Videofluoroscopy in cervical spine trauma: an interinterpreter reliability study.

OBJECTIVE

The primary purpose of this study was to determine if, when confronted with a series of cervical videofluoroscopic (VF) studies, board certified radiologists who are trained in the interpretation of VF spinal examinations would be in agreement with one another with regard to the findings of normal, hypomobile or hypermobile intersegmental motion. As a secondary objective, we polled participating physicians regarding demographic issues such as time in practice, self-assessment of proficiency, the type of training received in VF interpretation and the number of VF studies read per year, in order to determine if any of these factors had any predictive value in terms of interexaminer agreement.

DESIGN

Seven patients who had been exposed to cervical acceleration/deceleration (CAD) trauma from motor vehicle accidents were randomly selected from the practice of two of the authors (ACC and JSK). Three volunteers, who were asymptomatic and had no history of neck pain or injury to the neck, were also recruited. In all cases, informed consent was obtained in accordance with the Helsinki guidelines. Ten board qualified chiropractic radiologists were recruited to participate in this study. Blinded to the history and identity of the patients and volunteers, the participants were asked to view the 10 VF studies and, in each case, to report either "normal," "hypomobile" or "unable to determine" for all segments from OCC/C1 through C7/T1. The resulting data was analyzed for concordance using the kappa statistic. Kappa was calculated for all segments (OCC-T1) for agreement in "normal" vs. "abnormal" motion. The two possible choices for abnormal (i.e., hypermobility and hypomobility) were pooled together. We also compared the results of participants' responses to demographic questions with the results of their interpretations of the VF studies.

SETTING

An urban group practice. Four of the patients were those of an orthopedist (ACC) and three were those of a general practitioner (JSK). PATIENT/OTHER PARTICIPANTS: Seven patients were randomly selected from a larger group of patients satisfying our selection criteria. Patients were chosen who had been exposed to CAD trauma and had been provided with at least 6 months of conservative chiropractic care, yet remained symptomatic as a result of their injuries. All had evidence of intersegmental instability in the cervical spine as defined by other investigators and none had any history of injury or pain in the cervical spine prior to their motor vehicle accident. Four males and three females with a mean age of 16 yr comprised the patient group. Two asymptomatic atraumatic volunteers were males and one was female. They were age matched to the patient group with a mean age of 38 yr.

INTERVENTION

This study did not entail any form of intervention.

MAIN OUTCOME MEASURES

Our primary outcome measure was that of concordance or agreement between our group of 10 participant radiologists in regard to their interpretations of the 10 VF studies provided to them, i.e., that of interinterpreter reliability. Higher values of kappa suggested that agreement between the radiologists was not likely due to chance alone. Our secondary outcome measure was a correlation between the results of our demographic questionnaires, completed by participating radiologists, and the overall interinterpreter reliability concordance. The hypothesis tested was that radiologists who have institutional training in VF, more years of experience, and those who read more VF studies annually and self-rank their proficiency as "high" are more likely to agree with others and/or are more likely to correctly analyze intersegmental motion than those whose training was less formal, and who have less experience, read fewer VF studies and self-rank their proficiency in reading VF as only "adequate" or lower. (ABSTRACT TRUNCATED)