Running head: Epidemiology Article Critique

The objective of this study was to review the effect of helmet use on the frequency of cervical or thoracic fractures experienced in motorcycle crashes. This prospective cohort study was conducted at one hospital, St. Joseph’s Hospital Medical Center in central Phoenix, Arizona. Cases were 422 motorcycle crash victims treated during the 3 years of the study (July 1, 2002 through June 30, 2005), 190 of them had a traumatic brain injury and 75 of them experienced some form of a spinal injury. The exposed group consisted of 169 motorcyclists who were wearing helmets and the non exposed group was 253 motorcyclists who were not wearing helmets at the time of their crash. The incidence of either cervical or thoracic spine fractures for helmeted patients was 12% and for unhelmeted patients was 15%. The estimated odds ratio for helmet use and traumatic brain injury, 0.590 [0.396, 0.877] was statistically significant. Helmet use therefore reduced the chance of any form of traumatic brain injury by almost 50%. These results provide evidence that helmets reduce the incidence and severity of traumatic brain injury sustained by victims of motorcycle crashes but there was no significant relationship between location of spinal fracture and helmet use. Epidemiology Article Critique 3 Description of Evidence A. Exposure The intervention in this study was wearing a helmet. B. Outcome With wearing a helmet the motorcycle rider has an increased risk of cervical or thoracic spine fractures. With out wearing the helmet the motorcycle rider has an increased risk of a traumatic brain injury from which they would die. C. Design A prospective cohort study design was used to collect trauma data prospectively and recorded from paramedic sheets and trauma logs. Full follow up information was available for 3 years (July 1, 2002 through June 30, 2005). During this period, specific injury data on each patient were obtained from the St. Joseph’s Hospital Medical Center Trauma Center Registry (Trauma One) using International Classification of Diseases, 9 th Revision, Clinical Modification (ICD-9-CM) external cause and nature of injury coding as defined by National Trauma Data Bank Protocols Dataset. D. Study Population The population used in this study was motorcycle crash victims who were transported to the St. Joseph’s Hospital Medical Center Trauma Center during July 1, 2002 through June 30, 2005. E. Main Results When cervical or thoracic spine fractures were combined, 12% of helmeted patients and 15% of unhelmeted patients who arrived alive at the trauma room had fractures. This difference was not statistically significant. However, a motorcyclist not wearing a helmet and involved in a crash was two times more likely to sustain a traumatic brain injury and three times more likely to Epidemiology Article Critique 4 die from the crash than a helmeted individual. Helmet use therefore reduced the chance of any form of traumatic brain injury by almost 50%. Non-causal Explanations A. Observation Bias The need to pool data from trauma systems using varied screening protocols for potential spinal injury was avoided because all trauma registry data was retrospectively reviewed from a single Level I Trauma Center, where all patients are evaluated according to a standardized radiologic protocol for assessing spinal fractures. B. Confounding To identify potential confounding categorical variables, crash speed, collision or no collision, gender, and race were analyzed using Pearson’s X 2 tests to determine the association between spine fractures and death. Information on blood alcohol level was available from too few cases to be included in the analyses. Several covariates were considered for inclusion in the model-building process. Logistic regression models were used to examine the differential impact, if any, of helmet use (binary) and speed of travel (continuous) on cervical or thoracic fractures after adjusting for the covariates of gender, age, and type of collision. C. Chance Variation To control happenings by chance crash speed was analyzed. A multivariate analysis was performed using the maximum likelihood procedure to address the small number of patients (14 of 182 subjects) for whom speed was available and who had experienced a cervical or thoracic fracture. Epidemiology Article Critique 5 Features Consistency with Causation A. Time Relationship The temporality of this study was a retrospective review in that the information about the cause, the motorcycle crash, was gathered before the information about the effect, severity of traumatic brain injury and cervical or thoracic fractures. B. Strength/strong Association A strong association in this study was that a motorcyclist not wearing a helmet and involved in a crash was two times more likely to sustain a traumatic brain injury and three times more likely to die from the crash than a helmeted individual. The estimated odds ratio for helmet use and traumatic brain injury, 0.590 [0.396, 0.877] was statistically significant and therefore reduced the chance of any form of traumatic brain injury by 50%. However, a weak association does not mean that there is not a causal effect. In a review of a statewide trauma registry in Illinois, the speed of the crash was significantly related to the presence of spinal fractures. Although not statistically significant (p= 0.539), the model including speed showed that helmet use lowered the risk of a cervical or thoracic spine fracture by 30%. The higher the speed, the more likely victims were to sustain a cervical-thoracic fracture. Each 5 miles per hour of speed increased the odds of a spinal fracture by 3.7% (p = 0.0214). The mean rate of speed was 43.53 miles/h. Of the 14 patients with a cervical or thoracic fracture, 4 were traveling faster than 40 miles per hour. C. Dose Response There is not a dose response in this study because the incidence of traumatic brain injury and cervical or thoracic spinal fractures sustained during a motorcycle crash does not linearly increase with the exposure of wearing a helmet. Epidemiology Article Critique 6 D. Consistency of Association There is not a consistency within this study because it was conducted at one hospital in one city, St. Joseph’s Hospital Medical Center in Phoenix, AZ. Hence conclusions could only be drawn from patients that arrived at that hospital. E. Specificity Specificity does not apply to this study because motorcycle crashes can result in a multitude of other injuries besides traumatic brain injury and cervical or thoracic spinal fractures. Along with traumatic brain injury and cervical or thoracic spinal fractures being caused by other events besides motorcycle crashes. External Validity A. Eligibility St. Joseph’s Hospital Medical Center with Barrow Neurologic Institute is a Level 1, American College Surgeons-verified trauma center. Patients were involved in local traffic crashes at low or moderate speeds and in high-speed interstate and rural crashes requiring transport to the trauma center. Data on motorcycle injuries were imported from the St. Joseph’s Hospital Medical Center Trauma Registry (Trauma One) and included helmeted and unhelmeted motorcyclists. B. Source Population During the 3 years of the study, the total number of St. Joseph’s Hospital Medical Center trauma admissions was 11,075. Of these, 422 (3.8%) were cases involving motorcycle crashes associated with sufficient information to allow data analysis involving multiple variable. Of the 422 motorcyclists, 253 were not wearing helmets (60%) and 169 were wearing helmets (40%) at the time of their crash. There were 49 female patients and 373 male patients. Their mean age was Epidemiology Article Critique 7 33 years (age range, 6-78 years). Twenty-six patients were younger than 18 years (age range, 617 years). There were 101 patients between 18 and 25 years. C. Target Population The results of this study can not been generalized because the study was conducted in one hospital in one city and therefore not population based. Comparison with Other Evidence A. Consistency Other sources with similar results, recommendations, and incidence were mentioned multiple times to validate consistency of comparison with other evidence. However, only one example included the study’s actual results. From this study, a review of a statewide trauma registry in Illinois, the speed of the crash was significantly related to the presence of spinal fractures. Although not statistically significant (p= 0.539), the model including speed showed that helmet use lowered the risk of a cervical or thoracic spine fracture by 30%. The higher the speed, the more likely victims were to sustain a cervical-thoracic fracture. Each 5 miles per hour of speed increased the odds of a spinal fracture by 3.7% (p = 0.0214). The mean rate of speed was 43.53 miles/h. Of the 14 patients with a cervical or thoracic fracture, 4 were traveling faster than 40 miles per hour. B. Specificity Specificity in comparison with other evidence is not applicable to this study because specificity is the ability of a test to identify non diseased individuals who actually do not have a disease (Friis & Sellars, 2009, p.666). Epidemiology Article Critique 8 C. Plausibility This study does make logical sense in that the use of a helmet can reduce traumatic brain injury but possibly may also cause neck strain. Resulting in cervical or thoracic spinal fractures during injuries that do not involve head trauma when the weight of the helmet and force of impact are taken into consideration. D. Coherence This study does fit into current knowledge using epidemiological and laboratory findings such as the biomechanical effect of helmets. The biomechanical influence of helmets on spinal injury has been studied using cadavers and computer models. They have shown that the type of injury depends on whether the head strikes the ground or other barrier axially and on whether the neck is compressed anteroposteriorly, causing hyperflexion or hyperextension, or laterally. Computer models of the head and neck have been used to understand the relationship between injurious force and helmet padding and thickness. Summary of Assessment This study’s review of helmet use and associated spinal fractures in motorcycle crash victims unfortunately did not bring to light any new data in regards to a relationship between helmet use and cervical or thoracic fractures. However it did verify the protective effect of helmet use in traumatic brain injuries. Therefore, in future studies more research with a specific focus on spinal injuries not including the head, over a much larger population and in multiple hospitals could possibly offer different results. Epidemiology Article Critique 9 References: Friis, R.H. & Sellars, T.A. (2009). Epidemiology for Public Health Practice. Sudbury, MA: Jones and Bartlett Publishers. Goslar, P. W., Crawford, N. R., Petersen, S. R., Wilson, J. R., & Harrington, T. (2008). Helmet use and associated spinal fractures in motorcycle crash victims. The Journal of Trauma, 64(1), 190-6; discussion 196.