Biostatistics I: Basic for Public Health

Lecture No.: KUI 6011

Starting Date: 22/09/2015

Concepts of Probability, Random Variables, Discrete and

Continues Probability Distributions
Laboratory Session 3
 Copyright © 2015 Department of Public Health
Faculty of Medicine, Gadjah Mada University, Yogyakarta, Indonesia
Contents

1 Learning Objectives: 1

2 Activities: 1

3 Class Exercises: 1

4 Homework: 4

5 Output: 5

6 Required Reading: 5

7 Log sheet: 6

i
1 Learning Objectives:

Upon completion of the exercise unit, students should be able to:

1. describe the normal distributions and sampling distributions

2. describe the concept of random variables, continues and discrete probability distributions

3. inferences concerning the population mean, the median, and the proportion having some
characteristics

4. describe the definition of population, sample, variable, and statistics

5. applied the concept of probabilities for the diagnostic test (sensitivity, specificity, the
predictive value of a positive test or negative test)

2 Activities:

1. Lecture and discussion: random variables and probability distributions

2. Group discussion on the concepts of population, sample, variables, statistics, random

variables, discrete and continues probability distributions

3. Computation, application and presentation on those concepts

4. Exercise on the use of probability concepts applied to diagnostic test

5. Assignment reading journal on these topics

3 Class Exercises:

1. The following hypothetical table shows a cross-classification of marital status and “health
perceived satisfaction” (whether he or she satisfied with his or her health condition) from
the survey. For a person randomly sampled from this population,

1
(a) What is the probability of being dissatisfied with their health condition?
180/1142
(b) What is the probability of being dissatisfied with their health condition and being
married?
86/1142
(c) What is the probability of being dissatisfied with their health condition given that
they are married?
86/645 = 13%
(d) What is the probability of being dissatisfied with their health condition given that
they never married?
49/238 = 20%
(e) Can we conclude from (c) and (d) that the probability of being never married is
greater among individuals who are dissatisfied with their health condition?

Never married lebih tidak nyaman dengan status kesehatannya daripada yang sudah
msrried

Satisfied Dissatisfied Total

Married 559 86 645
Divorced or widowed 214 45 259
Never married 189 49 238
Total 962 180 1142

2
2. For each of the following scenarios:

• Decide if binomial, Poisson, or neither would be an adequate model for the random
variable described.
• If the binomial or Poisson model could be used, then identify the parameters of the
model (i.e. (N,p) for binomial, λ or µ for Poisson
(a) For men over the age of 65, the prevalence of Alzheimer’s disease is estimated to be
5% while for females over the age of 65 the prevalence is estimated to be 2%. Let X
be the total number of patients over the age of 65 that have Alzheimer’s in a sample
of 8 men and 22 women over the age of 65.
Jawab :
Binomial (laki laki dan perempuan)
Tidak bisa dihitung.
(b) In Yogyakarta there is an average of 8.5 cases/year of a rare form of cancer. Let X be
the number of reported cases in the first 6 months of 2008.
Jawab :
Distribution Poison  tidak tentu tiap tahun 8,5 kasus, cancer bermacam-macam
jenis
Probability org terkena cancer 6 bln pertama = 4,25
(c) The age-at-diagnosis for 296 women with breast cancer is recorded from a cancer
registry in Central Java Province. Let X be the time (in years) that a woman survives
beyond diagnosis.
Jawab :
Tidak bisa dihitung
Karena : tidak ada keterrangan yang lain, ex: tahun
(d) Same 296 women as in (c). Assume the 5 year survival of women diagnosed with
breast cancer is 80%. Let X be the number of women that survive more than 5 years
after diagnosis (assume that all 296 women were diagnosed at least 5 years ago).

3
 Jawab :

Binomial : karena survive atau tidak survive

Peluang survive 80% = 0,8

Peluang tidak survive 20% = 0,2

3. Suppose you are in charge of water quality control at the sewage treatment plant.
You routinely test the treated water for fecal coliform bacteria. If none are found then
you assume the plant is operating properly and the discharge water is safe. Unknown to
you, however, a valve controlling the flow of chlorine (which kills bacteria) into the
chlorine contact tank is malfunctioning. As a result, some fecal coliform survive and are
present in the treated outflow. Assume that the number of bacteria in the outflow can be
modeled as a Poisson random variable with λ = 3000/liter. You collect a 1 milliliter sample
(1 liter = 1000 milliliters). What is the probability that you will fail to find any bacteria in
your sample?

Jawab :

Poison

µ = λ = n.p = rata-rata

µ x . e−µ
P(x) =
X!

λ x . e− λ
=
X!

310 . 2,78−3
P(0)=
0!

= 1 - 2,78−3

4
 Masih dipikir lebih lanjut tutor mbak jessica pusing

4. Let Z be a standard normal. Compute the following:

(a) P(Z < 1.96)

(b) P(Z < -1.65)

(c) P(-0.8 < Z < 1.2)

Note: Use table normal distribution or display normprob(z) in Stata.

Jawab :

Value−mean
Z skor = SD

Liat tabel z skor

a. 0,9750
b. 0,0495
c. 0,8 Z Skor 0,2119
1,2  Z skor 0,8849
Jadi P(-0.8 < Z < 1.2) = 0,8849 - 0,2119 = 0,673

5. A new company has a new home HIV test that they hope to market in a population where
the sero-prevalence is 10%. They tested 100 known HIV positive and 100 known HIV
negative subjects to assess the performance of their home kit. The following results were
obtained:

Test + Test - Total

HIV + 93 7 100
HIV - 21 79 100
114 86 200
(a) What is the sensitivity of the test?

5
  Sensitivity = Positif benar diantara orang yang sakit
93/100 = 0,93

(b) What is the specificity of the test?

 Specificity = Negatif diantara yang tidak sakit
79/100 = 0,79

(c) What is the probability that a person randomly selected from the target population
who tests positive is actually HIV positive? (this is called the predictive value of a
positive test)
Jawab :
(Brp peluang seseorang yang dipilih secara acak dari populasi dari orang yang HIV+ yg
hasil tesnya positif)
93/114

4 Homework:

1. Health-Services Administration
The in-hospital mortality rate for 16 clinical conditions at 981 hospitals was recently
reported. It was reported that inhospital mortality was 10.5% for coronary-bypass surgery
and 5.0 % for total hip replacement. Suppose an institution changes from an academic
institution to a private for-profit institution. They find that after the change, of the first 20
patients receiving coronary-bypass surgery, 5 die, while of 20 patients receiving total hip
replacement, 4 die.

(a) What is the probability that of 20 patients receiving coronary-bypass surgery, exactly
5 will die in-hospital, if this hospital is representative of the total pool of 981
hospitals?
(b) What is the probability of at least 5 deaths among the coronary-bypass patients?

6
 (c) What is the probability of no more than 5 deaths among the coronary-bypass
patients?
(d) What is the probability that exactly 4 will die among the hip-replacement patients?
(e) What is the probability that at least 4 will die among the hip-replacement patients?
(f) What is the probability of 4 or fewer deaths among the hip-replacement patients?
(g) Can you draw any conclusions based on the results in Problems item a to g regarding
any effects of the change in hospital administration on in-hospital mortality rates?

2. Please read following article:

Yang, J. H., Huang, P. Y., Shie, S. S., Huang, C. G., Tsao, K. C., & Huang, C. T. (2012).
Diagnostic capacity of rapid influenza antigen test: Reappraisal with experience from
the 2009 H1N1 pandemic. Journal of Microbiology, Immunology and Infection, 45(2),
102-107. doi: http://dx.doi.org/10.1016/j.jmii.
2011.09.027

Please examine carefully on the tables 1 and 2 and then answer following questions:

(a) What is the sensitivity of the the rapid influenza antigen test ( RIAT)?
(b) What is the specificity of the RIAT?
(c) What is the probability that a person randomly selected from the target population
who tests positive RIAT is actually H1N1 positive? (this is called the predictive value of
a positive test or PVP)
(d) What is the probability that a person randomly selected from the target population
who tests negative is actually H1N1 negative? (this is called the predictive value of a
negative test or PVN )

Can you repeat their calculation for some numbers presented in the Table 1 of the article?
If yes please demonstrate use 2 by 2 tables! If you cannot recalculate, why?
Can you repeat the results of table 2 for the PVP as well as PVN? If you cannot recalculate,
why?

7
5 Output:

a. illustrate the concept of random variables and probability distributions having some
characteristics (normal, Poisson etc.)

b. apply the concept of continues and discrete probability distributions for health daya

c. apply the concept of probabilities applied to diagnostic test data

6 Required Reading:

1. Giavarina, D. (2012). Tools for critical appraisal of evidence in studies of diagnostic

accuracy. Autoimmunity Reviews, 12(2), 89-96. doi: http://dx.doi.org/10.
1016/j.autrev.2012.07.004

2. Kelly H, Bull A, Russo P, McBryde, ES (2008). Estimating sensitivity and specificity from
positive predictive value, negative predictive value and prevalence: application to
surveillance systems for hospital-acquired infections. Journal of Hospital Infection 69, 164-
168 doi: http://dx.doi.org/10.1016/j.jhin.2008.02.021

3. Siyam, N. Wilopo S.A. Hakimi M (2015). Asupan Vitamin D Rendah dan Keparahan Demam
Berdarah Dengue pada Anak Usia 1-14 Tahun. Jurnal Kesehatan Masyarakat Nasional Vol.
9, No. 1, pp: 87-93.
7 Log sheet:

No Activities Date Signature Comment

1. Group Discussion on the concept
of probability, random variable, discrete
and continues probability distributions

2. Assignment: application standard and

other types of statistical distributions in

8
 health practice,including application
Bayes Theorem, i.e.: estimation of
sensitivity and specificity, the predictive
value of a positive test or
negative test
Score : ____________________

Instructor,