Primer 15 - Biostats, Bias & Error

Question 1

What is Berkson’s bias?

Answer 1

Using hospital patients as study participants. It is a type of selection bias. “Studies performed on patients that have been hospitalized.”

Question 2

What is Recall bias?

Answer 2

A difference in recall between two study groups. “Parents of autism patients having a more detailed recall of events and illnesses in their child’s first two years of life compared to parent of healthy controls.”

Question 3

What is Sampling bias?

Answer 3

When the study participants are not a random selection of parent population. This leads to a biased sample and limits the generalizability of the results. “A study performed in China may not be generalized to the US population”

Question 4

What is Late-look bias?

Answer 4

Bias that results from information being gathered too late to draw conclusion about the disease or the exposure of interest in the entire intended study population. “Sending a survey out to people diagnosed with a fatal illness 5 years after diagnosis will preferentially sample those with a low grade disease (or fewer comorbidities).”

Question 5

What is Procedure bias?

Answer 5

When study groups within the same study are not treated the same. “The positive benefit of a new drug during a study may have been due to the fact that study participants were required to attend clinic monthly, where they received extra disease education and counseling compared with the controls”

Question 6

What is Confounding bias?

Answer 6

When the association between the exposure and the outcome is distorted due to some third variable, which is associated with the exposure and the outcome. “Are asbestos miners more likely to have cancer because they mine asbestos or because they are likely to smoke?” The other two factors are close related, the effect of one can be confused for the other. “Smoking causes cancer. Smoker just happen to drink coffee. Coffee does not cause cancer, smoking does.”

Question 7

What is Lead-time bias?

Answer 7

Results from early detection of disease, and some associated with increased survival, but just because it is detected early does not change the disease course. “While test PSA-xyz may detect prostate cancer (15 year survival) before it is detected by a traditional PSA by two years, early detection using PSA-xyz does not increase cancer survival compared to traditional PSA (it did not give the patient two extra years to live).”

Question 8

What is Pygmalion effect?

Answer 8

When the researchers expectation of an outcome affect the study results, a self fulfilling prophecy. “An orthopedic surgeon investigator who finds statistically significant benefit of arthroscopic surgery when compared to non-invasive therapeutic strategies. A chiropractor-led study that finds significant benefits of the effects of cervical manipulation when compared to traditional medicine strategies.”

Question 9

What is the Hawthorn effect?

Answer 9

Change in behavior in a study group that can result from knowing that they are being observed. “When studying the effects that infection control education has on physicians, the investigator notes that both the experimental and the control groups improved their hand hygiene.”

Question 10

What are the meanings of mean, median, and mode?

Answer 10

Mean is the average, median is the middle value, and mode is the most frequent value.

Question 11

When does mean equals median and equals mode?

Answer 11

In a normal distribution AKA Gausean distribution, or Bell distribution.

Question 12

What is a positive skew in a distribution curve?

Answer 12

It is when the distribution curve is asymmetric and has a tail to the RIGHT (a shark fin going to the Right).

Question 13

How are the values of mean and median and mode in a positive skew in a distribution curve?

Answer 13

The mean is greater than the median, which is greater than the mode.

Question 14

What is a negative skew in a distribution curve?

Answer 14

It is when the distribution curve is asymetric and has a tail to the LEFT(a shark fin going to the LEFT).

Question 15

How are the values of mean and median and mode in a negative skew in a distribution curve?

Answer 15

Mode is greater than the median, which is greater than the mode.

Question 16

What is a Null hypothesis?

Answer 16

Written as Ho (the 0 [zero] being much minuscule), it states that their is no association between exposure and outcome of disease; something is not related to something else.

Question 17

What is the Alternative hypothesis?

Answer 17

Written as H1 (the 1 [one] being much minuscule), it states that their is indeed an association between exposure and outcome of disease; something is not related to something else.

Question 18

What is the P value in hypothesis testing?

Answer 18

The P value is the value of interest that lets us know how compatible the study data are with the Null hypothesis. The P value is the probability that the study results occurred by chance alone, given that the Null hypothesis was true. A very small P value means that it is very unlikely that the study values occurred as a result by chance alone; a small P value is not compatible with the Null hypothesis, no association between exposure and the outcome of interest and that the all of the observed results were a result of chance.

Question 19

What is the P value cut off to reject the Null hypothesis?

Answer 19

A P value less than 0.05, the investigator will usually reject the Null hypothesis and accept the Alternative hypothesis, that there is indeed an association between exposure and certain outcome.

Question 20

What is a Type 1 error in hypothesis testing and what is another name for it?

Answer 20

Also know as alpha error, it rejects the null hypothesis and accepts the alternative hypothesis, when in reality there is no association between exposure and outcome. It is a False positive error.

Question 21

What is a Type 2 error in hypothesis testing and what is another name for it?

Answer 21

Also know as a beta error, it is when the investigator accepts the null hypothesis but in reality there is an association between exposure and outcome. It is a False negative error.

Question 22

What does power mean in a hypothesis testing?

Answer 22

Power refers to the probability that a test will correctly reject the Null hypothesis when indeed is false, and the alternative hypothesis is true.

Question 23

What affects power in a hypothesis testing?

Answer 23

Increased number of participants, increases statistical power.

Question 24

Study X shows that vitamin C can prevent coronavirus infections, but 10 other studies show no benefit. What type of error is found in the study X?

Answer 24

Type 1 (or alpha) error: you are finding an effect that does not really exist.

Question 25

Study Y shows that aspiring administration during an MI offers no improvement in patient morbidity or mortality. What type of error is present?

Answer 25

Type II (or beta) error: not finding an effect when there really is.

Question 26

What does it mean to say that a statistical distribution has a positive skew?

Answer 26

The curve is shifted to the left with a tail to the right. The mean is greater than the median, which is greater than the mode.

Question 27

What is Selection bias?

Answer 27

When there is a systematic difference in the way the study groups are chosen. “The referral centers for a trial of a new anticancer drug have more patients with end stage disease than early stage, so more patients with end stage disease are referred for the trial than early stage disease.”