Prospective evaluation of clinical assessment in the diagnosis and treatment of clavicle fracture: Are radiographs really necessary?

EM Advances

CJEM 2003;5(5):309-313

Abstract

Résumé

Introduction | Methods | Results | Discussion | Conclusions | References

Introduction

Radiography for known or suspected fractures and dislocations has, until recently, been routine. Over the last decade, studies have shown that routine radiographs for ankle injuries, knee injuries, C-spine injuries and (pre-reduction) shoulder dislocations are unnecessary. In place of indiscriminate radiography for all such injuries, clinical decision rules or guidelines have been established for determining when radiography will provide additional, relevant information to the clinical picture.^1-5 This rational approach to ordering radiographs selectively has resulted in a substantial reduction in unnecessary radiographs, with concomitant savings in costs and resources and, in some cases, reduction in pain and inconvenience to the patient.

Radiographs for suspected clavicle fractures may be performed first to identify the presence of a fracture, then to demonstrate the characteristics of the fracture that might affect treatment decisions. But if physicians can accurately determine by clinical examination alone that the clavicle is fractured, then a radiograph is necessary only if treatment decisions are changed by the particular characteristics of a given fracture. In a pilot survey on the use of radiography in suspected clavicle fractures, we found that 5 of 6 physicians indicated that they "never" or "rarely" omit clavicle radiography, even when they are certain of the diagnosis and certain that the radiographs will not affect the treatment. The most common reasons cited for performing radiographs were perceived patient expectations or "standard of care."

The primary purpose of this study was to determine whether emergency physicians can, by clinical exam alone, accurately diagnose clavicle fractures and differentiate middle-third fractures from medial- or lateral-third fractures. Our secondary purpose was to determine which, if any, clavicle injuries emergency physicians are comfortable managing without radiography.

Methods

Setting and patients

This prospective observational study was conducted at Mineral Springs Hospital, a rural community hospital that treats 15 000 patients annually and is staffed by 6 full-time specialty-trained emergency physicians, each of whom has been in practice for more than 5 years. The hospital is located in Banff, Alta., a mountain community of 7760 inside a national park. Because of its location, many people visit Banff to pursue outdoor recreational activities.

From Nov. 1, 2001, to Apr. 30, 2002, a convenience sample of patients presenting with isolated possible clavicle fractures were identified by the triage nurse. Patients were excluded if they had injuries involving non-musculoskeletal organ systems or if the treating emergency physician determined on initial exam that they "certainly" did not have a clavicle fracture.

Data collection

A data collection form was initiated by the triage nurse at the time of initial patient contact and completed by the attending emergency physician at the time of patient assessment. After evaluating the patient, and before obtaining clavicle radiographs, the physician documented clinical certainty of diagnosis on a 10-cm visual analogue scale (VAS) anchored by the terms "certain fracture" and "certain no fracture." When they were certain of fracture (i.e., a score of 10 on the VAS), physicians described the location of the fracture, the nature of the fracture and their hypothetical comfort in treating the injury without radiography. Following radiography, physicians documented the presence and location of any fractures, and whether the radiographs altered their treatment plan. Emergency physicians' radiographic assessments were compared to the gold standard of the radiologist's dictated report. Patient management was left to the discretion of the treating physician. The study was approved by the provincial Research Ethics Committee, which deemed consent not to be required.

Data analysis

Data were entered into a Microsoft Excel 2000 spreadsheet, and Excel was used to generate all descriptive statistics. Binomial 95% confidence intervals (CIs) for proportions were calculated using Stata (Version 5.0 MacIntosh).

Results

During the study period, 77 patients were identified as possibly having an isolated clavicle fracture and were enrolled in the study. After study completion, a hospital records search located 20 patients with clavicle fractures who had not been enrolled, and comparative data were collected on this group. Baseline characteristics were similar, and a medical record review did not reveal any significant differences between the unenrolled and enrolled groups. To our knowledge, no unenrolled patient was mistakenly excluded because the emergency physician had misdiagnosed the injury.

Clinical accuracy

Table 1 outlines the characteristics of the study population, and Table 2 summarizes the diagnostic accuracy of the physicians' clinical assessments. In 51 of 77 enrolled patients (66%; 95% CI, 54.6%-76.6%), the treating physician was certain of the diagnosis of clavicle fracture (i.e., 10/10 certainty on the VAS scale) prior to radiography. In these 51 cases, radiography revealed a fracture in 50 cases (98.0%; 95% CI, 89.6%-99.9%). The single incorrect diagnosis was in a patient with pain, tenderness and swelling over the medial third of the clavicle, which was subsequently diagnosed as a sternoclavicular sprain. Clinical judgement was also highly accurate for fracture location: the physicians were 100% accurate for 4 fractures clinically identified as lateral-third fractures (95% CI, 39.7%-100%) and for 41 fractures identified as middle-third fractures (95% CI, 91.4%-100%). They were correct on only 1 of 5 injuries (20%; 95% CI, 1%-72%) they clinically identified as medial-third fractures. Of these, 3 were actually middle-third fractures, 1 was a medial-third fracture and 1 was a soft-tissue injury. When asked about their ability to clinically identify fracture comminution and wide displacement, physicians were uncertain in 30 of 51 cases (59%) and in 23 of 51 cases (45%) respectively.

Impact of imaging

Table 3 illustrates the anticipated and actual influence of radiography on patient management. Physicians expected that radiography would affect their treatment in 13 of 51 fractures, including 11 of the 42 middle-third fractures. In fact, radiography helped guide treatment in 9 of these 13 cases, and in 1 case when they thought it would not affect management. The latter involved a patient whose clinical diagnosis was a medial-third clavicular fracture and whose x-ray was normal. Radiography did not lead to any unexpected management changes in patients with middle- or lateral-third fractures.

Physicians reported that they would be comfortable treating without radiographs in 17 of 51 patients in whom they were certain of diagnosis. This included 15 of the 42 patients in whom middle-third fracture was diagnosed clinically. In the 27 patients with middle-third fractures where physicians indicated they would be uncomfortable treating without a radiograph, reasons for discomfort included concerns about fracture displacement (n = 12), patient expectations for imaging (n = 6), degree of comminution (n = 5), the perceived standard of care (n = 4), the possibility of pneumothorax (n = 4), the need to clarify fracture location (n = 2) and other (n = 6). Note that, in several cases, physicians expressed more than one cause for concern.

Discussion

Clavicle fractures are common injuries accounting for approximately 5% of all fractures,⁶ 13% of injuries in mountain bikers,⁷ 11% of shoulder girdle injuries in skiers,⁸ and 10% of injuries in in-line skaters, roller skaters and skateboarders.⁹ Our study demonstrates that experienced emergency physicians are frequently "certain" of the diagnosis by clinical examination alone and that, when they are certain, their clinical assessment is highly accurate. Our physicians were also accurate in their determination of fracture location for lateral- (4/4) and middle-third (41/41) fractures; however, in 5 cases when physicians clinically identified medial-third fractures, they were correct only once. Despite high levels of certainty and accuracy, physicians frequently reported that they would be uncomfortable managing patients without radiography.

A recent pilot study also concluded that emergency physicians could accurately predict the presence and location of clavicle fractures,¹⁰ but this study did not determine whether the physicians were confident enough to treat without radiography or under what circumstances they might consider doing so. In the current study, physicians were less than "certain" about the presence of fracture in 26 (34%) of 77 patients and, in uncertain cases, only 11 patients (42%) had fractures. These data suggest that if imaging was limited to cases of diagnostic uncertainty, there is potential to reduce the proportion of patients undergoing radiography by 66%.

Treatment of clavicle fractures

Few controlled trials have addressed the treatment of clavicle fractures, and none have controlled for fracture character (i.e., degree of displacement), fracture site (dominant v. non-dominant limb) and patient characteristics (e.g., age, fitness, activity level).^11-15 There is no good evidence that manipulation, operative intervention or other aggressive modalities provide better functional or cosmetic outcomes than simple arm support alone.^6,16

Eighty percent of clavicle fractures involve the middle third,¹⁶ and treatment of middle-third fractures typically consists of rest and the use of ice and a sling, although some physicians may still use a figure-of-8 bandage or "clavicle strap" and some may attempt reduction for selected cases. Surgical intervention is rarely indicated, even when there are multiple or widely displaced fragments. Lateral and medial fractures, accounting for 15% and 5% of clavicle fractures respectively, have a higher incidence of complications and occasionally require surgical intervention;¹⁶ although even with these injuries, there are no clear guidelines indicating when to recommend rest, when to attempt reduction or when to surgically intervene.

X-ray utilization

Our data suggest that, in many cases, radiography adds little to clinical evaluation, but confirmation of diagnosis may not be the only reason to obtain x-rays. Radiographs may be helpful if they provide information that influences treatment or if they enable the physician to advise the patient regarding prognosis. Although our physicians were accurate in identifying and localizing fractures, they were uncertain about the degree of comminution in 30 of 51 cases (59%), and about the degree of displacement in 23 of 51 cases (45%). In these cases, radiographs clearly eliminated uncertainty and might have influenced treatment choices, but it is important to point out that radiographic appearance has not been shown to relate to prognosis.

In 27 (64%) of 42 middle-third fractures, physicians stated they were uncomfortable treating without a radiograph. Discomfort is understandable if radiographic findings are likely to influence treatment, but even in the 31 cases (74%) when physicians believed radiography would not affect treatment, they stated they would have been comfortable omitting radiography in only 15 (48%). Of note is the fact that radiographs altered treatment in 0 of 31 middle-third fractures when the physician expected that they would not. In 37% of cases, our physicians identified "patient expectation" or "standard of care" as obstacles to eliminating potentially unnecessary radiographs. These concerns could be overcome with appropriate evidence or decision rules, as has been demonstrated in the reduced use of radiographic imaging for ankle injuries.¹

Lack of certainty about the character of the fracture, coupled with uncertainty regarding best treatment, undoubtedly plays a role in physicians' discomfort in omitting a radiograph. It is possible that if the issue of appropriate treatment could be clarified, physicians might then feel confident that radiography of the fracture was unnecessary in selected cases.

Limitations

Our emergency department population includes a high proportion of fit young adults with sports injuries, and our data are most generalizable to similar patients. The emergency physicians who participated in this study have all practised for more than 5 years in this setting and may be unusually skilled in the clinical assessment of musculoskeletal trauma. Physicians who see fewer such injuries may be less comfortable and less able to identify and localize clavicle fractures by clinical exam alone. These limitations may affect the generalizability of our results.

Conclusions

This study shows that experienced emergency physicians can accurately identify and localize clavicle fractures, and that routine radiography (particularly for obvious middle-third fractures) is unlikely to improve diagnostic accuracy, treatment decisions or patient outcome. Despite this, physicians are often uncomfortable omitting radiographs for patients with clavicle fractures.

Study authorship: Dr. Shuster proposed the study. All authors contributed to the study design. Drs. Shuster, Boyd, Gauthier, Mergler, Shepherd and Turner collected the data. Drs. Shuster and Abu-Laban interpreted the data, and Dr. Abu-Laban performed the statistical analysis. Dr. Shuster wrote the initial draft of the manuscript, and Drs. Shuster, Abu-Laban and Boyd participated in revising the manuscript.

References

1. Stiell IG, Greenberg GH, McKnight RD, Nair RC, McDowell I, Reardon M, et al. Decision rules for the use of radiography in acute ankle injuries. Refinement and prospective validation. JAMA 1993;269:1127-32.
2. Stiell IG, Greenberg GH, Wells GA, McDowell I, Cwinn AA, Smith NA, et al. Prospective validation of a decision rule for the use of radiography in acute knee injuries. JAMA 1996;275:611-5.
3. Stiell IG, Wells GA, Vandemheen K, Laupacis A, Brison R, Eisenhauer MA, et al. Variation in emergency department use of cervical spine radiography for alert, stable trauma patients. CMAJ 1997;156:1537-44.
4. Shuster M, Abu-Laban RB, Boyd J. Prereduction radiographs in clinically evident anterior shoulder dislocation. Am J Emerg Med 1999;17:653-8.
5. Shuster M, Abu-Laban RB, Boyd J, Gauthier C, Shepherd L, Turner C. Prospective evaluation of a guideline for the selective elimination of pre-reduction radiographs in the clinically obvious anterior shoulder dislocation. Can J Emerg Med 2002;4(4):257-62.
6. Marx JA, Hockberger RS, Walls RM, editors. Rosen's Emergency medicine: concepts and clinical practice, 5th ed. Chicago: Mosby-Year Book, Inc; 2002. p. 581-4.
7. Jeys LM, Cribb G, Toms AD, Hay SM. Mountain biking injuries in rural England. Br J Sports Med 2001;35:197-9.
8. Kocher MS, Feagin JA Jr. Shoulder injuries during alpine skiing. Am J Sports Med 1996;24:665-9.
9. Orenstein JB. Injuries and small-wheel skates. Ann Emerg Med 1996;27:204-9.
10. Landine J, McGraw R, Pickett W. Clinical diagnosis of clavicle fractures: a pilot study. Can J Emerg Med 2001;3(2):95-8.
11. Anderson K, Jensen PO, Lauritzen J. Treatment of clavicular fractures: figure-of-eight bandage vs. a simple sling. Acta Orthop Scand 1987;57:71-4.
12. McCandless DN, Mowbray M. Treatment of displaced fractures of the clavicle. sling vs. figure-of-eight bandage. Practitioner 1979;223:266-7.
13. Stanley D, Norris SH. Recovery following fractures of the clavicle treated conservatively. Injury 1988;19:162-4.
14. Oroko PK, Buchan M, Winkler A, Kelly IG. Does shortening matter after clavicular fractures? Bull Hosp Jt Dis 1999;58:6-8.
15. Nordqvist A, Redlund-Johnell I, von Scheele A, Petersson CJ. Shortening of clavicle after fracture. Incidence and clinical significance, a 5-year follow-up of 85 patients. Acta Orthop Scand 1997;68:349-51.
16. Rockwood CG. Fractures of the clavicle. In: EV Craig, editor. Rockwood & Green's Fractures in adults, 4th ed, Philadelphia: Lippincott-Raven; 1996. p. 1109-61.