to screen adults over the age of 40 for unrecognized diabetes mellitus. No change in the incidence or prevalence of diabetes. …
REVIEW QUESTIONS FOR TEST 3
A pulmonologist who takes care of lots of patients with COPD always tells
them to quit smoking One day, one of his patients asks, Since I am
already sick, will it really help me live longer if I quit smoking? The
doctor goes to the literature, but cannot find any data about prognosis in
COPD So he decides to follow the next 100 patients in his practice who
quit smoking after their diagnosis and 100 others who will not quit
smoking
1 What kind of study is this?
a Case series
b Case control
c Cohort
d Randomized controlled trial
2 For this to be a valid study, which of the following would have to be
done?
a Patients would have to be randomized
b Patients in both cohorts would have to have similar stages of disease
c Patients would have to be blinded to their smoking status
d Drop-outs would have to be ignored because you wouldnt know if they
were alive or not
3 The physician does his study and comes up with the following data:
Smokers n100 Quitters n100 P value
Alive at:
1 year 95 93 055
2 years 84
85 085
3 years 71 78 026
4 years 43 62 0007
5 years 28 49 0002
What would be the best way to analyze these data:
a Plot them on a survival curve
b Run a t-test on the numbers for each year
c Do a 2X2 table to calculate predictive values
d Do a decision analysis
4 In examining the data, you can make the following conclusions:
a Smoking is not a poor prognostic factor for COPD
b Smoking is a bad prognostic factor, but only in the first 3 years after
diagnosis
c Smoking is a bad prognostic factor, but any benefit of quitting doesnt
occur until after 3 years
d Smoking is a bad prognostic factor with statistically significant
effects in every year of the study
You are seeing patients in a urology clinic and a 65 year old patient with
recently diagnosed prostate cancer tells you that the urology oncologist
has just told him that he has Stage 1 prostate cancer He has a choice of
doing nothing or having a radical prostatectomy He asks you which he
thinks he should do You do a literature search and find a paper doing
a
decision analysis of this same situation 65 year old with Stage 1 prostate
cancer In it is the following decision tree:
Die of something else 072 15 QALY
Do nothing Complication of prostate ca
012
8 QALY
Die earlier of prostate ca 016
7 QALY
Die at surgery 001
0 QALY
Success, no complications 050
15 QALY
Have surgery
Success, major complications 032
5 QALY
Success, minor complications 017
12 QALY
As the tree shows, assuming a 65 year old has a 15 year life expectancy, if
the patient does nothing and the cancer never bothers him, he has an 72
change of living another 15 years There is a 12 chance he will still
live 15 years, but have complications of the cancer, and a 16 chance
that
he will die an average of 8 years earlier from his cancer
If he has surgery, there is a 1 chance he will die during surgery He has
a 50 chance of a successful surgery with no complications and will live
his expected 15 years He has a 32 chance of major complications, though,
which will reduce his quality of life a great deal and a 17 change of
minor complications which will reduce his quality of life a little
5 If he decides to have the surgery and has a major complication, which of
the following is mostly likely to be true?
a He is likely to die in 5 years
b He is likely to live 5 years before the complications occur
c He is likely to live 5 years after the complications occur
d He is likely to live more than 5 years, but because of the complications
the quality of his life will be reduced
6 What is the approximate overall value of not having the surgery?
a 30 years
b 15 years
c 13 years
d 8 years
e 15 years
7 What is the approximately value of having surgery?
a 32 years
b 17 years
c 15 years
d 11 years
e 5 years
8 The study also says that in some very select hospitals complication
rates can be reduced to 10 for major complications and 15 for minor
complications In this case, 74 of the patients have good outcomes and
live their full 15 years of life expectancy What would you do with this
information?
a Change the QALYs to reflect the new values
b Perform a sensitivity analysis substituting these values for the
previous probabilities
c Try to get a residency at this very good hospital
d Add another probability node to the tree to incorporate the new
information
9 You are making rounds on the inpatient medicine service and hear an
overhead page that one of your patients has coded Your resident turns
to you and gives the following odds ratios and 95 confidence intervals
for which of your patients this might be:
|PATIENT |ODDS RATIO |95 Conf Interval |
|Mr Plum in dialysis with an MI |30 |07 - 52 |
|Mrs Apple in the OR with a PE |24 |15 - 35 |
|Ms Peach in radiology with a |02 |01 - 04 |
|stroke | | |
|Mr Banana in the cafeteria of |10 |05 - 20 |
|aspiration pneumonia | | |
Where should
you and the resident go to administer CPR:
A Dialysis to attend to Mr Plum
B The OR to rescue Mrs Apple
C Radiology to revive Ms Peach
D The cafeteria to perform the Heimlich maneuver on Mr Banana and get a
snack
E Nowhere because your heart just stopped
Questions 10 - 12 Match the following conclusion to the most correct
statement:
A The two groups are likely to be different statistically significant
B The two groups are likely to be the same
C The study is inconclusive
10 The P value is 005 and the Power is 80
11 The P value is 005 and the Power is 80
12 The P value is 005 and the Power is 80
13 After years of debate, a prevention expert panel decides that it is
worthwhile to screen adults over the age of 40 for unrecognized diabetes
mellitus The group decides that a rationale program for screening
should be testing all adults with a fasting blood glucose If the value
is over 120, then a 3 hour glucose tolerance test would be indicated
Over the next year, starting this program would likely cause:
A An increase in the incidence and prevalence of diabetes mellitus
B A decrease in the incidence and prevalence of diabetes mellitus
C
An increase in the incidence only
D An increase in the prevalence only
E No change in the incidence or prevalence of diabetes
Questions 14 -17 Match the following health screening/early detection
programs with the appropriate type of prevention category:
A Primary prevention
B Secondary prevention
C Tertiary prevention
D Quaternary prevention
14 Hypertension detection and management
15 Using aspirin to decrease mortality after a heart attack
16 A program to discourage pre-teens from smoking
17 Autopsy
ANSWERS AND EXPLANATIONS
1 C, cohort study In this study, the patients decide whether they are
going to quit smoking or not The investigator the doctor then
follows them ahead over time to see if the outcome of interest death
occurs
2 B, patients would have to have similar severity of disease For this to
be a valid study, you would want to be sure that the variable of
interest smoking was the primary difference between the two groups
If sicker patients kept smoking, then you wouldnt be sure if any
difference in death was due to smoking or their severity of disease If
sicker patient were the ones to quit smoking,
this difference in disease
severity might obscure any benefit conferred by smoking cessation
Answers A and C dont apply here because this is not a randomized trial
although a randomized trial would probably be a better design because
then confounders such as disease severity would be randomly distributed
in the two groups Answer D is not appropriate because you always have
to figure out whether lost-to-follow-up patients could have affected the
results
3 A, The best way to look at survival over time is a survival curve
Answer B, a t-testm compares means not proportions and gives you a P
value, which you already have 2X2 tables help you with diagnostic
tests Decision analysis helps figure out the values of different
decisions using known, or justifiably estimated, values for
probabilities out outcomes
4 If you draw a survival curve with the data, you should get something
that looks like this:
If you look at the curve, you can see that the two lines run together for
almost 3 years, then the smokers start dropping off compared to the
quitters So the answer is C You also could figure this out by looking
at the P values There is no
difference in the two groups for the first 3
years, but then in years 4 and 5, there is a statistically significant
difference
5 D, he is likely to live more than 5 years, but because of complications
the quality of his life will be reduced Quality-adjusted life years
refers not to an absolute number of years, but to the number of years
adjusted for how good your time is during those year The rationale
is that someone living with urinary incontinence, erectile dysfunction,
or chronic pain might not value those years as much as someone who is
healthy So if the person would live 7 years, because of the poorer
quality of life the QALY is reduced to 5 years The exact
determination of QALYs is complicated and usually involves asking large
numbers of people to make trade-offs in hypothetical situations
6 The overall value of not having surgery would be the sum of each
potential outcomes probability and value What that means in English
is:
15 QALY X 072 the value times the probability 108
8 QALY X 012 096
7 QALY X 016 112
1288
Which is closest to 13, answer C
7 We calculate this the same way as above;
0 QALY X 001 0
15 QALY X 050 75
5 QALY X 032 16
12 QALY X 017 204
1114
Which is answer D
8 When there are differences in some of the key assumptions, you perform a
sensitivity analysis Answer B to see if these differences change your
decision In this case, if the complication rates can be reduced, the
value of the surgery would go up making it a better option than no
surgery
9 First of all, the best thing to do in this case would be to get the
resident fired He or she should be on their way to the code and not
quoting the literature for you However, being the peon medical student
on the service without that option, you would look to see which of the
95 confidence intervals does not cross 10 which would make it
statistically significant You see that Ms Peach does not cross 1,
but her odds ratio is less than 1 making her less likely to be the one
to code However, Mrs Apple has an increased risk odds ratio greater
than 1 and a 95 CI which doesnt cross 1 So head to the OR
answer
B
10 When the P value is 005, the results are NOT statistically
significant, ie not different When the Power is over 80, the study
has sufficient power to say they are the same So it is Answer B
11 In this case the P value is 005, so the results are statistically
significant Answer A
12 Here the P value is not statistically significant ie, the two groups
are not difference, but the Power is not adequate to say they are the
same Consequently, the study is inconclusive Answer C
13 A new screening program would identify lots of people who are
symptomatic but have high blood sugar So in the first year of the
program there would be lots of new cases incident cases Also,
because more people would now be diagnosed with diabetes, the number of
people with the disease at this time prevalent cases also would
increase So the answer is A, an increase in both
14 B, secondary prevention Youve identified a problem before it has
complications
15 C, tertiary prevention The person has had a complication; you want to
make sure they dont do it again
16 A, primary prevention In this case, you want to stop exposure to a
risk factor
smoking before people even start doing it
17 D, quaternary prevention In this situation, its too late to prevent
a problem for the deceased, but by identifying a cause of death you
might be able to prevent others from dying from it
———————–
Source:musc.edu
