STAT 250 Dr Kari Lock Morgan Collecting Data

  • Slides: 33
Download presentation
STAT 250 Dr. Kari Lock Morgan Collecting Data: Randomized Experiments SECTION 1. 3 •

STAT 250 Dr. Kari Lock Morgan Collecting Data: Randomized Experiments SECTION 1. 3 • Randomized Experiments Statistics: Unlocking the Power of Data Lock 5

Exercise and the Brain �A study found that elderly people who walked at least

Exercise and the Brain �A study found that elderly people who walked at least a mile a day had significantly higher brain volume (gray matter related to reasoning) and significantly lower rates of Alzheimer’s and dementia compared to those who walked less �The article states: “Walking about a mile a day can increase the size of your gray matter, and greatly decrease the chances of developing Alzheimer's disease or dementia in older adults, a new study suggests. ” �Is this conclusion valid? (a) Yes (b) No Allen, N. “One way to ward off Alzheimer’s: Take a Hike, ” msnbc. com, 10/13/10. Statistics: Unlocking the Power of Data Lock 5

Exercise and the Brain �How would you design an experiment to determine whether exercise

Exercise and the Brain �How would you design an experiment to determine whether exercise actually causes changes in the brain? Statistics: Unlocking the Power of Data Lock 5

Exercise and the Brain �A sample of mice were divided randomly into two groups.

Exercise and the Brain �A sample of mice were divided randomly into two groups. One group was given access to an exercise wheel, the other group was kept sedentary �“The brains of mice and rats that were allowed to run on wheels pulsed with vigorous, newly born neurons, and those animals then breezed through mazes and other tests of rodent IQ” compared to the sedentary mice �Is this evidence that exercise causes an increase in brain activity and IQ, at least in mice? (a) Yes (b) No Reynolds, “Phys Ed: Your Brain on Exercise", NY Times, July 7, 2010. Statistics: Unlocking the Power of Data Lock 5

Randomized Experiments �If a randomized experiment yields a significant association between the two variables,

Randomized Experiments �If a randomized experiment yields a significant association between the two variables, we can establish causation from the explanatory to the response variable Randomized experiments are very powerful! They allow you to infer causality. Statistics: Unlocking the Power of Data Lock 5

How to Randomize? �Option 1: As with random sampling, we can put all the

How to Randomize? �Option 1: As with random sampling, we can put all the names/numbers into a hat, and randomly pull out names to go into the different groups �Option 2: Put names/numbers on cards, shuffle, and deal out the cards into as many piles as there are treatments �Option 3: Use technology Statistics: Unlocking the Power of Data Lock 5

Knee Surgery for Arthritis Researchers conducted a study on the effectiveness of a knee

Knee Surgery for Arthritis Researchers conducted a study on the effectiveness of a knee surgery to cure arthritis. It was randomly determined whether people got the knee surgery. Everyone who underwent the surgery reported feeling less pain. Is this evidence that the surgery causes a decrease in pain? (a) Yes (b) No Statistics: Unlocking the Power of Data Lock 5

Control Group �When determining whether a treatment is effective, it is important to have

Control Group �When determining whether a treatment is effective, it is important to have a comparison group, known as the control group �It isn’t enough to know that everyone in one group improved, we need to know whether they improved more than they would have improved without the surgery �All randomized experiments need a control or comparison group (could be two different treatments) Statistics: Unlocking the Power of Data Lock 5

Knee Surgery for Arthritis �In the knee surgery study, those in the control group

Knee Surgery for Arthritis �In the knee surgery study, those in the control group received a fake knee surgery. They were put under and cut open, but the doctor did not actually perform the surgery. Source: “The Placebo Prescription, ” NY Times Magazine, 1/9/00. Statistics: Unlocking the Power of Data Lock 5

Clinical Trials �Clinical trials are randomized experiments dealing with medicine or medical interventions, conducted

Clinical Trials �Clinical trials are randomized experiments dealing with medicine or medical interventions, conducted on human subjects �Clinical trials require additional aspects, beyond just randomization to treatment groups: Placebo Double-blind Statistics: Unlocking the Power of Data Lock 5

Placebo Effect �Often, people will experience the effect they think they should be experiencing,

Placebo Effect �Often, people will experience the effect they think they should be experiencing, even if they aren’t actually receiving the treatment. This is known as the placebo effect. �Example: Eurotrip �One study estimated that 75% of the effectiveness of anti -depressant medication is due to the placebo effect �A review of 15 clinical trials with different diseases found that 35% of patients were satisfactorily relieved by the placebo alone (The Powerful Placebo) �For more information on the placebo effect (it’s pretty amazing!) read The Placebo Prescription or listen to Radiolab’s podcast Placebo Statistics: Unlocking the Power of Data Lock 5

Study on Placebos �Blue pills are better than yellow pills �Red pills are better

Study on Placebos �Blue pills are better than yellow pills �Red pills are better than blue pills � 2 pills are better than 1 pill � 4 pills are better than 2 pills �And shots are the best of all! Statistics: Unlocking the Power of Data Lock 5

Placebo Effect Should doctors be allowed to prescribe placebos? Statistics: Unlocking the Power of

Placebo Effect Should doctors be allowed to prescribe placebos? Statistics: Unlocking the Power of Data Lock 5

Placebo-controlled studies �Control groups should be given a placebo, a fake treatment that resembles

Placebo-controlled studies �Control groups should be given a placebo, a fake treatment that resembles the active treatment as much as possible �This allows researchers to separate effect of treatment itself from effect of the placebo effect �Often instead of a placebo, new treatment is compared to the existing standard treatment (not ethical to not help a patient in need) Statistics: Unlocking the Power of Data Lock 5

Double-Blind Experiments �If possible, randomized experiments should be double-blinded: neither the participants or the

Double-Blind Experiments �If possible, randomized experiments should be double-blinded: neither the participants or the researchers involved should know which treatment the patients are actually getting �The participants should be blinded because of the placebo effect �The researchers should be blinded to avoid conscious or unconscious bias of the results Statistics: Unlocking the Power of Data Lock 5

Sildenafil and c. GMP The molecule c. GMP is a natural vasodilator: it increases

Sildenafil and c. GMP The molecule c. GMP is a natural vasodilator: it increases blood flow by relaxing muscle in the walls of blood vessels. Blood vessel c. GMP Sildenafil inhibits phosphodiesterase (an enzyme), which degrades c. GMP. sildenafil phosphodiesterase Statistics: Unlocking the Power of Data degrades c. GMP Lock 5

Sildenafil � 1980 s: Pfizer develops sildenafil citrate, which inhibits phosphodiesterase, hoping to increase

Sildenafil � 1980 s: Pfizer develops sildenafil citrate, which inhibits phosphodiesterase, hoping to increase blood flow in the coronary arteries of individuals with heart disease. � 1991: Phase I clinical trials are begun. � 1992: Pfizer gets reports of an unexpected “side effect” in men. � 1994: After sildenafil fails to show any cardiac benefit, Pfizer begins pilot studies for treatment of erectile dysfunction (ED). � 1998: FDA approves sildenafil, marketed as… Statistics: Unlocking the Power of Data Lock 5

Atenonol: A randomized experiment Silvestri et al assigned 96 men to take atenolol, a

Atenonol: A randomized experiment Silvestri et al assigned 96 men to take atenolol, a heart drug which can cause ED as a side effect: • 32 men were told they were taking either atenolol or placebo • 32 were told only that they were taking atenolol • 32 were told that they were taking atenolol, and that it can cause ED as a side effect. After 3 months, the researchers asked men to report any experience of ED. • Silvestri A, Galetta P, Cerquetani E, Marazzi G, Patrizi R, Fini M, Rosano GM. Report of erectile dysfunction after therapy with beta-blockers is related to patient knowledge of side effects and is reversed by placebo. Eur Heart J. 2003 Nov; 24(21): 1928 -32. Statistics: Unlocking the Power of Data Lock 5

They were all on the same drug! Statistics: Unlocking the Power of Data Lock

They were all on the same drug! Statistics: Unlocking the Power of Data Lock 5

Sildenafil: A Randomized Experiment The researchers then assigned all men who’d experienced ED to

Sildenafil: A Randomized Experiment The researchers then assigned all men who’d experienced ED to either 50 mg sildenafil or placebo. The men were blinded regarding this treatment. After a week the researchers asked the men if their ED had improved. Any guesses what the men reported? Statistics: Unlocking the Power of Data Lock 5

32 men had been randomized to atenolol and told it can cause ED 10

32 men had been randomized to atenolol and told it can cause ED 10 reported experiencing ED 5 randomized to take Viagra (blinded) 5 randomized to take placebo (blinded) All improved! Control, placebos and blinding are important! Statistics: Unlocking the Power of Data Lock 5

Trustworthy Clinical Trials �To assess the trustworthiness of results from clinical trials (or any

Trustworthy Clinical Trials �To assess the trustworthiness of results from clinical trials (or any medical result), ask… Was the treatment randomly assigned? Was the control group given a placebo or other treatment? Was the study double-blind? Statistics: Unlocking the Power of Data Lock 5

Green Tea and Prostate Cancer �A study was conducted on 60 men with PIN

Green Tea and Prostate Cancer �A study was conducted on 60 men with PIN lesions, some of which turn into prostate cancer �Half of these men were randomized to take 600 mg of green tea extract daily, while the other half were given a placebo pill �The study was double-blind, neither the participants nor the doctors knew who was actually receiving green tea �After one year, only 1 person taking green tea had gotten cancer, while 9 taking the placebo had gotten cancer Statistics: Unlocking the Power of Data Lock 5

Green Tea and Prostate Cancer A difference this large is unlikely to happen just

Green Tea and Prostate Cancer A difference this large is unlikely to happen just by random chance. Can we conclude that green tea really does help prevent prostate cancer? (a) Yes (b) No Statistics: Unlocking the Power of Data Lock 5

Matched Pairs In a matched pairs experiment, either each case gets both treatments, in

Matched Pairs In a matched pairs experiment, either each case gets both treatments, in random order, or cases are paired and each pair gets both treatments (randomly assigned). �Compare a case to itself or a similar case �Look at differences in response within the pair �Useful when the response variable varies a lot from case to case Statistics: Unlocking the Power of Data Lock 5

Gatorade and Running �Does drinking Gatorade make you run faster? �Variables: Explanatory: drink Gatorade

Gatorade and Running �Does drinking Gatorade make you run faster? �Variables: Explanatory: drink Gatorade or placebo before running Response: time to run a certain distance (a mile? ) �Completely randomized: Randomize half of the people to drink Gatorade half to drink a placebo (colored water? with sugar? ) �Matched pairs: have each person run twice (on different days? ), one time drinking Gatorade, one time drinking placebo. Randomize order. Statistics: Unlocking the Power of Data Lock 5

Matched Pairs: Identical Twins Statistics: Unlocking the Power of Data Lock 5

Matched Pairs: Identical Twins Statistics: Unlocking the Power of Data Lock 5

Why not always randomize? �Randomized experiments are ideal, but sometimes not ethical or feasible

Why not always randomize? �Randomized experiments are ideal, but sometimes not ethical or feasible �Ethics: Can’t force someone to partake in a treatment known or suspected to cause harm �Feasibility: Often, randomized experiments are simply not feasible (cost, time, and practical issues are limiting factors) �Often, you have to do the best you can with data from observational studies Statistics: Unlocking the Power of Data Lock 5

Randomization in Data Collection Was the sample randomly selected? Yes No Possible to generalize

Randomization in Data Collection Was the sample randomly selected? Yes No Possible to generalize to the population Should not generalize to the population Statistics: Unlocking the Power of Data Was the explanatory variable randomly assigned? Yes Possible to make conclusions about causality No Can not make conclusions about causality Lock 5

Two Fundamental Questions in Data Collection Random sample? ? ? Population Sample Randomized experiment?

Two Fundamental Questions in Data Collection Random sample? ? ? Population Sample Randomized experiment? ? ? DATA Statistics: Unlocking the Power of Data Lock 5

Randomization �Doing a randomized experiment on a random sample is ideal, but rarely achievable

Randomization �Doing a randomized experiment on a random sample is ideal, but rarely achievable �If the focus of the study is using a sample to estimate a value for the entire population, you need a random sample, but do not need a randomized experiment (example: election polling) �If the focus of the study is establishing causality from one variable to another, you need a randomized experiment and can settle for a nonrandom sample (example: clinical trials) Statistics: Unlocking the Power of Data Lock 5

Summary �Randomized experiments involve randomly determining the level of the explanatory variable �Randomized experiments

Summary �Randomized experiments involve randomly determining the level of the explanatory variable �Randomized experiments prevent confounding variables, so causality can be inferred �A control or comparison group is necessary �The placebo effect exists, so a placebo and blinding should be used �Matched pair experiments can be useful when the response variable varies a lot from unit to unit Statistics: Unlocking the Power of Data Lock 5

To Do �Read Section 1. 3 �HW 1. 3 due 10: 10 am Friday,

To Do �Read Section 1. 3 �HW 1. 3 due 10: 10 am Friday, Jan 30 th Statistics: Unlocking the Power of Data Lock 5