Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Health & Disease in Populations > HadPop > Flashcards

HadPop Flashcards

0
Q

How do you calculate the Crude Birth Rate? What are the advantages and disadvantages of measuring fertility this way?

A

No. of live births per 1,000 of the population

ADVANTAGE = easy; only info needed is no. of births and no. in population
DISADVANTAGE = does not take the population who cannot give birth into account: men, pre and post-menopausal women
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
1
Q

What is a census and what information can be gained from a census?

A

Simultaneous recording of demographic data by the government at a particular time, pertaining to all persons living in a particular territory.

Population size = calculate rates of change in size
Population structure = determine service needs
Population characteristics = determine measures of deprivation e.g. unemployment, overcrowding, lack of basic amenities

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

How do you calculate the General Fertility Rate? What are the advantages and disadvantages of using it to measure rates of fertility?

A

No. of live births per 1,000 females aged 15-44yrs

ADVANTAGE = removes population who cannot give birth from measure (more accurate)
DISADVANTAGE = ?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How do you calculate the Total [Period] Fertility Rate? What are the advantages and disadvantages of using this method to measure fertility rates?

A

Average no. of children that would be born to a hypothetical woman in her life = sum of age-specific fertility rates

ADVANTAGE = ? 
DISADVANTAGE = hypothetical; as women age the fertility rate changes
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the main determinants of fertility?

A

FECUNDITY = physical ability to reproduce
sterilisation/hysterectomies -> —fecundity

FERTILITY = realisation of fecundity as births
+sexual activity/economic climate -> +++births
contraception/abortion -> —births

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

How do you calculate the Crude Death Rate?

A

No. of deaths per 1,000 of the population

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

How do you calculate the Age-Specific Death Rate?

A

No. of deaths per 1,000 of a specific age group

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How do you calculate the Standardised Mortality Ratio (SMR)? What does the SMR do?

A

Compares “observed” no. of deaths in study population with no. of “expected” deaths of a hypothetical reference population (the age-sex distributions of the study and reference population are identical)

Observed/Expected x 100 = %

e.g. 136% = 36% higher mortality in study pop. than reference pop.

REMOVES AGE-SEX CONFOUNDING

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is the difference between population estimates and population projections?

A

POPULATION ESTIMATE = apply what is known about birth, death, and migration to the PRESENT

POPULATION PROJECTION = additional assumptions about birth, death, and migration in the FUTURE

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are the key variables affecting population estimates and projections?

A

Fertility rate and migration.

e. g. +migration due to political situation in another country
e. g. -fertility rate due to unforseen economic crisis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Describe how health information is used to identify healthcare needs.

A

“Amount” of Disease = focuses on new cases - describes rate of disease (epidemics)

No. of people affected by disease = existing cases (old & new) - describes burden of disease (health service needs)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is the definition of morbidity?

A

Any departure (subjective or objective) from a source of physiological/psychological well-being for any duration (acute or chronic conditions included).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What are some examples of methodological issues in data?

A

COMPLETENESS = births/deaths the only “complete” data
ACCURACY = incorrect/missing data (ease of recording)
VARIATION IN DIAGNOSIS = misdiagnosis/different definitions/”fashions” in diagnosis
NUMERATOR/DENOMINATOR MISMATCH = due to different definitions of either/both
INDIRECT ASSUMPTIONS = e.g. sick days does not always correlate to ill health
CONFIDENTIALITY = see Data Protection Act; informed consent necessary (unless Royal Assent has been granted

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

How do you calculate the incidence rate?

A

Measuring new cases in a pop. over time

New events/person-years

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What are person-years?

A

No. of people in pop. x period of time studied = sum of total time of everybody followed in study

e.g. 100 people studied for 1 year = 100p-y
10 people studied for 10years = 100p-y

Allows an estimate of actual risk to persons in a study pop.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

How do you calculate prevalence?

A

Proportion of existing cases in a population at a specific time (NOT RATE)

No. with disease at a point in time/Population

Prevalence ~ Incidence x Length of Disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

How do you calculate the incidence rate ratio?

A

Comparison between two populations

Rate B (exposed) / Rate A (unexposed)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What is the difference in the measure of risk when calculated as a rate or as a ratio?

A

RATE = measure of absolute risk, i.e. how many die per 1,000

RATIO = measure of relative risk, i.e. how many more die in Pop. A compared to Pop. B

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

How is healthcare performance monitored?

A

Quality Outcomes Framework (QOF) = voluntary reward/incentive programme for GP surgeries in England (given funding related to performance)

Hospital Episode Statistics = records every episode of admitted patient care delivered by the NHS

Hospital/Surgeon League Table = ranks hospitals/surgeons according to different variables e.g. no. of deaths

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What are some contentious issues with the methods of health care performance monitoring?

A 
PURPOSE = is it measuring the past or estimating the future? 
USERS = is the data meant for managers, academics, or the public?
QUALITY = is the data being collected in real-time; is it validated? 
COMPARABLE? = is the data designed to be comparable or has it been customised? 
RELATIONSHIP = is the data integral or independent? 
PUBLICATION = is the data NHS only, or can it be viewed by academics/the public? 
ACCESS = Data Protection Act v.s. Freedom of Information Act 
FUNDING = public or private funding?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is a trend, and how can it be influenced?

A

TREND = comparison of rates over time, place, and socio-economic groups (not influenced by population size)
Can be influenced by random and systematic variation.

Systematic variation = 
NUMERATOR errors: 
- death certification 
- disease diagnosis 
- classification/coding

DENOMINATOR errors:

  • population used
  • population definition
  • population count/estimate
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

How can incidence and prevalence be affected by changes in the population?

A

+cure +death (remove existing cases) = -prevalence

+incidence (add new cases) = +prevalence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

Explain why age & sex is standardised when measuring risk of disease.

A

Age & sex are strong determinants of health

i. e. Rate(old)/Rate(young) will almost always be >1.0
- > removes a CONFOUNDING factor

Age & sex are not useful factors to consider when considering prevention (non-modifiable factors)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Why can systematic variation be sometimes helpful?

A

Variations in risk of disease between groups of people can give clues about the aetiology (cause) of the disease, and determine what factors increase risk of disease/death.

(clinical trials = efficacy of the drug)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

Define and give examples for tendency and observation.

A

TENDENCY = expected probability (true or underlying tendency)
e.g. toss a coin ten times, will get heads half of the time.

OBSERVED = observed probability
e.g. toss a coin ten times, and get heads ten times
(can vary due to random or systematic variation, or confounding)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

Define hypothesis and explain how a hypothesis may be tested.

A

Supposition for a phenomenon/set of facts/scientific inquiry that may be tested, verified, or answered by further investigation or methodological experiment.

e.g. the coin is fair

Calculate the probability of getting an observation as or more extreme than the one observed, assuming that the stated hypothesis is true.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What can you say about a hypothesis if p<0.05?

A

Data is inconsistent with the stated hypothesis, therefore it is reasonable to reject the hypothesis (observations are statistically significant).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

What can you say about a hypothesis if p>0.05?

A

Failed to have rejected the hypothesis……..
Data is consistent with the stated hypothesis.

DOES NOT PROVE THE STATED HYPOTHESIS

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

What is the definition of the null hypothesis?

A

No significant difference between groups studied (any difference due to random variation and underlying rates).

IRR = B/A = 1 B-A = 0 SMR = 100%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

What is the 95% confidence interval referring to?

A

Range within which we can be 95% certain that the true value of the underlying tendency really lies.

(values within CI are consistent with the data)
Smaller range = more precise true value estimate

Inside CI -> p>0.05
Outside CI -> p<0.05

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

How can the 95% confidence interval be calculated?

A
  • Calculate the error factor (x = new cases)
  • Lower CI = observed value/error factor
  • Upper CI = observed value x error factor
  • Middle = point estimate (best guess)

Note: p = 0.05 just inside the confidence interval

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
31
Q

Describe the features of a cohort study.

A

Recruit disease-free individuals -> see who/how many develop disease.

Measures and records factors which may lead to inequality (aim of study does not have to be very specific).

Internal comparison (within pop.) or external comparison (to reference pop.)

Prospective (present -> future) or Retrospective (past -> present).

Data can be binary, categorised, or continuous

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
32
Q

What is the difference between a prospective and retrospective cohort study?

A

PROSPECTIVE = recruit disease-free individuals and classify according to their exposure status, start follow-up immediately/after a waiting period.

RETROSPECTIVE = recruit disease-free individuals and classify their initial exposure status and subsequent disease status (using historical records), follow-up using records (starts in the past and moves forward).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
33
Q

What is the difference between an internal and external comparison?

A

INTERNAL = IRR = r+/r-
If either sub-cohort are small -> error factor is large -> wider CI interval -> less precise data

EXTERNAL = SMR = o/e
Useful if you cannot use sub-cohorts
Error factor only has one x value, therefore does not have same precision problem as internal comparison.

34
Q

How do you adjust for the influence of ageing on disease rates during a cohort study?

A

Use a Lexis diagram: calculate the no. of expected cases/deaths for each age group in each calendar time period.

e.g. 50-54yrs in 1960-64 & 50-54yrs in 1965-69 (different sample of people)

35
Q

What are some of the limitations of external comparisons?

A
  • limited data for reference population (e.g. routine data only)
  • no incidence data for reference population (except mortality rates)
  • study and reference populations may not be comparable (SELECTION BIAS)
36
Q

Describe the healthy worker effect.

A

Employment is often restricted to healthy individuals.

Therefore, occupational cohorts yield SMRs < < 100% (ill people cannot work/may take more sick days).

37
Q

What are the advantages of cohort studies?

A
  • Study exposures not routinely collected e.g. by a census (Prospective)
  • Obtain more detailed info on outcomes & exposures (Prospective)
  • Collect data on potential confounders
  • Good for conditions which fluctuate with age (randomly or systematically) as you can adjust for it (Lexis)
  • Outcomes do not have to be as defined/restricted as case-control
  • Good for studying rare exposures
  • Good for establishing that exposure precedes outcome
38
Q

What are the disadvantages of cohort studies?

A
  • Large & time-consuming = expensive/resource-intensive
  • Definitions of disease/exposure can be expensive/invasive e.g. genetics, CT scans
  • High no. of drop-outs = SURVIVOR BIAS (those who remain in the study differ from those who dropped-outs)
  • Results take a long time to be compiled (changing political/ethical issues)
  • Not good for rare diseases (too small x value)
  • Unknown confounders e.g. genetics
39
Q

What are the features of a cohort study? How do you do a cohort study?

A
  1. Define outcomes (cohort = multiple outcomes)
  2. Cohort study = start with disease-free individuals at a point in time, split into two or more cohorts based on exposure.
  3. Decide whether to do an internal or external comparison (or both)
  4. Look out for random & systematic variation and confounders
  5. Decide how to collect and interpret data e.g. IRR or SMR
  6. Try and adjust for confounders (related to outcome AND exposure).
40
Q

What do the terms “national” and “historical” refer to when applied to cohorts?

A

National = different areas

Historical = different time periods

41
Q

How do you conduct a case-control study?

A

1) Identify a group of cases (diseased)
2) Identify a suitable group of non-cases (controls)
3) Ascertain previous exposure status of everyone
4) Compare level of exposure in cases of controls (calculate odds ratio)

42
Q

What is the rare disease assumption?

A 
A = exposed cases B = exposed controls 
C = unexposed cases D = unexposed controls

C + D ~ D A + B ~B

Allows you to make estimates of relative RISK with OR
Therefore, OR = ad/bc = measure of excess risk in cases compared with controls

43
Q

What is the advantage of nested case-control studies?

note: 6 controls for every 1 case at most

A
  • collect more detailed information for a minority of participants
  • can calculate incidence states i.e. absolute risk (cannot in a normal case-control study)
  • the population for sampling of controls is already defined (easy)
  • minimise recall bias (if you’ve already collected the info)
44
Q

What are some of the problems with case-control studies?

A

SELECTION BIAS = control cases are not truly exposed/unexposed (underestimated OR)
e.g. lung cancer & smoking, select controls from respiratory ward

Cases need to be representative of all cases. Controls need to be representative of the population where the cases came from.

INFORMATION BIAS =

Non-differentiated misclassification: randomly inaccurate measurement e.g. can’t remember if they’ve been exposed or not
(OR shrinks to null)

Systematic bias: over-reported exposure of cases = OR further from 1, over-reported exposure of controls = OR towards 1.

CONFOUNDING

45
Q

How can you reduce the confounding in case-control studies?

A

Minimise by matching using important confounders e.g. age & sex
e.g. match by DOB

Use logistic regression

46
Q

What are the features of a case-control study?

A

Compares exposures in groups based on disease status
Not good for rare exposures e.g. radioactive disasters
Can study a range of exposures for a single outcome
Conventionally retrospective
Cannot establish that exposure precedes the outcome (unless a nested-case control study is used)
Cannot directly measure the IRR (unless a nested case-control study is used)
Odds ratio used to calculate risk

47
Q

What are some of the possible explanations for association?

A

Unknown confounding: e.g. genetic effect on epidemiology

Common cause: lung cancer and bronchitis both caused by tobacco smoking, they do not cause each other

Reverse causality: bar work associated with alcoholism; do bar staff become alcoholic or do alcoholics become bar staff

True causal association

48
Q

What are Bradford Hill’s criteria for inferring causality?

A

ASSOCIATION:
Strength = causal link unlikely to be explained by unknown confounding/bias
Specificity = one specific factor only (however current disease models are multi-factorial)
Consistency = association observed in different studies/sub-groups (same bias/confounding unlikely to be present in all studies)

EXPOSURE/OUTCOME:
Temporal sequence = exposure to putative factor proven to precede outcome
Dose response = different levels of exposure correspond to differing levels of risk of acquiring the outcome (note: J-shaped curves are an exception)
Reversibility = removal/prevention of putative factor causes reduced/non-existent risk of acquiring the outcome

OTHER EVIDENCE:
Coherence of theory = observed association conforms with current knowledge (exception: publication bias)
Biological plausibility = biologically plausible mechanism likely/demonstrated
Analogy = analogy exists in other diseases/species/settings

49
Q

What are two examples of assumptions we make in epidemiology?

A
  • Disease does not occur at random

- Disease has causal and preventable factors that can be identified through systematic investigation

50
Q

What is the definition of a clinical trial?

A

Any form of planned experiment (i.e. not case-control & cohort studies; these are observational only) which involves patients and is designed to elucidate the most appropriate method of treatment of future patients with a given medical condition

51
Q

What are the three key points specific to clinical trials?

A
  • benefit for future patients, not for patients in the trial
  • patients allocated treatment or placebo
  • planned experiment (not observational only)
52
Q

What two things do clinical trials provide reliable evidence of?

A

EFFICACY = ability of health care intervention to improve the health of a defined group under specific conditions

SAFETY = ability of a health care intervention not to harm a defined group under specific conditions

Note: specific conditions refers to the setting of a clinical trial i.e. not in a real-life setting

53
Q

What three things do clinical trials need to be?

A
  • FAIR: unbiased without confounding
  • CONTROLLED: comparison of interventions only
  • REPRODUCIBLE: in experimental conditions
54
Q

Why do we randomise trials?

A

Reduce selection/allocation bias (by patient, clinician, or investigator)

Reduce known and unknown confounding (by randomising the groups you should get a similar population in each, thus limiting confounding factors, or differences, between the groups)

55
Q

What does “scientifically proven” mean?

A

More people receiving this treatment are cured than those on the other treatment

i.e. not that the majority of people receiving this treatment are cured

56
Q

Why don’t we compare patients on the new treatment with historical controls?

A

Selection of historical control was less defined/rigorous and there is less information available about potential bias and confounding

Therefore we would be unable to control for confounding in the standard treatment (historical) group, and thus it would be treated differently to the new treatment group.

57
Q

How do you conduct a randomised controlled trial?

A

1) Define the disease of interest, treatments to be compared, outcomes to measured (primary and secondary), possible bias and confounding, patients eligible for the trial, and patients to be excluded from the trial.
(all prevents data-dredging)
2) Identify source of eligible patients
3) Invite eligible patients to trial
4) Consent patients willing to be in the trial
5) Allocate participants to the treatments fairly (randomly)
6) Follow-up participants in identical ways
7) Minimise losses to follow-up
8) Maximise compliance with treatments
9) Compare outcomes (any observable differences? difference statistically significant? difference clinically important? difference attributable to treatments compared?)

58
Q

What is blinding?

A

Follow-up of participants is identical as one or more groups do not know the treatment allocation, therefore removing allocation bias, data collection bias, and the accounting for the placebo effect.

Can be single (patients only), double (patient and clinician/assessor), or triple (only designated pharmacy knows).

59
Q

What kind of situations are difficult to blind?

A
  • surgical procedures
  • psychotherapy v.s. anti-depressants
  • alternative medicine v.s. Western medicine
  • lifestyle interventions
  • prevention programmes
60
Q

What is the placebo effect?

A

Patient’s attitude to their illness, and possibly the illness itself, may be improved by a feeling that something has be done about it i.e. the idea of being given treatment

61
Q

Define a placebo.

A

An inert substance made to appear identical in any way to the active formulation with which it is to be compared

i.e. aims to cancel out the placebo effect

62
Q

What are the types of outcome? Give some examples.

A
  • PATHO-PHYSIOLOGICAL e.g. tumour size, thyroxine level, ECG changes
  • CLINICALLY DEFINED e.g. death (mortality), disease (morbidity), disability
  • PATIENT-FOCUSED e.g. quality of life, psychological well-being, social well-being, satisfaction
63
Q

At what times are measurements made in the study?

A

BASELINE = monitoring for inadvertent differences in groups

DURING TRIAL = monitoring for possible effects - efficacy (i.e. is one group being disadvantaged?) and monitoring for adverse effects - safety (i.e. are individual patients being harmed?)

FINAL = comparing final effect of treatments in trial

64
Q

What are the differences between explanatory and pragmatic RCTs?

A

EXPLANATORY (as-treated) = analyses only those who completed follow-up and complied with treatments
(hence loses effects of randomisation and has selection bias and confounding)
- larger size of effect, compares strict physiological effects of the treatments

PRAGMATIC (intention-to-treat) = analyses everyone regardless of whether they completed follow-up or complied with treatments
(hence preserves effects of randomisation and minimises selection bias and confounding)
- smaller, more realistic effect, compares effects of treatments in routine clinical practice

65
Q

What is difference between the collective ethic and the individual ethic in terms of the benefit of RCTs?

A

COLLECTIVE = all patients should have treatments that are properly tested for efficacy and safety (by RCTs)

INDIVIDUAL = RCTs do not guarantee a beneficial effect, may result in harm, do not allow for autonomous choice, and place burdens on the patients

66
Q

What is clinical equipoise?

A

There is reasonable uncertainty or genuine ignorance about the better treatment or intervention

(otherwise would have to stop the trial and allocate all patients the better treatment or intervention)

67
Q

Give some examples of qualities of scientifically robust trials.

A
  • addresses an important/relevant issue
  • asks a valid question
  • appropriate study design & protocol (addresses bias and confounding)
  • potential to reach sound conclusions
  • justifies use of placebo
  • acceptable risk of harm compared to anticipated benefits
  • provision for monitoring safety and well-being of participants
  • arrangements for appropriate reporting and publication
68
Q

Give some examples of qualities of ethical recruitment.

A

Prevent inappropriate inclusion of:

  • participants unlikely to benefit e.g. testing drugs in developing countries
  • participants with a high risk of harm with respect to potential benefits e.g. pregnant women
  • participants more likely to be excluded from analysis

Prevent inappropriate exclusion of:

  • people who differ from the ideal homogenous group (note: can lead to the elderly being excluded due to co-morbidities, but this affects the clinical importance)
  • people who are difficult to get valid consent from e.g. prisoners, mentally-disabled people, children
69
Q

What are the components of valid consent?

A
  • knowledgeable informant
  • appropriate verbal and written information + cooling off period + freedom to opt out
  • informed participant, competent decision-maker, legitimate authoriser
  • signed consent form (evidence of valid consent, but not valid consent itself)
70
Q

What is voluntariness?

A

Pre-requisite for valid consent = decision should be free from coercion or manipulation or the perception that coercion or manipulation may have taken place

71
Q

What are some examples of coercion and manipulation?

A

COERCION = non-access to the “best” treatment, lower quality of care, uninterested clinician

MANIPULATION = exploitation of emotional state, distortion of information, financial inducements

72
Q

How would you get approval for a clinical trial?

A

NHS trust Research & Development Office (research governance, financial management, resource implications)
+
Research Ethics Committee (protects patients’ dignity, rights, safety, and well-being)

73
Q

What is a systematic review? What are the steps involved?

A

An overview of primary studies that used explicit and reproducible methods and is transparent

1) Clearly focused question
2) Explicit statements about: types of study, participants, interventions, and outcome measurement
3) Systematic literature search
4) Selection of materials e.g. exclusion of some studies
5) Appraisal
6) Synthesis e.g. meta-analysis

74
Q

What is meta-analysis? What are the functions of meta-analysis?

A

Quantitative synthesis of the results of two or more primary studies that addressed the same hypothesis in the same way

  • facilitates synthesis of a large no. of study results
  • systematically reviews results
  • reduces problems of interpretation due to variations in sampling
  • qualifies effect sizes and their uncertainty as a pooled estimate
75
Q

When will meta-analysis not be used?

A

When clinical heterogeneity is too great i.e. the amount of difference between studies is too high

76
Q

How are forest-plots made?

A
  • calculate OR and 95% CIs for all studies
  • combine studies to give pooled estimate OR
  • weight studies according to size of study and uncertainty of OR
    (smaller e.f. —> greater weight)
77
Q

What is the difference between fixed-effect and random-effect model forest plots?

A

FIXED-EFFECT = assumes that all studies are estimating exactly the same effect size
(one true value; random errors cause variation of results from true value)

RANDOM-EFFECT = assumes that the studies are estimating similar, not the same, effect size i.e. can account for variation but cannot explain it
(each study has its own true value, and the distribution of true values has an average value)

note: smaller studies given greater weighting than with fixed-effect model, use when clinical heterogeneity to detect, wider CI

78
Q

What are some problems with meta-analysis?

A
  • heterogeneity between studies
  • variable quality of studies
  • publication bias
79
Q

What is sub-group analysis?

A

Stratification by study characteristics e.g. by study design, participant profile

80
Q

How can you solve issues with variable quality of studies?

A
  • define a basic standard of quality and only include studies that meet this criteria
  • score each study for its quality and then weight by score (higher quality studies have a greater influence)
  • use sub-group analysis to explore differences between studies
  • meta-regression analysis (effect size v.s. quality)
81
Q

What is publication bias?

A

Studies with statistically significant or favourable results are more likely to be published than studies with non-statistically significant or “unfavourable” results (particularly affects smaller studies)

82
Q

How can you identify publication bias?

A
  • check meta-analysis protocol for method of identification of studies (+ search for unpublished
  • plot results of identified studies v.s. measure of size (Funnel plot)
  • use a statistical test

Health & Disease in Populations (1 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions