Test 36 (statistics)

1. A 95 % confidence interval:

a. can only be used in parametric data

b. it is a test of the null hypothesis

c. it is calculated at ± 1.96 times the standard error of the mean

d. is useful when comparing data with another population

e. if zero difference lies within the 95% when comparing two groups to a 
    treatment, it indicates the treatment has no effect

Answera.F   b.F   c.T   d.T   e.T
A 95% confidence interval means that there is a 95% chance that the true population mean will lie within the range of values. 
It can also be used for parametric data but the calculation is different from that for the non-parametric data.

2. Meta-analysis of randomized controlled trials:

a. should only include peer-reviewed studies

b. is usually performed if individual trials are too small to give a reliable answers

c. should include trials in which patient selection is not randomized

d. provides a more stable estimate of the effect of treatment than individual trials

e. gives conclusion making further controlled trials unnecessary

Answera.F   b.T   c.F   d.T   e.F
Meta-analysis of randomized controlled trials is performed when the trials are too small individually to give reliable answers.
Meta-analysis also:
examines variability between trials
performs subgroup analysis
identifies the need for major trials
gives a more stable estimate of the effect of treatment.
To avoid bias, unpublished but properly controlled studies should be included.

3. The following are true about evidence-based medicine:

a. does not involve health economic assessment

b. is restricted to randomized placebo-controlled trials

c. is used to cut down waiting list

d. relies only on objective measurements of disease outcomes

e. combines clinical expertise and external evidence

Answera.F  b.F   c.F   d.T   e.T
According to the BMJ 1996, 812, 71-2, evidence-based medicine is the conscientious, explicit and judicious use of current best evidence in making decisions about care of individual patients. this means integrating individual clinical expertise with the best available external evidence.
It combines clinical expertise and current best evidence. The later comes from many sources, including randomized controlled trials, meta-analysis, national expert guidelines (eg. endophthalmitis and diabetic retinopathy treatment), patient-oriented studies and health economic assessment. 
It is not a method for cost cutting or reducing the waiting list.

4. The following are true about errors in clinical trials:

a. type I error is wrongly rejecting the null hypothesis

b. type II error is accepting the null hypothesis when it is invalid

c. errors are more common when big samples are used

d. type I error is more likely to occur when multiple t-test are used

e. type II error is reduced by the use of confidence interval.

Answera.T   b.T   c.F   d.T   e.T
Errors are more common when small numbers are used

5. The following are true about a normal (Gaussian distribution):

a. the mean, median and mode coincide

b. 95% of observation lie between the mean 2 standard deviation

c. 68% of observations lie between the mean 1 standard deviation

d. the 95% confidence interval may be calculated as the mean ±1.96 times 
    the standard error of the mean for population >30

e. data from a normal distribution is suitable for parametric tests without
    prior transformation.

Answera.T   b.T   c.T   d.T   e.T
In a normal distribution.
68% of observations lie within one standard deviation of the mean ±1SD
95% of observations lie within two standard deviation of the mean ±2SD
99.7% of observations lie within three standard deviation of the mean ±3SD
 
 
 

More MCQs