Week 7 USMLE Step 1 Review: Biostatistics and Nutrition

Download Report

Transcript Week 7 USMLE Step 1 Review: Biostatistics and Nutrition

Steven Katz MSIV
PART 1: BIOSTATISTICS

Terms:
 Independent variable: values that are controlled or
selected by the person experimenting to determine
its relationship to an observed phenomenon (the
dependent variable).
 Dependant variable: the observed phenomenon,
usually cannot be changed.
 In summary:
○ Independent variables answer the question "What do
I change?"
○ Dependent variables answer the question "What do I
observe?"
Types of Studies (p.60)

Case Control: Compares a group of people with
a disease to a group without.
 Asks “what happened?”
 Two types:
○ Observational and Retrospective
Famous example is lung cancer link to smoking
 Issues:

 Confounding: a variable that correlates to both
dependant and independent variables.
 Cannot prove cause and effect of risk factor to
variable
Types of Studies (p.60)

Cohort: Compares a group with a given risk
factor to a group without
 Assesses whether the risk factor increases the
likelihood of disease
 Asks “what will happen”
 Two types:
○ Observational and Prospective

Used to prove cause and effect of smoking to
lung cancer.
Types of Studies (p.60)

Cross-Sectional: Collects data from a
group of people to assess FREQUENCY
of disease (and related risk factors) at a
particular point in time.
 Asks “what is happening?”

Example: political polls
Types of Studies (p.60)

Twin Concordance: Compares the
frequency with which both monozygotic
twins or both dizygotic twins develop a
disease
 Measures heritability

Example: look at incidence of diabetes
in twins
Types of Studies (p.60)

Adoption: Compares siblings raised by
biologic v. adoptive parents
 Measures heritability and influence of
environmental factors

Famous examples are Swedish
adoption studies
Clinical trials (p.60)

Experimental study involving humans.
Compares therapeutic benefits of 2 or
more treatments, or of treatment and
placebo.
 Highest quality study is double-blind
randomized control trial.
Clinical trials (p.60)
Study Sample
Purpose
Phase I
Small number of pts,
usually healthy volunteers
Assess safety, toxicity,
and pharmacokinetics
Phase II
Small number of pts with
disease of interest
Assesses treatment
efficacy, optimal dosing,
and adverse effects
Phase III
Large number of pts
randomly assigned to either
the treatment under
investigation or to the best
available treatment (or
placebo)
Compares the new
treatment to the current
standard of care.
Is more convincing if
double-blind
Meta-analysis (p.60)

Pools data from several studies to come
to an overall conclusion.
 Achieves greater statistical power and
integrates results of similar studies
Highest echelon of clinical evidence
 May be limited by quality of individual
studies or bias in study selection

Evaluation of diagnostic tests (p.61)

2 x 2 table (TN = True neg, TP = True pos, FP
= false pos, FN = false neg)
DISEASE
TEST
+
-
+
TP
FP
-
FN
TN
Evaluation of diagnostic tests (p.61)

Sensitivity = TP/(TP+FN) = 1-FN rate
 Proportion of all people with disease who test
positive
 Value approaching 1 is desirable for RULING
OUT disease and indicates low false negative
rate.
 Used for SCREENING in diseases with low
prevalence
SNOUT = SeNsitivity rules OUT
 If sensitivity = 100% then all negative tests
are TN (TP/(TP+FN) = 1) because FN = 0

Evaluation of diagnostic tests
(p.61)

Specificity = TN/(TN+FP) = 1-FP rate
 Proportion of all people without disease who test
negative
 Value approaching 1 is desirable for RULING IN
disease and indicates a low FP rate
 Used as a CONFIRMATORY test after a positive
screening test
SPIN = SPecificity rules IN
 If specificity = 100% then all positive tests
are TP (TN/(TN+FP) = 1) because FP = 0

Evaluation of diagnostic tests
(p.61)

Positive Predictive Value (PPV) =
TP/(TP+FP)
 Proportion of positive tests that are true
positives
 Probability that a person actually has the
disease given a positive test result

Note: If the prevalence of a disease is
low then even tests with high specificity
or sensitivity will have LOW PPV
Evaluation of diagnostic tests (p.61)

Negative Predictive Value (NPV) = TN
/(TN+FN)
 Proportion of negative tests that are true
negatives
 Probability that a person actually is disease
free given a negative test result

http://gim.unmc.edu/dxtests/bayes.htm
Evaluation of diagnostic tests (p.61)
A = 100%
sensitivity
 B= most
accurate
 C = 100%
specificity

Prevalence v. Incidence (p.62)

Prevalence = TOTAL cases in a population at a given time
total population at risk at a given time

Incidence = NEW cases in a population over a time period
total population at risk during that time
Prevalence = incidence X disease duration
 Prevalence > Incidence for chronic dz’s
 Prevalence = incidence for acute dz’s

Odds ratio (p.62)

For case control studies
 (a/b)/(c/d) = ad/bc
 Odds of having disease in exposed group
divided by odds of having disease in
unexposed group
 Approximates the relative risk if prevalence
of disease is not too high
Relative risk (p.62)
For cohort studies
 Relative probability of getting a disease
in the exposed group compared to the
unexposed group

 [a/(a+b)]/[c(c+d)]
 Calculated as a percent of exposed pts with
dz to unexposed pts with dz
Attributable risk (p.62)
The difference in risk between exposed
and unexposed groups
OR
 The proportion of disease occurrences
that are attributable to the to the
exposure
 (e.g. smoking causes 1/3 of cases of
pna)
 [a/(a+b)] – [c/(c+d)]

Odds ratio, relative risk, attributable risk
(p.62)
Attributable risk = [a/(a+b)] – [c/(c+d)]
Odds ratio: (a/b)/(c/d) = ad/bc
Relative Risk: [a/(a+b)]/[c(c+d)]
Disease
Risk
Factor
+
-
+
a
c
b
d
Precision v. accuracy (p.62)

Precision:
 The consistency and reproducibility of a test
○ RELIABILITY
 The absence of random variation in a test
 Random Error: reduced precision in a test

Accuracy:
 The trueness of test measures
○ VALIDITY
 Systematic error: reduced accuracy in a test
Precision v. accuracy (p.62)
Neither Precise Nor Accurate
This is a randomlike
pattern, neither
precise nor accurate.
The darts are not
clustered together and
are not near the bull's
eye.
Accurate, Not Precise
This is an accurate
pattern, but not
precise. The darts
are not clustered,
but their 'average'
position is the
center of the bull's
eye.
Precise, Not Accurate
This is a precise
pattern, but not
accurate. The darts
are clustered
together but did not
hit the intended
mark.
Precise and Accurate
This pattern is
both precise and
accurate. The
darts are tightly
clustered and their
average position is
the center of the
bull's eye.
Bias (p.63)
Occurs when 1 outcome is systematically favored
over another
 Systematic errors:

 Selection bias: nonrandom assignment to study group
 Recall bias: knowledge of presence of disorder alters
recall by subjects
 Sampling bias: subjects are not representative relative to
general pop; therefore, results are not generalizable
 Late-look bias: information gathered at an inappropriate
time
 Procedure bias: subjects in different groups are not
treated the same
○ E.g. more attention is paid to treatment group, stimulating greater
compliance
 Lead time bias: early detection confused with increased
survival
Bias (p.63)
 Confounding bias: occurs with 2 closely associated
factors
○ The effect of the 1 factor distorts or confuses the effect of the other
 Pygmalion effect: occurs when a researcher’s belief in
the efficacy of the treatment changes the outcome of that
treatment
 Hawthorne effect: occurs when the group being studied
changes its behavior to meet the expectations of the
researcher

Ways to reduce bias:




Blind studies
Placebo responses
Crossover studies (each subject is its own control)
Randomization
Statistical distribution (p.63)

Normal, Gaussian, bellshaped curved
 Mean = mode = median


Bimodal = 2 humps
Positive skew—mean
>median>mode
 Asymmetry with tail on right

Negative skew—
mean<median<mode
 Asymmetry with tail on left

Mode is least affected by
outliers
Statistical hypotheses (p.63)

Null (H0): Hypothesis
of NO DIFFERENCE
Reality
 e.g. there is no

Alternative (H1):
Hypothesis that the is
some difference
 e.g. there is some
association between the
dz and the risk factor in
the population
Study Results
difference between the
dz and the risk factor in
the population
H1
H0
H1
Power
(1-b)
a
H0
b
Error types (p.64)

Type I error (a): Stating that there IS an
effect or difference when none exists (to
mistakenly accept the experimental
hypothesis and reject the null
hypothesis)
 p = probability of making a type I error
 p is judged against a, a preset level of
significance (usually <0.05)
 “False positive error”
Error types (p.64)

Type II error (b): Stating that there is
NOT an effect or difference when one
exists (to fail to reject the null hypothesis
when in fact H0 is false)
 b is the probability of making a type II error
 “False negative error”
Error types (p.64)
If p < 0.05 then there is a less than 5%
chance that the data will show
something that is really not there.
 a = you “saw” a difference that did not
exist
 b = you did NOT “see” a difference that
does exist

Power (1-b) (p.64)

Definition:
1.
2.

Depends on:
1.
2.
3.

Probability of rejecting a null hypothesis when it is in
fact false
The likelihood of finding a difference if one in fact
exists
Total number of endpoints experienced by the
population
Difference in compliance between treatment groups
(diff in the mean values of the groups)
Size of expected effect
If you increase sample size you increase power
Standard deviation v. standard error
(p.64)





n = sample size
s = standard deviation
SEM = standard error
of the mean
SEM = s/square root
(n)
Therefore, SEM < s and
SEM decreases as n
increases
t-test v. ANOVA v.

2
c (p.65)
t-test checks difference between the MEANS of
2 groups
 Mr. T is MEAN

ANOVA checks difference between the means
of 3 or more groups
 ANOVA = ANalysis Of VAriance of 3 or more
variables

c2 checks difference between 2 or more
percentages or proportions of categorical
outcomes (NOT mean values)
 c2 = compare percentages or proportions
Correlation coefficient (r) (p.65)
r is always between -1 and +1.
 The closer the absolute value of r is to 1,
the stronger the correlation between the
2 variables
 Coefficient of determination = r2 (value
that is usually reported)]

 Provides a measure of how well future
outcomes are likely to be predicted by the
model.
Disease prevention (p.65)
1o – prevent disease occurrence (e.g.
vaccination)
 2o – early detection of disease (e.g. Pap
smear)
 3o – reduce disability from disease (e.g.
exogenous insulin for diabetics)
 PDR:

 Prevent
 Detect
 Reduce disability
Important prevention measures (p.65)
Risk Factor
Services
Diabetes
Yearly eye exam, weekly self foot exams,
urine tests for microalbuminemia
Drug Use
Hepatitis immunizations, HIV, PPD for TB
Alcoholism
Influenza, pneumococcal immunization, PPD
for TB
Overweight
Fasting blood sugar test for diabetes
Homeless, recent immigrant,
inmate
PPD for TB
High-risk sexual behavior
Test for HIV, hepatitis B, syphilis, Gonorrhea,
Chlamydia
Reportable diseases (p.65)

Only some infectious diseases are
reportable in ALL states
 AIDS
 Gonorrhea
 Measles
 Rubella
 Shigella
Chickenpox
Hepatitis A and B
Mumps
Salmonella
Syphilis
 TB

Other diseases (including HIV) vary by
state
Reportable diseases (p.65)

Hep Hep Hep, Hooray, the SSSMMART
Chick is Gone!
 Hep A
 Hep B
 Hep C
 HIV
 Salmonella
 Shigella
 Syphilis
 Measles
 Mumps
 AIDS
 Rubella
 TB
 Chickenpox
 Gonorrhea
Leading causes of death in US by age
(p.66)
Infants
Congenital anomalies, short gestation/low birth weight,
SIDS, maternal complications of pregnancy, respiratory
distress syndrome
Age 1-14
Injuries, cancer, congenital anomalies, homicide, heart
disease
Age 15-24
Injuries, homicide, suicide, cancer, heart disease
Age 25-64
Cancer, heart disease, injuries, suicide, stroke
Age 65+
Heart disease, cancer, stroke, COPD, pneumonia,
influenza
Part 2: NUTRITION
Basal Metabolic Rate
Metabolism of the body at rest
 Heat production of the body when in a
state of complete mental and physical
rest and in the post-absorptive state.
 BMR can be estimated at 20-25
Cal/kg/day
 Varies between people and changes
throughout life.

 High when you are young, slows as you
age.
Resting Energy Expenditure
Energy expended in the post-absorptive
state and is approx 10% higher than
BMR
 Males: REE = 900 + 10W
 Females: REE = 700 + 7W

 W is weight in kilograms

REE is then adjusted for physical activity
by multiplying 1.2 for sedentary, 1.4 for
moderately active, or 1.8 for very active
individuals.
Caloric Requirement

Age and Caloric requirements:
 3 mo: 28 Cal/kg
 9-12 mo: 6 Cal/kg
 2-5 y/o: 2 Cal/kg
 9-17 y/o: 1 Cal/kg

10% reduction in energy allowance for
adults > 50 y/o.
Caloric Requirement
Unstressed hospitalized pts require 1.2
times their REE
 Stressed, febrile, catabolic pts require
1.5-2 times their REE

Question 7 of 40
A 79-year-old African-American female is admitted to the hospital for progressive shortness of breath. She
has no previous history of pulmonary insufficiency, and no history of emphysema, although she did smoke
one pack per day until she was 60. The symptoms started three weeks prior to admission, and were gradual
in onset. She has not had a cough, fever, or chest pain. She does have a history of hypertension, glaucoma,
arthritis, kidney stones, and hysterectomy. Medications at the time of admission include amlodipine,
ibuprofen, and eye drops. She is allergic to sulfur and penicillin, both of which caused a rash. Family history is
significant for colon cancer, breast cancer, arthritis, diabetes, and hypertension. Social history reveals that the
patient was married for forty years, but her husband died three months ago from heart failure. She lives
alone.
A chest x-ray at admission is suspicious for a mass in periphery of the left lower lung, and a follow up CAT
scan is suspicious for malignancy. Consultation is obtained from a pulmonologist, who performs a video
assisted thorascopic surgery (VATS) and biopsy. The pathology result reveals small cell carcinoma. An
oncologist is called for an opinion, and recommends chemotherapy since the tissue type indicates a good
chance of success. The problem is that the patient refuses treatment. She denies any depressive symptoms,
appears to be awake, alert, and oriented. She answers questions appropriately and does not appear to be
suffering from delirium or dementia.
As the patient's primary care physician, you would like to respect the patient's autonomy, but are concerned
about the consequences of her decision to forgo treatment. She has indicated to you that she understands
the proposed treatment options and that she understands how they relate to her situation. You decide to:
(A) Assess her competence by administering a bedside mental status examination
(B) Enlist the help of family members who may be able to change the patient's mind
(C) Respect her decision if she can demonstrate and communicate ability to reason
(D) Consult adult protective services because she is no longer able to care for herself
(E) Declare her incompetent and ask the oncologist to administer the chemotherapy
C Respect her decision if she can demonstrate and communicate ability to reason
Competence is a legal term, capacity is a medical term. Physicians are often called on to make a determination of
a patient's capacity to make medical decisions. The patient's primary care provider is an ideal person to make
the assessment as they have background knowledge of the patient's educational level, values, and medical
history.
A psychiatrist may be needed if overlying psychiatric problems make it difficult to determine capacity for judgment
or ability to reason. Courts make the ultimate determination of competence, although there is usually
concordance with the medical determination of capacity. Only lack of competence has legal ramifications,
however.
A bedside mental status examination may help to determine capacity, but in and of itself does not determine
competence. If the patient is deemed to have the capacity to make her own decisions, it may be detrimental
to encourage family member involvement in the decision making process.
Adult protective services are usually called to investigate cases of abuse or neglect, not issues of capacity or
competence. If still unclear, a psychiatrist or ethics board consultation could be utilized to help determine the
patient's capacity to make her own decisions.
Four main criteria should be used to determine a patient's capacity to make medical decisions.
1) They can demonstrate understanding of the treatment options.
2) They can demonstrate understanding of how the different options affect their own individual situation.
3) They can demonstrate ability to reason with the above information, using either evidence based in fact, or
personal beliefs rooted in their value system.
4) They are able to demonstrate 1-3 and can communicate a choice.
Question 37 of 40
A 47-year-old male presents to his primary care physician complaining of markedly
increased feelings of stress secondary to recent changes at his workplace. Which of
the following statements about stress and its health effects is true?
(A) Stress does not include emotionally negative responses such as anger and
hostility
(B) It is a factor in 10-20% of health problems
(C) Assertiveness training is unlikely to help an individual to avoid stress
(D) It is in the differential diagnosis for diarrhea
(E) It is the third leading cause of disability claims in California
D It is in the differential diagnosis for diarrhea
The definition of stress is an individual's negative emotional response to
a perceived inability to meet demands place on him or her. It may
express itself as anger, hostility, or feelings of helplessness, loss of
control, or victimization. It is believed to be a factor in 60-80% of all
health problems, and is the leading cause of disability claims in
California. Major symptoms include fatigue, exhaustion, tight back and
shoulders, insomnia, anxiety, anger, headaches, depression, sadness,
hopelessness, colds, indigestion, diarrhea, and ulcer symptoms.
Effective prevention and avoidance techniques include assertiveness
training and the development of communication skills. Treatment
methods include relaxation techniques, meditation, exercise, and
participation in enjoyable activities.
Question 1 of 40
A 20-year-old man arrives at the emergency room asking for a strong pain killer
because he is in serious pain. The attending physician notices that he is very
anxious and is sweating. The man states that he has no appetite, he has a runny
nose, nausea, stomach cramps and diarrhea. He said that he took his temperature at
home at it was 100F. While he talks he is continuously yawning and restless. The
attending physician recognizes that he is abusing a certain substance and is
experiencing withdrawal. Which substance is it?
(A) Alcohol
(B) Cocaine
(C) Amphetamines
(D) Barbiturates
(E) Opioids
E Opioids
The patient is experiencing the classic symptoms of withdrawal from opioids
which are anxiety, insomnia, anorexia, sweating, piloerection, fever, rhinorrhea,
nausea, stomach cramps, diarrhea, yawning. Symptoms usually appear within
8 to 10 hours after abstinence. The onset is longer if methadone has been
withdrawn. These symptoms peak within 48 to 72 hours and then disappear in
7 to 10 days. Methadone lessens the effects of withdrawal. It should be given
no more than 20-50mg/day.
Alcohol withdrawal appears within a few hours of stopping or decreasing
alcohol consumption. It lasts for three to four days and sometimes as long as a
week. The patient experiences tachycardia, tremulousness, diaphoresis,
nausea, orthostatic hypotension, malaise, anxiety, and irritability.
Benzodiazepine should be administered in a tapering dose over three days.
Cocaine withdrawal is classified by psychological symptoms such as
increased sleep, REM rebound causing nightmares, lassitude, increased
appetite, depression, and suicide attempts. Treatment would consist of an
antidepressant such as bupropion. Amphetamine withdrawal would include a
post use crash, including anxiety, lethargy, headache, stomach cramps, hunger,
severe depression, dysphoria mood, fatigue, and insomnia or hypersomnia.
Barbiturate withdrawal is characterized by anxiety, seizures, delirium, and life
threatening cardiovascular collapse.
Question 4 of 40
The city of Cancerville had a population of 10,000,000 (50% women) in
1995. In 1995, there were 80,000 women with previously diagnosed
ovarian cancer in Cancerville. Twenty thousand new cases of ovarian
cancer were diagnosed in 1995. What was the incidence rate of ovarian
cancer in Cancerville in 1995?
(A) 2000 per One hundred thousand population
(B) 4000 per One hundred thousand population
(C) 200 per One hundred thousand population
(D) 400 per One hundred thousand population
(E) 1,000 per One hundred thousand population
D 400 per One hundred thousand population
The incidence rate is the number of new cases
of a disease during a specific period per
population at risk. Twenty thousand divided by 5
million women gives a rate of 1 case per 250
women, or 400 cases per One hundred thousand
populations.
Question 8 of 40
A laboratory has developed a new test for rapid ascertainment
of serum parathyroid hormone levels. The test is repeated
twenty times on the same sample with a resulting coefficient of
variation of one percent. This is a measure of
(A) Accuracy
(B) Reliability
(C) Precision
(D) Validity
(E) Mode
B Reliability
-The mode is the most commonly occurring value in a series of
data.
-Reliability is a measure of the reproducibility of a test over
different conditions. The four most common types are interobserver reliability, intra-observer reliability, split-sample reliability,
and repeat testing reliability.
-Accuracy is a measure of the extent to which a test approximates
the real value of that which is measured. New tests are measured
against the gold standard, if one exists.
-Validity is the assessment of the degree to which a test measures
that for which it was designed. In other words, you need to
determine whether it reflect the outcome of interest or other
outcomes.
-Precision is the degree to which a measurement is not subject to
random variation.
Question 10 of 40
At a large university, a study of pulse rates at rest was conducted
on 5000 students. The mean pulse rate was 70, with a standard
deviation of 10. Which of the following statements is true?
(A) Approximately 95% of the students had pulses between 60
and 80
(B) Approximately 68% of the students had pulses between 60
and 80
(C) Approximately 99.7% of the students had pulses between 50
and 90
(D) Approximately 95% of the students had pulses between 40
and 100
(E) Approximately 68% of the students had pulses between 50
and 90
B Approximately 68% of the students had pulses
between 60 and 80
When a test is conducted on a normally distributed
population, 68% of the population will have values within
one standard deviation of the mean, 95% of the
population will have values within two standard
deviations of the mean, and 99.7% of the population will
have values within three standard deviations of the
mean. Therefore, in this population, 68% of the pulses
will be between 60 and 80, 95% between 50 and 90, and
99.7% between 40 and 100.
Question 13 of 40
A statistician analyzes data for several academic departments. She is free to
choose the appropriate methodology to her perform her analyses. Which of
the following data would best be analyzed by non-parametric statistical
methods?
(A) Results of a study on the effect of a new lipid-lowering drug on LDL
cholesterol
(B) Results of a study on the effect of asbestos exposure on forced vital
capacity
(C) Results of a study on the relationship between gender and lung cancer
(D) Results of a study on the differences in weight distributions between
children in different countries
(E) Results of a study on the relationship between hemoglobin and
reticulocyte count
C Results of a study on the relationship between gender and lung cancer
Parametric techniques can be used to analyze data where at least one of the
variables is quantitative (interval or ratio) and where the data is distributed
normally. If the data is not distributed normally or both variables are qualitative
(nominal or ordinal), non-parametric techniques must be used. Gender and
lung cancer are both qualitative variables, so non-parametric techniques, such
as chi-square, are used to determine the relationship between them. LDL
cholesterol, forced vital capacity, hemoglobin, and reticulocyte count are
quantitative ratio variables, so studies involving them can be analyzed using
parametric techniques, assuming they are normally distributed. The use of a
new lipid-lowering drug and the presence or absence of asbestos exposure is
qualitative nominal variables. Weight is a quantitative ratio variable, and various
parametric techniques can be used to compare the means, ranges, and
variances of distributions between populations.
Question 15 of 40
You are doing a research project on comparing the effectiveness of
cognitive-behavioral versus psychoanalytic therapy in depressed
patients. Your subjects consist of 60 outpatient females being seen at the
local college clinic. They are randomly assigned to three groups: those
who will receive cognitive-behavioral therapy, those receiving
psychoanalytic therapy, and a third group that receives no therapy to
serve as a control group. In your study what is the independent variable?
(A) the subjects participating in the different therapy groups
(B) the therapies being compared in the study
(C) the subjects receiving no therapy
(D) the level of depression in the participants at the end of the study
(E) the assignment of the participants into the separate groups
B the therapies being compared in the study
The independent variable is defined as the variable that is to be
manipulated or controlled, or that has been selected by the
researcher. In the study, as the researcher you are controlling the
type of therapy to be utilized in the study. You are also controlling
whether or not the participants are receiving any therapy at all.
The subjects that are participating in the different therapy groups
and that have been assigned to serve as the control group are the
sample being used in this study. The sample simply means the
participants chosen to represent the larger population. The level of
depression in the participants at the end of the study is considered
to be the dependent variable. The dependent variable is defined
as the response to the independent variable (or therapy), the
observed or measured behavior, or the outcome of the study.
Question 21 of 40
You are doing a study on the distribution of IQ scores in 15-year-old
adolescent males in a standard high school classroom. You have chosen
one school from Los Angeles, Seattle, Dallas, Miami, Chicago and New
York. The WISC-III is administered to all 15-year-olds in the schools
selected. After all tests have been administered, the scores are collected
and the distribution of the scores is analyzed. The IQ scores represent what
type of statistical measurement scale?
(A) Nominal
(B) Ordinal
(C) Interval
(D) Ratio
(E) Correlational
C interval
In statistical measurements, IQ is considered an interval scale because the
difference between an IQ of 90 and 100 is indistinguishable from the difference
in an IQ of 100 and 110. In interval scales, the difference between intervals is
relative. The difference between 1 and 2 is relative to the difference between 3
and 4. Nominal measurements are used for variables in which there are no
numerical values that can be compared, such as gender or ethnic background.
Ordinal scales are used for rank ordering. Ordinal scales can be used for such
variables as attractiveness, or grades in school. In each case one can state that
s/he is more attractive then, or an A or B is better than a C or D. Ratio scales
are based in measurements where there is an absolute 0. In IQ's there are no
absolute zeros, and one cannot state that an IQ of 50 is half as good as an IQ
of 100. Ratio scales can be used for variables such as the number of hours a
student spends studying, 2 hours of studying would be half as many hours as 4
hours of study. Correlations are not used as a method of statistical
measurements, but are used in research and statistics to define a relationship
between two variables.
Question 24 of 40
A researcher studied the levels of serum calcium in the U.S.
and Panama. The null hypothesis was proven. What does the
null hypothesis state?
(A) There is a significant difference between populations tested
(B) Difference between populations is not attributable to chance
(C) Difference between populations is due to a particular factor
(D) There is no significant difference between populations
tested
(E) Power of a study to detect a significant difference between
populations is nil
D There is no significant difference between
populations tested
The null hypothesis states that there is no significant
difference between the populations being tested, and
that any difference that is found is attributable to chance.
It is tested against the alternative hypothesis, which is
that there is a significant difference between the
populations tested.
Question 29 of 40
The public health officials of a particular city wish to evaluate the lead levels
of its constituents. In order to develop a sample population, they choose
every 10th family in the city for the study. This is an example of what kind of
population sample?
(A) Stratified selected sample
(B) Cluster selected sample
(C) Simple random sample
(D) Systematically selected sample
(E) Nonrandom selected sample
B Cluster selected sample
In cluster selected samples, the population of interested is
divided into subunits, such as families, and a random sample of
these units is used.
In simple random samples, each individual member of a
population has an equal probability of being chosen.
In stratified selected samples, individuals are chosen randomly
from within stratified groups, such as age groups.
In systematically selected samples, the population is ordered
by some characteristic, such as age, a starting point for selection
is randomly selected, and then the remainder of the sample is
collected by a predetermined scheme, such as choosing every x
number of people.
In nonrandom selected samples, some predetermined scheme
is used, such as the first x number of people presenting for a
certain disease to a clinic.
Question 31 of 40
A physician wants to learn more about prevalence rates for diabetes
mellitus in his local community. He has raw data from his town public
health department, but he is not sure how to determine the prevalence
rates. Which of the following comments is true of prevalence rates?
(A) Reflect a portion of specific illnesses in a population
(B) Include new and existing cases during a specific time period in the
numerator
(C) Denominator is the entire population, both those at risk and those not at
risk
(D) Include only cases prevalent at the start of the time period in the
numerator
(E) Are not influenced by the duration of disease
B Include new and existing cases during a specific time
period in the numerator
Prevalence rates are determined as the number of new
and existing cases of disease during a specific time
period in the numerator divided by the population at risk
in the denominator. They are influenced by both the
duration of disease and the incidence of new cases. By
measuring both existent and new cases of illness, they
reflect the total amount of specific illnesses in a specific
population.
Question 36 of 40
You are conducting an experiment on the effectiveness of behavioral therapy in
treating social anxiety. Your research hypothesis is that behavioral therapy is
effective in reducing social anxiety. The participant's in your study are 30
individuals who have been diagnosed with social anxiety. Each individual is
independently evaluated for social anxiety to confirm the diagnosis. After the
evaluation, 6 participants are found to not meet the set criteria for social anxiety
and are dropped from the study. The remaining 24 participant's are broken up into
two separate groups. Group A receives behavioral therapy and group B is put on
a wait-list to receive therapy after the experiment is over. At the end of the
experiment, you find that behavioral therapy was effective in treating social
anxiety. In your study what is the independent variable?
(A) The subjects participating in the treatment group
(B) The treatment administered to group A
(C) The subjects in the no treatment group
(D) The level of social anxiety in the participants at the end of the study
(E) The assignment of the participants into the separate groups
B The treatment administered to group A
The independent variable is defined as the variable that is to be
manipulated or controlled, or that has been selected by the
researcher. In this study, as the researcher, you are controlling
whether or not participants receive therapy. The subjects that are
participating in therapy and those that have been assigned to
serve as the control group are the sample being used in this
study. The sample simply means the participants chosen to
represent the larger population. The level of social anxiety in the
participants at the end of the study is considered to be the
dependent variable. The dependent variable is defined as the
response to the independent variable (or therapy), the observed
or measured behavior, or the outcome of the study.
Question 40 of 40
A researcher studied the relationship between childhood exposure to lead
and stature. The heights of the children measured at age 12 range from 4'8"
to 5'9", with a standard deviation of 5", a mean of 5'3", a mode of 5'2", and a
coefficient of variation of 7.9%. Which of the following statements is true?
(A) Variance is the square root of the standard deviation
(B) Range of a series of data provides information about the distribution of
the data
(C) Coefficient of variation is a measure of the spread of the data in regard
to the mean
(D) Standard deviation is an estimate of the standard error of a population
(E) Mode is a measure of central tendency of a data series
C Coefficient of variation is a measure of the spread of the data
in regard to the mean
The coefficient of variation is defined as the standard deviation
divided by the mean, expressed as a percentage. It is a measure
of the spread of the data with regard to the mean.
The standard deviation is the positive square root of the
variance.
The standard error is an estimate of the standard deviation of a
population.
The range of a series of a data is calculated as the highest value
in the series minus the lowest value, and it provides no
information about the distribution of data within the series.
The mode is the most commonly occurring value in a data series
and does not provide any information about the central tendency
of a data series.
Question 2 of 40
A researcher is preparing a paper for publication on characteristics of
hepatitis C infection in her local population. It includes exposure and
treatment information. She reports that female sexual partners of men with
hepatitis C virus are twice as likely than other women in the same population
to contract the hepatitis C virus. This is a measure of
(A) Type I (alpha) error
(B) Odds ratio
(C) Prevalence
(D) Attributable risk
(E) Bias
D Attributable risk
Attributable risk, which can be determined from cohort studies, is
a measure of the difference in occurrence of disease between
exposed and unexposed populations. The likelihood that a
positive result is due to chance is a measure of type I (alpha)
error.
Prevalence is the amount of disease existing in a population at a
certain point in time.
The odd ratio is a measure of the estimated relative risk occurring
due to certain factors. Confounding variables may cause bias in
studies.
Question 6 of 40
In reporting the results from a clinical study of a new anti-inflammatory drug for the
treatment of post-operative pain, the study's authors present data comparing the
total days of hospitalization for comparable groups of patients who have received
either the investigative anti-inflammatory drug or a placebo. The attached table
appears in their report. Which of the following would be a valid interpretation of the
data presented in this table?
(A) The p-value is greater than 0.05, indicating that there is no true treatment effect upon total days of post-operative
hospitalization
(B) The treatment group and placebo groups have unequal numbers of participants, and therefore the statistical test
results are not interpretable
(C) The results are suggestive of a true treatment effect, but the study has limited power to detect the effect due to
the relatively small number of study subjects
(D) Statistical testing of two group means yields a t-value, not a p-value
C The results are suggestive of a true treatment effect, but the study has
limited power to detect the effect due to the relatively small number of study
subjects
While the p-value for the differences between the mean days of post-operative
hospitalization is not below the conventional level of 0.05, it is relatively close to
that value. The values of the treatment group and placebo group means (3.0
and 4.5 days, respectively) do suggest that there is an effect of treatment. It is
likely that the statistical power of the study is rather limited, given the modest
number of people enrolled in each group. Ideally, this study would be repeated
with larger numbers of study subjects in each of the two groups. While it would
be a mistake to conclude that there was definitively a treatment effect, it would
also be a mistake to conclude that there was no evidence for a treatment effect,
as well.
In clinical trials, it is not necessary that the comparison groups have identical
numbers of subjects, although there should be a sufficient number of
participants in each study group to effectively evaluate the treatment being
considered. While statistical testing of two group means may use the t-test, it is
possible to derive a p-value from the use of this test.
Question 9 of 40
Suppose that a researcher is using hypothesis testing to determine whether two
treatments are equally effective. The hypotheses being tested are given below.
H0: Treatment A and Treatment B are equally effective
Ha: Treatment B is more effective than Treatment A
The study used an a-level of a = 0.05. The power of the test was 0.80. What is the
probability that H0 will be rejected if in fact the two treatments are equally effective?
(A) 0.05
(B) 0.20
(C) 0.80
(D) 0.95
(E) It is impossible to tell from the information given
A 0.05
When a researcher uses hypothesis testing, the researcher can never be certain that the conclusion he/she draws
is correct. The decisions a researcher makes versus the truth can be portrayed by the following table.
TRUTH
Ho True
RESEARCHER
Correct Decision
ACCEPTS Ho
RESEARCHER
ACCEPTS Ha
Type II Error
(Probability a)
Ha True
Type II Error
(Probability b)
Correct Decision
If H0 is true, but by chance the data suggested strong enough evidence against H0 to reject H0, then a type I
Error has been committed. The probability of a Type I Error is the a-level of the test. Therefore, if a = 0.01, then
only 1% of the time will data be strong enough to reject H0 when H0 is true, resulting in a Type I Error.
If Ha is true, but the evidence against H0 was not strong enough to reject H0, then a Type II Error has been
committed. The power of a test is defined as the probability of rejecting H0 when Ha is in fact true (the ability of
the test to correctly identify a significant difference). The power of a test is directly related to the probability of
committing a Type II Error. The probability of a Type II Error is b and the power of a test is given by (1 - b). One of
the most common reasons for a Type II Error is due to sample size being too small. In general, the larger the
sample size, the greater the power of the test.
Question 11 of 40
A trial is carried out to determine the impact of a new diet combined with exercise in addition to conventional therapy to further
reduce the risk of dying in patients recovering from heart surgery. Patients are assigned to one of the two study arms:
1- Conventional therapy only
2- Conventional therapy plus new diet plus new exercise program.
Patients are followed up every two months for the first year and then every six months for the next four years. Among other
factors, the following information is collected:
1) Sex
2) Age at time of surgery
3) Weight (at entry into trial and at each visit)
4) Percentage of body fat (at entry and at each visit)
5) Survival status and date of death where applicable
6) Need for further surgery and date where applicable
7) A grading for actual activity level (1 to 5 with 1=Sedentary & 5=Very Active)
Refer to the attached trial description. What study design is this?
(A) Case-Control study
(B) Cohort Study
(C) Randomized Clinical Trial
(D) Cross Over Study
(E) Cross Sectional
C Randomized Clinical Trial
Two study arms are present. In the first one, only the conventional therapy is
present. In the second, diet and exercise are added to conventional therapy.
This is, therefore, an experimental study. The patients are assigned to only one
of the two study arms. Due to the nature of the intervention (diet plus exercise),
patients are unblinded to their study group. This is a Randomized Clinical Trial.
In a cross-over study, patients are assigned to one of the study arms for a
period of time and then assigned to the other study arm for the same length of
time.
The other study designs mentioned are all observational studies. In casecontrol studies, people with and without a specific outcome are chosen. Then,
looking backward in time, one tries to detect possible causes or risk factors. In
cross sectional studies, data is collected at one time. Large governments
surveys are good examples of cross sectional studies. In a cohort study, people
are selected and followed over a period of time. At the beginning of the study,
people are defined as being exposed or not exposed to certain risk factors.
They are observed over time for the development of outcome. The outcome is
then compared to exposure to risk factors.
Question 14 of 40
A researcher wishes to compare the efficacy of a COX-2 inhibitor to that of
ibuprofen for treatment of pain in patients with osteoarthritis. Using a visual
analogue scale of 1-100, a difference of 15 points between the mean values
of the treatment arms is considered to be clinically significant. Given that a
true clinically relevant difference exists between the two therapies, which of
the following is most true about the probability that the statistical test used in
the study will fail to detect the difference?
(A) The probability decreases as a decreases
(B) The probability is determined by the type-II error of the study
(C) The probability decreases as the b increases
(D) The probability is impossible to determine without knowing the true
mean
(E) The probability decreases as the power decreases
B The probability is determined by the type-II error of the study
Before a study is conducted, the researcher must select the significance level
(a), which is the value used to interpret the result of the statistical test. The a
level represents the probability that the statistical test used will detect a
clinically significant difference due to chance alone. This is the chance of a
type-I error. The a level does not predict the response of an individual patient,
or the proportion of a sample that will have a particular therapeutic outcome.
The probability of a statistical test failing to detect a difference between means
of two samples when such a difference truly exists, is the b or type-II error. As
the level of significance increases, there is a greater chance of a type-I error,
but less chance of a type II error, therefore, b decreases as a increases.
The ability of a statistical test to detect a difference between two means is the
power of the test. Power is the probability that a statistical test will detect a
difference when such a difference truly exists and is not due to chance. Power
is the complement of b, and is equal to 1-b. Therefore, b decreases as power
increases. As the level of significance, and the chance of a type-I error
decreases, b increases. Power differs from a and b in that it is not a measure of
error.
Question 17 of 40
In a study of the effects of a new treatment for ovarian cancer
on mortality, the a level is 0.05 and the b level is 0.20. What is
the power of the study to detect a change in mortality from this
new treatment?
(A) 5%
(B) 20%
(C) 25%
(D) 80%
(E) 95%
D 80%
The power of a study is the ability of the study to detect
a significant change when one exists. It is calculated as
1 - b, where b is the Type II error. In this case, 1 - b =
0.80, or 80%. Therefore, there is an 80% surety that this
study has detected a change in mortality with this new
treatment when one exists. Or, in other words, 20% of
the time it will have missed a significant difference when
one exists.
Question 20 of 40
You and your colleagues are conducting a small clinical trial concerning the
management of pediatric asthma. The clinical trial involves three different
treatment arms and one placebo arm. The outcome of interest is
hospitalization for respiratory distress. In one treatment arm (n=31), there
are no patients that require hospitalization during the follow-up period (i.e., 0
events). What is the upper 95% confidence bound for the rate of
hospitalization for the 31 subjects in this treatment arm?
(A) The upper 95% confidence bound cannot be calculated from the data
provided
(B) 0
(C) 0.10
(D) 0.15
(E) 0.22
C 0.10
The answer to this question is derived using the "rule of three" (as
explained by Hanley and Hand, JAMA, 1983). When there are no
events of interest observed in a particular group, the upper 95%
confidence bound can be calculated by dividing 3 by the number
of subjects in the group (i.e., n). In the question, 3/n is equivalent
to 3/31 or 0.097. Rounding up produces the answer 0.10, and
thus the largest rate that we would expect (with 95% confidence)
would be 0.10 or approximately 3.0 events in this group of 31
study subjects. The 99% confidence bound can be obtained by
using the "rule of 4.6" (i.e., 4.6/n), and the 99.9% confidence
bound can be obtained using the "rule of 6.9" (i.e., 6.9/n). While
this explanation will not go into the derivation of this rule, the
calculations underpinning the convenient statistical device are
sound and well-tested.
Question 32 of 40
During a research rotation as a medical student, you spend several months
gathering data on the use of a new oral vaccine to prevent a serious gastrointestinal
disease in primates. Your research generates the attached table of data, and you are
interested in using the c2 test to statistically test the association between vaccination
status and the subsequent development of this particular gastrointestinal disease.
After calculating the c2 value, you are interested in looking at a table of c2 values to
determine the p-value that is associated with the c2 value that you obtained with the
numbers shown in the table above. What would be the correct "degrees of freedom"
associated with this table
(A) 1 degree of freedom
(B) 2 degrees of freedom
(C) 3 degrees of freedom
(D) 4 degrees of freedom
(E) 5 degrees of freedom
A 1 degree of freedom
The shape of the c2 distribution changes according to the number
of degrees of freedom (df) involved in a particular testing situation.
Thus, in order to determine the correct p-value associated with a
particular c2 value, it is necessary to know the correct degrees of
freedom. For contingency tables, the correct degrees of freedom
is obtained with the following formula: df = (r-1)(c-1), where r is the
number of rows, and c is the number of columns. In a table with 2
rows and 2 columns, the c2 test will have 1 degree of freedom.
Question 36 of 40
While doing morning rounds on the pediatric bone marrow
transplantation unit at a large university-affiliated medical center, the
attending hematologist-oncologist asks you about the allocation of
patients to treatment groups in pediatric marrow transplantation clinical
trials. How should you answer her question most correctly?
(A) Patients are allocated based on prognosis
(B) Patients are allocated based on parental preference
(C) Patients are allocated by random assignment
(D) Patients are allocated based on the attending physician's clinical
judgment
(E) Clinical trials cannot be done with pediatric subjects
C Patients are allocated by random assignment
To effectively evaluate experimental agents or
procedures, randomized clinical trials must be
performed. Randomized clinical trials should be doubleblinded in all but the most exceptional circumstances,
and patient allocation should be achieved by a random
process in which each patient has the same probability
of being allocated to a specific treatment or control arm.
Allocation based on prognosis, parental preference, or
clinical judgment can lead to seriously biased results
and flawed conclusions about the efficacy of the
experimental treatment.