exam 2

the two types of sampling are probability and nonprobability based

what are the 4 types of probability based sampling that exist?

1. simple random

2. systematicc

3. stratified

4. cluster

why is sampling important?

1. reduces selection bias

2. allows for statistical inference

3. improve validity of finidngs

4. essential for public health deciisons

5. ensures representativeness

Which selection biases are listed below?

1. selection process favors individuals with characteristics that are not representative of the population as a whole

2. There is a systematic difference between participants in the exposed vs unexposed groups

3. characteristics of responders and non-responders differ

4. some individuals lost to follow of differ from those that were not lost to follow up with respect to the exposure and outcome

1. prevelance-incidence bias

2. allocation bias

3. nonresponse bias

4. loss to followup bias

information bias is the disortion in the measure of ASSOCIATION caused by a lack of accurate measurements of key study variables

what are the 4 types of information bias?

1. recall - more likely to remember exposure if it had a significant outcome

2. observer- investigator’s prior knowledge of subjects disease status and treatment leads to asking questions or assessing the subject differently

3. interviewer- interviewer asks questions to confirm past beliefs

4. detection- differences between groups in how outcomes are determined

cros sectional studies where it cannot be determined if exposure proceded dissease since both are acertained at the ame time

antecendent consequent bias

cant tell if if patient was given Aspirin FOR CV risk or if they obtained CV risk for being on lisinopril

what are strengths of cross sectional studies?

• identifies relationships between variables

• multiple outcomes and exposure can be studied

• can be used as a preliminary step before further research

• convenient way to collect large amounts of data

• inexpensive

• no ethical difficulties

what are weaknesses of cros-sectional studies ?

• cannot infer causality (but you can estimate prevalance DOESNT NECESARILY MEAN ONE IS CAUSED BY THE OTHER BUT IT IS A PATTERN)

• unable to measure incidence (how many people have recently gotten the outcome from a specific time frame)

• require large sample size for accuracy (can be a stregth as well IF YOU HAVE ENOUGH PEOPLE)

• may not inclyde all the relevant groups of a populaton

• makes it dofficult to interpret identified relationships

• not suitable for long term trends

A cross sectional study found that many individuals with low cardiovascular risk were using aspirin. What might this suggest about the use of aspirin in the general population?

1. The use of aspirin in individuals with low cardiovascular risk suggests possible

overuse

2. Aspirin is only being used by those with a confirmed need for secondary

prevention.

3. Low-risk individuals benefit more from aspirin than high-risk individuals.

4. Aspirin has no side effects, so it is safe for everyone to use regardless of risk level.

1

a cross sectional study based on a questionanire/survey may have kewed results due to _____ bias

a. recall

b. interviewer

c. observer

d. selection

which study gives a snapshot of the frequency of diease or other health related characteristics in a population at any given point in time?

cross sectional study

cross sectional studies are observational studies that answer questions about ________ or _______________ of a condition, _____, or ________

incidence or prevelance of a condition, belief, or situation

proportion of persons in a population who have a partiuclar disease or attribute at a given time regardless of when they first developed the disease

prevelance

number of new cases that develop in a given period of time

incidence

how to calculate prevelance (amount of people exposed over a period of time)?

exposed with disease / total exposed divided by not exposed with disease / total not exposed

measure of how strongly an event is associated with exposure

an Odds Ratio of _____ means there is no association between exposure and outcome

1 NOT zero

how do you find odds ratio?

AD/BC

A= diseased and exposed

D= no disease and not exposed

B= no disease and exposed

C= disease and not exposed

odds of exposure and disease = A/B

odds of nonexposure and disease = C/D

A/B / C/D = AD/BC for odds ratio

Which type of cross-sectional study is used to

• characterize and assess the prevalence and distribution of one or many health outcomes in a defined population

• access how frequently, widely, or severely a specific variable occurs throughout a specific demographic

• MOST COMMON TYPE OF CROSS-SECTIONAL STUDY

• evaluates the proportion of a population with disease or with risk factors for disease

• used to ESTIMATE the occurrence of risk factors in segments of the population characterized by age, sex, race or socioeconomic status

DESCRIPTIVE cross sectional study (observing characteristics-

which cross sectional study is used to

• collect data for exposure and outcomes at one specific time to measure an association between an exposure and a condition with a defined population

• investigate an association between two parameters

• tries to answer HOW or WHY a certain outcomems might occur

• useful for establishing preliminary evidence for a casual relationship

ANALYTICAL (association and outcome vs exposure)

what are the three types of observational studies

1. cohort (expose—> see outcome)

2. case control (outcome —> identify what was common exposure)

3. cross sectional (outcome @ exposure)

_________ studies focus on one outcome but multiple exposusures

_________studies focus on one exposure but multiple outcomes

case-control = one outcome (disease) look back at exposure (what caused said disease)- retrospecitve

cohort= introduce one exposure see what outcomes arise from it (prospective)

true/false: case-control studies are ALWAYS retrospective

true

where do you find the information for case-control studies?

primary data or existing data

• questionaires (may have recall bias)

• Electronic health records or automated databases

case-control vs cohort?

when do you obtain odds ratio vs relative risk ratio?

which is prospective vs which is retrosspective?

which uses information from primary data existing data?

cohort =

relative risk ratio

prospective AND retrospective

case-control =

odds ratio

ALWAYS retrospective

for case-control studies are incidence cases or prevelance cases preffered and why?

incidence because you have better recall of the exposure to a new disease than one that has been present

what are the controls for case-control studies?

are case-control enrollement always 1:1 ?

patients from the same source population as the patients with cases

ex. case= patient with MI control = patient w/o MI from same population

NO can also be 1:5

A __________ can be conceptualized as a more effecient version version of a cooresponding ________ study when OUTCOME IS RARE

case-control > cohort when outcome is rare

why can’t you determine the absolute relative risk of case-control studies?

you can have different proportions of case:control which will change the absolute relative risk lacking consistency

Pros of case control study

• ideal for rare outcomes or outcomes with a long lag time

• fewer patients needed (good when chart review is needed)

• quick and innexpensive

Cons of case-control study

• cannot estimate incidence (except nested case control)

• cofounders

• selection of controls is difficult

• reliance on recall or records to determine exposure status

Pros of Cohort studies

• conceptually simmilar to randomized control trial

• can study multiple outcomes

• exposure is assessed before outcome (no potential for recall bias)

• less potential for bias

• absolute risks can be estimated

cons of cohort studies

less effecient for rare outcomes (especially prospective cohorts)

Requirements of Case-Control studies

• define the source ________from where cases and controls will be selected

• define exposure, __________, and outcomes of interest

• identify cases (diseased individuals) and controls (typically ____ or ______ ratios —- and determine exposure status)

• Calculate _____ RATIO

• population

• confounders

• 1:4 or 1:5 ratio of case vs control

• ODDS (case with exposure/ no case with exposure) / (case without exposure / no case without exposure)

cons of randomized control trials:

• unable to do blinding

• study population not generalizable

• underpowered

• short follow up

• measurement errors

• poor compliance

• randomization can fail

which experiemental study seeks to determine

• how does the drug preform in an ideal situation = efficacy

vs

• how does the drug preform in the real world = effectiveness

ideal situation = efficacy = RANDOMIZED

real world = effectiveness = nonrandomized

what are the 3 descriptive observational studies?

1. case-series

2. case reports

3. ecological

what is the biggest threat to observational studies?

lack of randomization- due to unmeasured confounding

while randomized control trials can balance both observed and unobserved characterstics

• blinding is not possible

What is a cohort?

• come from latin word “carhors” which is a _________

• what is a cohort in terms of medical studies?

• military unit (well defined group)

• drug-based inclusion criteria

what is the difference between retrospective and prospective cohort studies?

porspective cohort studies involve following up on patients after completing a questionairee or telephone interview to see which outcomes they have developed after a specifice period of time (limitation= very costly to follow up)

retrospective cohort studies involve looking at previously recorded datea such as electronic health factors and seeing which outcomes have resulted from a certain exposure (limitation= you only have the outcomes that are provided in the database)

the exposed and unexposed group are selected from different populations or the way they are selected differs and creates unbalanced comparison groups

_____ bias is also a type of ______bias

selection bias

attrition bias is also a type of selection bias

which bias may be present in this scenario:

• study compares GI risk of aspin vs naprozen

asprin users - identified from sales record at a local pharmacy

naproxen users= identified from medical records from a pain clinic

selection

what is the number 1 threat to internal validity in observational studies?

confounding

is the confounding factor associated with both the exposure AND the outcome?

yes BUT not everyone with the exposure has the confounding factor so it is not an intermediate

what are 3 ways to control confounding in observational studies?

1. restriction - limit analysis to patients without history of CV events

2. stratification - seperately conduct analysis for patient by exersize levels (confounder)

3. multivariate - adjut for pertinent confounders in regression model

what are two examples of way that compounding can manifest in an observational study?

confounding by indication and severity

what i the difference in fofrmula of risk ratio and odds ratio?

risk ratio = (case/exposed / case/nonexposed)

odds ratio = (AD/BC)

when woud you use adjusted relative risk in a cohort study?

if the relative risk may be due to counfounders.

it may appear that the risk of an outcome is due to the exposure of a drug, but there may be underlying factors such as the gender of the the patients that may decrease the outcome, which you should keep in mind before assuming it is the drug that is decreasing exposure

if there are multiple factor that can decrease the exposure make sure to acknowledge them through MULTIVARIATE regresion modles

what are examples of claims and EHR data that may be used in retrospective cohort studies?

claims = medicare, medicaid, BCBS (billing data from pharmacies, physicians, hospitals)

EHR data= CPRD, RWJH (EPIC data)

form of mutlivariate adjustment that can summarize contribution of risk factors to probability of exposure NOT considering outcome

propensity score

Cohort Study Summary:

• choice of ________ group can help a lot with confounding

• any method (restriction, stratification, multivariate adjustment) is only has good as the measurement of the _________

  • UNMEASURED ____________ can pose problems (healthy user effect)

• If you are concerned that ________were not properly addressed, think about the expected effect on the association

comparion

counfounder

confounder

confounder

Cohort Study Summary Continued:

• first choose between _______ and ______

• define _____ of interest and active ________

• define _____ and _______ criteria to prevent cofounders

• define study start and end of follow up

• operationalze exposure and __________, define ______

• estimate _________ score (and match)

• estimate _____ and ___________ ______________ ( )

1. retrospective vs prospective

2. define drug of interest and active comparator (CONTROL- instead of no treatment)

3. inclusion and exclusion criteria

4. cofounders define outcomes

5. propensity

6. incidence and adjusted rate (cox proportional hazard model)

SYSTEMATIC REVIEW:

advantages:

• limit _________

• improve ______ of results

• yield new ___________ about subgroups of the populations

• develop reliable and accurate conclusions

disadvantages:

• _________ and selection of _____

• _______ of publication

• loss of information on important ______

• ___________ of populaitons studied

advantages:

• limit bias

• consistency

• hypothesis

disadvantages:

• location and selection of studies

• duplication

• outcomes

• heterogeneity

steps of a systematic review:

1. select topic

2. conceptualize and create protocol

3. search for ________________

4. ___________ ________

5. ___________ ________

6. ______ data

7. _______ and _____ ______

seach for studies

screen studies

apraise studies

abstract data

synthesize and interpret results

what are the 3 components of a meta analysis?

1. combinaiton

2. estimation

3. standardization

what are components that are shared between meta analysis and randomized control trials?

1. predefined methods to evaluate a hypothesis

2. incluion/excluion criteria

3. predefined statistics

4. presentation and discussion of findings

what are the 6 steps to preforming a meta-analysis?

1. define the question

2. determine methodology

3. search strategy

4. data abstraction

5. data analysis

6. knowledge synthesis

whcih graph is used for meta analysises to allow readers to quickly:

1. See the individual effects of included studies

2. Evaluate the strength and precision of each estimate (weighted)

3. Understand the pooled result and its precision (diomand = pooled)

4. Judge heterogeneity visually

forest plot

META ANALYSIS:

advantages:

• high _____ _______

• ______ data analysis

• _____ validity increased

• top tier _________-based resource

disadvantages:

• study ________ can be cumbersome

• inadequate//inconsistent ______ between studies

• advanced statistical technique required

• ________ of populations studied

advantage:

• high statistical power

• confirmatory

• external

• evidence

disadvantages:

• identification

• data

• heterogeniety

meta analysis allows researchesrs to increase ______ _____ and quantitatively ______the available literature

increase sample size

evaluate

what is I² ?

measure of heterogeniety

if I² less than or equal to 25% then homogenous

or fixed effect model (studies are very simmilar, differences only occur in sampling)

if I² greater than or equal to 75 percent then heterogenous or random effect modle (studies have differences)

what is publication bias?

publication with positive results are more likely to be published then those with negative results

check using funnel blot for symetry to see if studies were used that have differing results using EGGERs weighted statistic

what are the components of the data analysis of meta analysis?

• evaluate heterogeniety

• select meta- analytic model (PRISMA or Prospera)

• compute outcome measures

what are the components of PRISMA which is used to determine methodology of meta analyises?

Identification (where did you get your sources)
Screening (get rid of duplications)

Eligibility (exclusion/ inclusion - evaluate quality)
Inclusion (which sources involved in analysis!)

I See Every Item

what are the componenets of data extraction?

1. evaulate quality of study

2. collect and extract data

3. pool data

what is the hierchy of research and scientific evidence

1. systematic review and meta analysis

2. randomized control trial

3. non-randomized control trial (quasi or pre-post intervention study)

4. observational studies (case control vs.cohort)

5. qualitative studies

what are the 6 types of quasi experiments (non randomized experiemental trials) ?

1. pretest-posttest

2. non-equivalent control group

3. time series

4. interrupted time series/multiple time series

5. regression discontinuity

6. retrospective ex-post facto

what is the difference between the following quasi experiements?

1. pre-test/post-test

2. nonquivalent control

3. time series

4. interupted time series/ multiple time series

5. regresion discontinuity

6. retrospective ex-post facto

1. record obervation before and after

2. nonrandom assignment of participants that recieve exposure and control that does not recieve exposure

3. many observations taken on the same variable consecutively over time

4. many observations taken before exposure and after exposure

5. determine casual effects by asigning threshold

6. cause- and effect relationsjopns through RETROSPECTIVE search

when and why you would ue a quasi-experimental design?

• clasic study designs are not _____ or _____

• if the investigator cannot implement a ____ group or _____ the study groups

• small/large sample size

• inadequate _______

• feasible or ethical

• if you CANT implement contol group or randomize the study groups

• small sample size

• funding

internal validity examines the extent to which _____ _____ (____) is present

systematic error (bias)

_______ validity is limited in short-term studies of patients who need to be treated for months to years

external

which is decribing DUR (Drug Utilization Review) vs which is describing MUE?

1. ongoing, systematic criteria based, drug or disease specific assesment that ensures appropriate medication utilization at the individual patient level

2. emphasis on improving patient outcomes and quality of life through assesment of clinical outcomes via a multidiciplinary approoach

1. individual patient = DUR

2. overall - MUE

preformance imrpovement method used to optimize patient or system moutcomes

medication use evaluation

who are the MUE team members?

1. core team members (pharmacists)

2. extenders (students)

3. subject matter experts (specialists)

4. other contributions (case managers)

What are the two main types of MUE?

retrospective and prospective

life cycle of a MUE:

1. design the MUE

2. collect and analyze ______

3. implement _____ plans

4. assess ________

5. follow up- with _______ / further ______

2. data

3. change

4. effectiveness

5. changes/ research

Match the following threats to validity:

1. undermine the validity of predictions by over-emphasizing features close to the outcome of interest

2. Participants change their behavior knowing they are being studied/examined

3. differences in how outcomes are measured in each group

4. systematic error in the way data is collected/ creates false associations that impact the results of the study

5. publishing only positive results and neglecting negative results

6. The patients/participants selected have distinct differences between groups

7. events that occur during the study period could influence the outcomes of the results

1. temporal ambiguity

2. hawthorne effect

3. detection bias

4. confounders and control techniques

5. reporting bias

6. selection bias

7. historical bias

Selection bias leads to —→ _____ __________

• maximizes the likelihood that measured and unmeasured confounding variables are distributed equally

selection bias —> random allocation

Breaking down the MUE Changes:

• ________ of broad-spectrum antibiotics

• _______ changes

• _______ to improve compliance

• new ________/ ________procedure

overuse

formulary changes if drug not used often

education

new monitoring/ dispensing procedure

what are the appropriate steps of following up after instilling a Medication Utilization Evaluation?

1. DOCUMENT the change and write new processes for continual reference and onboarding training

2. ensure it is SHARED

3. set timelines to MONTIOR change

4. ensure the culture DOESNT REVERT to old ways

What are limitations of Medication Utilization Revews?

• lack of ______, ________ , ______, _____, or _______

• poor ____________

• data ______, difficulty retrieving data

• no ______ _______

• evaluation method disruspts _____ _______

• lack of _______

• lack of _________and training of the MUE team

• lack of hard-wired _______ actions

• lack of authority, organization, structure, leadership, involvement

• poor communication

• integrity

• follow-through

• patient care

• scope

• education

• corrective

what are some example of MUEs that have been used in experimental studies?

1. usage restriction to severe scenarios

2. formulary changes

3. system alignment/standardization and cost saving evaluation

4. how well does a health system align with national metrics

what is the process of a MUE?

1. design

2. evalueate

3. implement

4. assess and follow up

5. repeat

data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources

real world data

clinical evidence about the usage and potential benefits and risks of a medical product derived from analysis of real world data

real world evidence

Answer the following regarding randomized control trials (RCT):

• purpose

• settiing

• population

• population size

• patient follow-up

• treatment

• comparator

• attending physician

• generalizability

• control for bias

• costs

• purpose= efficacy/safety

• setting = research

• population = homogenous

• population size = small- moderate

• patient follow-up = fixed (design)

• treatment = fixed

• comparator = placebo / selective alternnative interventions

• atttending physician = investigator

• generalizable= low-moderate

• control for bias = design and conduct

• costs = high

Is a randomized control trial or real world evidence being described?

• purpose = Efficacy/Safety

• setting = research

• population = homogenous

• population size = small- moderate

• patient follow up = fixed (design)

• treatment = fixed

• comparator = placebo/ select alternative interventions

• attending physcian = investigator

• generalizability = low-moderate

• control for bias = design and conduct

• costs = high

randomized control trial (RCT)

Is a randomized control trial or real world evidence being described?

• purpose = Effectiveness/safety

• setting = real- world

• population = heterogenous

• population size = large-huge

• patient follow up = variable (actual practice)

• treatment = variable

• comparator = MANY alternative interventions

• attending physcian = practitioner

• generalizability = moderate - high

• control for bias = analysis

• costs = low

real world evidence

Answer the following regarding real world evidence (RWE):

• purpose

• settiing

• population

• population size

• patient follow-up

• treatment

• comparator

• attending physician

• generalizability

• control for bias

• costs

pupose = effectiveness/ safety

setting = real-world

population = heterogenous (not everything is simmilar/controls very different patients)

population size = large - huge (not controlled amount of people)

patient follow-up = variable (actual practice)

treamtment = variable (you don’t get to chose which patient gets which drug, its up to their provider)

comparator = many alternative intervention (not just one drug/ placebo vs drug you are observing)

attending physisian = practitioner (actually in practice they are giving the patients the medicaitons, not an investigator)

generalizability = moderate-high

control for bias = analysis

costs = low

the 21st centrury cures act defines Real World Evidenceas data regarding the ______, potential ______, or _____ of drug derived from sources other than traditional clinical trials

usage

potential benefits

risks

Potential Uses of Real World Disease in Clinical Studies

• non-________, non-__________ (___________)

• prospective data collection

  • _______ studies

  • prospective ______ studies

• retrospective (existing data)

  • ____________

  • ____________

• non-randomized, non-interventional (observational)

propective data collection:

• registry studies

• prospective cohort studies

retropective (existing data)

• cohort studies

• case-control studies

what are the sources of real world evidence?

1. social media

2. EHR

3. claims

4. registeries

5. software apps

6. clinical labs

7. Internet of Things (IOT) - glucose monitors

Generation of Real World Evidence from diverse sources of Real World Data:

• claim data

• lab results

• EHR

• registeries

• cost of care

• patient reported outcomes

what is the difference between a nonrandomized interventional study vs a randomized noninterventional story?

interventional is an externally CONTROLLED trial whereas noninterventional studies are OBSERVATIONAL including cohort, case-control, and case-crossover

Fitness for Use is assesing data reliability and relevance:

what are some things that should be taken into consideration regarding the data?

• minimal _______ data

• sufficient data ______ and ________

• established data ______ _______ procedures

what are some things that should be taken into consideration regarding the methods?

• _______ vs ____________

• ________ vs _____________ vs _________

• credibility established (_______ developed and _______of results achieved/planned)

after fitness for Use

data:

• minimal missing data

• data reliability and validity

• quality assurance

methods:

• observational vs interventional

• retrospective vs prospective vs hybrid

• protocol developed, replication of results achieved