CAUSE AND EFFECT Flashcards
What is a sufficient cause?
A factor or combination of several factors that will inevitably produce disease.
What is a component cause?
A factor that contributes towards disease causation but is not sufficient to cause disease on its own
What is a necessary cause?
Any agent (or component cause) that is required for the development of a given disease (like specific infectious agent)
In the book ‘Who sank the boat’ what type of cause involved the sinking of the boat?
Each animal was a component cause and together they created a sufficient cause. e.g. thrombosis leading to heart attack usually occurs if there is already a blockage in blood vessel or it is damaged.
What is the Bradford Hill criteria?
Temporality Plausibility Consistency Strength Dose Response Reversibility
What is temporality?
Exposure occurs before the disease can occur (this is definite)(e.g. tb, exposure to the bacteria occurs before the disease) (in any infectious disease)
What is reversibility?
- If you remove the exposure then you reduce the outcome e.g. If you remove saturated fats from diet you reduce risk of heart disease, reduce alcohol consumption and reduce risk of liver problems
What is dose response?
- Change dose of something and it changes the outcome (so if you increase dose, then increases chance of outcome, or if you increase dose of antibiotics disease decreases and health increases) (also radiation poisining)
What is plausibility?
Consistent with other knowledge (reinforces the body of evidence)
What is consistency?
Many studies give the same findings (e.g. trial of a new drug- multiple trials showed that the drug is effective)
What is strength association?
Relative risk. The higher or lower the risk, the greater the relationship is causal.
What is the condition for the Baseline Risk being the same in the intervention group as the control group?
That the groups are selected RANDOMLY
What is prevalence?
What PROPORTION of the population has the disease at a SPECIFIC POINT IN TIME (SNAPSHOT)
e.g. 160 of every 100 000 people in Europe has HIV at the end of 2007
Can prevalence also be measured in a specified period?
YES!! This is called period prevalence (combines prevalence and incidence)
What does prevalence measure the amount of?
DISEASE in a population at a given time
What is the equation for prevalence?
(number of people with disease at given point in time)/(total number of people in population) - it has no units; just proportion or percentage
Is prevalence rate a true rate?
NO because true rates include units of TIME
What is incidence?
Measures HOW QUICKLY people are catching the disease
“RATE of occurrence of NEW cases in a given PERIOD in a SPECIFIED POPULATUION.”
- Express it as either’ Rate (incidence rate) or proportion (culmulative incidence)
Is time a factor in incidence?
YES!! (so is a true rate)
What does incidence consider?
Only NEW infections!
What is the formula?
Number of people who develop the disease in one year/ Average number of people in population
What is an incidence rate good for and what is formula?
Underlying forces driving a disease
- It is ‘number of new cases/total person time of observation/risk”
What is prevalence good for?
Measuring diseases that have gradual onset and long duration (e..g type 2 diabetes and osteoarthritis_
What types of factors are part of descriptive epidemiology?
Prevalence, incidence and cumulative incidence
What type of data measure do we use for PREVALENCE OF DISEASE?
Survey (cross sectional study) such as a random sample cross section- questioned to see if they have a certain condition at that POINT IN TIME
How is the incidence of a disease measured?
Start with a cohort that don’t have the disease BUT are AT RISK for developing it
- Then follow over time to see who ACTUALLY DEVELOPS the disease
What is cumulative incidence also known as?
ATTACK RATE (particularly in short period of time like food poisoning)
What is absolute risk?
Percentage of people with outcome within group
- Event rate
- Calculated for intervention and control groups then compared
What is relative risk?
Takes into account the baseline risk for outcome among people in the INTERVENTION group COMPARED to those in the CONTROL group
RR= AR (INTERVENTION)/ AR (control group)
When can relative risk (RR) be calculated?
- For a study with only 2 POSSIBLE OUTCOMES
- If research designs follow both groups through time
What is another formula for RR?
RR= Re(exposed)/Ru (unexposed)
What is Attributable Risk?
Measure of the absolute difference between the two measures of disease frequency.
- Excess risk of disease in people exposed compared to those not exposed (Re-Ru)
What is indicative of a bias that influences the ‘cause and effect’ outcome?
Confounder
What bias may occur if the investigator distorts the assessment of the participant in the study?
Detection
What is a method that can overcome selection bias?
Randomisation
What type of reporting bias occurs when there are differences in accuracy of events from participants?
Recall bias
What type of bias may occur if researchers manipulate the enrolment of participants into a study?
Selection bias
What bias is introduced when healthy volunteers enrol in a study about healthy living?
Sampling bias
What is a type of error that happens by fluke?
Random error
What other name can ASSESSMENT and DETECTION bias go by?
Ascertainment bias
What type of bias is present if participants can identify which group they’ve been allocated to during a trial?
Performance bias
What type of bias can occur when participants withdraw from a study?
Attrition bias
What type of error can be methodologically controlled?
Systematic error
What is a method that may overcome performance or detection bias?
Blinding
What is a bias that may occur when allocation of participants to groups is compromised?
Allocation bias
What type of reporting bias may occur with patient reported data?
Response bias
What is response bias?
Where a patient may answer a survey untruthfully because they don’t want to reveal possibly embarrassing or private information about them self.
What is selection bias?
Study cohort isn’t representative of population due to sampling techniques
- Appropriate spectrum of patients not included in study (inadequate allocation concealment)
How do we minimise selection bias?
RANDOM ALLOCATION when assigning patients -computer generated
What is performance bias?
Systematic differences in care provided to comparison groups OTHER than intervention being studied
How do we minimise performance bias?
BLINDING ensures patients are unaware of interventions of study (important for subjective things like pain)
Which level of blinding protects against the placebo effects?
Double blinding (participant and researcher blinded to treatment allocation)
What is attrition bias?
Losses to follow ups or dropouts whixh are systematically different between comparison groups
How do we account for attrition bias?
ITT (intention to treat analysis) - analyse ALL participants recruited from randomisation process EVEN IF they failed to complete study (avoids effects of crossover and dropout)
When can the level of attrition threaten study?
If it reaches above 20%
How can you PREVENT attrition bias (5 things)?
- Minimal communication between study staff and participants
- Follow up interviews frequent and long
- Clinics hard to access (distance)
- No incentives on offer (like money)
- No reminders sent to participants
- Only a few contact details available for each participant
What is recall bias?
Type of measurement/information bias
- Systematic error when participants do not remember an event or previous experiences accurately or in the same detail
- Can be UNINTENTIONAL or INTENTIONAL
Can recall bias be corrected for after a study?
NO IT CANNOT so it is important to have good study design
How can we minimise recall bias?
- Define the research question carefully
- High quality questionnaires
- Highly trained research staff with good interview training
- Choose and implement appropriate data collection methods
What is measurement bias?
Systematic error in measurement of information on the exposure or outcome.
- Part of a broad category including recall bias and observer bias
What is an example of measurement bias?
Information was collected using different methods among two groups (one interviewed and the other got questionnaire)
In measurement bias, what is non differential misclassificaiton?
Information is INCORRECT but is the SAME ACROSS BOTH GROUPS (two groups will be more alike and underestimates strength of association between exposure and disease
In measurement bias, what is differential misclassification?
Information errors DIFFER between two groups
What is confounding?
A variable that alters measure of association between exposure and outcome
What is negative confounding?
When the confounding variable biases towards the null hypothesis
What is detection bias?
If outcome assessment differs systematically between comparison groups
e.g. if assessor knows which drug person is on and their response reflects that ‘wow I’m surprised’
What is recall bias?
- A type of measurement/information bias
How is recall bias minimised?
- Define research question carefully
- High quality questionnaires
- Staff with good interview training
- good study design
What is measurement bias?
- Systematic error in measurement of information on exposure or outcome
- Part of broad category including recall bias and observer bias
e. g. info among the two groups was collected differently (one with questionnaire and other interviewed)
What is observer bias?
- When RESEARCHER is aware of the disease or exposure status of the participants
What is the broad classification of misclassification?
- Incomplete medical records, recording errors, misinterpretation, errors in questionnaire (recall bias)
(can be differential and non differential)
What is positive confounding?
- When confounding variable biases towards the alternative hypothesis
Which methods help control confounding?
- Randomise individuals to different groups
- Matching in case control studies
- Filter for certain groups
- Analysis of data through stratification
- Limit participation of certain subgroup who share confounding factor
The difference in risk between the exposed and
unexposed groups in a population is known as the:
Attributable risk
The percentage of disease in a whole population that
can be attributed to a particular exposure is known as
the
Population Attributable Risk
In which one of the following circumstances will
the prevalence of a disease in the population increase, all else being constant?
If survival time with the disease increases
The only one of the Bradford Hill criteria for establishing
causation that is absolutely ESSENTIAL is
Temporal relationship
Which of the following is not true about morbidity?
1) Is a measure of disease burden
2) Useful for high-case fatality
3) Takes into account disability
4) Useful for low-case fatality
2) Useful for high-case fatality
Poorly worded questions; a misunderstanding in interpreting an individual answer from a
particular respondent; or a typographical error during coding; a poorly functioning machine measuring
blood levels of some parameter like sodium.
IS THIS RANDOM ERROR, SYSTEMATIC, OR BIOLOGICAL VARIATION:?
RANDOM ERROR
RANDOM ERROR, BIOLOGICAL VARIATION, OR SYSTEMATIC ERROR? : In human pregnancies, the day of the onset of pregnancy is usually unknown hence the first day
of the last menstrual period is used as a surrogate variable. Pregnancy duration is therefore counted from
the last menstrual period. This introduces a random component to the length of a human pregnancy.
BIOLOGICAL VARIATION
RANDOM ERROR, BIOLOGICAL VARIATION, OR SYSTEMATIC ERROR?
The machine used to measure oxygen saturation of a patient’s blood is calibrated incorrectly so
you get consistently high measures. Since the error happens in every instance, it is systematic and
conclusions made will be incorrect. The measuring device could be precise but not accurate. There is
therefore a problem with the validity of the measurements
SYSTEMATIC ERROR
How do you control for confounding at the DESIGN STAGE?
RANDOMISATION: Equal distribution of groups
RESTRICTION: (e.g. only non-smokers, or males)
MATHCING: match for age, sex, social class in case-control study (time consuming and expensive)
How do you control for confounding in the ANALYSIS stage?
-STRATIFICATION: Two tables of exposure versus outcome-one for each level of confounder
STATISTICAL MODELLING: Can adjust for multiple things e.g. late antenatal care by aboriginal infant + 9 risk factors
A study of the relationship between contact lens use and the risk of eye ulcers. The crude relative risk is
3.0 and the age-adjusted risk is 1.5. Is age a confounder in this study?
YES
A case-control study of the relationship between cigarette smoking and pancreatic cancer. In this study,
coffee drinking is associated with smoking and is a risk factor for pancreatic cancer among both smokers
and non-smokers. Is coffee drinking a confounder in this study?
YES
A study of the relationship between exercise and heart attacks that is conducted among men who do not
smoke. Is gender a confounder in this study?
NO
A cohort study of the risk of liver cirrhosis among female alcoholics. Incidence rates of cirrhosis among
alcoholic women are compared to those among non-alcoholic women. Non-alcoholics are individually
matched to alcoholics on month and year of birth. Is age a confounder in this study
NO
What is reliability?
Demonstrates stability and reproducibility over time
- Stability: Consistent answers to questions over time?
- Reproducibility: Different interviewers= same answers from the questions?
What is Validity and what are the two types?
Validity: Test capable of measuring what it is intended to measure
- No systematic error
- Random Error small
- Internal Validity: Are the results correct for the group being studied?
- External Validity: How do results generalise to the greater population/apply to those not in studied population?
A study of 3000 children in selected rural areas of Ethiopia looked at the levels of disease and death caused by diarrhoea (children were free of diarrhoea at
the beginning of study). The finding was of 4 deaths of diarrhoea per 1,000 children per year. The same study found new 360 episodes of diarrhoea per 1,000 children per year:
PREVALENCE OR INCIDENCE?
INCIDENCE
The proportion of the population in Australia that have fractures today
INCIDENCE OR PREVALENCE?
PREVALENCE
1% of women having a baby in a maternity hospital in Sao Paulo Brazil, were
infected with HIV.
PREVALENCE
What is case fatality?
- Deaths within those diagnosed with the disease
- Usually expressed as percentage
- Number of deaths from diagnosed cases/no of diagnosed cases in same period
What is person years?
- Estimate of the actual time at risk in years of all persons that contributed to the study
e. g. years of treatment with a certain drug - Someone can then calculate how quickly people are developing the disease
What is the mortality?
- Deaths
- useful for high case fatality
- generated from death certificates, gross pathology autopsy, verbal autopsy, psychological autopsy
What is the morbidity?
- Burden of disease
- Useful for low case fatality
- Takes into account impairment, disability, handicap
What does a necessary cause mean?
- That in the absence of this factor, disease will not occur (disease is never present if this factor is absent)
A new treatment developed to cure people with an infectious
disease will reduce the incidence of disease in the population. T/F?
TRUE
those with infectious disease will be cured and not infect others
Mortality measures are a useful indicator of low case-fatality
chronic conditions.
• E.g. Mental illness, rheumatoid arthritis (T/F)?
FALSE!
Death rates useful for high case fatality (not for low case chronic conditions)
If an agent is “causative” this means it is able to cause disease by itself (T/F)?
FALSE! Not all causative agents are capable of causing disease by themselves, some only achieve this in combination with other factors or at certain levels
A necessary cause means that in the absence of this factor
disease will not occur (T/F)?
TRUE!
Necessary cause = disease is never present if this
factor is absent
Potential years of life lost (PYLL) is a measure of health
expectancy (T/F)?
- TRUE!
Number of years of potential life they lost if they die
before a certain age*
*Most reports use death before 65 years premature deaths
The “common cold” is an example of a condition with low incidence and high prevalence (T/F)?
FALSE!
Colds tend to have high incidence and low prevalence as they are short-lived
Low incidence, high prevalence (e.g. diabetes; long time course)
High incidence, low prevalence (e.g. cold; short duration)
• Life expectancy is the average number of years someone alive
today is expected to live and is not related to the year of their
birth (T/F)?
FALSE!!
Life expectancies are determined by many
factors, including year of birth
Mortality rates are a better measure for investigating the impact
of disabling accidents on the health of the population (T/F) ?
FALSE!
Mortality rates are a better measure for investigating the impact
of disabling accidents on the health of the population
What is Descriptive Epi?
- Prevention approach
- Looking at describing disease distribution by characteristics relating to time, place, and person
- “Who, What, When , Where””?
What does the population at risk include?
ONLY people who are susceptible to the disease being studied
e.g. For occupational injuries can have workforce as population
