Final Help
Final Help
An investigator wants to assess whether the use of a specific medication given to infants
born prematurely is associated with developmental delay. Fifty infants who were given the
medication and fifty comparison infants who were also born prematurely but not given the
medication will be selected for the analysis. Each infant will undergo extensive testing at
age 2 for various aspects of development. Identify the type of study proposed and indicate
its specific strengths and weaknesses.
Strengths: This type of study can assess the temporal relationship. Bias are not as big of
an issue in cohort studies as they are in case-control studies.
Weaknesses: This type of study is difficult if the chosen disease is rare, and confounders
may be an issue.
2. In the study of a new anti-hypertensive medication, systolic blood pressures are measu
treatment is administered). The data are as follows:
120 112 138 145 135 150 145 163
148 128 143 156 160 142 150
3. Total cholesterol in children aged 10-15 is assumed to follow a normal distribution with
and a standard deviation of 22.4.
a. What proportion of children 10-15 years of age have total cholesterol between 180 an
0.171
5. The following data are collected in a clinical trial evaluating a new compound designed to
healing in trauma patients. The new compound is compared against a placebo. After treatm
the new compound or placebo the extent of wound healing is measured and the data are sh
Suppose that clinicians feel that if the percent reduction in the size of the wound
50% then the treatment is a success.
Is there a difference in the extent of wound healing by treatment? (Hint: Are treatment and
healing independent?) Run the appropriate test at a 5% level of significance.
Reject since 14.25 > 4
a. Generate a 95% confidence interval for the percent success in patients receiving the new
(0.50, 0.67)
b. Generate a 95% confidence interval for the difference in the percent success between th
placebo
(0.11, 0.35)
c. Generate a 95% confidence interval for the relative risk of treatment success between tre
(1.25, 2.19)
d. Generate a 95% confidence interval for the odds ratio of treatment success between trea
(1.55, 4.31)
Reject since 14.25 > 4
a. Generate a 95% confidence interval for the percent success in patients receiving the new
(0.50, 0.67)
b. Generate a 95% confidence interval for the difference in the percent success between th
placebo
(0.11, 0.35)
c. Generate a 95% confidence interval for the relative risk of treatment success between tre
(1.25, 2.19)
d. Generate a 95% confidence interval for the odds ratio of treatment success between trea
(1.55, 4.31)
New Compound
5 9 9 14
(n=125)
Placebo (n=125) 9 8 9 59
Total 14 17 18 73
Is there a difference in the extent of wound healing by treatment? (Hint: Are treatment a
healing independent?) Run the appropriate test at a 5% level of significance
New Compound
15 37 32 41
(n=125)
Placebo (n=125) 36 45 34 10
Total 51 82 66 51
Step 1: H0: reatment and percent wound healing are independent of each other
H1: H0 is false
a = 0.05
2
(O - E )
Step 2: 2
E
Step 3: Reject H0 if X^2 > 7.81
Std Dev
Diet Program Sample Size Mean Cholesterol
Cholesterol
Is there a significant difference in mean weight reduction among the 4 treatments? Use a
SST=1,889.
A B C Placebo
Mean (SD)
6.4 (4.1) 8.9 (4.0) 2.2 (3.9) 2.5 (4.3)
Weight Reduction
2
S S B = n j(X j - X )
2
between treatments S S B = n j(X j - X ) k- 1
625.2 3
2
error S S E = ( X - X j) N-k
1263.8 76
2
total S S T = (X - X ) N-1
1889 79
N= 80
k= 4
df1 = 3
df2 = 76
|1 2| 15
ES= = =0 . 30 ,
50
Z 1 /2 + Z 1- 2
( ) (
2
1 . 96+0 . 84
ni = 2
=174.2.
ES
=2
0 . 30 )
Need n1=175 and n2=175 with complete outcome data.
p - p 0 .0 5
E S p1 1 - p 0 0 0 .0 5 0 .1 1
E S p 0 ( 1 - p 0 ) 0 .2 6 ( 1 - 0 .2 6 ) 0 .1 1
p 0(1 - p 0 ) 0 .2 6 ( 1 - 0 .2 6 )
2
Z 1 - /2 Z 1 - 2 1 . 9 6 1 . 2 8 2 2 2
n Z 1 - / 2 Z 1 - 1 . 9 6 1 . 2 8 2 8 6 8 . 6
n
E S 0 .1 1 8 6 8 .6
E S 0 .1 1
|p1 - p0| = |difference in mean|
= SD Need n1 = 200 and n2 = 200 with c
Z (1- a/2) = 1.96 outcome data.
Z (1 - B) = 0.84 (for 80%)
= 1.282 (for 90%)
E = % (convert to decimal!!!)
* always round up n *
12. Suppose we want to design a new study to assess the implications of childhood obe
health outcomes in adulthood. In the study we will consider total cholesterol at age 50
cardiovascular disease by age 50 as the key outcomes. We will compare normal versu
overweight/obese children. Assume that the standard deviation in total cholesterol is 3
a. How many normal and overweight/obese children are required to ensure that a 95%
interval estimate for the difference in total cholesterol has a margin of error not exceed
376.5
2
Z 2 1 . 96( 35 )
n
Need n1 = 377 (normal
complete outcome data.
i = 2 =2 ( ) (
weight children)
E
and n2 = =
5 )
377 (overweight/obese childr
b. How many normal and overweight/obese children are required to ensure 80% power
4% difference in incident cardiovascular disease between groups? Assume that approx
of normal weight children will develop cardiovascular disease by age 50 and that 7% o
overweight/obese children will develop cardiovascular disease by age 50 and that a tw
with a=0.05 will be used.
Need n1 = 377 (normal weight children) and n2 = 377 (overweight/obese childr
complete outcome data.
b. How many normal and overweight/obese children are required to ensure 80% power
4% difference in incident cardiovascular disease between groups? Assume that approx
of normal weight children will develop cardiovascular disease by age 50 and that 7% o
overweight/obese children will develop cardiovascular disease by age 50 and that a tw
with a=0.05 will be used.
c. How many normal and overweight/obese children should be enrolled to satisfy (a) an
assuming that 20% will be lost to follow-up over the course of the study?
A total of 928 children (464 * 2) will suffice for both parts (a) and (b).
Y = Bo + B1 * X
Bo = 143.26
B1 = 1.344
Y = 143.26 + 1.344 X
Y = 143.26 + 1.344 * 50
Y = 210.46
Y = 143.26 + 1.344 * 40
Y = 197.02
Y = 143.26 + 1.344 X
Y = 143.26 + 1.344 * 50
Y = 210.46
Y = 143.26 + 1.344 * 40
Y = 197.02
ation given to infants
who were given the
ly but not given the
extensive testing at
roposed and indicate
e, and confounders
pressures are measured at baseline (or the start of the study before any
3b)
P(x>200) 200
c. 80 / 631
0.1267828843
e of Wound
76-100% Total None
9 4.5673758865
46
10 9.4326241135
95
19 141
df = (5-1)(2-1)
X^2 =
lower Answer
0.1160710259 (0.50, 0.67)
p1/p2
p2 z RR (relative risk)upper lower
0.352 1.96 1.6590909091 0.78639147 0.22614814
Expected
Total 0-25% 26-50% 51-75%
25.5 41 33
125
25.5 41 33
125
250
df = (4-1)(2-1) =3 -->
dent of each other X^2 = 28.32
Step 1. H0: u1 =
s2 n2 sp Z or t H1: u1 u2
21.6 75 22.7995310892 5.18905726
Step 2.
Step 3. Reject H0
Step 4. z = 5.19
Step 5. Reject H0
mean squares F
mean of means = 5
Step 4: F = 12.53
nd (b).
u SD normdist(true)
191 22.4 0.312
191 22.4 0.482
0.171 (Subtracting 0.482 - 0.312)
inal outcome
=4 --> Table 3
14.25 9.49
Answer
(1.55, 4.30)
76-100%
25.5
25.5
Table 3
7.81
Step 1. H0: u1 = u2 = 0.05
H1: u1 u2
Step 2. X1 - X 2
Z
1 1
Sp
n1 n2
Step 3. Reject H0 if z -1.96 or if z 1.96. (Table 1c)
Step 4. z = 5.19
2.73
we have to substract by 1 at the end!!
bstract by 1 at the end!!
each other