Measuring the Disease: Study Design and Statistical Analysis

ADHD and Vision Problems in the National Survey of Children's Health

- A study that looked at ADHD and vision problems - Cross-sectional study - The vision problems were problems that could not be corrected with normal glasses - The risk factor was vision problems, the disease was ADHD; and they found that 15% of pts with ADHD had vision problems compared to 8% of those with ADHD had normal vision - they found more statistical significance for mild and moderate problems - Concluded: That the prevalence of ADHD was greater among children with vision problems not correctable with glasses or contacts

TQ) Blue Mts Study

- Cohort study in Australia that looked at risk factors for branch retinal vein occlusion (BRVO) and central retinal vein occlusion (CRVO) who is likely to get this; they followed ppl over 10 yrs (looked at the incidence) - they got a bunch of Odds ratios, not considering age, they wanted to just look at these risk factors -Arteriovenous nicking —> odds ratio is 4, so the odds are 4xs greater that a person with a vein occlusion has AV nicking than without a vein occlusion has AV nicking, so its an exposure odds ratio *TQ) The other thing you have to understand about the odds ratio is that if you see 1.49 we can round it to 1.5 thats a 50% greater chance. 1.19 thats a 19% greater chance, even with a 1.19 while that number may look low, thats still a significant greater chance; if the odds ratio is 2.0 thats 100% chance;

2011 Study: Risk factors Associated with Childhood Strabismus

- Study that looked at some Odd ratios - They looked at hyperopia and the prevalence of esotropia - They are looking at refractive error as a risk factor for strabismus (eso or exo-tropia) - You should know why: Hyperopia is a risk factor for esotropia; b/c when you are a hyperope uncorrected you have to accommodated to see clearly; when you accommodate it brings the eyes inward and too much you cannot exert negative fusional vergence and one eye crosses and you have esotropia - Exotropia was associated with prematurity, maternal smoking during pregnancy, family history of strabismus, female sex, astigmatism *Because you know the odds ratio you can do something about the risk factors - Conclusion: B/c refractive error is correctable, these risk associations should be considered when developing guidelines for the screening and management of refractive error in infants and you children.

Rhopressa is a new type of glaucoma med called a ROCK inhibitor.

- its a drug that lowers IOP by increasing outflow through the meshwork

OHTS Results

- risk ratio is .46 ~ .5; so now reducing IOP is protective, b/c anything below 1 is protective - 90% of the untreated group didn't develop Glaucoma; to explain this they came up with the Risk Difference (they came up with this by subtracting the two risks, then take the reciprocal = the number needed to treat (NNT))

Rhopressa- More Phase III

- they changed the pts they were looking at from those that had an IOP of 20-26 to those that had an IOP of 20-25 and said is Rhopressa "non-inferior" to Timolol

In Phase 3 of clinical trials for new drugs, you are comparing a new medication with _____ used treatments.

1. Commonly - you hear a lot about phase 3

TQ) There are two types of odds ratio; they are 1.______ odds ratio and a 2._____ odds ratio.

1. Exposure 2. Disease

Sampling: 1._____ random sample each person has an equal chance of being selected. 2._____ random sample dividing the population into subgroups according to important characteristics—-> age, sex, race, location, economic status: then selecting random sample out of each subgroup.

1. Simple 2. Stratified

If the accuracy of information is not the same for each group than this would be what type of bias? A) Information B) Selection C) Confounding

A! - also if you don't do the same thing for the cases vs. controls

Masking

AREDS study - was randomized, placebo-controlled, and double blinded - so when evaluators looked at the retinas of patients that had taken vitamins and those that had taken placebos they didn't know which had taken which

Concerned with causes and effects of diseases within populations. This is referring to....

Analytic Epidemiology

For AMD smokers have a relative risk 3.29. This means A) The risk of developing AMD among smokers in Beaver Dam is 3.29x greater than that of developing AMD among the overall population in Beaver Dam. B) The risk of developing AMD among smokers in Beaver Dam is 3.29x greater than that of developing AMD among the population of non-smokers in Beaver Dam.

B!

If you are doing a study and only select individuals based on a particular means such as internet ads, phone calls, in person-street locations then this would be what type of bias? A) Information B) Selection C) Confounding

B!

Keratoconus is a progressive thinning and bulging of the cornea, leading to loss of vision. It is higher in what races?

Blacks and Hispanics

TQ) bias due to other factors that influence the outcome besides those of the risk factor that you are measuring is called what type of bias? A) Information B) Selection C) Confounding

C! - remember you are measuring risk factor with respect to disease, and in this case there are other factors that can affect the disease besides the risk factors- those are called confounders - Ex: Suppose you were looking at whether coffee drinking leads to heart disease; studies have shown that there is a higher prevalence of smokers that drink coffee, they tend to be associated; you realize that when you measure heart disease smoking definitely increases the risk of heart disease; so if you find a higher number of coffee drinkers that have heart disease you may not be measuring coffee drinking, you may be measuring smoking; so you have to control for those confounders and eliminate smokers

Null hypothesis: the idea that there is no real difference, and that any statistical differences are due to _____ alone.

Chance Ex of Null Hypothesis: AV nicking has nothing to do with vein occlusions despite the study and thats the null hypothesis - we want to be able to reject the Null Hypothesis to show that there is a true affect of the study

A group of people with given characteristics followed over time (longitudinal study) is referring too..... There are two types prospective and retrospective.

Cohort

TQ) The risk ratio can only be used in a _____ study.

Cohort - b/c a risk ratio is the only metric that directly measures the risk; in order to directly measure the risk you have to follow people from when they do not have the disease until they get it; and thats the incidence (new cases) *only way to directly measure new cases, which is a way to directly measure the risk, is by measuring the risk ratio and that is done in cohort studies (can be done prospectively and retrospectively) *The risk ratio is a way of measuring the risk of getting a particular eye disease among those who are/are not exposed to a risk factor

____ ______ is used to address the repeatability of the result, or how much it will vary from one study to another.

Confidence interval - Every result varies, every test you do can change from one day to another - what the CI represents is the precision of the study and it indicates a chance (95% chance) that the interval includes the true value

In which of the following studies can the odd ratio be used: A) Case-control B) Cross sectional C) Cohort Studies D) All of the above

D! - Cases = ppl with the disease; controls = ppl without the disease - you can take any risk factor

____ study designs include case reports, case series, cross sectional, and ecological.

Descriptive

Describes the amount and distribution of disease within a population, without regard to causality. This refers to.......

Descriptive Epidemiology

TQ) The odds in favor of DISEASE among exposed/the odds in favor of the DISEASE among unexposed refers to....

Disease Odds ratio - having the disease NOT the risk factor - Ex: The risk factor for chlamydia caused inclusion conjunctivitis is age (18-25); You might say that if a pt walks into the room and is that age group the odds are greater that they would have inclusion conjunctivitis, than the odds that a 60 year would have it - the odds are less b/c the risk factor is the age

TQ) Odds in favor of exposure among cases/ odds in favor of exposure among controls refers to the _______.

Exposure odds ratio - its a ratio so its comparing two different odds - it is not asking about the odds of having a disease; IT IS A MEASUREMENT OF THE ODDS OF HAVING A RISK FACTOR - Ex "If a pt comes into the office with a corneal ulcer its asking about the odds of whether they wear contact lens or not; whether they have had the exposure or the risk factor - odds that a pt with an ulcer wears CLs (cases)/ odds that a pt without an ulcer wears CLs (control) Pt comes in saying their mother had AMD (disease), would ask if their mother was a smoker (risk factor)

Ecological Studies do not have _____ information.

Individual Correlation between dietary fat intake and breast cancer by country - Looking at aggregate data, not individual data - looking at total fat intake in the country and total breast cancer in the country (not saying occurring in any particular individuals)

TQ) In prospective cohort studies ____ of the individuals have the disease in the beginning.

None - You have 4 categories 1) Those without the disease with the risk factor in the beginning 2) Those without the disease without the risk factor in the beginning 3) Those with the disease with the risk factor 4) Those with the disease without the risk factor - you look at incidence (new cases) *Cohort studies are the only studies where you can look at the incidence - in cohort studies no one has it in the beginning so your able to look at the incidence rate - direct measurement of the incidence

TQ) In case control studies we use the _____ ratio.

Odds - the odds of risk factors in occurring with those that have and those who do not have the disease - is an estimate of the relative risk TQ) The odds ratio can be used in any study vs. the risk ratio can only be used in the cohort study; the risk ratio cannot be used in a case-control study b/c you can not directly measure risk in a case-control study b/c your taking pts who already have the disease, not a longitudinal study, so you cannot directly measure incidence

Congenital Cataract following German Measles in the mother

One of the first major case-control studies

The _____ _____ allows us to measure the null hypothesis. It is the probability that the effect is due to chance alone.

P value - In order to reject the Null hypothesis you can do any P value you want. Usually its P(less than or equal to) 0.05, but not for all - <5% probability that the results are due to chance; or 95% probability that the results are not due to chance (you have a true relationship between the risk factor and the disease *If its not statistically significant the P would be greater than 0.05- cannot reject the null hypothesis

TQ) Identify individuals with and without disease in the present, so they already know who has the disease, so they start in the present and go backwards when all of the people were free of disease... this refers to what type of cohort study?

Retrospective - So you have been seeing a patient for 20 years and they have cataracts, so you go back 20 yrs to when you first started seeing them when they didnt have cataracts. Can look at risk factors; important thing is when you go back in the chart no one has the disease in the beginning—-> basically a chart review - still a cohort study, taking people from when they dont have the disease and seeing years later if they develop the disease; looking at incidence

Study done by record review: Looked at 50,000 pts in a national database without cataract or statin use. They wanted to answer was statin medications causing cataracts? In the beginning no one had cataracts they looked at the charts for 10 years. This would be a prospective or retrospective cohort study?

Retrospective!

TQ) The most important study ever done in Glaucoma- Ocular Hypertension Treatment Study

Risk factors for conversion to glaucoma - Larger c/d, older age, higher IOP, thinner central corneal thickness (CCT), higher PSD on VF——> lead to a change in protocol when testing for glaucoma; when you have a glaucoma suspect you have to measure CCT

Absolute risk is the risk with regard to the overall population Excess risk or relative risk refers to.....

Risk in people exposed/ risk in people not exposed

TQ) The DREAM Study

The results where: that they found that there was not very much difference. Those that were randomly assigned to receive supplements did not have significantly better outcomes than those who were assigned the placebo.

Acanthamoeba Keratitis in Soft Contact Lens Wearers, a case-control study

When you do a case-control study you use an odds ratio