EBM

What are quantitative methods used for in epidemiology?

Used to study the occurrence, distribution and causes of health and disease in the population

How is data collected in studies?

In the form of variables

What are the 2 main variable types?

Numerical and categorical

What are the two types of numerical variables?

Continuous and discrete

Describe continuous numerical variables:

Measurements are made on a continuous scale and data is often displayed on a histogram

What are examples of continuous variables?

Height
BMI
Blood pressure

Describe discrete numerical variables:

Measurements are counts

What are examples of discrete variables?

Number of cigarettes smoked in a day

What are the two types of categorical variables?

Ordered and unordered

How is categorical data usually displayed?

In frequency tables, bar or pie charts

What are examples of an ordered variable?

Social class
Severity of disease

What are ordered variables?

Used to rank observations according to an ordered classification

What are unordered variables?

Used to class observations into a number of named groups

What are examples of unordered variables?

Marital status
Ethnic group

What is a dichotomous/binary variable?

When observations are divided into two groups

What kind of variable is a binary variable?

An unordered categorical variable

What kind of variable is BMI?

Continuous

What kind of variable is ethnicity?

Unordered categorical

What kind of variable is number of relapses in depressions?

Discrete

What kind of variable is social class?

Ordered categorical

What kind of variable is blood type?

Unordered categorical

What kind of variable is number of admissions to hospital?

Discrete

What kind of variable is platelet count?

Continuous

What kind of variable is birth weight?

Continuous

What kind of variable is blood cholesterol level?

Continuous

What kind of variable is number of psychotic episodes?

Discrete

What kind of variable is post treatment mortality?

Binary

What kind of variable is educational level?

Ordered categorical

What kind of variable is whether an individual has a disease or not?

Binary

What is the equation for proportion?

number with disease / total number of individuals

What are proportions often expressed as?

Percentages

What are the two types of proportion?

Prevalence and risk

What is prevalence?

The proportion of individuals with the disease at a particular point in time

What is the equation for prevalence?

number with the disease at a particular time / total number in population at that time

What is risk?

The proportion of new cases of diseases occurring in a specified time period

What is another word for risk?

Cumulative risk

What is the equation for risk?

number of new cases in time period / number initially disease free

What does risk tell you?

How many new cases will occur in a time period

What does prevalence tell you?

About how many cases exist at a particular point in time

What is the difference between risk and prevalence?

Risk tells you how many new cases will occur in a time period, whilst prevalence tells you about how many exist at a particular point in time

What does incidence rate tell you?

How fast new cases of disease are occurring

What is the difference between incidence rate and mortality rate?

Incidence rate is about cases of disease, mortality is the same as incidence rate but with deaths

What is the equation for incidence rate?

number of new cases of disease / (number initially disease free * time interval)

What is the equation for prevalence in terms of incidence rate?

incidence rate * average duration of disease

How can prevalence of a disease increase?

If the incidence rate of that disease increases and/or because the average duration of that disease increases

Why do more lethal diseases have a lower prevalence than less lethal ones?

Because they kill the patient faster, so there are less cases in the population at that time

What do most medical, biological, social, physical and natural phenomena display?

Variability

What is used to express the variability of phenomena?

Frequency distributionsn

How are frequency distributions summarised?

By measures of central tendency (mean, median and mode) and variability (range, inter-quartile range, standard deviation)

What can measures of central tendency and variability describe?

The distribution across the population

What do most medical statistics display?

Normal distribution

What is identical in normal distribution?

Mean, median and mode

What determines the shape of a normal distribution curve?

Standard deviation

What is the shape of a normal distribution curve is SD is small?

Tall and narrow

What is the shape of a normal distribution curve is SD is large?

Short and wide

What is SD if the shape of a normal distribution curve is tall and narrow?

Small SD

What is SD if the shape of a normal distribution curve is short and wide?

Large SD

What is the 95% reference range?

mean(-1.96×SD) to mean(+1.96×SD)

When is the 95% reference range mean(-1.96×SD) to mean(+1.96×SD)?

When the data is normally distributed

What does the 95% reference range mean?

In a normal distribution, 95% of all the data will lie in this range

What does a frequency distribution look like in a negative skew?

Long tail to the left

What does a frequency distribution look like in a positive skew?

Long tail to the right

What is the relationship between mean, median and mode in symmetrical data?

mean \= median \= mode

What is the relationship between mean, median and mode in negative skew?

mean < median < mode

What is the relationship between mean, median and mode in positive skew?

mean \> median \> mode

What are the measures of central tendency?

Mean, median and mode

What is the mean?

The average

What is the equation for mean?

sum of all values / number of observations

What is the median?

The middle value when all the values in a set are arranged in order

When can the median be more useful that the mean?

When there are few extreme values

What is the mode?

The most frequently occurring value in a set

What are the measures of variability?

Range, inter-quartile range and standard deviation

What is the range?

The difference between the largest and the smallest values in a set

Why can range be quite unrepresentative?

Because it depends solely on two extreme values

What is the inter-quartile range?

The difference from the end of the 1st quartile to the end of the 3rd quartile

What is standard deviation?

The measure of the spread of the observations about the mean based on the deviations of each observation from the mean value

What is 1 standard deviation?

68.3%

What is 2 standard deviations?

95.5%

What is 3 standard deviations?

99.7%

What are the two types of epidemiological study designs?

Interventional/experimental and observational

What is an example of an interventional/experimental study design?

Randomised controlled trial (RCT)

What is an example of an observational study design?

Cross-sectional study

Describe interventional/experimental study designs:

Investigator tests whether modifying/changing something in the treatment exposure alters the course of a disease

What is an example of an RCT?

Trials of medical therapies

Describe observational study designs:

Involve the investigator collecting data on exposures (potential risk factors) associated with the occurrence/progression of disease, without attempting to alter participants' exposure status

What is another name for cross-sectional studies?

Prevalence studies

What do cross-sectional studies measure?

The prevalence of a disease in a population at a particular time

What are cross-sectional studies useful for?

Measuring the burden of disease in a population as they usually obtain more accurate information than routine data from hospitals/primary care
For identifying risk factors associated with a disease by comparing prevalence of disease in individuals who are exposed to certain factors with those who are not

Why are cross-sectional studies more useful than routine data from hospitals/primary care?

Because of the clinical iceberg, which refers to the fact that doctors are only aware of the relatively small proportion of disease that is presented to them

Why is there a clinical iceberg?

Because depending on the disease, people may not have sought medical advice, not known that they were ill, or been afraid of stigma

What are the advantages and disadvantages of cross-sectional studies?

Useful for measuring the prevalence of health behaviours and conditions in the population
Can be used to generate hypotheses if comparing over time or across different geographies

Information on the exposure is obtained at the same time as the outcome, making it difficult to determine which came first (reverse causality)
The prevalence of health behaviours and conditions may differ in the study sample compared with the target population (selection bias)

What is reverse causality?

A situation in which one variable is said to cause another variable, when in reality the reverse is true. (don't know what comes first in some cross-sectional studies)

What is the possible selection bias in cross-sectional studies?

The prevalence of health behaviours and conditions may differ in the study sample compared with the target population

What is the target population?

The collection of individuals for which we wish to draw inferences or be able to generalise to

When can results from a cross-sectional study be generalised to the target population?

When there is a high level of response (study sample is a high proportion of selected sample)

Define generalisability:

The extent to which research findings can be applied to a setting other than the one being tested

What is the difference between selection bias and generalisability?

Selection bias relates to internal validity, whilst generalisability relates to external validity

What does internal validity represent?

The degree to which the results from a study are accurate, and the results can be applied to the target population

What does external validity represent?

The degree to which the results of the study are applicable for populations other than the target population

What does accuracy relate to?

How representative the sample is of the population