Eihp Exam Review Questions

RCT can show causality or confidence in study because of lack of bias
Quasi is an experiment/ intentional study to take

It is a false positive. Where you reject the null hypothesis, when it is actually true. You observe a relation when there is none

It is a false negative. You accept the null hypothesis when you shouldnt have. You do not see a relation when there is one.

Increase sample size

No. It is too large and must be of a smaller category

The mean is sensitive to outliers and is not resistant to the median

Yes, the median is resistant to outliers

Negative

Positive

Positive

Negative

The average of all the numbers divided

The 'middle' value in the list of numbers

The number that occurs most often

Consistency and repeatability - ie same results, over and over

Whether the construct you are using really measures what you are using it to measure

The result is significant and you accept the null hypothesis

Descriptive describes the nature of the study

Relational are questions that question the relationship between two variables

Casual questions refer to how one variable will affect another

- The independent variable is what is changed. It is the exposure or the intervention
- The dependent variable is the variable changes because of the independent variable. Essentially it is the outcome

- The null hypothesis is a statement that declares that the variables will not have a relationship. Eg: 'it does not', 'there will not', etc
- The hypothesis is a statement that declares that there will be a relationship between the variables. Eg: 'if, then',

- Simple random: Every unit has an equal chance of selection
- Stratified random: Units divided into clusters and a % of each cluster is chosen
- Cluster: Units divided into clusters and all of a few selected clusters are chosen
- Systematic random: Every 'X' unit is chosen
- Multi stage: A combination of above stages

- Convenient: Whichever units are convenient
- Purposive: Specific units sought out
- Modal instance: Most frequently occuring unit
- Exerpt: High quality units sampled
- Diversity: Wide range of units sampled
- Snowball: More units are added as the study progresses

There is a bias in measurement which will affect the mean

Is a human error which will affect the standard deviation

The attribution of quantities and qualities to characteristics of people, objects or events

Quantitative research

How the findings of the study can be generalised to other studies

Validity of the study design characteristics

Objective is physical quantities or qualities
Subjective is human ratings or judgements

Observed value = True value +/- measurement error

1. Well defined
2. Same information creates same answer
3. Different information creates different answer
4. Information is understood

Nominal: Categorical information. Where you can not get an average
Ordinal: Intervals between each unit are not the same
Interval & ratio: Intervals between each unit are the same

Experimental: Scientific interventions = holds high external validity
Observational: No intervention

1. Ecological
2. Cross sectional
3. Cohort
4. Case control
5. Intervention

1. Focus groups
2. Interviews
3. Questionnaires

It is a systematic review of multiple pooled studies

1. Beneficence - Benefit to patient
2. Non Maleficent - Do no harm
3. Justice - fairness
4. Autonomy - Respect for rights

Guided notation for what is morally acceptable/unacceptable

1. Evaluates risks and benefits
2. Evaluates value of research
3. Evaluates biases of study

Mode, mean, median and standard deviation

What it does not contain "1". Eg: a 95% confidence rating would be 0.05

( A / C ) / ( B / D )

( A / A + B ) / ( C / C + D )

<0.05 = Rejection of the null hypothesis
>0.05 = Acceptance of the null hypothesis

1. Patient values
2. Clinical Expertise
3. External evidence

1. Systematic review of RCT
2. Random control trials
3. Quasi controlled trial
4. Case control and cohort group (observational studies)
5. Case studies, cross-sectional, Anecdotes

Quasi controlled trials are not random

Less expensive than cohort studies
Suitable for rare exposures
No attrition problems

Less

C - More
B - Positive
A - Protective
E - Less

3. -0.30 is stronger than a correlation of 0.25

Interval / ratio data

Who you want to generalise the study findings to

Accessible population to draw the sample from

1. Theoretical population
2. Study population
3. Sample frame
4. Sample

By changing the sample size

0.01 - 0.25 = weak
0.25 - 0.50 = weak to moderate
0.50 - 0.75 = moderate to strong
0.75 - 1.0 = strong

1. Focus groups
2. Interviews
3. Questionnaires

They can infer correlation but cannot infer causation

<1 = Exposure decreases the odds of having an outcome
>1 = Exposure increases the odds of having an outcome

It is 1.7 times more likely to have that outcome

It is 31% less likely to have that outcome

1. "My dog does not bite"
2. "My dog bites, but not you"
3. "My dog did bite but it didnt hurt"
4. "My dog bit and it hurt but its not my fault, its your fault"

The case will be more personal, rather than direct data

Three

It is a form of indirect protection from infectious disease that occurs when a large percentage of a population has become immune to an infection, thereby providing a measure of protection for individuals who are not immune

1. All actions must benefit the patient
2. Patient confidentiality must always be present
3. Doctors must not seek self-benefit

That there must be voluntary consent

Deductive: Many ideas are funneled down into simple statements

Inductive: Simple statements create larger concepts

Levels of exposure

A study with no preconceived ideas. The data flow, influences the direction of the study

It is a study of cultural differences

It is the study of lived experience vs representative experience and is conducted through observation

When it is the baseline demographic: >0.05 represents between-group differences.

Baseline demographics are good so we can know if any differences occur because of the study and not extraneous variables

Variables outside the study that can effect the results of the study

Randomly Controlled Trials (RCT's)

RQ is a question. The hypothesis is a statment

X is the casual
Y is the affected

1 Extraneous variables: variables outside the study affecting the results of the study
2. Directionality problem: whether or not the cause truly prefaced the effect

Directionality problem

Extraneous variable problem

It is the process of converting theoretical ideas into a tightly defined & quantifiable statement

How the researcher chooses to study a variable

1. Reliability
2. Validity
3. Practicality
4. Applicability

1. Test-retest: administering the same test twice
2. Inter-observer: having multiple clinicians assess patient
3. Internal consistency: How each unit within the study correlates with one another

Systematic error affects the mean
Random error affects the spread

Population: information regarding the population
Intervention: treatment, diagnosis, prognosis
Comparison: comparing interventions
Outcome: Desired outcome

They help structure a solid research question

To what degree information can be trusted

Pros:
- Resolves contradictory findings
- Easy to interpret
- Minimal bias
- Low confounding variables
Cons:
Can gloss over individual flaws

Probability: Accurate
Non-probability: Inaccurate sample frame

Probability: Can not be calculated
Non-probability: Can be calculated

Probability: Determined by sample theory
Non-probability: Matter of convenience

Probability: Representative
Non-probability: Illustrative

Face/content: Difficult to find external measures to correlate with the test
Sensitivity/specificity: Proportion of people who test positive/negative
Predictive: Ability to test predictive values of research

- Repeat several times
- Measure under controlled conditions
- Calibrate equipment
- Know error of the tool
- Use same person to measure

Relative risk: Ratio of probability of an event occurring in an exposed group, compared to an unexposed group
Odds ratio: Measure of association between exposure and outcome

Test that compares the average scores between 3 or more outcomes

Single blind: Participant doesnt know what group they are in
Double blind: When neither researcher nor participant knows what group they are in

Single: Reduces personal bias
Double: Reduces researcher bias

Observational

Attrition

Changing the research question in light of the data

Positive: Exposure increases offs of outcome
Negative: Exposure does not increase odds of outcome

The relative risk is generally less than the null value

SD = 68%
SD2 = 95%
SD3 = 99%

SD2 = 90 and 110
SD3 = 85 and 115

SD2 = 300 and 400
SD3 = 275 and 425

Smoking = Independent variable
Cancer = Dependent variable

1. A positive correlation

As sleeping hours increase, so does mental alertness. Showing a graph in a positive direction upwards

It provides direction and strength. Correlation is always between -1 and +1