Figure 1

Absolute difference in adjusted probability of receiving a head computerized tomography (CT) scan when a head CT-item template was used compared to when a head CT-item template was not used. Error bars indicate 95% confidence intervals. ^a Absolute difference calculated as follows: Adjusted probability of receiving a Head CT when a template with a pre-printed head CT item is used MINUS adjusted probability of receiving a head CT when a template with a pre-printed head CT item is not used. ^b Probability of head CT was derived from a logistic regression model with head CT performance as the dependent variable and the following independent variables: socio-demographic variables (age, gender, race-ethnicity, insurance status), hospital, type of dizziness symptom, dizziness presentation type, number of medical symptoms, number of other neurological signs or symptoms, number of stroke risk factors, admission status, head CT-item template, and the interaction terms of head CT-item template with dizziness presentation type. Probabilities were calculated with all other variables in the model held constant at the population means. ^c Probability of head CT was derived from a logistic regression model with the dependent variable of head CT performance and the following independent variables: head CT-item template, dizziness presentation type, the interaction terms of head CT-item template with dizziness presentation type, and the propensity score quintile. The propensity scores were derived from a logistic regression model with head CT-item template as the dependent variable and the same independent variables as in the first model.