Designing Epidemiological Studies (Part 1)

Question 1

What are the two types of epidemiological approaches?

Answer 1

Analytical epidemiology

Descriptive epidemiology

Question 2

What is a descriptive epidemiological investigation?

Answer 2

Investigation that describes the problem at an aggregated (generalised) level

Question 3

What is a use of a descriptive study if epidemiological investigation?

Answer 3

Can be used to inform analytic research

Question 4

Define analytic study.

Answer 4

Type of study that deploys and tests hypotheses at a person-based level through which association can be measured and causation inferred

Question 5

What is the purpose of a analytic study as a form of epidemiolgical investigation?

Answer 5

Deploy and test hypotheses through which association can be measured and causation inferred.

Question 6

Differences between analytic and descriptive study as forms of epidemiological investigations.

Answer 6

Descriptive is aggregated level based whereas analytical is person level based

Descriptive can help formulate theories and hypotheses and analytic tests these theories

Question 7

List 5 types of epidemiological investigations that deploy a descriptive approach

Answer 7

Case report

Case series

Cross-sectional studies

Longitudinal studies

Ecological studies

Question 8

Which 2 descriptive studies can sometimes have hypotheses and adopt some analytical approaches?

Answer 8

Cross-sectional

Longitudinal

Question 9

What are the 3 epidemiological dimensions in which diseases are described in relation to for descriptive studies?

Answer 9

Person

Place

Time

Question 10

Give examples of ways diseases are described in relation to person, place and time

Answer 10

Person: Age, gender, occupation, disease status

Place: Hospital, communities, geographical area

Time: within a specifc point in time or within a specified time period

Question 11

Give 2 examples of measures in descriptive studies

Answer 11

Incidence

Prevalence

N.B. Often point estimates with a confidence interval

Question 12

Define parameter

Answer 12

A fixed often unknown value which describes an entire population

Question 13

Define statistic

Answer 13

A fixed value derived from a sample that is meant to estimate the value of a population

Question 14

Give an example of a statistic and a parameter

Answer 14

Statistic: Gets a large sample of mens height to estimate the averge mans height in the entire population

Parameter: The true value of average mans height (Which is unknown as they would have to measure every single mans height as a result)

Question 15

What is the relationship between a statistic and a parameter?

Answer 15

A statistic can be used to estimate a parameter (estimate as a paramater is an unknown value unless every person in the population was in the estimate)

Question 16

A statistic which seeks to estimate a paramater is known as?

Answer 16

A point estimate

Question 17

What are 2 potential reasons for a large 95% confidence interval?

Answer 17

Volatility in low number samples

Question 18

A point estimate is a value within?

Answer 18

A 95% point interval

Question 19

Define case report

Answer 19

Detailed report of the symptoms, signs, diagnosis, treatment, and follow-up of an individual patient

Question 20

What are 3 main purposes of a case report/series?

Answer 20

Communicate a potentially new disease

Communicate new and unusual presentations of a disease

Communicate new findings of a disease

Question 21

What are case reports/series used mainly for in medicine?

Answer 21

Used as a learning opportunity for continued medical education (CME) or UK continuing proffesional development (CPD)

Question 22

In order for case report to identify a potentially new disease, what is required?

Answer 22

Needs to have support from other similar case reports showing similar findings, collectively known as a case series.

Question 23

Give 2 examples of diseases in which a case series helped to idenitfy their existence

Answer 23

Covid-19

MERS

Question 24

What is a cross-sectional study?

Answer 24

Type of study which describes the prevalence of a condition across a population in a single point in time

N.B. Essentially a snapshot of disease

N.B.

unlike longitudinal you only have 1 set of data collected from tht point in time

Question 25

Give an example of a cross-sectional study.

Answer 25

Survey

Question 26

Define longitudinal study

Answer 26

Type of study which describes the prevalence or incidence of an exposure or outcome over time

Question 27

What is the relationship between cross-sectional and longitudinal?

Answer 27

Cross-sectional data at different time points can be used collectively to form a longitudinal study.

N.B. This is example of aggregated data as unlikely to get the same exact participant

Question 28

Longitudinal studies may use aggregated data or person-level data. Explain what this means.

Answer 28

Aggregated data can be data collected from a collection of cross-sectional studies

Person-level data involves taking data from the exact same participants from different periods of time

Question 29

What do ecological studies look to compare?

Answer 29

Compares groups versus individuals (e.g. prevalence of a disease between different countries/schools/hospitals)

Question 30

Does ecological studies use aggregate or person-level data?

Answer 30

Aggregate level data

Question 31

Assumption that information within groups hold for individuals is known as?

Answer 31

Aggregation bias (aka ecological fallacy)

Question 32

Define aggregation bias/Ecological fallacy

Answer 32

Assumption that information between groups accounts for individuals within the group

Question 33

Despite some negatives around ecological studies, what are 2 benefits?

Answer 33

Can generate hypotheses for further research to be conducted from an anlaytical perspective

Can use secondary data sources which are cheap and quick as readily available

Question 34

Give 2 examples of when ecological studies are suitable to conduce

Answer 34

When the variable between exposure is limited

Question 35

Give two examples of weaknesses of ecological studies

Answer 35

Cannot tell if exposure was before outcome and vice versa (e.g. did chocolate consumption increase due to celebrating nobel prize win or did chocolate consumption increase the chances of winning?)

Subject to ecological fallacy/aggregate bias

Question 36

What is primary and secondary data?

Answer 36

Primary data: Collected by researcher first hand

Secondary data: Data that was collected for a different purpose other than the research itself

Question 37

What is a strength and 2 weaknesses of primary data?

Answer 37

Strength: type of data collected can be tailored to fit purpose of research

Weakness: Time consuming and costly

Question 38

What is a strength and 2 weaknesses of primary data?

Answer 38

Strength: Cheap and less time consuming

Weakness: Data not collected specificaly for study so researcher will have to make assumptions which can introduce limitations on how the study is interpreted.

Question 39

Public healthcare term primary and secondary data as?

Answer 39

Primary: Non-routinely collected data

Secondary: Routinely collected data

Question 40

Give 2 examples of routinely collected data

Give 1 example of non-routinely collected data

Answer 40

Census (every ten years so can be outdated)

HES (Hospital episode statistics) [data of all hospital administrations (in-patient, out-patient and emergency admission) from laast ten years]

Non-routine: Surveys

Question 41

What is data linkage and its purpsoe?

Answer 41

Combination of 2 or more different data

Helps in finding out more than was possible by analysis of either original dataset alone.

Question 42

Data linkage could be incredibly helpful in healthcare (e.g. identify association between frequent attenders of A&E and GP attendances to take pressure off of A&E). However it is often underused. Suggest 2 reasons why?

Answer 42

Technical issues (hard to join healthcare records between different hospital trusts for example)

Privacy concerns

Question 43

Cross-sectional studies can measure prevalence but not what else and why?

Answer 43

Cannot measure incidence (like longitudinal) as there is no follow up and is rather a snapshot in time

Question 44

Cross-sectional is not good for looking at exposures that change over time. Give an example of exposures in which a researcher doing a cross-sectional study can be confident that there is no difference in exposure over time.

Answer 44

E.g. Ethnicity, Gender

Question 45

What are the two types of study designs taking an analytical approach?

Answer 45

Observational

Experimental

Question 46

Give an example of a observational study design in analytical approaches of reasearch.

Answer 46

Case control study

Question 47

What is a case control study?

Answer 47

Study which looks at a target population (same characteristics) and compares the cases (e.g. disease) to the control (e.g. no disease), looking at differences in outcome to the exposure and unexposed

E.g. Look at male (target population) lung cancer patients (case) and compare to non-lung cancer patients who either smoke (exposure) or dont smoke (unexposed) and look at differences

Question 48

In case control studies, from a medicine perspective cases should represent?

Answer 48

Should represent everyone with disease under investigation (e.g. dont just collect data from lung cancer patients who are going into surgery)

Question 49

In case-control studies, in order for a sample to truly be controls, what 2 things must be considered?

Answer 49

Source of control (Must come from same population and should be representative of the population at risk)

Assesment of exposure (They should have same degree of exposure as the cases) [e.g. both case and control are heavy smokers]

Question 50

When determining that control and case had same measures of exposure in a case-control study, what type of bias must be taken into account when determining exposure?

Answer 50

Recall bias

E.g. A person who has lung cancer may recall that they smoked way more than they did as a means of trying to find a reason for why they developed the disease)

Question 51

In a case control study, if the number of cases are small, what can be done to increase statistical power?

Answer 51

Can assign more than one control (as exposures are past events, can investigate a multitud of different exposures to determine cause)

Question 52

What are 3 strengths and 4 disadvantages of case control studies?

Answer 52

Strengths:

Good for studying rare diseasea

Inexpensive

Quick to obtain data (can assess exposure and outcome all at same time)

Disadvantages:

Bias associated with exposure assesment (e.g. recall bias)

Difficulty selecting control group that is representative

Can investigate different exposures but only one outcome

Cant determine temporal relationship between exposure and disease (did lung cancer start developing before the patient smoked?)