Public Health Final

What are the components of the Evidence-based Medicine (EBM) triad?

- individual clinical expertise

- best scientific evidence

- patient values & expectations

What are the steps to applying EBM to clinical practice?

1. Define a clinically relevant question

2. Search for the best evidence

3. Critically appraise the evidence

4. Apply the evidence

5. Evaluate the performance of EBM

What type of study provides the strongest research evidence?

Meta analysis & systematic reviews

What type of study is the gold standard for clinical medicine?

randomized control trials

How do traditional clinical practice and evidence-based clinical practice differ?

traditional clinical practice is based on experience & clinical practice guidelines

evidence-based clinical practice is a process that combines clinical expertise with the best scientific evidence to make patient care decisions

Why is searching medical literature important?

staying current is essential for quality patient care

How does searching the medical literature apply to clinicians?

- what test to order

- what med dose

- second & third line options vs experimental modalities

- predictors of efficacy

- predictors of morbidity & mortality

What are sources of medical literature?

- journal articles

- reviews of primary studies

- clinical practice guidelines

- medical textbooks

- CME lectures

- new media

Name some point-of-care resources.

- Clinical evidence (best practice)

- DynaMed

- UpToDate

How does a clinician decide which sources to read or use?

- most current

- most useful

- most widely accepted

- easiest to find

- secondary sources

What are the 5 fundamental types of research questions? What are the recommended studies to evaluate each type of question?

- Therapy → RCT

- Harm → RCT

- Differential diagnosis → prospective blind comparison

- Diagnosis → prospective blind comparison

- Prognosis → cohort study

What does PICO(T) stand for?

Patient/problem/population

Intervention

Comparison

Outcome

Timing

What type of study is best for determining prognosis?

cohort study

What type of study is best for determining the best treatment option?

RCT

What are the limitations of medical research?

- limited/narrow participant inclusion criteria

- inappropriate outcome measures

- external constraints

- reduced reliability, generalizability, or real world applicability

- introduction of unintended biases

- statistical errors

- poor date quality

- commercial interests

What is the purpose of performance improvement models in healthcare? Are there major differences between the major models?

What are biostatistics?

application of statistical principles to the analysis of biological & health-related data to solve problems in medicine, public health, & biology

Differentiate between population and sample.

population:

-entire group of individuals having a disease or characteristic of interest

sample:

-subset of the population; group selected and assumed to be a fair representation of the population

Why are samples used in research?

- studied quicker

- less expensive

- more feasible

- probability methods can be used to estimate error

What is random sampling?

selecting a group of individuals from a larger population in a way that gives every member an equal chance of being chosen

Define the types of random sampling.

Simple → entire population is accessible

Stratified → population is divided into subgroups according to demographic factors (age, gender)

Cluster → group into clusters then randomly select a cluster from there

What are the types of variables?

Quantitative (discrete vs continuous, nominal vs ordinal)

Qualitative

Independent

Dependent

Confounding

Control

What are descriptive statistics?

quantitative summaries of results without drawing inferences

How are descriptive statistics typically represented in literature?

tables or graphs

What are the measures of central tendency?

mean, median (best), mode

What is skew of central tendency?

What are the measures of variability?

range

variance

standard deviation

What do low and high variability indicate?

Low → values more consistent, better prediction about the population

High → values are less consistent, harder to make predictions

What information is provided by the interquartile range?

provides spread of the middle of the distribution

range is highest to lowest overall

What information does the standard deviation provide?

data variability

low SD → values clustered around mean

high SD → values widely spread

What is the Empirical Rule for standard deviation?

67% lie within 1 SD

95% lie within 2 SD

99.7% lie within 3 SD

What are some common tests used in inferential statistics?

- t tests

- z test

- ANOVA

- correlation

- regression

What is correlation?

strength of an association between 2 variables

correlation coefficient ranges from -1 to +1

What information does linear regression provide?

cause & effect

What is the null hypothesis?

default assumption that there is no association

What are the steps in hypothesis testing?

What is the alpha value, and what does it mean?

significance level (probability of rejecting the null hypothesis)

What information does the p-value provide?

probability value

How is the p-value calculated?

calculated from a statistical test (t test or z test)

What does a low p-value indicate?

low considered p<0.05

observed results are unlikely to have occurred by chance

What does the p-value NOT indicate?

size or importance of an effect

need a confidence interval to understand

When is the null hypothesis rejected?

when p ≤ alpha (p<0.05)

What is a Type 1 Error in hypothesis testing?

false positive (null is rejected when it is actually true)

What is a Type II Error in hypothesis testing?

false negative (failure to reject the null hypothesis when it is actually false)

What information does the t-test provide?

tells us whether the differences between members of a group occurred randomly or systematically

When should the t-test be used?

when the sample size is small (<30)

What does a small t-value indicate?

groups are similar

What is the correlation between the null hypothesis and a t-value far from zero?

a t value far from 0 indicates that the null hypothesis is less likely to be true (groups are very different)

What would be the degrees of freedom (DF) for a sample size of 23?

DF = N-1

22

When is the z-test used?

when sample size is large (>30)

11.What z-value(s) indicate the null hypothesis should be rejected?

What z-value would indicate the null hypothesis should not be rejected?

it is between the critical values

When is analysis of variance (ANOVA) used?

there are ≥2 groups

What values are produced by ANOVA?

F-statistic with p-value or CI

What does a large f-value indicate?

greater variance between groups than within groups

When does the f-value indicate statistical significance?

if it is greater than the critical value

What information does the confidence interval provide?

estimation of an unknown population parameter

Which studies are observational studies?

- case series

- retrospective case control studies

- cross sectional studies or surveys

- historical cohort studies

- prospective cohort studies

Which studies are experimental studies?

- control trials

- studies with no controls

What is the purpose of experimental studies?

draw conclusions about the intervention

What type of observational study has the simplest design?

case-series study

What information is included in a case-series?

interesting characteristics observed in a group of patients (no control, no hypothesis)

What are the limitations of a case-series?

cannot establish causality & are not generalizable

Are case-control studies retrospective or prospective?

retrospective

What criteria are used to select cases and controls for case-control studies?

cases → presence of a disease or outcome

controls → without the disease or outcome

What is matching?

selecting the controls so that they are similar to the cases in certain characteristics (age, sex)

What is the purpose of case-control studies?

look backward in time to try to detect possible causes or risk factors

Which type of study is also known as a prevalence study?

cross-sectional study

What is the timeline for a cross-sectional study?

one point in time

What are the disadvantages of a cross-sectional study?

cannot determine causality

rare outcomes cannot be studied

What is the timeline for a cohort study (retrospective, prospective, point-in-time)?

What is the purpose of a cohort study?

evaluate the future consequences of a risk factor or exposure (what will happen?)

What factors would cause a cohort to get smaller over time?

- attrition

- migration

- death

What was the purpose of the Framingham Heart Study?

investigate factors associated with the development of atherosclerotic & hypertensive cardiovascular disease

What are some findings from the Framingham Heart Study?

What is unique about historical cohort studies?

there is no control group (based on events that have already occurred)

Which type of study design is best for investigating incidence of disease?

cohort study

What is the purpose of the CONSORT Statement?

provides standardization for reproducibility of RCTs

What is the basis/design of a concurrent control study?

a control group is enrolled and followed during the exact same time period and from the same source population as the group receiving the experimental treatment

basis of this design is to ensure maximum comparability between the groups by minimizing bias & confounding variables

Which type of study best identifies intervention causality?

RCTs

What were the findings of the Physicians' Health Study?

low-dose aspirin reduced the risk of first MI by 44%

no benefit or harm associated with beta carotene

What are other names for non-randomized trials?

clinical trials or comparative studies

What is a disadvantage of not randomizing subjects?

results may be questionable due to introduction of biases

What is a sequential controlled trial?

pre-planned, repeated series of data comparisons (can recruit subjects gradually)

How do investigators know when to halt a sequential control trial?

when treatment efficacy or non-efficacy is demonstrated

What is a self-control trial?

the same group of subjects is used for both the experimental and control options

What is the Hawthorne Effect?

subjects change their behavior simply because they are being observed or are participating in a study

What differentiates a crossover study from a simple randomized control trial?

in a crossover study, both groups receive both treatments just at different times

What are external controls, and when would this type of study be useful?

external controls replace placebo arms in clinical trials (control may be historical data)

useful in studying rare disease or diseases for which a cure does not yet exist (cannot randomize & cannot withhold treatment)

When might uncontrolled studies be utilized?

- phase trials for new drugs

- when it is ethically difficult to conduct a trial with a comparable control group (rare diseases)

- establish efficacy for further investigation

What are some disadvantages of uncontrolled studies?

- risk of bias (no randomization)

- potential overestimation of effect (no comparison group)

What are some advantages of performing meta-analysis?

combines results of multiple similar studies to form an overall conclusion

can be useful for studies with small subject numbers or when different conclusions are drawn

What is it called when neither the subject nor investigators know if the subject is assigned to the treatment or control group?

double blinding

What are nominal data?

data that is categorical without inherent order (gener, hair color, sports preference)

What are the Experimental Event and Control Event Rates?

Experimental Event Rate is the proportion of people with the risk factor who have or develop the disease (A/A+B)

Control Event Rate is the proportion of people without the risk factor who have or develop the disease (C/C+D)

What information is provided by the risk ratio?

tells whether there is an increased risk, decreased risk, or no risk of a particular outcome

How is the risk ratio calculated?

RR=EER/CER

calculated from cohort study or clinical trial

Calculate the relative risk for a meningitis vaccine study with an EER of 0.3224 and CER of 0.3838. What does this value indicate?

RR = 0.3224/0.3838 = 0.84

since RR < 1, there is a decreased risk of meningitis for those who received the vaccine

What information does the odds ratio provide?

provides the odds that an outcome will occur given a particular exposure compared to the odds of that outcome occurring without the exposure

What is the significance of an OR of 2.35

the exposure is associated with higher odds of the outcome (OR>1)