal MR signal intensity. Radiology 1994;191:193–197. 5. Redeker J, Meyer-Marcotty M, Urbanek F, Hankiss J, Flügel M. [Diagnostic value of unspecific requested and implemented MRI for detecting intracarpal lesions, compared to arthroscopic findings at 217 patients][Article in German]. Handchir Mikrochir Plast Chir. 2009;41:129 –134. 6. Walton MJ, Mackie K, Fallon M, Butler R. The reliability and validity of magnetic resonance imaging in the assessment of chronic lateral epicondylitis. J Hand Surg 2011;36A:475– 479. 7. Torstensen ET, Hollinshead RM. Comparison of magnetic resonance imaging and arthroscopy in the evaluation of shoulder pathology. J Shoulder Elbow Surg 1999;8:42– 45. 8. Bradley MP, Tung G, Green A. Overutilization of shoulder magnetic resonance imaging as a diagnostic screening tool in patients with chronic shoulder pain. J Shoulder Elbow Surg 2005;14:233–237. 9. Tocci SL, Madom IA, Bradley MP, Langer PR, DiGiovanni CW. The diagnostic value of MRI in foot and ankle surgery. Foot Ankle Int 2007;28:166 –168. 10. Henn CM, Weiss AP, Akelman E. Utility of magnetic resonance imaging obtained before evaluation by the hand surgeon. J Surg Orthop Adv 2010;19:159 –161. 11. Shreibati JB, Baker LC. The relationship between low back magnetic resonance imaging, surgery, and spending: impact of physician selfreferral status. Health Serv Res 2011;46:1362–1381.
JHS 䉬 Vol A, April
Magnetic Resonance Imaging in Evaluating Workers’ Compensation Patients Daniel Babbel, MD, Ghazi Rayan, MD
Purpose We studied the utility of magnetic resonance imaging (MRI) studies for workers’ compensation patients with hand conditions in which the referring doctor obtained the images. We compared the MRI findings with the eventual clinical findings. We also investigated the approximate cost of these MRI studies. Methods We retrospectively reviewed the charts of all workers’ compensation patients seen in a hand and upper extremity practice over the course of 3 years. We selected patients who had MRI studies of the affected upper extremities before referral to the senior author (G.R.). We reviewed the charts for information regarding demographics, referral diagnoses, MRI diagnoses made by the radiologist, the area of the upper extremity studied, and eventual clinical diagnoses by the senior author. We made a determination as to whether a hand surgeon could have adequately diagnosed and treated the patients’ conditions without the imaging studies. We also investigated the cost associated with these MRIs. Results We included 62 patients with a total of 67 MRI scans in this study. The MRI studies did not contribute to clinically diagnosing the patients’ conditions in any of the cases we reviewed. The hand surgeon’s clinical diagnosis disagreed with the radiologist’s MRI diagnosis in 63% of patients. The MRI was unnecessary to arrive at the clinical diagnosis and did not influence the treatment offered for any of the 62 patients. The total cost for the 67 non-contrast MRI studies was approximately $53,000. Conclusions Costly imaging studies are frequently done to determine the validity of a patient’s reported problems; unfortunately, these tests are frequently unnecessary and waste resources. Magnetic resonance imaging scans may not be the standard for accurate diagnosis and can misdirect care. (J Hand Surg 2012;37A:811–815. Copyright © 2012 by the American Society for Surgery of the Hand. All rights reserved.) Type of study/level of evidence Therapeutic III. Key words MRI, upper extremity, workers’ compensation. frequently present to a hand surgeon for treatment of a variety of upper extremity disorders. According to the Survey of Occupational Injuries and Illnesses conducted by the U.S. Department of Labor, the upper extremity is the most common site of work-
W
ORKERS’ COMPENSATION PATIENTS
From the Hand Surgery Section, INTEGRIS Baptist Medical Center; and the Orthopedic Surgery Department, University of Oklahoma, Oklahoma City, OK; and Private Practice, Wichita Falls, TX.
related injuries.1 Many workers’ compensation patients are referred to hand surgeons after being evaluated by primary care physicians. The initial treating physician may order a magnetic resonance imaging (MRI) study as part of the patient’s evaluation. Magnetic resonance imaging can be a useful diagnostic tool in the evaluation No benefits in any form have been received or will be received related directly or indirectly to the subject of this article.
Received for publication August 25, 2011; accepted in revised form December 6, 2011.
Correspondingauthor:GhaziRayan,MD,3366N.W.Expressway,Suite700,PhysiciansBuildingD, Oklahoma City, OK 73132; e-mail: [email protected].
The authors thank Dr. David M. Thompson for statistical help and Annette Kezbers for administrative assistance.
0363-5023/12/37A04-0029$36.00/0 doi:10.1016/j.jhsa.2011.12.008
© ASSH 䉬 Published by Elsevier, Inc. All rights reserved. 䉬 811
812
MRI IN EVALUATING WORKERS’ COMPENSATION
of certain upper extremity conditions such as rotator cuff tears and soft tissue tumors. Upper extremity MRI findings can support a clinical diagnosis or yield new information that may not be detected by the clinical examination alone. The utility of the MRI in these settings may guide the hand surgeon toward providing appropriate treatment. Despite its benefits, MRI of the upper extremity has limitations, and its diagnostic accuracy has been questioned. For instance, a review article2 reported the diagnostic accuracy of MRI to be inconsistent in evaluating ulnar-sided wrist pain. In a study3 examining the wrists of asymptomatic volunteers, ganglions were identified in 51% of subjects. Another study of healthy asymptomatic subjects4 found 50% to have abnormally high triangular fibrocartilage complex signals. In a comparative study of MRI and arthroscopy for evaluating wrist pain without a specific diagnosis,5 the MRI was found to have the following diagnostic values for tears of the scapholunate ligament: sensitivity, 19%; specificity, 95%; positive predictive value, 71%; negative predictive value, 66%; and accuracy, 66%. A recent study assessed the reliability and validity of MRI in diagnosing lateral epicondylitis and found that the degree of tendinosis manifest on MRI has no correlation to the clinical symptoms.6 The lack of validity with clinical symptoms suggests that the results of MRI should be interpreted with caution in assessing lateral epicondylitis and should be used only as part of a comprehensive clinical assessment. Other studies also have questioned the utility of MRI as both a screening and diagnostic tool for shoulder symptoms compared with arthroscopic findings.7,8 These studies underscore the potential for overuse of pre-referral MRI and for treating hand surgeons to encounter discrepancies in MRI and clinical diagnoses of hand conditions. These issues may adversely alter patient management including those in workers’ compensation settings. Overreliance on MRI diagnoses may misdirect care and affect treatment outcomes. The purposes of this study were to determine the utility of pre-referral upper extremity MRI in helping the treating surgeon arrive at a clinical diagnosis and to decide on a course of treatment among a group of workers’ compensation patients who presented with upper extremity conditions. We determined the utility by comparing the MRI and eventual clinical diagnoses. In addition, we investigated the approximate costs of the MRI studies and the areas of the upper extremity for which they were most commonly obtained.
MATERIALS AND METHODS We retrospectively reviewed the charts of all workers’ compensation patients seen in a single hand and upper extremity practice over the course of 3 years (2008 – 2010). We examined each chart to determine whether the patients had an MRI of the affected extremity before referral. The inclusion criterion was patients who had no surgical intervention on the affected part of the upper extremity between the date of the MRI study and the referral to the treating hand surgeon and the availability of the official MRI report. We did not review actual images for this study. Patients who underwent at least 1 MRI evaluation of the upper extremity before referral were the subjects of this study. Patients were excluded if the MRI did not image the affected area. Clinical information We reviewed patients’ charts for information regarding demographics and diagnoses including the age, gender, mechanism of injury if any, chief problem at initial visit and referred preliminary diagnosis, definitive clinical diagnosis by the senior author (G.R.), treatment offered to the patient after clinical examination, and specialty of the physician ordering the MRI. MRI information We obtained information about the MRI diagnoses from radiologists’ reports and specific related information, if applicable, from each chart, including stated reason for ordering the MRI, anatomical area studied by the MRI, and MRI diagnosis or diagnoses. MRI cost To estimate the cost of the MRIs obtained, we contacted 4 local MRI providers and queried them about the total cost of obtaining a non-contrast upper extremity MRI study and of generating the radiologist’s report. We then averaged those figures to arrive at an estimated cost for all MRI studies obtained. MRI utility After extracting the needed information from charts, we evaluated the utility of each patient’s MRI by answering the following questions: (1) Was the MRI available for review at the initial visit? (2) Did the MRI and clinical diagnoses agree? (3) Did the MRI contribute to the hand surgeon’s arriving at a clinical diagnosis? (4) Did the MRI and referral diagnoses agree? (5) Was the MRI necessary for the hand surgeon to make the clinical diagnosis? (6) Did the MRI alter or influence the treatment prescribed?
JHS 䉬 Vol A, April
MRI IN EVALUATING WORKERS’ COMPENSATION
Only the senior author determined each patient’s clinical diagnosis after collecting a careful history and thorough upper extremity examination. In addition, the surgeon reviewed available radiographic studies, including each patient’s MRI report after clinical evaluation. Statistical analysis We used chi-square tests to compare subject groups on the proportion of patients for whom diagnoses were in concordance. We calculated exact confidence intervals on estimates of certain binomial proportions. In some cases, a yes/yes-but/no variable was merged into a yes/no variable so that the appropriate tests could be carried out. RESULTS Clinical information A total of 62 patients met the inclusion criteria and were the subject of this study. Of these, 34 were women, which was statistically consistent with equal gender groups. Mean age was 44 years (range, 23– 67 y). Twelve patients reported no accident or trauma, but rather repetitive use as the cause of the symptoms. The remaining 50 patients reported a specific action or accident that caused the symptoms. Within this group, a fall was the most common cause of injury, occurring in 14 patients. The remaining patients reported injuries from different mechanisms including twisting, motor vehicle accidents, and strikes by heavy objects. The most often reported problem at the time of evaluation by the hand surgeon was pain in the affected extremity. Other symptoms included stiffness, swelling, paresthesias, and weakness. Although the clinical diagnoses of the hand surgeon varied within this group of 62 patients, the most common diagnoses were triangular fibrocartilage complex injury, lateral epicondylitis, median nerve compression, scapholunate ligament injury, and trapeziometacarpal joint injury. The hand surgeon offered 8 patients surgery after evaluation. MRI information There were 67 MRI examinations used in evaluating this study group. The MRI studies were ordered by physician assistants and physicians including 23 orthopedic surgeons, 15 occupational medicine physicians, 12 hand surgeons other than the senior author, 4 family practice physicians, 2 physical medicine or rehabilitation physicians, and 1 internist. There were a total of 5 physician assis-
813
tants from different disciplines, including family practice, general medicine, and occupational medicine. Although a few of these studies were ordered by surgeons who treat hand and upper extremity conditions, all reports were generated by radiologists. For 48 patients, MRI studies were ordered primarily to evaluate pain. In the remaining 14 patients, other reasons for ordering an MRI were to rule out fractures, ensure fracture healing, and evaluate wrist sprains, thumb sprains, and lateral epicondylitis. Of the MRI studies, radiologists read 12 as normal. In 27 patients, radiologists gave a single MRI diagnosis. In the remaining 28 MRI reports, 2 to 6 different diagnoses were given. A few examples of multiple diagnoses were tendinosis, increased signal suggesting soft tissue inflammation, scapholunate ligament tears, abnormally high triangular fibrocartilage complex signals, and triangular fibrocartilage complex tears. The MRI studies included wrist (n ⫽ 37), elbow (n ⫽ 13), shoulder (n ⫽ 4), hand (n ⫽ 4), finger or thumb (n ⫽ 4), forearm (n ⫽ 3), cervical spine (n ⫽ 1), and brachial plexus (n ⫽ 1). MRI cost The average cost for each non-contrast MRI study of 1 anatomic area of the upper extremity including a radiologist report was approximately $787. We did not determine the cost of an MRI study excluding the radiologist fee. One patient had 3 MRI studies, 3 patients had 2 MRI studies, and the remaining patients each had 1 MRI study. The total cost for all these MRIs was estimated to be $53,000. MRI utility Only 16 patients brought their MRI images for review at the time of evaluation by the senior author. All patients brought the official MRI report without the actual or digital images. The senior author determined the diagnosis in all cases before consulting images or reports and did not modify the diagnosis after reviewing MRI reports in any case. He did not find that the MRI contributed to the diagnosis in any case. The MRI diagnosis disagreed with the senior author’s diagnosis in 39 patients (95% confidence interval, 51% to 75%). There were 23 cases of agreement between the MRI and the clinical diagnosis. In only 11 of the 23 cases of concordance was there complete agreement between the MRI and clinical diagnoses. In the remaining 12 cases there was only partial agreement. In 8 of these 12 cases, the MRI report listed several diagnoses, not every one of which was clinically
JHS 䉬 Vol A, April
814
MRI IN EVALUATING WORKERS’ COMPENSATION
concordant. There were also 3 cases of partial agreement in which the clinical examination established an additional diagnosis that was not part of the MRI report. As an example of only partial agreement, an MRI report stated that the patient had a tear of common flexor tendon and a partial tear of ulnar collateral ligament. The clinical diagnosis agreed with the finding of a partial tendon tear but did not confirm ulnar collateral ligament tear. The surgeon additionally diagnosed cubital tunnel syndrome. The proportion of cases in which the MRI and clinical diagnoses agreed did not differ when the referring physician was a hand surgeon (agreement in 4 of 12 cases) or not a hand surgeon (agreement in 19 of 50 cases) (P ⬎ .9999, Fisher exact test). We examined the referring diagnosis and its agreement with the MRI diagnosis. Because the referring physician’s notes were unavailable for 1 patient, we were able to compare agreement between the referring and the MRI diagnoses for 61 patients. There was agreement between the MRI and referring diagnoses in 32 cases (95% confidence interval, 40% to 65%). The referring diagnosis often lacked a specific pathology but described the symptoms. Examples were wrist pain, shoulder pain, hand stiffness, finger injury, wrist sprain, and soft tissue injury. When we separated the referral diagnoses for MRIs ordered by the 12 hand surgeons from those ordered by the 50 other referring professionals, we found a disparate concordance rate. For MRIs ordered by hand surgeons, the referring and MRI diagnosis agreed in 9 of 11 cases. For other referring professionals, the agreement rate of 23 of 50 cases was substantially lower (P ⫽ .046, Fisher exact test). When we compared acute versus chronic injury, the concordance rate of MRI diagnosis versus referring or clinical diagnosis did not change dramatically (P ⫽ .520, Fisher exact test). The MRI diagnosis was in agreement with the referring diagnosis in 50% of acute and 64% of chronic cases. There was agreement between the MRI and definitive clinical diagnosis in 40% of the acute cases and in 27% of the chronic cases. These differences did not meet statistical significance (P ⫽ .510, Fisher exact test). DISCUSSION This study determined the utility of obtaining upper extremity MRI studies among a group of workers’ compensation patients before referral to the treating hand surgeon. In addition, it assessed the agreement of the MRI diagnosis by the MRI reports in comparison with a definitive clinical diagnosis. We found that the
clinical diagnosis agreed with the MRIs diagnosis in only 37% of this group of workers’ compensation patients. This degree of agreement, although low, is misleading. This is because even when the MRI and clinical diagnoses agreed, in no patient was the MRI necessary to establish the patient’s definitive diagnosis. The MRI studies in this group of patients neither contributed to the diagnosis of the problem nor influenced the treatment prescribed. We found disparate concordance between the diagnoses for MRIs ordered by the 12 hand surgeons (82% agreement) and by the 50 other referring professionals (46% agreement). This suggests that the hand specialist has a tendency to order MRI studies for more specific reasons. Our results are in agreement with previous studies that questioned the utility of obtaining MRI study before to referral to a specialist for shoulder symptoms and for foot and ankle problems. Bradley et al8 evaluated 101 consecutive patients (104 shoulders) with chronic, atraumatic shoulder pain and found that routine pre-evaluation with MRI did not have a significant effect on the treatment or outcome. They concluded that MRI should not be used as a screening tool for atraumatic shoulder pain before completing a comprehensive clinical evaluation of the shoulder. Tocci et al9 reviewed 221 consecutive patients referred for treatment of lower extremity conditions. They found that 27 of 31 prereferral MRI scans were unnecessary and 15 of 31 yielded radiographic interpretations that were immaterial to the patient’s clinical diagnosis or inconsistent with the specialist’s interpretations. In contrast to our study, a recent article by Henn et al10 looked at 396 consecutive patients referred to a hand surgery practice. A total of 14 patients presented with an MRI. The authors found that of those MRIs, 10 were useful. They did state, however, that they did not exclude subsets of patients, and that excluding patients with acute problems may effectively decrease the reported usefulness of MRI. A recent article evaluating the use of MRI for low back pain in a Medicare population found that the acquisition of MRI equipment was strongly correlated with patients receiving MRI scans.11 We were not able to evaluate whether financial interest in an MRI scanner was a factor in obtaining a scan among the patients we studied. This factor should be considered when attempting to evaluate the utility and usefulness of MRI scan for all regions. One drawback of our study is its retrospective nature. In our patient population, however, we found that obtaining MRI studies before referral to a hand surgeon is not likely to be useful and may be a misuse of re-
JHS 䉬 Vol A, April
MRI IN EVALUATING WORKERS’ COMPENSATION
sources. It is also disconcerting that the total cost of these studies was approximately $53,000. In an era of seeking to reduce the cost of health care, this is a monetary resource that can be better used. In some instances, the workers’ compensation insurance carrier may require an objective test before authorizing treatment, which becomes a factor in increasing the cost of patient care. Our study suggests that in the population of workers’ compensation patients referred to a hand surgery practice, MRI studies ordered before referral are not likely to be useful for definitive diagnosis or treatment and may in fact be a waste of resources. Hence, this practice trend should be reconsidered. REFERENCES 1. Survey of Occupational Injuries and Illnesses. Counts, rates, and characteristics, 2007. Bureau of Labor Statistics, US Department of Labor; 2008. 2. Faber KJ, Iordache S, Grewal R. Magnetic resonance imaging for ulnar wrist pain. J Hand Surg 2010;35A:303–307. 3. Lowden CM, Attiah M, Garvin G, Macdermid JC, Osman S, Faber KJ. The prevalence of wrist ganglia in an asymptomatic population: magnetic resonance evaluation. J Hand Surg 2005;30B:302–306.
815
4. Sugimoto H, Shinozaki T, Ohsawa T. Triangular fibrocartilage in asymptomatic subjects: investigation of abnormal MR signal intensity. Radiology 1994;191:193–197. 5. Redeker J, Meyer-Marcotty M, Urbanek F, Hankiss J, Flügel M. [Diagnostic value of unspecific requested and implemented MRI for detecting intracarpal lesions, compared to arthroscopic findings at 217 patients][Article in German]. Handchir Mikrochir Plast Chir. 2009;41:129 –134. 6. Walton MJ, Mackie K, Fallon M, Butler R. The reliability and validity of magnetic resonance imaging in the assessment of chronic lateral epicondylitis. J Hand Surg 2011;36A:475– 479. 7. Torstensen ET, Hollinshead RM. Comparison of magnetic resonance imaging and arthroscopy in the evaluation of shoulder pathology. J Shoulder Elbow Surg 1999;8:42– 45. 8. Bradley MP, Tung G, Green A. Overutilization of shoulder magnetic resonance imaging as a diagnostic screening tool in patients with chronic shoulder pain. J Shoulder Elbow Surg 2005;14:233–237. 9. Tocci SL, Madom IA, Bradley MP, Langer PR, DiGiovanni CW. The diagnostic value of MRI in foot and ankle surgery. Foot Ankle Int 2007;28:166 –168. 10. Henn CM, Weiss AP, Akelman E. Utility of magnetic resonance imaging obtained before evaluation by the hand surgeon. J Surg Orthop Adv 2010;19:159 –161. 11. Shreibati JB, Baker LC. The relationship between low back magnetic resonance imaging, surgery, and spending: impact of physician selfreferral status. Health Serv Res 2011;46:1362–1381.
JHS 䉬 Vol A, April