Research Design, Biostatistics, and Literature Evaluation

Practice Case

Questions 4–6 pertain to the following case.

The neurosurgery physician group at your institution would like your input in evaluating the usefulness of pro-

calcitonin for distinguishing neurogenic fevers from true infectious etiologies of fever. An additional goal of the

study would be to evaluate risk factors and possible confounders for positive and negative procalcitonin values.

You decide on a predictive cutoff of 0.5 ng/mL (positive 0.5 mg/mL or greater; negative less than 0.5 ng/mL).

Which statistical test would be most appropriate to evaluate any difference in creatinine clearance (milliliters

per minute) between patients with positive and negative procalcitonin?

Which statistical test would be most appropriate to determine an independent association between positive

procalcitonin and true infection in the presence of possible confounders?

You enroll 201 patients who have a true infection and 200 patients who are not infected. Of the 201 with

infection, 189 have a positive procalcitonin and 12 have a negative procalcitonin. Of the 200 patients without

infection, 100 have a positive procalcitonin and 100 have a negative procalcitonin. Which best describes the

sensitivity and specificity of procalcitonin in this population?