Statistics Flashcards
Define what a type 1 error is [1]
What increases the chance of a type 1 error? [1]
A type I error, also known as a false positive, is the incorrect rejection of the null hypothesis when it is true.
A type I error is when the null hypothesis is true and was wrongly rejected.
The greater the p-value, the greater the chance that the null hypothesis will be rejected. - e.g. if p value is 20%
What is the defintion of a type 2 error? [1]
A type II error refers to accepting the null hypothesis when we should have rejected it i.e. there is a difference.
The power of a study is the ability of a study to reject the null hypothesis when it is false. It is dependent on []
The power of a study is the ability of a study to reject the null hypothesis when it is false. It is dependent on sample size
A [] study investigates participants over a sustained period of time. The participants each have a common characteristic
A cohort study investigates participants over a sustained period of time. The participants each have a common characteristic
e.g
A study investigating the long-term use of opioid painkillers following an initial, 1-week prescription for post-surgical pain was conducted.
600 patients who had undergone elective surgical procedures were followed up weekly for a period of 12 weeks after their initial prescription. The aim of this was to determine whether or not they had continued to request prescriptions for opioid painkillers.
These studies examine an association between an outcome and exposure to a risk factor.
Recruited participants based on the presence of an outcome from the outset of the study. Once cases are recruited, suitable controls are identified and selected for the study.
These studies are more suited to investigate rare conditions as individuals with the outcome of interest are identified at the outset
What type of study is being described?
Case-control
What is the sensitivty of this test? [1]
This is calculated by dividing the number that tested positive amongst those with RA (168) by the total number of participants with RA (168 +32 = 200). Therefore sensitivity is 168 / 200 = 0.84 or 84%
What is the positive predicted value of this test? [1]
PPV is calculated by dividing the number testing positive amongst those with RA (168) by the total number of participants that tested positive (168 + 28 = 196). PPV is 168 / 196 = 0.857 or 85.7%
A new drug which is said to reduce the chance of patients with acromegaly developing carpal tunnel syndrome is introduced. You review the paper so that you can decide how to advise your patients. The study recruited 3000 patients. A total of 1700 received the new drug, of which 170 patients developed carpal tunnel syndrome. The remaining patients received a placebo, of which 300 developed carpal tunnel syndrome.
What is the Absolute Risk Reduction (ARR) of developing carpal tunnel syndrome
The absolute risk of the exposure group is 10% (170 / 1700). The absolute risk of the control group is 23% (300 / 1300). Therefore, the ARR is 13% (23 – 10).
A study on a new drug for controlling diabetes pre-surgery is being conducted. Participants need to be selected in a way that reduces confounding.
What method can be used to do this?
Blinding
Standardising variables
Decrease variability in measurements
Randomisation
Increase sample size
A study on a new drug for controlling diabetes pre-surgery is being conducted. Participants need to be selected in a way that reduces confounding.
What method can be used to do this?
Randomisation
If participants are randomly allocated to a group that receives the treatment and one that does not, they could have approximately the same baseline characteristics. This would reduce confounding.
To analyse the effectiveness of the CBT program, the researchers must choose an appropriate statistical test that can handle the data’s non-normal distribution.
What is the most appropriate statistical test?
Analysis of variance
Kruskal-Wallis test
Mann-Whitney U test
Pearson correlation
Two-sample T-test
To analyse the effectiveness of the CBT program, the researchers must choose an appropriate statistical test that can handle the data’s non-normal distribution.
What is the most appropriate statistical test?
Analysis of variance
Kruskal-Wallis test
Mann-Whitney U test
Pearson correlation
Two-sample T-test
A team of paediatricians is planning a study to explore the possible link between maternal occupation during pregnancy and the incidence of microcephaly in their children. The cases will include children diagnosed with microcephaly, and the controls will be age and sex-matched individuals without the condition. Data on maternal occupation throughout pregnancy will be gathered from interviews and medical records.
What measure should they use to report the results?
Confidence interval
Hazard ratio
Odds ratio
P-value
Relative risk
A team of paediatricians is planning a study to explore the possible link between maternal occupation during pregnancy and the incidence of microcephaly in their children. The cases will include children diagnosed with microcephaly, and the controls will be age and sex-matched individuals without the condition. Data on maternal occupation throughout pregnancy will be gathered from interviews and medical records.
What measure should they use to report the results?
Confidence interval
Hazard ratio
Odds ratio
P-value
Relative risk
The correct answer is odds ratio. Within the context of a case-control study aimed at exploring the association between maternal occupation and the development of microcephaly, a rare disease, odds ratios are particularly pertinent. They serve as a statistical measure to quantify how the likelihood of an event occurring in one group compares with that in another group. Essentially, odds ratios provide insight into whether an event is more probable within a specific cohort relative to another.
How do you calculate NNT? [1]
The NNT is calculated by dividing 1 by the ARR
A 16-year-old patient asks you about a study she read investigating the efficacy of a new form of emergency contraception in preventing pregnancy. You review the study so that you can answer her questions effectively. The study found there were fewer cases of pregnancy when patients took the new emergency contraception compared with those who did not. The absolute risk of pregnancy in those taking the new emergency contraception was 0.15 versus 0.70 in those who did not.
What is the NNT in this study to prevent pregnancy?
2
55
181.8
5.5
0.55
A 16-year-old patient asks you about a study she read investigating the efficacy of a new form of emergency contraception in preventing pregnancy. You review the study so that you can answer her questions effectively. The study found there were fewer cases of pregnancy when patients took the new emergency contraception compared with those who did not. The absolute risk of pregnancy in those taking the new emergency contraception was 0.15 versus 0.70 in those who did not.
What is the NNT in this study to prevent pregnancy?
2
The NNT is calculated by dividing 1 by the ARR. The ARR is calculated by subtracting the difference in absolute risk between the two groups.** In this case, the ARR was 0.70 – 0.15 = 0.55**. Therefore, the NNT is 1 / 0.55 = 1.80. subsequently rounded to 2.
A randomised control trial about a new antihypertensive drug is being conducted. In the trial, 400 patients with hypertension are given the new drug and 500 patients with hypertension are given the placebo. After 5 years, 100 patients with the new drug have a cardiovascular event and 250 patients with the placebo have a cardiovascular event.
If this drug is given to 600 patients with hypertension how many cardiovascular events can be prevented?
200
300
400
150
120
First, the risk of a cardiovascular event in the placebo group needs to be calculated which is: 250/500= 0.5.
Next, the risk of a cardiovascular event in the new drug group needs to be calculated which is: 100/400= 0.25.
The difference in risk or the absolute risk reduction in this trial is 0.5-0.25= 0.25.
The number needed to treat is 1/absolute risk reduction which in this case is 1/0.25=4. This means that for every 4 patients treated, 1 cardiovascular event will be prevented.
This means that for 600 patients, (600/4= 150) 150 cardiovascular events will be prevented.
How do you work out the likelihood ratio for a positive test result?
e.g
.
A 35-year-old female presented to your clinic with complaints of abdominal pain relieved with eating for three months. You suspect a duodenal ulcer. She has no history of use of NSAIDs. She lives in London where the prevalence of Helicobacter pylori is 40%. You plan on testing for the infection with a urea breath test. The sensitivity of the test is 96%, and the specificity is 92%.
The likelihood ratio for a positive test result (LR+) is calculated by dividing the sensitivity of the test by (1 - specificity). In this case, the sensitivity is 96% or 0.96 and the specificity is 92% or 0.92.
So, LR+ = Sensitivity / (1 - Specificity)
= 0.96 / (1 - 0.92)
= 0.96 / 0.08
=12
A study is investigating whether caffeine actually increases the risk of hypertension. 80 people who regularly drink coffee are matched to 80 people who do not drink coffee. 60 of the coffee drinkers had developed hypertension at some point compared to 10 of the non-caffeine drinkers who had also developed hypertension.
What is the odds ratio of developing hypertension for people who regularly drink coffee?
Think (a/b)/(c/d) = (60/20)/(10/70) = 21 is the formula that should be provided in the explanation
