Reference: Jolson, HM, et al. (1992). Journal of the National Cancer Institute, 500 - 505.
The drug cytarabine is used for bone marrow ablation in preparation for bone marrow transplantation in patients with malignant, non-responsive leukemia and lymphoma. Cerebellar toxicity is a severe, therapy-limiting adverse reaction to this drug when given in high doses. Even with the best of care, high-dose cytarabine treatment is associated with a serious form or cerebellar toxicity that occurs in approximately 8% of patients.
The current investigation was prompted when the Food and Drug Administration received a report of an increased frequency of cerebellar toxicity associated with cytarabine from the University of Wisconsin Hospital. This hospital had recently switched from using the cytarabine product manufactured by innovator company (The Upjohn Co., Kalamazoo, MI, called "Jones" in this exercise because of the medico-legal concerns at the time of the investigation) to a generic manufacturer (Quad Pharmaceuticals Inc., Indianapolis, IN, the "Smith" company). To address this concern, you are asked to complete a chart review in which information is abstracted on patients and their treatment characteristics.
Variables in the data table NEW (59 records) in the Project TOXICITY.MDB:
Name Type Len Description
---------- ------- --- -----------------------------------------------------
AGE Integer 2 Age at time of treatment (years)
SEX Real 5 1 = male; 2 = female
MANUF Alpha 10 Manufacturer: Jones (Quad = Generic); Smith (Upjohn = Innovator)
DIAG Real 7 1 = leukemia; 2 = lymphoma
STAGE Real 10 1 = relapse; 2 = remission
TOX Real 11 1 = yes; 2 = no
DOSE Real 11 Dose in gms/M2
SCR Real 6 Serum creatinine level (mg/dl)
WEIGHT Real 8 Body weight (kg)
AGEGRP Integer 1 1 >= 50 yrs, 2 < 50 yrs
The hospital now asks you to determine whether there is a higher risk of toxicity in patients treated with the Jones (Quad) product than with the Smith (Upjohn) product and to quantify this difference. They also ask you to decide whether differences in patient or treatment characteristics might help explain any such finding. (For example, if the patients treated with the Jones product had poorer renal function, this might explain higher than expected rates.) As a side matter, you also want to evaluate other potential risk factors for the outcome.
(1) Compare the two groups (Smith and Jones):
(A) Compare the two groups with respect to their continuous variables (AGE, DOSE, SCR, WEIGHT). Report means and standard deviations, and perform
tests of significance in each instance. Comment on your findings.
(B) Compare the groups with respect to their categorical variable (SEX, DIAG, STAGE, TOX). Report counts and proportions. Perform a test of
significance in each instance. Comment on your findings.
(2) Identify risk factors for the outcome. Calculate the relative risk of toxicity associated with each of the following variables:
(A) AGEGRP
(B) SEX
(C) DIAG
(D) STAGE
(E) MANUF
(3) Consider the potential for confounding. Based on the analyses in (1) and (2), which factors might confound the relationship between MANUF and TOX? (For an extraneous variable to confound the relationship between E and D, it must be associated with E and must be an independent risk factor for D.)
(4) Adjusted Relative Risks. Compute adjusted the relative risks of TOX associated with MANUF, adjusting (stratifying) for each of the following factors individually:
(A) AGEGRP
(B) SEX
(C) DIAG
(D) STAGE
(E) MANUF
(Let us assume no significant interaction by any of these factors.)