HaDPop glossary Flashcards
1
Q
Define census
A
The simultaneous recording of all demographic data by government at a particular time peregrinating to all the persons who live in a particular territory
2
Q
What are the features of a census?
A
Regular intervals - e.g. every 10 years
Every household - every member or each household
Bound by law to cooperate
Information relating to individuals confidential and can’t be used for sampling
Run by government
Covers a defined area
Personal enumeration - or delegated person in each household fills in form
Simultaneous
Universal coverage
3
Q
Define crude birth rate
A
Number of live births per 1,000 population.
For describing the impact of births (-deaths, +/- migration) on the population size
4
Q
Define general fertility rate
A
Number of live births per 1,000 of females aged 15-44 years.
Affected by age specific fertility rates and distribution of families in each age group (unlike TPFR).
For comparing the fertility of fertile female populations.
5
Q
Define total (period) fertility rate
A
The average number of children that would be born to a hypothetical woman in her life - if she experiences the average fertility at each age throughout her life, how many children would she have?
Sum of the current age-specific fertility rates, therefore standardises populations (they’re no longer influenced by the size of different age groups, as each age-group gets an equal weighting)
6
Q
Fertility
A
The realisation of biological potential as births.
Therefore higher sexual activity, better economic climate -> higher fertility.
Higher use of contraception, higher abortion rate -> lower fertility
7
Q
Fecundity
A
The physical ability to reproduce.
Biologically able.
Therefore higher sterilisation and hysterectomies, lower fecundity.
8
Q
What is a trend?
A
A comparison of rates - implies a comparison over time only, but can also be a comparison between places, across socio-economic groups etc… or a combination/
9
Q
What is the purpose of the crude birth rate?
A
For describing impact of births (-death, +/- migration) on the size of a population.
10
Q
What is the purpose of the general fertility rate?
A
For comparing fertility or fertile female population.
11
Q
What is the purpose of the total (period) fertility rate?
A
For comparing fertility of fertile female without being influences by age group structure.
12
Q
Define crude death rate
A
number of deaths per 1,000 population
13
Q
Define age-specific death rate
A
Number of deaths per 1,000 in age group
14
Q
Define standardised mortality ratio (SMR)
A
Compares death rates by applying them to the same population structure.
It is an adjusted ratio (usually for age and sex), therefore it is better when considering the likelihood of death
15
Q
Define confounders
A
A factor that is associated with the exposure and is also independently associated with the outcome but is not on the causal pathway. E.g. age and sex can be confounding factors.
16
Q
Define rate
A
A measure of absolute risk. Rates have a unit of time in the denominator (compare to a proportion like prevalence which can have a unit of time in its numerator)
17
Q
Define ratio
A
A measure of relative risk. E.g. comparing risk in group A with group B.
18
Q
What is an incidence rate ratio?
A
Compares the incidence rate of two groups.
IRR= RateA/Rate B
If the IRR is greater than 1, it suggest group A is at greater risk than group B.
If the IRR is less than 1, it suggests group A is at lesser risk than group B
19
Q
What is a control group?
A
A group of subjects that is matched as closely as possible with an experimental group, but is not exposed to any experiential treatment/ exposure of interest. The results are then compared to determine the changes that may occur due to the exposure.
20
Q
Cross-sectional study
A
A type of observational study that involves the analysis of data collected from a population, or a representative subset, at one specific point in time e.g. prevalence point study.
21
Q
Define incidence rate
A
The number of new cases of a disease arising in a given time (where the focus is on events). It is a rate because it has time in the denominator.
22
Q
Define prevalence
A
The total number of people in the population that currently have a particular disease. This is different from incidence which measures the number of new cases in a population. Incidence and prevalence are related as all prevalent cases will have been an incident case at some point. This is a proportion, not a rate.
23
Q
Define statistically significant
A
The likelihood that a result or relationship is caused by something other than mere chance. Statistical hypothesis testing or estimation can be used in order to determine this:
- When a p-value is less than 0.05 it is considered statistically significant against the null hypothesis.
- When the null hypothesis value (i.e. IRR=1 or SMR=100) is outside the 95% confidence intervals of an observed value, then we say it is statistically significant evidence against the null hypothesis
24
Q
Define clinically significant
A
Clinical significance is the practical importance of a treatment effect or mortality rate - whether it has a real, genuine, palpable, noticeable effect on daily life/ the mortality rate is so high that even if there is insufficient data for it to be statistically significant, if true, would show such an increased risk measures should be taken to solve it.
25
Define the 95% confidence interval
The range within which we can be 95% certain that the 'true' value of the underlying tendency really lies.
26
What is a cohort study?
A study which starts with disease free individuals, that have been exposed to something and then follows them to see if they develop the disease.
Can be prospective or retrospective.
27
What is a longitudinal study?
A longitudinal study is an observational research method in which data is gathered for the same subjects repeatedly over a period of time.
28
Define person-years
The sum of the total time of everybody followed up in a study, e.g.
1 person followed up for ten years +
3 people followed up for 2 years + ...
29
What is a prospective cohort study?
Recruit disease-free individuals and classify them according to their exposure status and then follow them up over a time period.
E.g. study starting now may begin collecting information and continue following-up until 2020.
30
What is a retrospective cohort study?
Recruit disease-free individuals and classify their exposure status and subsequent disease status USING HISTORICAL RECORDS.
E.g. study taking place now may go back and collect information from records 2000 onwards.
31
Secondary analaysis
Secondary analysis involves the use of existing data, collected for the purposes of a prior study, in order to pursue a research interest which is distinct from that of the original work; this may be a new research question or an alternative perspective on the original question.
32
Selection bias
A bias in the selection of individuals for analysis such that they are systematically different from the population being analysed.
This undermines the external validity of a test.
E.g. A case-control study of lung cancer and smoking. Cases of lung cancer from the respiratory medicine ward. Controls are a random sample of patients from the same wards who do not have lung cancer. BUT smoking cause other respiratory diseases, so the patients on the ward are not a representative sample of the general population.
Healthy worker effect is a type of selection bias that occurs when participants are recruited from a workplace in an occupational study and compared with the general popuatlion - as generally those that are working are healthier than the general population
33
Internal validity
Refers to how well a study is designed, especially whether it avoids confounding factors. The less chance of confounding in a study the higher its internal validity.
34
External validity
How able you are to generalise the results of a study to the rest of the population. Selection bias effects external validity
35
Survivor bias
Occurs when those that remain in the study are systematically different from those that left. E.g. If the patients in the worst medical condition all leave the study.
In general those that are most health-conscious stay in health studies and those that aren't leave.
36
Observational studies
Like experiments, observational studies attempt to understand cause-and-effect relationships. However, unlike experiments, the researcher is not able to control (1) how subjects are assigned to groups and/or (2) which treatments each group receives.
For example, a sample survey, does not apply a treatment to survey respondents. The researcher only observes survey responses. Therefore, a sample survey is an example of an observational study.
37
Case-control study
An observational study that compares patients who have a disease or outcome of interest (cases) with patients who do not have the disease or outcome (controls), and looks back retrospectively to compare how frequently the exposure to a risk factor is present in each group to determine the relationship between the risk factor and the disease.
38
What is the rare disease assumption?
IRR aproximately = AD/BC
The rare disease assumption is a mathematical assumption in epidemiologic case-control studies where the hypothesis tests the association between an exposure and a disease. It is assumed that, if the prevalence of the disease is low, then the odds ratio approaches the relative risk (IRR).
39
Odds ratio
An odds ratio (OR) is a measure of association between an exposure and an outcome. The OR represents the odds that an outcome will occur given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure.
OR=ad/bc
Under the "rare disease assumption" IRR aproximately = AD/CB, but AD/CB always equates to the OR (=AD/BC) which is a valid measure of the excess risk in cases compared with controls.
40
How can the terminology retrospective and prospective be confusingly used to two different issues in a study?
1. How the analysis is conducted - forwards or backwards in time?
2. How the data is collected - in the present or from a historical review?
Cohort studies always looks forwards in time in terms of their analysis, but the data can be collected prospectively or retrospectively.
Case-control studies always look backwards in time in terms of their analysis.You can NEVER have a prospective case-control study, with the exception of a nested case-control study - which collects data from the evolving outcome and exposure database of a prospective cohort study.
41
Nested case-control study
A case-control study 'nested' within a cohort study. IRR can be calculated as the sampling fraction is known
A nested case control (NCC) study is a variation of a case-control study in which only a subset of controls from the cohort are compared to the incident cases. In a case-cohort study, all incident cases in the cohort are compared to a random subset of participants who do not develop the disease of interest.
42
Information bias
A bias occurring due to measurement error - observational bias or misclassification bias.
E.g. systematic misclassification - recall bias, assessor bias, data collection methods may differ
43
Recall bias
A systematic error caused by differences in the accuracy or completeness of the recollections retrieved ("recalled") by study participants regarding events or experiences from the past.
44
Define epidemiology
The study of the distribution and determinant of health related states or events in specific populations, and the application of this study to the control of health problems.
45
Define clinical trial
Any form of planned experiment which involves patients and is designed to identify the most appropriate method of treatment of future patients with a given medical condition
46
Non-randomised clinical trial
Involves the allocation of patients receiving a new treatment to compare with a group of patients receiving the standard treatment.
These are prone to allocation bias and confounding
47
Allocation bias
A systematic difference introduced by patients/researchers/clinicians on the study group that participants - already involved in a study - are allocated to.
This differs from selection bias as people have already been selected into the study.
48
Comparison with historical controls in a randomised controlled trial
When a group of patients who had the standard treatment are compared to a group receiving a new treatment. Both groups otherwise experience the same experimental conditions e.g. level of care, protocols, professionals etc...
49
Define random allocation
This refers to the how experimenters divide participants into each experimental condition, to reduce any bias in the distribution of participant characteristics e.g. allocation bias.
In the long run randomisation leads to treatment groups that are likely to be similar in size and characteristics by chance (both known and unknown confounding factors)
50
Open label
a type of clinical trial in which both the researchers and participants know which treatment is being administered.
51
What is a non-treatment effect in a clinical trial?
E.g. a clinician may alter their treatment, care and interest in the patient. Patient may alter their behaviour, other treatment or even expectation of outcome
52
Measurement bias
Occurs when measurement (exposure or outcome) differs amongst study groups. That is there are non-comparable information sources that could influence the study.
E.g. If study group A use one BP machine and study group B used a different one, but one of the BP machines read off higher BP values
53
Blinding (masking)
blinding prevents people – researchers, assistants, patients, even statisticians involved in the study from knowing the treatment a subject is receiving. .
54
Single blind
One of patient, clinician or assessor does not know the treatment allocation (usually patient)
55
Double blind
Two of patient, clinician, assessor does not know the treatment allocation (usually patient + clinician/assessor)
56
Triple blind
Term rarely used - double-blind normally used to imply that all do not know the allocation - but means the patient, clinician AND assessor do not know the treatment allocation.
57
The placebo effect
Even if therapy is irrelevant to the patient's condition, the patient's attitude to his or her illness, and indeed the illness itself, may be improved by a feeling that something is being done about it
58
Placebo
An inert substance made to appear identical in every way to the active formulation with which it is to be compared e.g. appearance, taste, texture, dosage regime, warning etc...
The aim of the placebo is to cancel out any placebo effect that may exist in the active treatment
59
'as-treated' analysis
Analyses only those who completed follow-up and complied with treatments, therefore compares the physiological effects of treatment BUT loses the effects of randomisation - non-compliers may be systematically different from compliers therefore introduces selection bias and confounding again!
60
'intention-to-treat' analysis
Analyses according to the original allocation to treatment groups (regardless of whether they completed follow-up or complied with treatment). This preserves randomisation and compares the likely effects of using the treatments in routine clinical practice
61
Narrative reviews
A type of literature review. It summaries different primary studies from which conclusions may be drawn into a holistic interpretation contributed by the reviewer's own experience, existing theories and models. Result are of a qualitative not quantitative meaning.
It has: implicit assumptions, opaque methodology, not reproducible and therefore is biased and subjective
62
Systematic reviews
An overview of primary studies that used explicit and repoducible methods
They have: explicit assumptions, transparent methodology, reproducible and therefore are unbiased and objective
63
Implicit assumption
an assumption that includes the underlying agreements or statements made in the development of a logical argument, course of action, decision, or judgement that are not explicitly voiced nor necessarily understood by the decision maker or judge.
64
Explicit assiumption
Fully and clearly expressed; leaving nothing implied
65
Meta-analysis
A quantitative synthesis of the results of two or more primary studies that addressed the same hypothesis in the same way
66
publication bias
When studies with statistically significant or 'favourable' results are more likely to be published than those studies with non-statistically significant or 'unfavourable' results - this particularly applies to smaller studies
67
Crude mortality rate ratio
Compares number of deaths per population per time in subgroups of people.
Not adjusted for characteristics of the population (e.g. age, sex).
For similar sized populations it compares the burden of mortality.
68
Sampling frame
The group of people in the population that we wish to study.
Needs to relate to the objectives of the study.
Needs to be a relevant group
69
Reference population
The sample population which we compare to when carrying out a study.
The rate for this group would be the denominator in the rate ratio.
Rate ratio of 1 means that the population under study is no different to the reference population.
70
Response bias
This occurs when the characteristics of non-responders are different to those that respond.
These characteristics may also explain the reason for non-response.
E.g. the non-responders to a postal survey may be systematically different to responders in the fact that they may not speak English
71
Sensitivity analysis
An analysis of the effect that non-response has had on the results of a study.
Involves making assumption about the missing responses in our data.
Assumes extreme values for the missing responses and sees how different the result is when including these values.
72
Due to chance
The value we observe from our sample may be different from the true population estimate just due to the make-up of the sample that we have chosen.
This means that we may take another sample and end up with a different estimate, even though we are trying to estimate the same thing.
Small sample sizes are more likely to be influenced by chance
73
null hypothesis
The null hypothesis represents a theory that has been put forward, whether because it is believed to be true or because it is to be used as a basis for argument, but has not been proved
We carry out hypothesis tests to see if there is any evidence AGAINST the null hypothesis
74
Observer bias
This refers to the cultural assumptions which all researchers and data collectors bring to their work and which help determine their method of research and their observations.
In a non-blinded research trial this could be introduced by researchers expecting to see a better outcome in the individuals trialling a new drug
75
Deterministic approach
Validation of hypotheses by systematic observations to PREDICT WITH CERTAINTY FUTURE EVENTS e.g. coronary atheroma and CHD
76
Stochastic approach
Assessment of hypothese by systematic observations to give the LIKELIHOOD OF FUTURE EVENTS.
Significant association does not mean causality exists
E.g. multifactorial risk and CHD
77
Point prevalence
is the number of persons with disease in a time interval (eg, one year) divided by number of persons in the population; that is, prevalence at the beginning of an interval plus any incident cases.
78
Period prevalence
Period prevalence is the number of persons with a specific disease at one point in time divided by total number of persons in the population.
79
What does the p value mean?
The probability that, given that the null hypothesis (ho) is true, the result is due to chance.
If p 0.05 there is not sufficient evidence to reject Ho as the result is not statistically significant. We never accept Ho
80
What is a cohort?
A defined group of people with a known exposure or characteristics
81
What is an explanatory trial?
This follow the 'as treated' analysis and only uses the data from those participants who proceeded with the treatment and discards the data of those who dropped out.
It therefore measure physiological action of the treatment
82
What is a pragmatic trial?
This follows the 'intention to treat' analysis and interprets the data from all participants, regardless of whether or not they stopped taking the treatment. The advantage of this is that it shows more accurately how well the treatment would work in a real-life scenario such as a clinical setting. It recognises that even if a treatment is intrinsically better, it is not worth having if lots of people drop out for reasons such as bad side effects
83
What is a fixed effect model?
A model used for calculation the pooled OR and its 95% CI in meta-analyses that assumes all studies are estimating exactly the same effect size.
84
What is a random effect model?
A model used for calculation the pooled OR and its 95% CI in meta-analyses that assumes that studies are estimating a similar, but not the same, effect size. The weighting given to studies is more equal and the confidence interval is often larger
85
What are systematic reviews?
Compilation of primary studies (usually RCTs)
Use a meta-analysis to combine the results of primary studies to give an overall result.
Allows larger sample size to reduce the impact of random variation.
Best known source is the Cochrane library
86
What is meant by analogy in the Bradford Hill criteria?
A causal link is more likely is an analogy exists with other diseases, species or settings.
Inappropriate analogy - BSE won't transmit to humans because scrapie doesn't
But epidemiology of Hepatitis B virus was used to predict how HIV would spread
87
What is meant by biological plausibility in the Bradford Hill criteria?
A causal link is more likely if a biologically plausible mechanism is likely or demonstrated.
However this is limited by current knowledge and if such a mechanism exists it does not guarantee an association or effect
88
What is meant by coherence of theory in the Bradford Hill criteria?
A causal link is more likely if the observed association conforms with current knowledge.
However this leads to inappropriate rejection of 'unfavoured' associations e.g. publication bias
89
What is meant by reversibility in the Bradford Hill criteria?
A causal link is VERY likely if removal or prevention of putative factor leads to a reduced or non-existent risk of acquiring the outcome.
This is probably the strongest evidence for a causal link but is often difficult to demonstrate. Thic can be because of time lags between exposure and disease (e.g. mesothelioma takes 20-40 years after exposure to asbestos) and also ethical considerations of e.g. smoking cessation prevention programmes.
90
What is meant by dose response in the Bradford Hill criteria?
A causal link is more likely if different levels of exposure to the putative factor leads to different risks of acquiring the outcome.
Lack of biological gradient does not rule out a causal effect e.g. threshold effect.
Example is the risk of leukaemia increased in survivors of Hiroshima and Nagasaki atomic bonds the nearer they were to the epicentre
91
What is meant by temporal sequence in the Bradford Hill criteria?
A causal link is more likely if exposure to the putative factor has been shown to proceed the outcome.
A causal link cannot exist if outcome proceeds the exposure!
This is why RCTs and prospective cohort studies are stronger studies and prevalence and case-control studies are weaker
92
What is meant by consistency of association in the Bradford Hill criteria?
A causal link is more likely if the association is observed in different studies and different sub-groups. Consistency of association between studies or groups is unlikely ro be due to the same confounding or bias
93
What is meant by specificity of association in the Bradford Hill criteria?
A causal link is more likely when an outcome is associated only with a specific factor and vice-versa. Specificity of association strengthens the case for a causal link, however lack of specificity does not weaken the case e.g. smoking -> many diseases and current models of disease causation are multifactorial.
Example, mesothelioma is associated with exposure to asbestos - no other significant causes of mesothelioma (which is a reportable occupationally related disease)
94
What is meant by strength of association in the Bradford Hill criteria?
A causal link is more likely with strong associations (commonly measured by rate ratios or odds ratios). Strong associations are unlikely to be explained by undetected confounding or bias. However not always the case and weak associations can still be causal e.g. environmental tobacco smoke and ischaemic heart disease
95
In a prospective cohort study, what does the term prospective mean?
That participants were recruitred/ followed-up by researchers not from historical data
