Computer generated transcript

Warning!
The following transcript was generated automatically from the content and has not been checked or corrected manually.

a little everything to frdicine atdym cambridge tuesdays ndthursdays 7pm-8pm Anushka Irodi STATISTICS:ALITTLEREFRESHER Earth Tachatirakul Megan HodgeSTUDYDESIGN Ø There are several types of study designs Ø Each has pros and cons Ø Some are better than others Ø It is not always possible or ethical to use the type of study that would provide the best evidence, can you think of some situations where this might be the case? Ø What are confounding variables and why are they important to consider?SAMPLING § population – the pool of individuals from which a statistical sample is drawn fora study § sampling frame – the actual set of units from which a sample has been drawn § sample – a set of observations drawn from the population § random sampling – a sampling technique where each sample has an equal probability of being chosen § systematic sampling – a sampling technique where the samples are selected from an ordered sampling frame § quota sampling – a non probability sampling method where individual samples are selected to represent the population § Measures of central tendency: Mean, Median & Mode. STATISTICAL § Measures of dispersion: Range, Deviation.ile Range (IQR), Standard TESTINGSTUDY RESULTSSTUDYEFFECTIVENESSQUIZ Has disease Does not have disease What is the sensitivity of this test? Positive test 194 337 A 34 B 64 Negative test 6 63 C 77 D 84 E 97QUIZ Has disease Does not have disease What is the sensitivity of this test? Positive test 194 337 A 34% B 64% Negative test 6 63 C 77% D 84% E 97% 194/(194+6) = 0.97RELATIVERISKANDODDSRATIO Positive Negative incidence incidence Not all people exposed to a risk will develop a disease... Exposure 50 50 How much does the exposure to the risk increase the chance of developing a disease, that is what relative risk explores.o exposure 30 70 Relative risk = (50/100) / (30/100) = 5/3 approx 167%RELATIVERISKANDODDS Positive Negative incidence incidence Another way to quantize this is the odds ratio. Instead of dividing the positive incidence Exposure 50 50 (diseased) group by the respective total number for exposure and non-exposure we divide by the negative incidence (healthy people). No exposure 30 70 In this case the odds ratio is (50/50) / (30/70) = 1 / (3/7) = 7/3 approx 2.33 (relative risk is divide by total number, odds is by diseased (negative incidence)RISKANDNUMBERTOTREAT When we look at the effectiveness of a medical treatment, we must also look at the context of how much it reduces the incidence in groups exposed to the risk. One measure is number needed to treat which is the inverse of the attributable risk (absolute risk reduction). This number is used for when you treat an unknown group of people which you do not know about the disease or exposure and how much they benefit from it. It is calculated as how many people needed to be treated to prevent ONE bad outcome. How to think about it is if the treatment prevents the effect of exposure to the risk, what is the remaining chance that several people would still develop the said disease.QUIZ Positive Negative incidence incidence What is the number needed to treat? Exposure 50 50 A 1 B 3 No exposure 30 70 C 5 D 7 E 9QUIZ Positive Negative incidence incidence What is the number needed to treat? Exposure 50 50 A 1 B 3 No exposure 30 70 C 5 D 7 E 9 NNT = 1/absolute risk = 1/((50/(50+50))-(30/(30+70)))= 1/(0.5-0.3)= 1/0.2 = 5THANKYOU!