| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Evidence in Practice |
Britt Smith, PT, MSPT, OCS, FAAOMPT, is a private practice physical therapist, S.O.A.R. Physical Therapy, Grand Junction, Colo
Joshua A Cleland, PT, DPT, OCS, is Assistant Professor, Physical Therapy Program, Franklin Pierce College; Physical Therapist, Rehabilitation Services of Concord Hospital, Concord, NH; and Fellow, Manual Therapy Program, Regis University, Denver, Colo
| The purpose of "Evidence in Practice" is to illustrate the literature search process to obtain evidence that can guide clinical decision making. This article is not a case report. The examination, evaluation, and intervention sections are purposely abbreviated.
|
The father of a 9-year-old girl contacted our clinic to schedule an examination. He reported that his daughter injured her left knee during gymnastics practice the previous evening. He stated that, while attempting to perform a forward flip maneuver, she heard a "pop" in her left knee as she initiated the jump. According to witnesses, the girl landed awkwardly on her heels, with her knees bent, and she subsequently fell backward into a sitting position. The girl reported immediate anterior left knee pain following the maneuver. Because of her pain, she had been unable to walk since the time of the injury. Her father spoke to a physical therapist (Smith) at our clinic to relay this information and to express urgency in the case. The patient had not seen another medical professional for the injury. The father was a former client at our clinic and, therefore, sought our opinion about his daughter's injury.
The girl had not yet undergone an examination or diagnostic imaging; therefore, we would be serving as a point of entry to the health care system. We planned to assess the extent of the injury in order to determine whether referral to another medical specialist was warranted. The patient's history of trauma during a high-velocity maneuver (ie, forward flip) and her inability to walk the following day indicated the possibility of a fracture, a growth plate injury, or some other serious injury (eg, ligamentous or meniscal injury). The prevalence of ligamentous or meniscal injury in knee trauma is much larger than the prevalence of fractures; less than 7% of traumatic injuries to the knee result in a fracture,1,2 and the prevalence of knee fractures in a primary care clinic is only 1.2% of all patients with knee injuries.3 The frequency of patients presenting with knee fractures in an outpatient physical therapy clinic is probably even lower; however, we decided to investigate the literature before she arrived for her examination to identify evidence that could aid our clinical decision making and help us decide whether radiologic examination was warranted in this case.
Although the incidence of knee fracture is low, we needed to ascertain whether the patient had a fracture prior to implementing a physical therapy plan of care. Therefore, we were interested in determining the factors or conditions that would necessitate taking radiographs of the knee to rule out the possibility of a fracture in the patient.
 | Database used for search: |
|---|
| Initial keywords: knee injury, diagnostic imaging, knee fractures, sensitivity |
|---|
 TopDatabase used for search: Initial keywords: knee injury,... Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
We now wanted to combine our searches to reveal citations that included all previously used keywords. On the main PubMed search screen, we clicked on Preview/Index, located in the Features bar underneath the query box. (For more information on the Preview/Index page, see the sidebar on p. 1094.) The Preview/Index page then displayed our most recent searches under the heading "Most Recent Queries" (Fig. 1). We combined the 2 searches by typing #1 AND #2, which combined search lines 1 and 2 in our search history, into the query box at the top of the page. We then clicked on the Preview button. This search identified 204 citations.
| Preview/Index Page PubMed's Preview/Index page has two functions. First, the Preview function allows users to build and refine search strategies by adding keywords one at a time using a query box at the bottom of the page under the heading "Add Term(s) to Query or View Index." Once the user clicks on the Preview button, PubMed displays the number of citations retrieved and a search line number rather than a list of citations, allowing the user to combine separate searches (Fig. 1). Second, the Index function allows users to build a search in the same entry box using Medical Subject Heading (MeSH) terms. To do this, the user selects MeSH Terms in the Fields dropdown menu next to the lower query box, types in the keyword, and clicks on the Index button. PubMed then displays an index of MeSH terms, which the user can scroll through to select the proper MeSH term. The user can then preview the results or combine the term with other terms using the AND, OR, or NOT buttons.
|
|
We added the keyword "sensitivity" to the previous search by typing sensitivity into the query box located under the heading "Add Term(s) to Query or View Index." We left the Fields dropdown menu to the left of the box on its default setting of All Fields. (This dropdown menu allows users to limit the search of a keyword to a specific field in MEDLINE records, such as "Author," "Journal," "MeSH [Medical Subject Headings] Terms," and "Title.") We then clicked on the AND button, which added the term to our search string. Our final search string was #1 AND #2 AND sensitivity. We then clicked on the Preview button next to the main query box at the top of the Preview/Index screen. This resulted in 23 citations (Fig. 2). We clicked on the Go button at the top of the page, and PubMed displayed a list of these citations. We read all the titles in the list of citations to decide which articles might provide evidence for the diagnostic accuracy of signs and symptoms associated with a knee fracture.
|
| Selection of articles for review: |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
The development of a CPR is a 3-step process: deriving predictor variables, validating the CPR, and analyzing the impact of the CPR on clinician behavior. The first step is identifying the possible predictor variables that can increase the likelihood of identifying the outcome of interest.6 The OKR was originally derived by Stiell and colleagues7 and identified 5 variables that predicted the ordering of radiographs with patients with acute knee injuries. These 5 predictor variables were:
Once a CPR has been derived, it must be validated before it can be incorporated into clinical practice, because the predictor variables identified during the derivation study may have occurred simply by chance.6 Validation of the rule should occur in different settings, with heterogeneous patient populations, and should be carried out by different examiners to improve the rule's generalizability.6 The OKR has been validated in a number of studies.2,8–10 Once a study has been validated for widespread clinical use, it must be subjected to the scrutiny of impact analysis to determine whether the rule changes clinician behaviors or results in improved efficiency and patient care.6
As we read through our list of 23 citations, we selected 2 articles that reported evidence on the diagnostic accuracy of the OKR. We were aware of the potential usefulness of CPRs in determining a diagnosis and felt the OKR might be most beneficial in determining appropriate evidence to guide our clinical decision making. Both articles were obtained from a local medical library.
The first article of interest and the one we expected to be of greatest value was Bachmann et al11 (Fig. 2, citation 1) because it was directly related to our clinical question and because it was a systematic review. Systematic reviews rank high on the hierarchy of evidence proposed by Sackett and colleagues.12 The creators of a systematic review attempt to limit the inherent selection bias often associated with review articles by following a strict methodology to enhance the review's validity. This is accomplished by performing an exhaustive search of the literature using explicit inclusion and exclusion criteria and by analyzing and scoring the identified articles to determine their methodological quality and strength of the evidence for the benefit and harm of an intervention or, in this instance, the diagnostic accuracy of a test.13 The citation and abstract for the systematic review by Bachmann et al11 are reproduced below.
Bachmann LM, Haberzeth S, Steurer J, ter Riet G. The accuracy of the Ottawa knee rule to rule out knee fractures: a systematic review. Ann Intern Med. 2004 Jan 20;140(2):121-4.BACKGROUND: The Ottawa knee rule is a clinical decision aid that helps rule out fractures and avoid unnecessary radiography. PURPOSE: To summarize evidence about the accuracy of the Ottawa knee rule. DATA SOURCES: Relevant English- and non-English-language articles were identified from PreMEDLINE and MEDLINE (1966-2003), EMBASE (1980-2003), CINAHL (1982-2003), BIOSIS (1990-2003), the Cochrane Library (2002, Issue 3), the Science Citation Index database, reference lists of included studies, and experts. STUDY SELECTION: Articles were included if they reported enough information to determine the sensitivity and specificity of the Ottawa knee rule for detecting fractures confirmed either radiologically or in combination with follow-up. DATA EXTRACTION: Two reviewers independently extracted data on study samples, the ways that the Ottawa knee rule was used, and methodologic characteristics of studies. DATA SYNTHESIS: Of 11 identified studies, 6 involving 4249 adult patients were considered appropriate for pooled analysis. The pooled negative likelihood ratio was 0.05 (95% CI, 0.02 to 0.23), the pooled sensitivity was 98.5% (CI, 93.2% to 100%), and the pooled specificity was 48.6% (CI, 43.4% to 51.0%). CONCLUSION: A negative result on an Ottawa knee rule test accurately excluded knee fractures after acute knee injury. However, because the rule is calibrated toward 100% sensitivity and actual fracture prevalences are usually low, large-scale, multicentered studies are still needed to establish the cost-effectiveness of routinely implementing the rule.
[© 2004 American College of Physicians. Abstract reprinted with permission of the American College of Physicians.]
This article was a systematic review analyzing the accuracy of the OKR in determining the need to order radiographs following an acute knee injury. The systematic review initially identified 104 articles; however, only 11 met the review's strict inclusion criteria. The articles included in the review were cohort studies comparing the OKR with the reference standard of plain film radiography. The authors ranked a study as high quality (level 1) if patient enrollment was consecutive, if all participants received the reference standard, and if radiographs were assessed by experts blinded to the OKR results. (A level 2 ranking was given if 2 of 3 criteria were met, a level 3 ranking was given if only 1 of 3 criteria was met, and a very low ranking [level 4] was given if none of the criteria were met.) Five studies were excluded from the pooled data because of low methodological quality. The pooled results of the remaining 6 studies revealed a sensitivity of 98.5% (95% confidence interval [CI]=93.2%–100%), a specificity of 48.6% (95% CI=43.4%–51%), and a negative likelihood ratio of 0.05 (95% CI=0.02–0.23). In the remaining 6 studies reviewed, the probability of a patient presenting with a patellar fracture after a negative result on the OKR was 0.37% (95% CI=0.15%–1.48%). As we read through the article, however, we realized that studies investigating the accuracy of the OKR in children were excluded from the review. The results of the systematic review, therefore, could not be generalized to our adolescent patient.
The second article of interest from the initial search was Cohen et al14 (Fig. 2, citation 15) because it dealt specifically with clinical criteria for using radiographs with children with acute knee injuries.
Cohen DM, Jasser JW, Kean JR, Smith GA. Clinical criteria for using radiography for children with acute knee injuries. Pediatr Emerg Care. 1998 Jun;14(3):185-7.OBJECTIVE: To evaluate clinical criteria for selective radiography for knee injuries in children. DESIGN: Retrospective chart review. SETTING: Emergency department (ED) of a children's hospital. PARTICIPANTS: All patients evaluated by radiography for an isolated, acute knee injury during 12 months. Patients were excluded for injuries: >1 week; isolated to superficial lacerations/abrasions; with prior knee surgery; being reassessed. RESULTS: Two hundred fifty-four patients (60% male; 12.7 years median age) were included. Twelve patients (4.7%) sustained a fracture. Evaluated criteria were point tenderness, inability to bear weight in the ED, and inability to flex the knee to 90 degrees. Point tenderness was not statistically associated with fracture, P = 0.7. Inability to bear weight in the ED (37% fracture rate, P = 0.001) and inability to flex to 90 degrees (52% fracture rate, P < 0.001) were associated with the presence of fracture.[table in text] Applying a rule combining nobearwt and noflex90 would decrease the number of x-rays by 73%, with no missed fractures. CONCLUSIONS: Point tenderness was not a good predictor of knee fracture in children. Using the clinical criteria to select patients requiring knee radiography may greatly reduce the number of unnecessary x-rays.
[© 1998 Lippincott Williams & Wilkins Inc. Abstract reprinted with permission of Lippincott Williams & Wilkins Inc.]
This study revealed 2 variables that were associated with the presence of fractures in children—the inability to bear weight and the inability to flex the knee to 90 degrees. The study design, however, was a retrospective chart review, which presents a number of inherent limitations, including missing data points and the lack of standardized methods of patient examination. Considering this, we were not confident that the study by Cohen et al14 provided sufficient evidence to guide our decision making in this particular case, so we decided to expand our search.
| Second keywords: knee radiographs, rule* |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
The results of our search using the search string knee radiographs AND rule* produced 38 citations. We further refined the search by clicking on the Limits button and selecting the All Child: 0-18 years option in the Ages dropdown menu. This further narrowed the number of citations retrieved to 14 (Fig. 3). Of the 14 citations identified, 2 (Fig. 3, citations 1 and 5) appeared relevant to our clinical question because they both contained "Ottawa Knee Rule" and "children" in their titles. Based on the abstracts of the 2 citations, we decided to retrieve both articles from our local medical library. The first article we read was the study by Khine et al17 (Fig. 3, citation 5).
|
Khine H, Dorfman DH, Avner JR. Applicability of Ottawa knee rule for knee injury in children. Pediatr Emerg Care. 2001 Dec;17(6):401-4.OBJECTIVE: Previous studies have shown that the application of the Ottawa knee rule (OKR) reduces the need for radiographs in adults with acute knee injuries. Our objectives were to describe the epidemiology and incidence of knee injuries in children with acute knee trauma and to validate the OKR in a pediatric population. DESIGN: A prospective, consecutive study. SETTINGS: Two urban pediatric emergency departments. METHODS: All children 18 years of age and under who presented with acute traumatic knee injury of less than 1 week's duration, excluding patients with a normal knee examination, superficial skin injuries, prior history of knee injury, underlying bone disease, serious injuries involving two or more organ systems, or altered mental status were enrolled. Physicians assessed each patient for 22 standardized clinical findings prior to radiography. The OKR was applied to each patient by the investigating physician. RESULTS: All 234 patients eligible for the study had radiographs of the affected knee. The median age was 13 years with a range of 2 to 18 years. Using the OKR criteria for obtaining knee radiographs, 12 of 13 patients with fractures were identified (sensitivity 92%; 95% CI= 64-99). The missed case was an 8-year-old male who had sustained a nondisplaced fracture of the proximal tibia after a fall. If the OKR were applied to the pediatric population, it would have reduced the need for radiography in 46% of children. CONCLUSIONS: In the pediatric population studied, the OKR did not identify all patients with knee fractures. Future studies may consider modifying the OKR to accommodate the differences between pediatric and adult patients to improve the sensitivity of the rule while maintaining its specificity, before it can be applied routinely in clinical practice.
[© 2001 Lippincott Williams & Wilkins Inc. Abstract reprinted with permission of Lippincott Williams & Wilkins Inc.]
Khine and colleagues17 investigated the accuracy of the OKR in 234 patients under 18 years of age (mean age=13 years) who were seen in 1 of 2 emergency departments following a traumatic knee injury. The prevalence of fractures in this sample was 6%. According to the results of the study, the OKR exhibited a sensitivity of 92% (95% CI=64%–99%) and a specificity of 49% (95% CI=42–56). The OKR failed to identify one child (8%) with a knee fracture. The small sample size and limited prevalence of patients with knee fractures resulted in wide CIs associated with the specificity and sensitivity. Considering these factors, we were not confident that this study provided adequate evidence to guide our clinical decision making. The authors acknowledged the limitations of their study and suggested that further validation is necessary before the OKR can routinely be used in a pediatric population.
Bulloch et al18 (Fig. 3, citation 1) was the second OKR validation study identified in our latest database search.
Bulloch B, Neto G, Plint A, Lim R, Lidman P, Reed M, Nijssen-Jordan C, Tenenbein M, Klassen TP; Pediatric Emergency Researchers of Canada. Validation of the Ottawa Knee Rule in children: a multicenter study. Ann Emerg Med. 2003 Jul;42(1):48-55.STUDY OBJECTIVE: The main objective of this study was to determine the sensitivity and specificity of the Ottawa Knee Rules when they were applied to children. The secondary objective was to determine post hoc whether use of the rules would reduce the number of knee radiographs ordered. METHODS: This prospective, multicenter validation study included children aged 2 to 16 years who presented to the emergency department with a knee injury sustained in the preceding 7 days. Children were assessed for the variables comprising the Ottawa Knee Rules, and physicians ordered radiographs at their discretion. A positive outcome was defined as any fracture. A negative outcome was defined as children who did not have a fracture on radiograph or, if no radiograph was obtained, were asymptomatic after 14 days. RESULTS: A total of 750 children were enrolled. The mean age was 11.8+/–3.1 years, and 443 (58.7%) were male patients. Seventy children had fractures. Radiography was performed for 670 children, whereas 80 children had only a structured telephone interview. The Ottawa Knee Rules were 100% sensitive (95% confidence interval [CI] 94.9% to 100%), with a specificity of 42.8% (95% CI 39.1% to 46.5%). Only 460 children would have required a radiograph if radiographs had been performed according to the Ottawa Knee Rules, which would have resulted in an absolute reduction of 209 (31.2%) radiographs. CONCLUSION: The Ottawa Knee Rules are valid in children and have the potential to decrease the use of radiography in children with knee injuries.
[© 2003 American College of Emergency Physicians. Abstract reprinted with the permission of the American College of Emergency Physicians.]
This validation study included a greater sample size (750 children who were less than 18 years of age) and larger prevalence (9%) of fractures than the study by Khine et al.17 The average age of patients was 11.8 years, and data were collected from 5 pediatric emergency departments. The sensitivity and specificity of the OKR in this group of patients were 100% (95% CI=94.9%–100%) and 42.8% (95% CI=39.1%–46.5%) respectively. These results gave us increased confidence in using the OKR to guide our decision making with our patient, considering the sensitivity of 100% (no missed fractures) and the narrower confidence intervals than the study by Khine et al.17
Bulloch et al18 further classified their findings on the accuracy of the OKR into age groups, which allowed us to generalize the findings to our patient more adequately. In the group of patients 6 to 12 years of age (n=375), the sensitivity was determined to be 100% (95% CI=88.8%–100%), and specificity was reported as 47.1% (95% CI=41.8%–52.4%). In addition, the study reported the interrater reliability of the application of the OKR in a pediatric population. In the 6–12 age group, the kappa statistic was .81 (95% CI=.55–1.0).
| Clinical decision: |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
| Examination: |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
| Applying the CPR: |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
| Results: |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
|
| References |
|---|
|
TopDatabase used for search:Initial keywords: knee injury,...Selection of articles for...Second keywords: knee...Clinical decision:Examination:Applying the CPR:Results:References |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
