Developing a predictive model for vertigo using demographic and laboratory data: An evidence-based medicine approach

Conclusion. The approach described in this paper may be helpful in establishing an early-warning, evidence-based mechanism for diagnosing vertigo, which can be utilized in medical education to reduce medical uncertainty. Objective. To use an evidence-based medicine approach to evaluate the probability of having vertigo using laboratory and demographic data. Material and methods. The study was conducted on 22 working days during July 2002. Targeted cases who visited a general hospital in southern Taiwan for routine physical examinations were asked to participate in the study and agreed to take additional tests during their visits. A total of 200 subjects were systematically and randomly selected from this data pool. We ran binary logistic regression on all these cases. Results. The logistic regression model explained 71.3% of the variance in having vertigo or not. The equation for having vertigo was as follows: -21.855+(1.132 xmale gender)+ (0.071 xage)+(-0.023 xsystolic blood pressure)+(0.057 xdiastolic blood pressure)+(0.048 xfasting glucose)+ (0.051 xcholesterol)+(-0.005 xtriglycerides)+(-0.361 xpresence of cardiovascular diseases).